Federal Regis'er / Vol. 50, No. 229 / Wednesday, Noverllber 27, 1n~)1 / Proposed Rules- - .___________ - ---------_--0-- __ _ . ... _. .. .__..._..._

60825

21 CFR Part 101

JOocket No. 91N-00951

RIN 0905-AB67

Food Labeling: Health Claims andLabel Statements; Sodium/Hypertension

AGENCY: Food and IJrug f\drninistralion.IIIIS.ACTION: Proposed rule.

SUMMARY: The Food and DrugAcIIninistnltion (FDA) is proposing toauthorize health clain1s on food labelsand labeling that state that a lowsodium diet is associated with lowerblood pressure in son1e people. Theagency reviewed the relationshipbetween dietary sodiun1 intake andhypertension under provisions of theNutrition Labeling and Education Act of1990 (the 1990 amendments). On thebasis of this revie\v, the agencytentatively concludes that there issignificant scientific agreement amongexperts qualified by scien tific trainingand experience to evaluate such claims,and that the strength and consistency ofthe publicly available scientificevidence supports such claims. Theagency's tentative conclusion is basedon its review of the scientific literatureand on review of conclusions andrecommendations provided in Federalgovernn1ent and other authoritativedocuments. \

DATES: Written comments by February25,1992. The agency is proposing thatany final rule that may issue based uponthis proposal become effective 6 monthsfollow ing its publica tion in accordancewi th requirements of the 1990amendn1ents.ADDRESSES: Written comments to theDockets Management Branch (HFA­305), Food and Drug Administration, I'm.1-23, 12420 Parklawn Dr., Rockville !YID20857.FOR FURTHER INFORMATION CONTACT:Ellen M. Anderson, Center for FoodSafety and Applied Nutrition (HFF-266),Food and Drug Administration, 200 C S1.SW., Washington, DC 20204,202-472­5375.SUPPLEMENTARY INFORMATION:

I. Background

A. The Nutrjtion Labelj]zg andEducation Act of 1990

On November 8,1990, the Presidentsigned into law the 1990 amendnlents(Pub. L. 101-535), which amended theFederal Food, Drug, and Cosmetic Act(the act). The 1990 anlendments, in part,authorize the Secretary of }-Iealth and

IIUIl),I!1 SerTic(~s (the SecrctHry) to iS~';lle

('(~glll;lti()ns illllh()riziIlg nutricnt c()nl(~nt

~ Ind h(~ill the Iil inl~; 0 nth e IClI)( d 0 r

Llheling of foods. \Vith !TSl)(!ct to hCilllhc!ilirns, th(~ ncvv provisions provide that(I product is n1isbranded if it bears acLlin1 lhut chari.lcterizes the J'(~I(ltionshjp

of a nutrient to a disease or hcallh-rc Ia t f~ d con d i tion, un Ipss the c I. a in1 islnade in accordance with the proceduresand standards established under section40::3(r)(1 HB) of the act (21 lJ.S.C.~14~~ (r)( 1)(B)J.

Published elsewhere in this FederalRegister is a proposed rule to establishgeneral reql1irenH~ntsfor health claimsthat characterize the relationship ofnutrients, including vitamins andminerals, herbs, or nutritionalsubstances (referred to generally as"substances") to a disease or healthrelated condition on food labels and inlabeling. In this companion documentFDA has tentatively deternlined thatsuch claims would be justified fordietary supplenlents as well asconventional foods only if the agencydetern1ines based on the totality of thepublicly available scientific evidence(including evidence from well-designedstudies conducted in a manner which isconsistent with generally recognizedscientific procedures and principles)that there is significant scientificagreement among experts qualified byscientific training and experience toevaluate such claims, that the clainl issupported by such evidence.

The 1990 an1endments also require(section 3(b)(1)(A)(ii), (b)(l)(A)(vi), and(b)(l)(A)(x)) that, within 12 months oftheir enactment, the Secretary shallissue proposed regulations to implementsection 403(r) of the act (21 U.S.C.343(r)), and that such regulations shalldeternline. an10ngother things, \vhethcrclain1s respecting 10 topic areas,including sodium and hypertension,meet the requirenlents of section403(r)(3) of the act (21 U.S.C. 343(1')(3)).In this document, the agency willconsider whether a claim on food orfood products, including conventionalfoods and dietary supplements, on therelationship between sodium andhypertension would be justified underthe standard proposed in the companiondoculllent entitled "Food Labeling:General Requirements for Health Clainlsfor Food: Proposed Rule."

B. Sodhlln/flypertensjol1 Relationship

1. I-Iypertension

IIypertension, commonly referred toas high blood pressure, is a seriouspublic heal th concern. One in threeadults in the United States ishypertensive (Ref. 85) approximately ,58

rnil! ion adults (Ref. 2~)). Individ lIal s \tvi thhigh blood pressure have an incre<ls(~d

risk of developing stroke, heart dise(ls(~.

<lncl several types of kidney disease(Refs. 43 and 62). I-feart disease aiHIstroke are 2 of the 10 leading causes ofdeath in the United States (Ref. 43). In1nBB, 35.3 percent of all deaths wereattributable to heart disease and 7.0percent to stroke (Ref. 82).

In spite of inlprovements in thea warcness and control of hypertensionand a decline in related mortality ratesfor heart disease and stroke, "hypertension continues to be a seriouspublic health problem. Developingstrategies to lower blood pressure in thegeneral population remains an irnportantpublic health goal (Ref. 74).

2. Sodium and Salt

Sodium is an essential nutrient with avariety of physiological functions (Ref.63). It is the major electrolyte of bloodplasma and other noncellular fluid andis essential for maintenance of fluid andelectrolyte balance vvithin the body.Sodium is also necessary for normalkidney function, nerve conduction, andmuscle contraction (Ref. 7).

Sodiunl requirenlcnts vary wi th age,physical activity, environn1ental factors,and pregnancy status. Estimates havebeen- made for safe minimum dailyrequirements for sodiunl in healthypersons taking into account widevariations in climate and physicalactivity but not including an allowancefor large or prolonged sweat losses.These estinHltes range fromapproxin1ately 300 milligrams (mg) perday for ch ildren 2 through 5 years of ageto 500 nlg per day for adults over 18years of age (Ref. 63). In the UnitedStates, sodium is generally consumedwell in excess of bodily needs. Dietaryintake estin1ates range from 3,000 to6,000 Ing per day (refs. 18, 34; 35, and 43.1.

3. Relationship Betvveen Sodium andI-Iypcrtension

An association of salt intake with highblood pressure \NaS first observed in1904 (Ref. 1). Since then, considerableexpcrin1ental evidence linking sodiumintake to hypertension has accuffiula ted(Ref. 14). This increasing body ofevidence resulted in public healthconcerns about the high levels of sodiumintake in the U.S. population (Refs. 3, 9,11, 22, 43, 62. 63, and 85). Consequently,a series of recomnlenda tions for .'Americans to moderate or reduce theirsodium consumption have been made(Refs. 43, 62, 63, and 85).

Despite widely acceptedrecommenaH tions to reduce or modera tesodiunl intake, estinHlting the

60826 :Federal Register IVoL·.. 56,'No. 229 I VVednesday, November 27, 1991 I Propo8ed Rules

effectiveness of sodium restriction inreducing blood pressure has provendifficult because high blood pressurehas many causes, and blood pressurelevels are affected by many factors. The1990 amendments require FDA to revie\vand evaluate the data on sodium andhypertension to determine whetherhealth claims on this topic areappropriate.

C. Sodium: Regulatory History

Sodium qnd salt have long regulatoryhistories. Salt (sodium chloride) hasbeen regulated as an ingredient (21 CPR100.140) and a flavoring (21 eFR 101.22}.It has traditionally and historically beenregarded as a generally recognized assafe (GRAS) substance {21 CFR 182.1}.Sodium has been regulated as anessential nutrient (21 CFR 107.10, 21 CFR107.100, and current 21 CPR 101.9}.However, in the early 19801s, concernover high sodium consumption led to theGRAS safety review of sodium chloride(June 18, 1982, 47 FR 26590} and to FDAregulations Uune 18, 1982, 47 FR 26580;April 18, 1984" 49 FR 15510} to includesodium content information on nutritionlabels (current 21 CFR 101.9},to define,descriptive, terms for ulow sodium" and"'reduced sodium" foods· (current 21 CPR101.13}.and to permit sodium labelingwithout full nutrition labeling on foodsused to regula te sodium intake (21 CFR105.69).

The intent of these regulations was toprovide guidelines for sodium and saltlabeling on Coods, to establishdefinitions for descriptor terms useful inl~be1ing foods low in sodium and salt,and to encourage manufacturers toprovide a greater number and variety oflow sodium foods. The emphasis was ondeveloping and maintaining policiesappropriate for the general public sothat consumers could structure theirdiets to meet individual health needs,and so that medical professionals couldbetter manage those patients requiringcontrol of dietary sodium intake .. Twoquotes summarize the agency position in1982 to 1984. The first refers to thegeneral public:

Adult intake of sodium in the United Statesis in excess of physiological needs, and itwould be prudent for the general populationto reduce sodium intake whenever possible.The role of excess dietary sodium in thedevelopment of hypertension needs to bedefined more clearly. but there is no evidencethat a moderate reduction in sodium intakefor the general public would have anyadverse effects,- and there' is a strongindication that such a reduction would bebeneficial to a large segment of thepopulation. (47 FR 26580 at 26581.)

The second quote refers to tha tportion of the U.8. populaHonpredisposed to hypertension:

Although many epidemiological studiesindicate a relationship between sodiumintake and the prevalence of hypertensionsthe evidence that sodium consumption is amajor factor in causing hypertension is notfully conclusive~ Nevertheless, the evidenceis strong enough for most members of themedical and scientific community to concludethat a substantial portion of the U.s.population which is predisposed tohypertension would benefit from a reductionin dietary sodium. (47 FR 26580 at 26581.}

In the Federal Register of June 18, 1982(47 FR 26590), FDA reviewed the GRASstatus of sodium chloride. Regulatoryaction lovas deferred until the agencycould assess the impact of sodiumdescriptor and labeling regulations andvoluntary efforts of manufacturers toreduce the salt and sodium content oftheir products. It was recognized thatsalt occupies a unique place in the foodsupply because it occurs,naturally infoods, has 8 wide variety ofmanufacturing uses, and has a longhistory of direct consumer use in foodpreparation and at the table. In addition,the leveJof dietary sodium.recommended"for different individualsvaries widely, from severe sodiumrestriction for some hypertensivepatients, to moderate restriction forothers, to general recommendations toreduce sodium intake for the generalpublic. FDA concluded that it would beimpractical to set upper safe limits forsalt in indi.vidual foods, and that it wasmore appropriate to provide sodiumcontent information'than to try torestrict sodium use. In the yearsfollowing the sodium labeling initiatives"FDA has taken no further action on theGRAS status of salt.

Consideration of health claims for asodium and hypertension relationshipwas first proposed by FDA in areproposed rule on health messagespublished on February 13, 1990 (55 FR5176). Sodium and hypertension wasproposed as one of six possible topicsmost likely to be suitahIe for healthclainls.

Elsewhere in this issue of the FederalRegister, FDA is publishing asupplementary proposal on mandatorynutrition labelingt Reference DailyIntakes (ROrs), and Daily ReferenceValues(DRV'-s) for nutrients. Theproposed DRV for sodium is 2~400 mg.Also in this issue of the FederalRegister, FDA is proposing a revision ofnutrient content claims t.l}at includesodium content cl~irns.

D, El/idence Consjdered in Reaching theDecjsjon

'fhe agency has reviewed relevantscientific evidence on sodium andhypertension. Federal governmentdocuments considered include theSurgeon General's Report on UNutritionand I-Iealth" (Ref. 43), the u.s.Department of Agriculture's (USD.l\) andthe Department of I-Iealth and HumanServices, (DHI--IS) "Nutrition and YourHealth-Dietary Guidelines forAmericans" (Ref.. 85)t the NationalInstitute of Health (NIH), NationalHeart, Lung, and Blood Institute's(NHLBI) til'he 1988 Report of the JointNational Committee on Detection,Evaluation, and Treatment of HighBlood PressureU (Ref. 38), and the NIl-IINHLBI Hypertension Workshop (Ref.103).

The agency also reviewed additionaldocuments prepared by recognizedscientific bodies:,The National AcadenlYof Sciences/National Research Counci1~s

(NAS/NRC) "Diet and Health­Inlplications for Reducing ChronicDisease Risk n (Ref. 62)~ and theNAS/NRC "Recommended DietaryAllowances" (Ref. 63). FDA recentlycontracted with the Federation ofAmerican Societies for ExperimentalBiology (FASEB), Life Sciences ResearchOffice (LSRO) to prepare anindependent evaluation of the availablescientific evidence on the relationshipbetween sodium and hypertension. Theagency has also considered.theresul tsof this "Sodium and Hypertension"review {Ref.108}. These reportsconsidered the weight ofthe publiclyavailable scientific evidence up untiltheir pubhcation J and they provided afoundation for studies publishedsubsequently. The agency consideredthe results of animal studies to theextent that they clarified human studiesor suggested poss.ible mechanisms ofaction. FDA updated the evidence inthese documents by reviewing relevanthuman studies that have becomeavailable since 1988. The agencyevaluated one major, nlultinationalinvestigation (Ref. 37), four clinical trials(Refs. 44, 70, 79, and 109), and· threemeta-analyses {Refs. 100, 106, and 107}.

To ensure that its review of relevantevidence was complete, FDA requested,in the Federal Register of lVlarch 28, 1991(56 FR 12932), scientific data andinformation on the 10 specific topicareas identified in section 3(b)(1)(A) ofthe 1990 amendments.. The topic ofsodium and hypertension was amongthe 10 subjects on which the agencyrequested information.

Federal Register / Vol. 56, No. 229 I Wednesday, November 27, 1991 I Proposed Rules 60827

E. COl1llnents Receh'ed in flespollHo 10FDA.. Request foJ' Scienll/ic Dolo ondIn fortna lion

IiIJ!\. received 13 cornrncnls inresponse to the Federal Register req ues lfor data and information about therelationship between sodium andhypertension (56 FR 12932). Oneronlillent was a request for an extensionfor additional time for comments, andthis request was denied because of thelimited time available. Several providedconlnlents about the general process ofwri ting health claims. Others expressedopinions in support of or in opposition tosodium reduction or health claims forsodlurn and hypertension. Among thosetaking positions, a manufacturer and atrade association opposed reducingsodium intake and sodium/hypertensionhealth claims. Reduced sodium intakeand sodium/hypertension health clainlswere supported by a professional healthassociation, a distributor of healthfoods, and a foreign government.

Comnlents from a trade associationstated that health claims wereinconsistent with the statutoryrequirements of the act. However, thiscomment was contained in a letter thatwas written before the enactment of the1990 amendments which explicitlyauthorize health claims.

Comments from a State departnlent ofhealth, an association of State andterritorial public health nutritiondirectors, a trade association, and adistributor of health foods includedsupport for the 1990 amendnlents andthe Surgeon General's report (Ref. 43).The comments favored requiringsignificant scientific agreement as aprecondition to a heal th claim andsuggested that FDA should authorizesuch claims only if other nutrient levelsdo not contradict the health benefitsfrom the substance. These conlmentssaid that such claims should emphasizethe total diet rather than individualfoods, supplementation, or fortification.Some expressed concern that industrycould abuse health claims or that thegeneral public could misinterpret thenl.One suggested that FDA should do aliterature search to obtain an impartialselection of data for review. Anotheremphasized that the public shouldcontinue to rely on modern medicine forthe cure and mitiga"tion of diseases. FDAbelieves that the proposed rule isresponsive to these concerns.

Comments from a health fooddistributor and a professional healthassociation made recommendationsabout levels of daily sodium intake. Thehealth food distributor advised thatadult sodium intake should not exceed1,"600 mg per day, while the professional

h(~Cllth Hssocintion recornn1ended thatadult souiun1 intake should nol exceed 3gral11S (gl (~j,OOO mg) per day. In thisissue of the Federal Register, as statedabove, FDA is proposing a DRV of 2,400mg of sodium per day. Commentsconcerning recommended daily sodiun1intakes are more appropriatelydiscusscd in response to theestablishment of a DRV for sodium.Copies of these t\'\lO comments havebeen placed under Docket No. 90N--D134.

A distributor of health foodsrecommended a two-tiered approach toestablishing the maximum amount ofsodium that a food could contain andstill bear a health claim. It suggested anabsolute value (less than 100 mg ofsodium per 100 calories) andrecomlllcnded a secondary criteriabased on the na turally occurring sodiun1levels in the various food categories.The health food distributor emphasizedthe importance of maintaining standardlevels of other important nutrients andsuggested that sodium/hypertensionhealth claims would be misleading onlow sodium foods if other ingredients inthe food caused increased hypertension.These issues have been addressed in theproposed regula tion on generalrequirements for health claimspublished elsewhere in this issue of theFederal Register.

Comnlents from a trade associationsuggested that health claims should benational in scope and uniformnationwide, and that FDA should notproceed without the resources toadequately enforce any new regulation s.Under the 1990 amendments, regulationsestablished by FDA on health claims arenational in scope. FDA is required toprepare appropriate regulations inresponse to the congressional mandate,The agency will enforce the foodlabeling regulations to the best of itsability with the resources available.

Comments from a trade associationsuggested that model label statementsshould be created by expert advisorycommittees, evaluated throughconsumer testing, and published in theFederal Register for pLblic comment.Manufacturers will have the latitude todevelop claims that meet therequirements of the rule. FDA hastentatively decided that, under the act,the appropriate course is for the agencyto determine the requirements that ahealth claim must meet. In this and otherdocuments, FDA is proposing toauthorize health claims and is proposinga model claim. FDA is inviting publiccomm~nt on that model claim as well ason the proposed rule.

Conlments from both a State healthdepartrnent and a health food distributor

suggestcd that health claims shouldre~~ognize the populations affected, reff'rto other factors that contribute to thedi~case, and emphasize the overall dip~

and lifestyle and not overstate theeffectiveness of the nutrient or allo\tvshort descriptive statements separatefrom the total health claim. As discussedbelow and in the document on generalprinciples for health claims, FDA'sproposal is responsive to theseconcerns.

Several organizations sent inreferences for scientific studies. Allrecent and pertinent studies andcomments concerning the scientificevaluation are included in the scientificreview and sunlmary elsewhere in thisdocument.

A comment frOITI a trade associationincluded detailed objections to theSurgeon General's report (Ref. 43) andthe NAS report (Ref. 62) and suggestedthat the documents were outdated,incorrect, incomplete, and biased. TheCODlnlent concluded that the reportsshould not be given specialconsideration. FDA disagrees with thesecomments and believes that thedocuments are appropriate forconsideration.

The Canadian Government alsosubmitted a comment, outlining itsposition on the relationship of diet andnutrients to disease. The positionreflects the work of the CanadianScientific Review Committee (theCommittee) (Ref. 84). The Committeereviewed the scientific data andrecommended that the sodium contentof the Canadian diet should be reduced.1"'he report stated that there wereinsufficient data to support aquantitative recommendation. However,it concluded that a reduction in currentsodium intakes of the Canadianpopulation \vould involve no risk.Canada also pointed out that its Foodand Drug Act expressly prohibits thesale or advertisement of foodsrepresented to treat, prevent, or curehypertension and other diseases.

II. Review of the Scientific Evidence

A. Introductjon

Defini tions of hypertension arerelated to both contracting, or systolic,blood pressure (SBP) and resting, ordiastolic, blood pressure (DBP)measurements, are based oncorrelations with risk of heart diseaseand stroke, and differ by organiza tionand purpose (Refs. 4, 17, 27, and 38).Currently, individuals with SBP greatelthan or equal to 140 millimeters ofmercury (lllm Hg) or DBP grea tel' than orequal to 90 mm 1-lg or currently taking

60828 Federal Register I Vol. 56 t No. 229 I Wednesday, November 27, 1991 I Proposed Rules&&Iil ..1AWiiSiL rsa:USi:A&

antihypertensive medication areconsidered hypertensive. ·rhose withSBP less than 140 mm Hg and DBP lessthan 90 mm fig are consideredf1ormotensive (Refs. 17, 38, and 83)."High normal H DBP is defined as DBPbetween 85 and 89 mm Hg. Alldefinitions are currently under reviewby the NIH{NHLBI Joint NationalCommittee.

In considering the scien tirie evidenceon the relationship between dietarysodium intake and hypertension, FDAreviewed three Federal governmentdocuments (Refs. 38, 43, and 85), aFederal government workshop (Ref.103), and three other documents fromrecognized scientific bodies (Refs. 62, 63~

and 108). FDA also reviewed the humanstudies that have become availablesince these documents were written.The agency included in its reviewEnglish language reports of primaryhuman studies involving sodium andhypertension specifically. FDAconsidered review articles and issuesinvolving hypertension or othernutrients only as they related to theprimary relationship between sodiumand hypertension.

B. Federal Government Documents

1. "The 1988 Report of the Joint NationalCommittee on Detection, Evaluation,and Treatment of High Blood Pressure"

"The 1988 Report of the Joint NationalCommittee on Detection, Evaluation,and Treatment of High Blood Pressure"(Ref. 32.) noted that research onhypertension prevention was inprogress, and that recommendations forways to prevent hypertension could notyet be made. It concluded thatpapulaHon studies suggest that lowsodium intake, weight reduction, andmoderation of alcohol consumption maycontribute to prevention of age-relatedincreases of blood pressure. The reportnoted that Uhigh sodium intake plays acritical role in maintaining the elevatedblood pressure of some hypertensive

_patients and in limiting the effectivenessof certain antihypertensive drugs, tt andthat "some patients with mild ormoderate blood pressure elevation mayachieve control through moderatesodium restriction.'t The report observedthat there is no easy way to identifyspecific individuals who would profitfrom sodium restriction and indicatedthat moderate sodium intake(approximately 1t 500 to 2,500 mg perday) produced no serious adverseconsequences.

2. wfhe Surgeon Generalts Report onNutrition and Health,tt 1988

liThe Surgeon Generars Report onNutrition and Health" (Ref. 43) observedthat epidemiological studies have shownthat in populations with low sodiumintake, blood pressure does not rise withage, and that populations with lowblood pressure do not generallyconsume much salt. The report notedthat the correlation between salt intakeand blood pressure is not consistent inpopulation studies, and that theassociations among individuals within apopulation have been less consistent,"'Thich may be due to methodologicalreasons.

The report observed that long-termclinical studies have shown that 40percent of hypertensive patients and 30percent of mildly hypertensive patientscould control their blood pressures byreducing sodium intake below 1,150 and1,720 mg per day, respectively. It furthernoted that the effect of sodiumrestriction has been less well studied innormotensive populaHans as comparedto hypertensive populations. There arefewer studies of normotensiveindividuals, and the studies have beensmall in size and short in duration. Afew studies have indicated that dietarysodium restriction in normotensiveadults or infants can result in smallblood pressure decreases.

The report observed that interventionstudies have suggested that sodiumrestriction and weight control can bebeneficial in helping control .hypertension in mildly hypertensiveindividuals who have discontinued theirantihypertension medication.

The Surgeon General's Report onNutrition and HealthU concluded thatU[d]ietary factors that clearly contributeto high blood pressure include obesityand excessive intake of sodium andalcohol, tt and that h[sJtudies indicate arelationship between a high sodiumintake and the occurrence of high bloodpressure and stroke."

The report observed that the averagesodium consumption by U.S. adults(4,000 to 6,000 mg per day} significantlyexceeds the range that NRC estimated in1980 as would be a safe and adequatedaily intake (1,100 to 3,300 mg). It notedthat there is no easy way to identifyindividuals who would profit fromsodium restriction, and that someindividuals appear to.respond to sodiumrestriction and are considered "salt­sensitive" and others do not respondand·are considered usalt-resistant. u Thereport observed that there is nopractical way of distinguishing the t NO

groups other than by measuring theblood pressure response itself. It

concluded that modera te reduction ofdietary sodiunl would not be harnlfuland might be of significant benefit tothat portion of the population at risk ofdeveloping hypertension. The reportsuggested that most Americans shouldconsider reducing their sodium intake bychoosing foods with less sodiufn, usingless sodium in food preparation, andadding less sodium at the table.

3. "Nutrition and Your Health~DietarvGuidelines for Americans," 1990

In 1990, "Nutrition and Your Health­Dietary Guidelines for Americans" (Ref,85) made seven nutritionrecommendations for the U.S.population. Among other suggestions, itstated that Americans should "[uJse saltand sodium in moderation" andrecommended that Americans choosefoods \lvith less sodium, use less sodiulTIin food preparation, and add lesssodium at the table.

4. Summary

These three Federal governmentdocuments acknowledged a relationshipbetween sodium intake in excess ofphysiological need and the prevalenceof hypertension. There was agreementthat limiting dietary sodium may benefita portion of the population withelevated blood pressures, Le., be ofbenefit for some hypertensiveindividuals. Dietary Guidelines and theSurgeon General's report also indicatedthat in addition to benefittingindividuals identified as hypertensive,moderation of dietary sodium might alsobenefit the portion of the normotensivepopulation at risk of developinghypertension.

C. Federal Governnlent "Workshop Oil

Salt and Blood Pressure," 1989

On November 1 and 2, 1989, NHLBIsponsored a "Workshop on Salt andBlood Pressure" to review the scientificevidence on the relationship between'sodium and blood pressure, to considerthe variability in human response, toreview research findings relative toclinical and public health policies, andto provide recommendations for futureresearch (Ref. 103). Three articles thatresulted from this workshop (Refs. 109,111, and 114} are discussed elsewhere inthis document. Positions and opinionsexpressed at the meeting were highlypolarized on the value of salt restriction.A wide range of topics was presented,and the scientific discussions reflectedthe controversy surrounding this topic.Some participants at the conferencesupported reducing sodium intake andargued that the relationship isscientifically supported (Refs. 94, 97, and

Federal Register I Vol. 56, No. 229 / Wednesday, November 27, 1991 I Proposed Rules 60829itt

113), that many hypertensives are "salt­sensitive" (Ref. 95), th3 t there are nonegative consequences of decreasedsodiulll intake (Ref. gel, and that ~daC2

the target population cannot beidentified easily or cheaply (Ref. 104;, ;J

population approach, which is oftenused for nutrition policies (Ref. 99), isnecessary (Ret 104). Some indicatedthat reductions in sodium intake an~

possible because interventions ha\ (~

been successful and have madesignifica.nt contributions to treatn1e(,\~

and prevention (Ref. 98). Otherscontended that only expensive, labor·,intensive interventions with highlyITlotivated participants have beensuccessful (Ref. 105), and that the n'o~~ft.

pragmatic approach \vould be to alterthe sodium content of the entire foodsupply (Refs. 102 and 105)~

Other participants opposed reducingsodium intake and contended that more:research is necessary becauseelectrolytes other than sodium mayaffect hypertension (Ref. 110).. SOITIB

indicated that sodium restriction affectspeople in very different waYSt and thatsome individuals might be closer to acritical deficit of extracellular fluid ormight have more difficulty reconstituUnglosses after acute salt-depleting streSSgThey argued that sodium reductionshould be used only for individuals atrisk and for those in whom it has proveneffective (ReL llZ}. Some asserted thatlong-termt substantial reductions insodium intake have not beensuccessfully achieved in comp.arativ{~

trials (Ref. 113]'.

D. Other Documents and Statemellts

1. "Diet and Health-Implications forReducing (~hronicDisease Risk." 1989

The NAB "Diet and Health-ImplicaHons for Reducing ChronicDisease Rjsk" (Ref. 62) observed thatcross-culturat epidemiological studiesshow that blood pressure doe·s notincrease with age, and that there is a10\\1 prevalence of hypertension inpopulations with low sodium intake.However~ the relationship behMeen lowsodium intake and lo\v blood pressureor low incidence of hypertension hasbeen less consistent in epidemiologi.calstudIes within individual cultures. The.report noted that INTERSALT, a large t

multinationat poole.d study. showedboth a small but significant positivecorrelation between sodium excretionand mean SBP and also a significantpositive correlation betwe'en sodiumexcretion and increases in bloodpressure tha t occur with age [Ref. 3.7}~

The report obs.erved that small,. short~term clinic,al studies suggest that sodiumrestriction is related to reductions in

blood pressure in normotensiveindividuals. l·lo-vvever, these results iiavenot been confirrncd in long-term,pr'ospeclivc, controlled trials innormotensive populations.

The report noted that animal stuoie~

support the conclusions from hUluans t i.Idies. IIigh salt intake appears toprornote the development of high b!oodpressure in SOlne anirnalrllodeht,especially ,,\Then renal defects redtH.;e tIElibility of the kidney to excrete salt Th~~

E'eport noted that these hndings sugges~

that high-salt diets in cornuinatien vvi threduced sodium excretion may berelated to the deveIDOlllent ofhypertension in hum;lns" It further' no!pdtha t, once high blnod pt~':~ssure isinduced by high SO\.HUfll intake~ it C~}(i~Hyl

necessarily be reversed by resunlpUnnof a moderately low intake, dueprobably to irreversible changes in Chr'kidney.

tcDiet and Health-ImplicaHons forReducing Chronic Disease Risk~~

concluded that: "[b]lood pressure levelsHre strongly and positively correIated

, w·ith the habitual intake of salt:' andthat "the weight of evidence supportsthe contention that intake of sodium isan important factor in the occurrence ofhypertension. U The report recomn1endzdthat total daily salt intake should be notgreater than 6 g (2,400 mg sodiunl)~ ~]ith

a possible future goal of 4.5 g salt (1,800nig sodium). It suggested reducing saltand sodium intake by choosing lowsodium foods and using less sodium infood preparation and at the table. Thereport observed that there is a widevariability in genetic susceptibility tosalt-induced hypertension, that somepeople are more salt-responsive ('*salt­sensitive") than others, and that there isno reliable way to identify individuals inthe popuIalion who \vould benefit fromSOdiunl restriction. It concluded thatlimiting dietary somUIn nlay be ofsignificant benefit to that portion of thepopulation at risk of developinghypertension and noted tha t therecommended intake levels would notbe harmful to the general public.. _

2. uRecomrnended Dietary AUo\¥ances, H

1989

"Recommended Dietary AUowances"t'>loth Edition (Ref.. 53} noted that:"(slustained overconsUt-nption ofsodium, particularly as salt. has, beenrelated to development of hypertensionin sensitive individuals:t It supportedthe recommendation of the NAS Reportto limit daily sodium intake to 2,400 mg.It noted that 500 mg sodium per day is asafe minimum intake for adults" and thatthere is no known advantage inconsuming large amounts ot·sodium~

3. "Dietary Sodium Chloride and BloodPressure," 1991

Fi\SEB recently prepared anindependent evaluation of the availablecri~ntif1r au:rlonl"o nn tho raJ,) ~;nn['hinVO'-lA"-".& .... "' ....... .A'J '"" Y "'-AvA.I.,,",,""" '-J." \.11""" .. \...oJ.Lil.1l1V.l.l.~I.llP

between sodium and hypertension (Ref.108). The FASEB report concluded thatthe association behveen increasedsodium or salt intake and increasedblood pressure is due to sodium andchloride in cambination, and tha t theincrease is mitiga ted by the presence ofpotassium and calcium ions" It indicatedtna t the most convincing evidencecomes both from studies acrosspopulations and from controlled clinicaltrials which have shown a sHlaH,significant positive correlation behveendietary sodium chloride intake andblood pressure for hypertensive andnormotensive individuals.

The FASEB report noted that studi~s

"vithin populations have beeninconclusive or have shown a lowcorreIation. The report noted that there"vas little long term information' aboutthe effect of dietary sodium intake onthe development of hypertension. andthat the available data have beeninconclusive. The report concluded thatobservational data and interventiontrials document a small, but consistenteffect of dietary sodium chloride onblood pressure.

4. Summary

There' is general agreement among thethree authoritative documents that thereis a relationship between sodium intakeand hypertension.

E. Review of the Scientific EvidenceSince the Authoritative Reports

1. INTERSALT, 1988

INTERSALT (Ref.. 37) was a large t

multinational investigation of therelationship between electrolytes,including sodium, and blood pressure(Table 1). The intent was to apply highlystandardized methods across variedpopulationst to examine the majorconfounding factors t and to evaluate therelationships in individuals (Ref. 64).The study involved 10t 079 adults in 52population centers around the world(Refs. 37,50 through 54. 58, 59, and 54).Within-individual variability in sodiumexcretion was estimated using data froma random sampling (8 percent) ofindividuals who provided two 24-hoururine conections~The within-center datawere pooled, and a statisticallysignificant relationship between sodiumintake and increased SBP was reported.A relationship between sodium intakeand DBP was significant under someanalysis conditions and not others.

60830 Federal Register I Vol. 56, No. 229 / \!\Jednesday, November 27, 1991 / Proposed Rules --Siillilar results weJ'(~ found when thp,data "vere .lnalyzcd hy gender and byage [Ref.' 51), and vvhcn thenOfillotensive P()pulcllion vvasconsidel'ed: indcpenlh~ntly (Ref. 50).

The across-center data analysisconsiderecl rela tionsh ips. betweensodiurn intake and blood pressure andbetvveen soditun intake and trends inblood pressure \vith age. The data "vcreanaiyzed with and without four isolatedpopula tion cen ters, two Brazilian Indian(Yanonuln1o and Xingu), the Papua NevvGuinean, and the Kenyan. rrhese fourcenters had exceptionally low mediansodium intakes (ranging from 5 to 1,180mg per 24 hours) and the lowest averageblood pressures of all 52 centers (SBP of103 mm Hg, DBP of 63 rorn I-Ig) (Ref. 58).The relationship between sodium intakeand blood pressure, across centers, wasstrongly dependent on the inclusion orexclusion of these four popula tions.When these populations were included,the relationship between sodium intakeand blood pressure was positive andsignificant. Results were negative andsignificant or inconclusive when thesefour papula tions were excluded from theanalysis. The relationship betweensodiurn intake and trends in bloodpressure with age was positive andsignificant under all analysis conditions.The four centers with exceptionally lowsodium intakes had little or no upwardslope of blood pressure with age andlow prevalence of hypertension (5percent in Kenya, absent in remainingthree centers) (Ref. 58). The YanomanuIndians consumed as little as 1 mg ofsodium in 24 hours and appearedhealthy and physically active with noevidence of malnutrition or proteindeficiency (Ref. 59).

The INTERSALT CooperativeResearch Group analysis includedadjustments for age, sex, potassiumexcretion, body mass index, and alcoholintake. The group estimated that anaverage sodium reduction of 100millimole (mmal) per day (2,300 mgsodium) would correspond to an averagereduction in SBP and DBP of 2.2 mm Hgand 0.1 mm Hg, respectively, on apopulation basis. In addition, assuminga cumulative effect over tirne, the groupestimated the difference that this 2,300mg reduction in sodium would have onthe age-related increase in bloodpressure that is characteristic ofWestern populations. It calculated thatthe average blood pressure wouldincrease more slowly and, after 30 years(from 25 to 55 years of age), would be 9.0mm Hg (SBP) and 4.5 mm Hg (DBP)lo\ver than it would have been with adiet higher in sodium. The INTERSALTCooperative Research Group.c<?ncluded

that even these small changes in bloodpressure could result in important publicheal th beneHts vvhen applied to thepopulation as a "vhole.

In rece.ntYeart;, there have been manypubli'shed opinions on theINTERsAt/r:~findings. In reviewing the totality ofpublicly available scientific evidence,FDA also included these articles andconsidered the INTERSALT findings inthis total context. The arguments "'leresimilar to those expressed at thegovernment wotkshop discussed above.Several authors supported sodiumrestriction and emphasized thepredicted benefits on a popula tion basis(Refs. 52, 60, 69, 75, 111, and 114). 1'woauthors objected to sodium restriction,contended that it is unclear whether therelationship is nonexistent or small withnegligible benefit, and expressedconcern about potential adverse effectsof sodium restriction (Refs. 90 and 120).

The Stamlers, et a1. (Refs. 69 and ~14)used the INTERSALT data (Ref. 37) toestimate that the 2.2 mm Hg reduction inSBP would correspond to a 4 percentreduction in coronary mortality and a 6percent reduction in stroke mortality,. or12,000 fewer u.s. deaths each year forpeople in the: age range from 4'5 to 64.They estimated that the 9 mm Hgreduction in the expected increase inblood pressure from age 25 to 55 wouldcorrespond to a 16 percent reduction indeaths from coronary heart disease(CI-ID) and a 23 percent reduction indeaths from stroke. R. Stamler estimatedthat 85 percent of the Americanpopulation have some risk for mortalityassociated with blood pressure levels(Ref. 114).

2. Clinical Trials (Table 2)

Many of the studies consideredinvolved hypertensive subjects. Dustanand Kirk (Ref. 121) investigated sodiumdepletion (210 mg sodium per day) and.loading (varied by body weight, added90 mg sodium per kilogram (kg) pet day)in 31 hypertensive and 84 norrIlotensivesubjects. The authors reported that inhypertensives and some normotensives,mean arterial blood pressure fell withsodium depletion and rose with sodiumloading. In other normotensives, bloodpressure remained stable throughout.The study phase was very short (4 dayssodium depletion, and 3 days sodiumloading), and the sodium loading wasadministered intravenously whichintroduced additional uncontrolledvariability. In addition, the sodiumdepletion regime was very extreme,allowing·only 210 mg sodium per day.

Lasaridis et a1. (Ref. 55) studied theresponses of 18 (10 male, 8 female)hypertensive patients to controlled dietslow (1,150 mg per day) and high (4,600

n1g per day) in sodium. Average supineblood pressurerose significantly (6.7mrn I-ig). r'\verage standing bloodpressure rose (5.0mmllg), but theincr~~sew(,lS l)ot sigriiHcant.Thestudysize (18 suhlectsl\vassD1all. "

The Australian"Nationall-Iealth and~1cdicalRe'searchCouncil Dietary SaltStudy Managen1cnt COffiD1i ttee(Ref. 44)conducted an B-week, double blinctplacebo-control1cd intervention studywith "103 (B6 male, 17 female) D1ilcllyhypertensive subjects (DBP: 90 to 100mm I-Ig). Lower and statisticallysignificant decreases in SBP (averagedecrease of 6.1 versus 0.6 n1ffi I-Ig) andDBP (average decrease of 3.7 versus 0.9mm I-Ig) were observed in the lowsodium intake group (1,840 mg sodiun1per day) as compared to the normalsodium intake group (3,680 mg sodiumper day). A large range of variation inindividual response was observed butnot confirmed.

The Australian NationalI-Iealth andMedical Research Council Dietary SaltStudy Management Committee (Ref. 45)continued the intervention study into acrossover design. Eighty eight (73 male, 15female) subjects continued into thesecond phase of the study. 'Similardecreases in SBP (average decrease of6.0 versus 0.1 mm Fig) and DBP (averagedecrease of 4.1 versus 0.4) \'vereobserved for the low and high sodiumintake groups when the data wereanalyzed as a parallel design identicalto that of the first study (Ref. 44). Whenindividual response was considered inaccordance with the crossover design ofthis second study, the average reductionwas 3.6 m.m Hg (SBP) and 2.1 mm I-Ig(DBP) in the placebo phase (1,840 mgdietary sodium per day) versus the dietphase (1,840 mg dietary sodium plus1,840 mg sodium chloride tablets perday).

Koopman et a1. (Ref. 76) conducted anintervenHon trial in 28 mild to maderatehypertensives (average initial SBP of144.5 mm I-Ig and DBP of 95.4 mm Hg) toencourage reduced sodium diets throughdietary counseling and feedback fromresults from urinary sodium excretion.

. At the end of 18 months, the averagesodium had decreased by 510 mg per 24hours (from 3,590 to 3,080 mg),accompanying average decreases in SSPof 3.7 mm Hg and in DBP of 4.0 mm Hg.In general, over the 18 months, thesodiunl intake and blood pressuredecreased over the first 6 months andthen remained at the lower levels for therestof the trial period. Four subjectsdropped out because of high bloodpressure. 'fhis was a small· study (18subjects) with no untreated controlgroup (eG).

Federal Register I Vol. 56, No. 229 I Wednesday, November 27, "1991 I Proposed Rules 60831

In another study of mildlyhypertensive subjects, Luft et al. (Ref.79) used a placebo controlled, crossoverstudy:design to investigate sodiumeffects on blood pressure of 10 mildlyhypertensive (SBP > 140 mm Hg or DBP> 90 mmHg) and 10 normotensive (SBP< 140 mm Hg and DBP < 90 mm fIg)suhjects (10 male, 10 femal~l (10 black,10 white).' Sodium chloride (1,810 mg~odium perday) or sodium, bicarbonate(1,810 mg sodium per day) s~pplements

were supplied with a controlled basaldiet (1.380 mg sodium per day). Duringthe sodium chloride intake period, nostatistical'ly significant chaI1ge in bloodpressure was observed in either themildly hypertensive or the normotensivegroup.TheSBP of the mildlyhypertensive group was decreased by5mm Hg during the sodium bicarbonateintake period. The population size wassmall (2 groups: of 10' subjec~s).

In a:flother study involving 20 (11 male,9female) (5 black, 15 white) mildhypertensives (DBP: 90 to 110 mmHg),MacGregor eta!. (Ref. 122) investigated­blood' pressure response in a crossoverstudy involving three levels of sodiumintake determined by urinary 'excretion'£1,130 mg, 2,480mg, and 4,370 mg sodiumper 24 hours).;Blood pressure increasedstepwise with sodium intake (SBP: 147,155, ,and 163mm Hg, DBP: 91, 95, and 100rom Hg). The differences werestatistically significant and were notaffected by the order of sodium intake.

Several studies involvednormotensive, subjects. In addition to thetwo studies considered ahove (Refs. 79and 121), Mas:cioliet al. (Ref. 109)conducted a double blind, placebocontrolled, crossover study involving 48(79' percent male) (1 black, 47 white)normotensive (SBP < 150 mm Hg;DBP:80 to 89 mm,Hg; not on antihypertensivemedication or diagnosed ashypertensivel subjects, randomized intotwo g~oups, ingesting sodium capsules(2,21.0;mg sodium perday) or a placeboin addition to a low sodium diet(monitored as less than 805 mg sodiumper 8-hour overnight urine collec'tion). In65 percent of the participants, SBP washigher during the sodiulll chloride intakeperiod than during the placebo period(Group 1: 4.3 mm Hg higher, 126.4 versus122.1 mm I-Ig;Group 2: 2.8 mm Hg higher,121.4 versus 118.5 mm Hg) .. In 69 percentof the participants, DBP was higherduring the sodium chloride intake periodthan during the placebo period (Group 1:2.7 mm Hg higner, 78.8 versus 76.1 mmHg; Group 2: 1.8 mm Hg higher, 78.5versus 76.6 mm Hg). The study usedtimed, overnight, 8-hour urine excretionto as~ess adherence to low sodium diet.

Mtabaji et al. (Ref. 80) investigatedblood pressure response to salt intake in30 normotensive, black male'fanzanians. In the group on the lowsodium diet (1,200 mg per 24 hours), theaverage mean arterial blood pressuredecreased from 87 to 81 mm I-Ig,whereas, in the group on the highsodium diet (7,750 mg per 24 hours), theaverage mean arterial blood pressureincreased from 86 to 89 mm lIg. The highsodium diet phase was excessively highin sodium (7,750 mg per day).

Three studies, from Scotland (Ref. 41),Japan (Ref. 71), and Belgium '(Ref. 42),were cross. sectionaL The Scottish hearthealth studY(Ref. 41) investigated therelationship of blood pr'essure to sodiumin 7,354(3,754 male, 3,600 female) free­living subjects from 22 districtsinScotland. The study concluded tha tthere was ~ weak, positive correlationbetween sodium and SBP (males: 0.025,females: 0.055) and between sodium andD'BP (males: 0.026, females: 0.052) inboth sexes. Sodium intake was notindependently significant aftermultivariant analysis. Single sodiumrneasurements in cross sectional. s,tudiesdo not assess previous or habitualsodium intake habits.

Takemori et a1. (Ref. 71) consideredsodium intake and blood pressureresponse in 7,441 Japane~e females from88 urban (3933 subjects) and 81 'rural(3,508 subjects) ,municipalities includingall prefectures in Japan. The authorsconcluded that an increase .of 2,300 mgsodium per day 'was related to anincrease in SBP of 4.5 mm Hg (urban: 4.1mm Hg; rural: 4.9 mmHg) and to anincrease in DBP ofl.6 mm Hg(urban: 1.2mm Hg; rural: 2.0 mmHg]. Spot urineand predictive equations were used toestimate 24-hour sodium which addeduncertainty to the results.

Staessen et al. (Ref. 42) conducted a 5­year, cross sectional, intervention trialin two Belgian towns (12,000 and 9,000inhabitants). A mass media campaign toavoid salt was implemented in one ofthe two towns, and the second townreceived no information and served as acontroL Data from a random sampling of777 males and 733 females wereanalyzed. There were decreases inaverage urinary sodium, SBP, andDBPfor men in the·interventiontown, andthe trends in the control town were notsignificantly different. In women,sodium decreased in the interventiontown and increased in the control town;"wAwThereas SBP and DBP decreasedsimilarly in both towns. No conclusionsabout the relationship between sodiumintake and blood pressure could bemade. There was a large range ofvariability in the results, and no

independent assessment \vas made ofwhat information was available toinhabitants in the control town.

,Three of the studies were interventiontrials. Stamler et al. (Ref. 70) conducted,a 5-year. dietary, multiple interventiontrial involving 201 subjects with highn9rn1al blood pressure (DBP: 80 to 89m:m Hg). The intervention group (IG)w,as encouraged to reduce alcohol andsodium,intakes (goal: 1,800 mg sodiumper day or less), reduce weight, andincrease physical activity. Theintervention group significantly modifiedtheir behavior in three of these fourcategories relative to the control group(CG), increased frequent, moderatephysical activity, weight reduction, andsodium reduction (IG: drop of 25 percentfrom 3980 to 3040 mg sodium per day;CG: drop of 6 percent from 4,300 to 4,060mg sodium per day). Both groupsshowed similar reductions in alcoholconsumption. After 5 years, theincidence of hypertension (IG: 9 percent;CG: 19 percent], the average SBP (IG:decrease. of 2.6 mm Hg from 122.5 to119.8 rnm Hg; CG: decrease of 1.3 mm Hgfrom 122.7 to 121.5 mm Hg), and theaverage DBP(IG: decrease of 1.3 mm JIgfrom 82.5 to 81.2 mm Hg;CG: decrease of0.1 mmHg from 82.6 to 82.5 mm Hg]were significantly lower in the IG ascompared to theCG. After multipleregression analysis, the independenteffect of reduced sodium intake onlowering bloo'd pressure\vas notstatistically significant. Appropriatestatistical tools were used to assess theeffect; however, the. analysis wascomplicated due to the foursimultaneous interventions.

The Hypertension Prevention TrialResearch Group (Ref. 124) conducted adietary counseling interventioninvolving 841 subjects randomized intofour intervention groups and a control.The four interventions involved dietarycounseling to encourage reducedcalories, reduced sodium, reducedsodium and calories in combination, andreduced sodium and increasedpotassium. Sodium and blood pressurewere reduced in all groups, including thecontrol group. In the sodium onlyintervention group, sodium was reducedsignificantly at 6 months and marginallyat 3 years. Blood pressure was generallylower in the sodium only interventiongroup than in the control group, but thedecreases vvere not sta tisticallysignificant.

The Trials of Hvnertension Prevention(TOHP) Collabor~t-iveResearch Group(Ref. 123) investigated sevennonpharnlacological interventions(weight loss and exercise; sodiumrestriction'; stress management; and

60832 F~ederal Register I Vol. 5£», No. 229 I "W"ednesday, Novernber 27 t 1991 I Proposed Rules

supplementation with four nutrients:calcium, ruagneSiUI11, potasslunl, and fishoil) relative to a control population. in2.'182 subjects v~ith high nornlal bloodpressure (DBP: 80 to 89 "min lolg). Aftp.r 18months, there was 8 39 percentreduction in sodiurn ~n the sodh.HUrestriction population, and SBP and DBP\vere reduced by 1.5 and 0.8 mm ~Ig,

respectively. The authors concluded thatweight loss and sodiun1 restriction "werethe most promising oonpharrnacologicalinterventions.

3. Meta-Analyses

FDA evaluated five lnetawanalyses(Refs. 94, 97, 100, 106, and 107) whichanalyzed the effect of sodium intake onblood pressure (table 3). Meta-analysescombine data collected using a widevariety of nlethodologies, and thiscomplicates data analysis andassessment.

Cutler et a1. (Ref. 94) considered 23randomized clinical trials involving 1536subjects. Net reductions in sodiumranged from 370 to 3,910 mg sodium~ andaverage pooled reductions in bloodpressure were 2.9 mm Hg (SBP) and 1.6nlffi Hg (DBP). When hypertensive andnormotensive subjects were consideredseparately, the reductions were large,rfor the hypertensive subjects (SBP: 4.9mm Hg: DBP: 2.6 mm Hg) and smaller forthe normotensive subjects (SBP: 1.7 mmHg; DBB: 1.0 mm Hg). The results werestatistically significant except for DBPresul ts in normotensive subjects afterincluding adjustments for inversevariance weights.

Elliott et a1. investigated the combinedresults of 14 observational studiesinvolving 12,503 (7,099 male, 6,136fernale) subjects in 1~ populations. "oofheauthors concluded that an averagereduction of 2,300 mg sodium per daywas related to average"reductions inSBP and DBP of 3.7 and 2.0 mm Hg.respectively. Regression coefficientswere somewhat larger in women than innlen.

Three meta-analyses from orie group(Refs. 100, 106, and 107) considered therelationship of sodium intake to bloodpressure among populaHons, withinpopulations, and from clinical trials ofsalt reduction. In the analysis amongpopulations (47,000 subjects) (Ref. 106),12 economically undeveloped and 12economically developed communitieswere considered separately. The authorsdeveloped a model to analyze therelationship of blood pressure to sodiumintake. The variability in blood"pressureincreases with age was controlled byage-stratified. analysis. On- a populationbasis, the ,ana.lys·is showed small butconsistent increases in",blood pressurewith increases in sodium intake for both

econolnicaUy developed andundeveloped populations. l'hen1agni tude of the increase was grea lerfor older people and for those wi thhigher initial blood pressures. Adifference of sodium intake of 100 mmolper day (2,300 mg) was associated withan average change in SBP of 5 ronl Hg(ranging from 3 to 7 rom l-Ig) for those 15to 19 years of age and of 10 mm fig(ranging from 6 to 15 mIn fIg) for those60 to 69 years of age. The magnitude ofthe change was greatest for those withhigher initial blood pressure. Smallerchanges were observed for those with!o\\I"er initial blood pressure, but sorriCchange was observed in even the lowestblood pressure range.

'The within-population analysis of 14studies (Ref. 100) tested the modeldeveloped in the first paper. Using aconcept previously applied to 24-hourdietary recall data (Ref. 5), the analysisdemonstrated that there is considerableday-to-day variability in sodium intake,de termined tha t a single 24-hourexcretion study underestimates the truevariance, and used two studies of dailyvariation to estimate the magnitude ofthe bias. After adjustment for bias, themagnitude of the correlation betweenblood pressure and sodium excretion forthe within population data agreed withthe estimates of the correlation for theamong population data in the first paper.

The authors noted that for a smalleffect, such as the change in bloodpressure with sodium intake, very largesample sizes are required to producestatistically significant results becauseof the substantial random error inmeasuring sodium intake and the widera"nge of blood pressures associatedwith each level of sodium intake. Theauthors estimated that a study wouldneed to include 400 hypertensivesubjects and 400 normotensive subjectsto have a 50 percent probability ofdetecting such a small effect. Doublingthe sample size would increase theprobability to BOpercent. The authorsconcluded that, when estirnates of thecorrelation of sodium intake and bloodpressure are based on 24-hour dietaryintake data, the estimates of the truecorrelation are too low~ and therelationship is stronger than previouslyreported.

The third analysis included data from68 crossover trials and 10 randomizedcontrolled trials (Ref. 107). The authorsconcluded that 16wersodium intake wasassociated with reduced blood pressure :in those with high and normal initialblood-pressure levels. The authors'estimated that. in people between 50 .and 59 years ofage, "a 50mmal per day(1,150 mg] reduction in sodium intakewould lower SBP by an average of 5 mm

l-Ig in the total population and by 7 mnlI-Ig in those with initially high bloodpressures. l·hey also estimated thatthese lower blood pressure levels for av~

entire populaHon would result in a 26percent reduction in stroke and a 15percent reduction in heart disease in\Vestern populations.

Sodium intake was associated \'\/ithblood pressure. Studies of 4 ~veeks orless showed smaller differences thanstudies that lasted 5 weeks or longer.

'Taken together, the three meta­analyses concluded that the correIa Uonbetween sodiunl intake and bloodpressure is stronger than previouslyestimated, and that the INTERSALTstudy, among others, underestimated thomagnitude of the correlation. The meta..analyses supported the conclusion thatmodest sodium intake is related to lowerblood pressure on a popula tion basisand suggested a beneficial effect on anindividual basis, the magnitudedepending on the age and the existingblood pressure of the individual.

4. Summary

FDA reviewed the totality of availablehuman studies published since theauthoritative documents. One of thestudies showed a decrease in·bloodpressure ",tith increased sodiumbicarbona te intake in 10 mildlyhypertensive subjects (Ref. 79). For theother subjects in the study and for allsubjects during the sodium chlorideintake period, the results wereinconclusive. The results of the 5-yearstudy involving four simuHaneou~interventions, the results of the 5-yearintervention in two Belgian towns. andthe results of the 3-year dietarycounseling intervention were ,alsoinconclusive (Refs. 42, 70,: and 124).However. the large, multinationalINTERSALT study (10,079 subjects) (Ref.37), 11 other recent studies (Refs. 41, 44,45,55, 71, 76, 80,109, 121, 122, and 123),and 5 meta-analyses (Refs. 94, 97, 100,106, and 107) supported the relationshipbetween sodium intake and bloodpressure levels.

F. Sumlnary and Conclusions

l·here was significant scientificconsensus among the three Federalgovernment documents (Refs. 38, 43, and85), rnost of the position paperspresented at the Federal governmentworks"hop (Re"f.103), and the otherdocuments of recognized scientificbodies (Refs. 62, 63,-andl08) that highdietarysodh.lffi intake, particularly as :sodium chloride, is relaled" to theprevalence of-hyPertension, and thatdiets that are low in sodium will be

Federal Register / Vol. 56, No. 229 / Wednesday, November 27, 1991 / Proposed Rules 60833

associa teu wi th low occurrencps ofhypertension.

FDA updated the evidence in thedocuments described above byrevic\ving the totality of availablehuman studies published since thesedoc~nH~nts.One study (Ref. 79) "vasnegative in l11iIdly hypcrtensi'vc subjects v

and four studies IRcfs. 42, 70, 79~ and124) shovved no effect or "VQ;:--e

inconclusive "vith I\;Spcct to arelationship betvvcen sodiuln intake a:n.dlblood pressure. The o~hcr studies (Refs.3~41,44,45? 55, 71, 76. 8~94)9~ 10~

106, 107, '109, 121, 122, and 123)supported the conclusions r:~acrv~d Inearlier goverrlOicnt and authoritatjverevievvs which recognized a linkbehveen sodium intal<e andhypertension. Based on its review, FDi\tentatively concludes that thecontradictory or inconclusive studies areinsufficient to affect the consensusamong the governnlent documents andother reviews discussed above.

In sun1mary, the effect of changes indietary sodium. on blood pressure issmall but statistically significant.Changes in sodium intake areassociated with changes in bloodpressure across a wide ~ange ofnormotensive and hypertensive bloodpressures. Thus, reductions in sodiumintake have broad applicability. Themagnitude of the effect varies widely,with benefit for some but not for allindividuals. This variability is typical ofnutrient and chronic diseaserela tionships. The responsiveness ofsome individuals is tho~ght to be theresult of a "salt sensitivity"; however,the difficulty in identifying theseindividuals makes it impractical topredict those individuals most likely tobenefit by moderation or reduction insodium intake. There is some indicationthat different sodium salts may produce~ifferentblood-pressure responses, andthus, increasing emphasis is beingplaced on the potential importance ofthe chloride ion in combination with thesodium ion in producing blood pressureincreases. Additional research is neededin this area. However, because mostsodium in foods is in the form of sodiumchloride, this issue has little practicalimpacton public health policies.

G. Tentative Decision To Authorize aHealth Claim Relating Sodium andHypertension

FDA reviewed the publicly availablescientific data and authoritativedocuments on the association betweendietary: sodiunl. intake and hypertension.On the-basis of this review, the agencytentatively concludes. that there issignific.ant scientific agreement amongexperts who by training and experience

arc qualified to evaluate such evidenceto support health clairns that highsodiunl intake is reLiteu to thepioevalcnce cf hYPf:r~ension.The basisfor this dcc~~;!on is threefold: (1) Thestrength acd th2 scientific evidencerclaiing high sodium intakes to thepU~~lah~nce of hypertension; (2) theextent anG significance of tht~ likely

healdl: b:::nefit; and (3} the safetyeXf-,(;ctcd dieta.ry changes.

1. Sch~;nt~fic Evidence Is Suffjcient toSupport lhc~ Relationship

Proposed § 'I01.14(c) states that ahealth cl3illl HldY Le illude if theSecretury determines, "based on thetotality of publicly availab~e scientificevidence (including evidence from v\TeUdesigned studies conducted in a mannerwhich is consistent vvith generallyrecognized scientific procedures andprinciples), that there is significantagreement, among experts qualified byscientific training and experience toevaluate such claims, that the claim issupported by the evidence." Acon1panion document, publishedelsewhere in this issue of the FederalRegister, considered this requirementand is proposing this standard for healthclaims for both conventional foods anddietary supplements.

In the case of sodium andhypertension, the "totality of publiclyavailable scientific evidence" includedthree Federal government documentsderived through consensus-buildingprocesses (Refs. 38, 43, and 85), aFederal government workshop (Ref.103), three other documents prepared byrecognized scientific bodies (Refs. 62, 63,and 108), one major internationalepidemiological investigation (Ref. 37),17 clinical trials (Refs. 41, 42, 44, 45,55,70,71,76,79,80,94,97,109,121 through124), and five meta-analyses (Refs. 94,97, 100, 106, and 107).

In determining whether there was"significant scientific agreement," FDAfirst looked for consistency' in theconclusions' and recommendations ofthe relevant', Federal governmentdocuments. The agency then consideredthe contribution of the Federalgovernment workshop,.otherrecentauthoritative documents, and allpertinent human studies available since1988. In considering' the value ofparticular studies and assessing thequality of the research that 'produced thedata, FDA took into consideration therelevance of study objectives f.orexamining the relationship of sodium tohypertension, the experimental design ofthe study, the treatment of resultantdata, and the statistical significance ofthe conclusions.

In revievv"ing the recent primaryresearch, the agency looked for generalagreement or disagreement '\Ivith theconclusions and policy of the Federalgovernrnent and other cOlnprehensi'v'e,authoritative documents and evaluatedwhether inconsistencies in results fronnne~jer studies were sufficient to causethe agency to reverse or rnodify theconclusions reached in those earlierrevievv documents.

Throughou.t its evall!ation~FDl\focused primarily on hurnan sludicsbecause the public health issue ishyperten~;ionin hUJuans, and especiallyin i\mel'icans. In addition. FDI\concentrated on the relationshipbetv..Neen the I1utrier:t, sodiuill, and thedisease, hypertension. FDi\ is a \t\'arethat a wide range of variahles, inaddition to sodium intake, have beenreported to affect hypertension. Amongothers, these include chloride, calcium,and magnesium ions; chemical forms ofsodium other than sodium chloride; theratio of serum sodium to serumpotassium; alcohol consumption; andobesity. Given the severe timeconstraints and other specificrequirements of the 1990 amendments,FDA limited its evaluation of thescientific data to, the relationshipbetween "sodium" and "hypertension."The agency considered these otherissues to be peripheral, and they wereaddressed only if they related directly tointerpretation of the relationshipbetween sodium and hypertension.

In general, the Federal governmentdocuments (Refs. 38, 43, and 85), theFederal government workshop (Ref.103), and the other documents (Refs. 62,63, and 108) were in agreement thatsodium intake specified as sodiumchloride in the FASEB document) isrelated to the prevalence ofhypertension. While the effect of theaverage change in blood pressure inresponse to sodium restriction is "small"in magnitude, much larger benefit can beexpected for persons at greater riskbecause of already elevated bloodpressure .levels or because of apredisposition or sensitivity to theadverse effects of salt. Many of thedocuments noted that there is someindication that, in addition to benefitingmany hypertensive individuals, reducedscdiutn levels may reduce bloodpressures and associated risks in somenormotensive individuals as well.

In research published subsequent tothe documents described above, a fewof the human· studies showed no effect.However,·most of the studies supportedthe previous conclusions of a linkbetween sodium and hypertension.Thus, the more'recent studies were

60834•

Federal Register /Vol. 56., No.-229 'I Wednesday, November 27, 1991 I Proposed Rul~s- -generally consistent with theconclusions reached by earliergovernrnent and authoritative reviews.

FDi\ tentatively concludes that,having reviewed the relevant, pubHclyavailable, scientific evidence, there issignificant scientific agreement amongexperts quaHfied by scientific trainingand experience to evaluate clabns on arelationship between sodium andhypertension tha t such clairas aresupported by the evidence.

2. Public }-fealth Impact

"fhe prevalence of hypertension in the'u.s. population is very high, v/ith aboutone in three adults classified ashypertensive (Ref. 85). As ma.ny as 58million people in the United States haveeleva ted blood pressure (SBP equal to orgreater than 140 mm Hg and/or DBPequal to or greater than 90 mm I-Ig)(Refs. 23 and 38), and only one-quarterto one-third of these individuals havetheir blood pressure under control (Ref.74).

Uncontrolled high blood pressure is aserious public health problem because itis associated with mortality from heartdisease and stroke, which were rankedas the first and third leading causes ofdea th respectively in the United Statesin 1987 (Ref. 43). In 1988,35.3 percent ofall deaths were attributable to heartdisease and 7.0 percent to stroke (Ref.82). Individuals with uncontrolled highblood pressure have seven times the riskof developing a stroke and three to fourtimes the risk of developing CHD aspersons with normal blood pressurelevels (Ref. 74).

Though mortality risk is greatest forhypertensives. normotensives are also atrisk, and the higher the blood pressure,the greater the risk (Refs. 69 and 114). Arecent, followup, surveillance study ofthe men screened for the Multiple RiskFactor Intervention Trial (Refs. 68 and114) showed age-standardized deathrate among middle-aged (35 to 57 'yearsof age) U.S. men to be directlyproportional to SBP across all bloodpressure ranges. Not only didhypertension appear to be a risk factorfor premature death, but below averageblood pressure appeared to have abeneficial effect on survival. The deathrate among hypertensive men (SBPgreater than 160 mm Hg) was 41.7deaths per 1,000; the death rate amongthose with high normal blood pressure(SBP from 135 to 139 mm Hg) was 2045deaths per.l,ODO; and the death rateamong those with low normal bloodpressure (SBP from 115 to l19mm Hg)was 14.9 deaths per 1.000. Because thereis a continuum of risk across all bloodpressure levels. reducing blood pressure

has the potential to benefit the entirepopulaHon.

In the adult U.S. populaHon. theprevalence of hypertension varies withage, gender, and race (Refs. 27, 43, 57,and 62). ffigh blood pressure and rela tedrisks increase sharply \vith age. Less~han 1 percent of individuals under 18years of age are hypertensive, whereas23 percent of those from 45 to 64 yearsof age and 38 percent of those over 65years of age a.re hypertensive (Ref. 81).lHypertension commonly occurs in nlalesat a younger age than in females..IIowever, as people age, the prevalenceof hypertension increases more rapidly.in women and eventually surpasses thatof men (Ref. 57). The group with thehighest prevalence of hypertension isnon-Hispanic blacks, and both malesand females are at risk (Refs. 27 and 57).In those over 65 years of age, 52 percentof blacks and 37 percent of whites arehypertensive (Ref. 81).

Changes over time in mean bloodpressure and in the prevalence ofhypertension have been estimated usingdata from three large national healthsurveys. These changes were estimatedusing an earlier definition ofhypertension: SBP equal to or greaterthan 160 mm Hg and!or DBP equal to orgreater than 95 mm l-Ig and/or currentlytaking antihypertensive medication (Ref.27). Although the data from thesesurveys show that, between 1960 and1980, the prevalence of hypertensionamong black adults decreased from 34to 29 percent, this difference was notstatistically significant. There was nodecrease of hypertension a'mong ",hiteadults during the 20-year period.Average SBP decreased by 5 and 10 mmHg 1n white and black adults,respectively. The greatest improvementwas among older adults. The datasuggest a trend toward lower averageblood pressure in the U.S. populationthat has been attributed to increasedpublic awareness, diagnosis, andtreatment. The prevalence ofundiagnosed hypertension decreasedfrom 52 to 29 percent. medical treatmentof hypertension increased from 30 to 45percent. and the proportion ofindividuals with hypertension whosecondition was medically controlledin'creased from 39 to 52 percent.

Recognition of the continuum ofmortality risk across all blood pressuresprompted recent changes in the clinicaldefinition of hypertension. The currentdefiniHan identifies hypertension as SBPgreater than 140 mm Hg or.DBP greaterthan 90mm Hg or currently takingantihypertensive medication. Based onthis definition. DHHS, in its uYear 2000I-IealthObjectives for the Nation tt (Ref.

74), established a goal for reducinguncontrolled high blood pressure suchthat at least 50 percent of people withhigh blood pressure would have theirblood pressure under control, a 108percent increase. Achievement of thisgoal is expected to have a major effecton reducing the number of dea ths fromCfiD and stroke, t~vo other Year 2000objectives.

Blood pressure is regulated by acomplex process involving multiplefactors that are not well understood.Sodium intake, alcohol consumption.and obesity are considered the majordietary factors that influence thedevelopment of hypertension ingenetically susceptible individuals (Refs.38, 43. and 62). Nonpharmacologicalapproaches to controlling hypertensionhave included sodium restriction,alcohol restriction, and weight control(Ref. 29). Thirty to 60 percent ofhypertensives and 15 to 45 percent ofnormotensive individuals respond tosodium reduction and are consideredUsalt sensitive" (Ref. 116).

The most common source of dietarysodium in the U.S. food supply is sodiumchloride or common table salt. Theterms "salt" and "sodiumH havefrequently been used interchangeablyalthough salt (sodium chloride) is only39 percent sodium by weight. Additionalfood sources of sodium include sodiumbicarbonate or baking soda, bakingpowder, monosodium glutamate, sodiumnitrite. and sodium citrate. Additionalsources of sodium include drinkingwater and sodium-containing drugs (Re[16).

In addition to providing sodium toDleet nutrient needs of individuals, salthas important uses in foods. Salt isadded to a wide variety of foods toenhance and improve flavor. In picklingbrines and salted meats, salt helpsretard spoilage by inhibiting bacterialgrowth. In food processing, sodiunl saltspromote curd formation in cheeses,serve as leavening agents in chemically­leavened baked goods, control thegrowth of yeast in yeast-leavened bakedgoods, and help to solubilize muscleproteins in some processed meatproducts (Refs. 6, 7, 8, and 10). Some ofthe 'sodium used for these functions canbe reduced without unduly affecting thefinal food product (Ref. 13).

Sodium intake is a small, butsignificant risk factor for high bloodpressure. It has been estimated thatreducing sodium intake by 100 mmol perday (2,300 mg) would correspond to anaverage reduction in SBP and DBP of 2.2mm Hgand 0.1 mm Hg respectively,- on apopulation basis (Ref. 37), resulting in a4 percent reduction inCHD.mortali"

Federal Register I Vol. 56, No. 229 I Wednesday, November 27, 1991 J Proposed "Rules"...,

60835

and a 6 percent reduction in strokemortality each year (Refs. 69 and 114).Assuming a cumulative effect over time,it has been estimated that reducingsodium intake by 100 mmo} per day(2,300 mg), frarn the age of 25 to 55,~Jould correspond to a 16 percentreduction in CHD mortaH ty and a 23percent reduction in stroke rnortality(Refs. 09 and 11.4). Other estio1atessuggest that a 50 mDlol per day (1,150mg) reduction in sodium inta!(e, fronl the:age of 50 to 59, would result in a 15percent reduction in heart disease Hnd a26 percent reduction in stroke (Ref.. 10:7).,

Based on the weigh t of the evidencethat high dietary sodium intakeincreases the prevalence of high bloodpressure, several authoritative groupshave recently recommended thatAmericans reduce or modera te theirsodium intake: HNutrition and Yourliealth-Dietary Guidelines forAmericans" (Ref. 85), "The SurgeonGeneral's Report on Nutrition andI·IealthH (Ref. 43), and "Diet andHealth-Implicalions for ReducingChronic Disease Risk" {Ref. 62). "Dietand Health-Implications for ReducingChronic Disease Risk" recommendedlimiting daily salt intake to 6 g or less(2,400mg sodium). This recommendationserve-s as the basis for the proposedDRV for sodium in the proposal onmandatory nutrition labeling publishedelsewhere in this issue of the FederalRegister. Current consumption isestimated at between 3,000 and 6,000 mgsodium per day (Refs. 18, 34, 35, and 43),approximately 25 to 150 percent abovethe maximum level recommended.

It has been esti,mated that 90 percentof the sodium in foods is from added salt(75 percent added during processing andmanufacturing and 15 percent addedduring preparation and consumption).Therefore, only 10 percent of dietarysodium is attributable to the natural saltand sodium content of foods (Refs. 34and 35). Because approximately 90percent of the sodium in foods is added,reduction in sodium consumption is anachievable goal.

FDA has long recognized that sodium,s 'Brisk factor contributing to high bloodtlressure, ·and the agency first outlinedits position concerning sodium andhypertension in 1982 to 1984 in theproposed-sodium claim and labelingregulation {47 FR 26580) and in thediscussion on the GRAS status of salt(47 FR2-6590]. FDA con{;luded thatsodium consumption should be reducedin the general population becausesodium intake was in exce'Ss ofbiological requirements, that moderatesodhtm intake would have no adverseeff r,ls, and that a large portion or the

population would benefit. The agencyemphasized that the policy wasintended for the general public so thatconsumers could make informeddecisions about their diets.

FDA has monitored sodium labelingsince the first Food Label and PackageSurvey (FLAPS) in 1976 to 1978 (ReL 25).)and sodium education initiatives wereincluded as part of The National IIighBrood Pressure Education Program v

begun in 1981 by FDA and NIILBn (Re~f.

4:7J. The labeling and educationinjti~1tives resulted in more sodiurncontent labeling on foods (an increase ofnearly 60 percent betvv'een 1978 and1988) (Ref. 46), the introduction of moreproducts with 10\,\Ter sodium levels bylnanufacturers {Ref. 56), greater publica'iNareness of the relationship betweensodium and hypertension (up frofi112 to34 percent between 1979 and 1982) (Refs.,47 and 56), an increase in the number ofconsun1ers who have seen sodiumreduced products and in the number\\tho have purchased such products (Ref.55), lower sales of table salt (down by 13percent) (Ref. 47), and an increase insodium avoidance dieting (practiced byapproximately 40 percent of surveypopulation) (Ref. 78).

In conjunction with sodium content inthe nutrition label and the use of sodiumcontent claims, the sodium!hypertension health claims, descrihed inthis proposal, win provide additionalassistance to consumers inimplementing the dietary guidelines andin understanding the nalure of therelationship between sodium andhypertension. These re.gulaHons will besupplemented by extensive, educationalinitiatives. Such efforts have proveneffective in the past in encouragingresponsive actions by manufacturers(Refs. 46 and 56). in increasing conSUlnerawareness '[Refs. 47 and 56), and inaffecting consumer s purchasing habitsand behaviors (Refs. 56 and 78).

In summary, because high sodiunlintake is related to the prevalence ofhigh blood pressure, and because highblood pressure is related to increasedrisk of heart disease and stroke.•reduction and moderation in sodiumintake have the potential for having asignificant impact on the health of thegeneral u.s. population. Althoughaverage changes in salt and sodiumintake are associated with changes inaverage blood pressure that are small inmagnitude, the overall potential effecton health care costs and morbidity andmortality rates is quite significant.Persons who are sensitive to sodiumwould be expected to benefitsjgnificantly, and at the recommendedlevels, there is no apparent risk for those

who are not sensitive to sodiumReductions in sodium intake Hrc feasiblewithin current dietary patterns, boththrough potential changes in foodformulations and through the potentialfor altered consumer awareness andbehavior in food selection and indecreased use of discretionary salt

3. Safety

~1inimum average adult requireOlentsfor sodium, under condi tions ofH1aximum adaptation and without activesvveating, have been estimated to be 115Ing per day (Ref. 63). A safe rninimumintake has been estimated to be 500 mgper day (Ref. 63), more than three timesthe minimum requirements. Thisestima te takes into account widevariations in patterns of physicalactivity and climatic exposure but doesnot include an allowance for largeamounts of sodium loss from sweating.CUfrent sodium intakes in the u.s.population are thought to be 5 to 10times higher., well in excess ofphysiological needs (Refs. 18, 34, 35" and43).

TheDRV proposed for sodium (2,400mg per day) represents a 20 to 60percent reduction below currentestimates of sodium intake (3,000 to6,000 lllg per day) (Refs. 18, 34, 35, and43). The DRV is well in excess of thesafe minimum intake, and NAS hasnoted that there is no known advantagein consuming large amounts of sodium(Ref. 63J. Reductions in sodium intake, inresponse to Dietary Guidelines to usesalt and :sodium in moderation [Ref. 8S),to the proposed DRV for sodium., or tosodium/hypertension health claims inthis proposed rule. are unlikely to pose asafety risk given the large g.ap betweencurrent intakes {310OO mg to 6,000 mgsodium per day) and minimum safelevels {500mg sodium per day) and thewide margin between the proposed goal(2,400 mg sodium per day) and theminimum safe intake levels {500 mg-sodium per day}.

Sodium is naturally present in manyfoods, albeit frequently in smalla:mounts.A diet that includes a varietyof foods is likely to remain above theminimum safe intake level even withoutadditional salt or sodium being added.Moderate sodium intake, below currentconsumption levels, is a reasonablepublic health objective for the generalpopulation. This policy would benefit alarge segment of the popuAation andwoulp maintain ad-equate sodium intakefor biological functions.

Recommendations to reduce ,sodiumintake are likely to result in reducedchloride intake because sodiumchloride~ or i··salt,'" is the most commnn

60836 Federal Register I Vol. 56, No. 229 I Wednesday, November 27, 1991 .' Proposed Rules=

form of dietary sodium. Reducedchloride intake is not likely to pose asafety concern because dietary chloridedeficiencies do not occur under normalcircumstances, and the safe minimumintake of chloride was formulated jnintlywith sodium and salt minimum intakes(Ref. 63).

Ofsome concern is the loss of sodiumas salt during periods .of heavy slNeatingfrom high tempera lures or vigorousphysical activity (Ref. 30). Sodium lossescan be significant under such c.onditionsand tend to be more severe inindividuals who are not acclimated tothe ten1perature or conditioned to thelevel of activi ty (Refs. 20 and 30).

Illness can result from heatexhaustion, primarily as a result of saltdepletion, and if accompanied byunreplaced fluid losses, can lead topotentially fatal heatstroke (Ref. 30).The concerns over excessive sweatlosses led to recent experimentsinvestigating the impact of dietarysodium on the adaptation of soldiers tohigh temperatures and vigor~us exercise(Refs. 117, 118, and 119). Subjectsconsumed either 4 g salt (t,600 mgsodium) or 8 g salt (3,200 mg sodium) perday, and fluid losses were replacedfrequently. Heat acclimation was safelyachieved by all subjects, though subjectson the lo\ver salt diet reported moresynlptoms of heat illness during the firstfew days, and Johnson (Ref. 118)recommended that higher sodiumintakes may be beneficial during thefirst few days of heat acclimation.

Reports of heat exhaustion tend toinvolve isolated situations withexcessive temperatures or extremeactivity levels (Ref. 30). When makinghealth policy recommendations, FDAmust balance concerns abouthypertension, which affects one third ofthe U.S. population. against safetyconcerns under conditions of extremesweat losses. Heat acclimation wassafely achieved on the controlled, lowsalt diet (1,600 mg sodium per day). andFDA is recommending a DRV for sodium(2,400 mg per day) that is well in excessof 1,600 mg per day. FDA's policy toencourage moderation in sodium intakeprovides for a wide safety margin. It isthe agency's position that concernsabout excessive sweat losses should bepart of educational efforts aimed atgroups that experience heavy physicalexertipn and especially at those whowork with people under conditions ofhigh temperature or vigorous exercise.such as military personnel, s-portscoaches, and officials involved inexerGiseprograms in hot regions of thenation.

A few studies suggest that someindividualsmaY'respond to sodiulll

reduction with blood pressure increasesinstead of decreases (Refs. 33 and 7.2).As \tvith many physiologicalmeasurements. a· heterogeneousdistribution may be the·result ofrandom 'variatidn,' especially because the 'magnitude of the blood pre:ssurelowering effect is small. Additionalstudies are needed under controlledconditions to determine whether theseresults are significant and reproducible.

There are a few studies in whichplasma lipids were associated withincreased sodium restriction (Refs. 40, .49, and 89) and another study that \vasinconclusive (Ref. 2). The interventionperiods in these studies were very short(1 week or less). and the sodiumrestriction was extreme (460 mg and 780mg as compared with the 2,400 mg DRVrecommended by FDA). FDA believesthat these studies are so few in number,so short in duration and conductedunder such extremely restrictedconditions that they have no bearing onpublic 'health recommendations for thegeneral public.

Between 1982 and 1984, FDAconcluded that moderate sodium intakewould not have any adverse effects onthe general public (47 FR 26580). Afterreviewing the scientific evidence relatedto sodium and hypertension and thesafety issues relevant to moderatedietary sodium, FDA reaffirms thatmoderate sodium intake is unlikely topose a safety concern in the U.S.population. Recommendations tomoderate sodium intake have been partof public health policy guidelines formore than 10 years (Refs. 9, 22. and 85)with no adverse effects. There issignificant agreement among theauthoritative documents that moderatesodium intake would not be harmful(Refs. 38. 43. and 62). and seriousproblems have not been observed inpopulati.ons that traditionally consumelow amounts of salt (Ref. 69). Inaddition. the review of the scientificevidence indicates that high sodiumintakes pose a significant health risk toa large number of people (Refs. 43 and62).

FDA welcomes any additionalinformation or data on the safety ofsodium and salt intake and will continueto monitor the safety implications of allpublic policy recommendations.

III. Provisional Requirements for HealthClaims

A. RelationshipFDA is proposing in § 101.74 to

authorize health claims on therelationship of dietary sodium andhypertension on food labels andlabeling. The· agency .has identified

several key points that it considersessential for helping consumers tounderstand this relationship. Thesepoints are made in § 101.74(a).

The definition of hypertension used- in'.§ 101.74(a) is taken from U.S. DHHS! .PHS/NIH reports (Refs'.! 23. and 38).: Itdefines hypertension as SBP ofmorethan 140 m;m HG or DBP of more than 90mm HG. The regulation alsodistinguishes sodium from salt.

Proposed § 101.74(a) describes therelationship between sodiuIll andhypertension. Based on its review of theavailable scientific evidence, FDA statesthat high sodium intake is related to theprevalence of hypertension and to theincrease of blood pressure with age. Theagency also states that low sodiumintake is related to low prevalence ofhypertension and to a low rise or noincrease of blood pressure with age.

A substantial amount of human andanimal data indicate thathigh potassiumintake may be related to reduced bloodpressure levels (Refs. 38, 43, and 62). Inaddition, high sodium-potassium ratioshave been positively correlated withblood pressure levels (Refs. 43 and 62),and NAS (Ref. 62) noted that lowsodium intake in combination with highpotassium intake "is associated with thelowest blood pressure levels and thelowest frequency of stroke in individualJand populations." FDA consideredincluding potassium intake informationin sodium/hypertension health claims.However, because of time and resourceconstraints, the lack of evidence for aquantitative ratio. and safety concernsinvolving potassium supplementationand fortification (21 CFR·201.306), FDAat this time has limited the relationshipstatement to sodium and hypertension.This is the topic that FDA was directedto address in section 3(b)(1)(A)(vi) of the1990 amendments.

B. Significance

In summarizing the significance ofreductions and moderation in sodiumintake relative to the reduction in theprevalence of hypertension in thegeneral U.S. population and \vithin thetotal dietary context, FDA has identifiet..in proposed § 101.74(b) several keypoints that it considers .essential forhelping consumers in understanding thisnutrient and disease relationship.

This section states that hypertensionis a public health concern because it is arisk factor for CHD and stroke. Thisstatement is based on the SurgeonGeneral's Report (Ref. 43) and the NASReport (Ref. 62). The recogni tion thatthere is a continuum of risk across therange of blood pressures, which isreflected in this provision,. wa.s

Fed~ral Register / Vol. 56, No. 229 I Wednesday, November 27, 1991 / Proposed Rules', 60837

d0cumented in the followup surveillancestudy of the men screened for theMultiple Risk Factor Intervention Trial(Refs. 68and 114). The agency hasincluded a statement from DietaryGuidelines on the prevalence of highblood pressure in the United States inthis section to provide some indicationorthe magnitude of the problem and thenumber of Americans currently affected(Ref. 85).

Based on FDA's evaluation of thesci.entific evidence, proposed § 101.74(b)goes on to state that reduced sodiunlintake may benefit some but not allhypertE~ris1vesand possibly some but notall normotensives. The range 0'£percentages in § 101.74(bJof responsivehypertensive and normotensiv~

Individuals that respond to sodiumreduction was taken from the Sullivanreview (Ref. 116). The regulationrecognizes, however, based on theSurgeon General's report (Ref. 43). theNAS report (Ref. 62), and "DietaryGuidelines" (Ref. 85) that there are nopractical biological markers foridentifying responsive individuals.

The regulation goes on to list thepopulations most at risk forhypertension and most likely to benefitfrom sodium reduction. Thesepopulations were identified in theSUfgeon General's report (Ref.

i43), the

NAS report (Ref. 62), and the NationalHe:alth and Nutrition ExaminationSurvey (Ref. 27). It then lists the riskfactors for hypertension other thansodium intake. These factors arementioned in the Surgeon General'sreport (Ref. 43) and the NAS report (Ref.62). The statement that the magnitude ofthe effect is "small" but statisticallysignificant is based on FDA's evaluationof the scientific evidence, which issummarized in section II. of thisdocument. Proposed § 101.74(b) goes onto cite theestimated magnitude of thechange in blood pressure in response toa change in dietary sodium intake. Theagency took this information from theconclusions of the INTERSALT study(Ref. 37)~ The estimated reductions inmortaH ty cited in proposed § 101.74(b)were taken from the Stamler's analysisof the impact that the change in bloodpressure would have on a population­wide basis (Refs. 69 and 114). Thissection concludes withrecommendations for ways to reducesodium intake, which were taken from"Dietary Guidelines" (Ref. 85), theSurgeon General'sreport (Ref. 43), andthe NAS report (Ref. 62).

In discussing the magnitude of theeffect of a change in sodium intake, theagency uses the words "estimate" and"approximate" to indicate that the

values cited are based on the bestinformation available and are close tobut not identical to the actual and truevalues. FDA would consider changingthese estimates only if newer estimatesthat were based on better data and thatwere significantly different from thesevalues were presented to it.

C. General Requirenlenls

In § 101.74(c)(1), FDi\ is requiring thatfor a food to bear a health rola im on thetopic of sodium and hypertension, itmust meet the general requirernents forheal th clain1s set forth in proposed§ 101.14, published elsewhere in thisissue of the Federal Register. Under thisregulation, a sodium/hypertension I

health claim is prohibited if any' of :thespecified disqualifying nutrient levelsare exceeded. Thi,s requirement assuresthat sodium/hypertension health clainlsmay not. appear on foods and. foodproducts that contain 11.5 g or more offat per reference amount cOlnnlonJyconsumed~ per label serving size, or per100 g, 4 g or more of saturated fat perreference amount commonly consumed,per label serving size, or per 100 g, and45 mg or more of cholesterol perreference amount COTIIIDonly consumed,per label serving size, or per 100 g. Thereare also disqualifying criteria forsodium: 360 mg or more of sodium perreference amount commonly consumed,per label serving size, or ,per 100 g. :However, to qualify to make a sodIum!hypertension health claim under :proposed § 101.74(c)(2), it must contain140mg or less of sodium per serving andper 100 g. A more'thorough discussion ofthe criteria for identifying risk nutrientsand the levels of these nutrients allowedin foods that bear health claims isincluded in the document on generalrequirements for health claims,published elsewhere in this issue of theFederal Register.

The requirement that a food mustmeet the "low sodium" definition tobear a sodium/hypertension healthclaim assures that such claims willappear only on foods and food productsthat contain 140 mg or less of sodium perserving and per 100 g.. A more thorough'discussion of the. "low sodium" criteriaand the rationale for the establishedsodium content levels is presented in theadjectival descriptor document,published elsewhere in this issue of theFederal Register. Should additionalconsiderations or evidence prompt theestablishment of a different definitionfor "low sodium," only the descriptordocument will require revision.

FDA used the qualifying criteria andthe disqualifying criteria, describedabove, to identify foods that wouldlikeiy be allowed to bear sodium/

hypertension health claims (Ref. 93).Examples of foods qualifying forsodium/hypertension health clainlsinclude tuna and salmon without addedsalt: most fruits and vegetables. exceptfor canned and frozen vegetablesprocessed with salt; lowfa tmilk (2percent or less,fat], evaporated Inilk,lowfat yogurt with fruit cottage cheese,ice n1ilk, sherbet~ and nondairy desserttoppings and creanl substitutes; mostflours, nleals, grains, and pastas (exceptfor egg pastas): and breakfast cerealssuch as shredded ''\Theat, low sodiun1corn flakes, frosted shredded (mini­sized) wheat, puffed rice, sugar crisp,wheat·germ~ .and many prepared cerealssuch as cream of wheat. cream of rice,and grits. In addition to these types offoods, several other food types wouldqualify for sodium/hypertension healthclaini.s including beverages such ascarbonated soft drinks, coffee~ tea, somefruit juices, drinks. and punches; sonH~

candies.. cookies, baked goods, andicings; jams, jellies, and othersweeteners; and margarines and saladdressings without added salt. ·Given, theminimal nutrition value of many of thesefoods, FDA requests comments as towhether they should be allowed to beara health claim.

D. Relationship Statement

In the companion document ongeneral principles for health claimspublished elsewhere in this issue of theFederal 'Register, FDA is proposing torequire thatclaims present an accuraterepresentation of the nutrient/ diseaserelationship. Consequently, based on thescientific evidence regarding therelationship between sodium andhypertension, in § 101.74(c)(2], FDA isproposing that sodium/hypertensionhealth claims must state that a lowsodium diet is associated with !olverblood pressure in some people, or that ahigh sodium diet is associated withhigher blood pressure in some peopie.Because sodium reduction helps lowerblood pressure in some but nut allindividuals, FDA is proposing thathealth claims acknowledge this fact. It isthe agency's position that, without suchan acknowledgelnent, the health clainlwould be Dlisleading· to those peoplewhose blood pressures do not respondto sodium reduction.

E. Populations at Greatest Risk andDietary Risk Factors

In § 101.74(c) (3), FDA is proposing torequire that health claims acknowledgethat many factorsare associated withthe development of high blood pressure.Thus, under this proposal, claims will berequired to identify high riskp'opulations

60838 .~ederal Register I VoL 56, No. 229 I Wednesday~ Noverrlber 27~ 1991 I ,Proposed Rules'W

and dietary risk factors associated \vithhypertension. Those most at risk ofdeveloping hypertension, andconsequently most likely to benefit fromsodium restriction, include the elderlyand those with family histories of highblood pressure, which may encompassindividuals in specific ra.cial or gendergroups (Refs. 27, 43, 57, and 62). Inaddition to dietary sodium intake,alcohol consumption and obesity areidentified, modifiable, dietary riskfactors for hypertension (Refs. 43 and02,). Consequently, achieving weightcontrol and reducing alcoholconsumption have been recommendedto assist in lowering blood pressurelevels in the general population (Ref. 85).T'his additional information onpopulations and risk factors provides abroader context for the nutrient/diseaserelationship. Presentation of thisinformation will ensure that consumersare aware that, in addition to sodiumintake, there are many other factors thatcontribute to the development andcontrol of hypertension.

IV. Optional Health Claim Information

A. Sodium as an Essential Nutrient

Sodium is an essential nutrient, and itis important that consumers includesorliurn in their total diets. On the otherhand, NAS has recommended a safe,minimum level of 500 mg sodium per day(Ref. 63) and an upper limit of 2,400 mgsodium per day (Ref. 62). Elsewhere, inthis issue of the Federal Register, FDA isproposing to establish a DRV for sodiumof 2,400 mg per day for use in nutritionlabeling. Yet, while some sodium isrequired for good health, excessiveintake of sodium is unnecessary andmay be harmful. For consumers tounderstand the significance of thesodium contained in a food that isqualified to bear a sodium/hypertensionhealth claim in relation to the total dailyintake goal, FDA considered requiringthat sodium/hypertension health claimsstate that adults should consume atleast 500 mg but not more than 2,400 mgsodium per day. However, in an attemptto keep health claims short and notoverwhelm consumers with information,FDA is tentatively proposing in§ 101.74(dJ (1) to allow, but not torequire, quantitative limits for sodiumintake. The agency requests commentson whether this additional informationwill be beneficial to consumers, andwhether it should be required on healthclaims or remain optional.

B. qonsultation ofPhysicians, Many ,people are now aware of the

dangers :of high hI,ood pressure (Ref. 5,6).With the ready: ayallapHity; qf ,"dQit

yourself~machines to measure bloodpressure levels in grocery stores andshopping malls and the commonpractice of having blood pressure levelschecked each time an individual visits aphysician or health professional, manypeople now know what their bloodpressure levels are. FDA is concernedthat some individuals may attempt touse the ready-availability of sodiumlabeling, and in particular sodium/hypertension health claims, to self­medicate or treat their hypertensionwithout consulting a physician. For thisreason, the agency considered requiringthat health claims state that individualswith high blood pressure should consulttheir physician for specific medicaladvice and guidance.

Health claims that result from thisregulation are intended for the generalhealthy public, however. Hypertensionis a serious medical condition. It isFDA's view that any individual with anidentified medical problem should beunder the care of a physician, and thathealth claims are not intended as asubstitute for individu,al patient!doctorcare and especially not for individualswith identified medical diseases orhealth-related conditions. The agencyhas tentatively decided to include thisinformation as anoptiorial element,§ 101.74(d) (2), and requests comments.

C.' Sodium and Salt

FDA is proposing in § 101.74(d) (3), toallow manufacturers to use the termHsalt"'in addition to the term "sodium,"both of which have been incorporatedinto "Dietary Guidelines" to use salt andsodium in moderation. Salt, which is 39percent sodium by weight, is the mostcommon source of dietary sodium and isa more familiar term to the generalpublic than sodium. A recent FDAsurvey found that approximately 70percent of the survey populationgenerally understood that sodium andsalt are related (Ref. 1982). Respondentsfrequently used "sodium" and "salt"interchangeably, which is technicallyincorrect but functionally effectivebecause reducing salt intake alsoreduces sodium intake. The availableevidence suggests, however, that sodiumis the nutrient most clearly implicated inhypertension. Furthermore, in theproposed nutrition labeling documentpublished elsewhere in this issue of theFederal Register, FDA is proposing touse the term "sodium" in the nutritionlabel to inform consumers of the sodiumcontent of a food. Therefore, allowinguse of the· term "salt" in a sodiuni/hypertension health claim, rather thanthe term "sodium," would be potentiallyconfusing to consumers becaus~ itW~)\~Id p~ inco~sistentboth\yUh the :

nutrition label and with the strongestscientit1c evidence foI' linking dietaryfactors to hypertension. Conversely,using the term "salt" in addition to theterm "sodium" would seem less likely tobe misleading and may actually he "useful to those consumers who areunfamiliar wi th the more technica I ternlbut who wish to reduce their sodiumintake. Therefore, the agency isproposing to allow the use of the term"salt" if the term 6lsodium" is a.lso used.

e-fhe agency is aware that a few recentstudies and reviews suggest that thechloride ion, rather than or in additionto the sodium ion, may be important inthe development of high blood pressure(Refs. 31, 48, 79, 87, and 92). Earlystudies "'\Tith sodium chloride attributedblood pressure increases to the chlorideion; ho\vever, in the 1950's the sodiumion was considered to be moreimportant (Refs. 14 and 43). Becausemany of the studies that investigated therelationship between sodium andhypertension used sodium chloride asthe source of dietary sodium, thesestudies do not distinguish the effects ofsodium from the effects of sodiumchloride.

In the early and mid-1980's, studieswith various sodium salts found thatwhile sodium chloride raised bloodpressure levels in sensitive individualsand animals, other sodium salts had noeffect (Refs. 31 and 43). The recentstudies (Refs. 79 and 87) areinconclusive with respect to chloride orindicate tha t the sodium and chlorideions have different roles. Recent reviews(Refs. 48 and 92) suggest that sodiumand chloride together produce largerblood pressure changes, and that the ionthat is associated with the sodium maygreatly influence the subsequent bloodpressure response. To date, the studiesip.volving humans have been few innumber and small in size. Consequently,at this time, there is insufficientinformation available for drawingsubstantive conclusions or for changingpublic health policy recommendations.Nonetheless, these results raiseimportant questions, and FDAencourages additional research todetermine the independent andcombined effect of sodium and ofchloride on blood pressure.

Sodium chloride is the major source ofdietary sodium. Because FDA's policy ofencouraging sodium reduction will ,alsoresult in. chloride reduction,. the policyremains prudent regardless ofwhethersodium,chloride, or so.dium~hlorideisdetermined to he important in' ,

, relationship to, hypertension. In, addition,compliance .issifIlplified bec;Ruse sodium~on~ent is}den~ified.on the la:beliI)g.~d

Federal Register / Vol. 56, No. 229 / Wednesday, November 27, 1991 / Proposed Rules 60839~........ ZRWa.<"!77777?III ti&c~~,~l1Il

verification would involvf~ detection ofsodiurn alone. The agency requests dataand comrrH~nis on the lJ ppropria teness ofselecting sodhF~n r~!ther than sCidiuHlchloride as the spectfied nutrient {=xnd onthe appropriatcnpss of allovving th(:~ t(~rnl\

"881f' in (.~dditi['n to the tc~rrn "r-;od,;urn"on sodium/hypertcntion hCt.dth ch;:~;~,1S"

nWln~jlla,ctucr2!rs to usc the tc~rr;]

i:1 addi Uon to [he tc·'.~n1

blood pre:Jsurc,," is 8

broader tc;nn V'l1hicb encomp[~sses

perscns VJith un trea ted high bleedr,,"i~~"C"llil'"'D levels 3S vie!! as person'] 'ilvHh--.".."~,,,, ..·.-.,,,,!I·· levels as a result of effectivetrentrnent Hy'pertension is also thedisease specified in section3(b)(1)(A)(vi) of the 1990 amendments.The tenn Hhigh blood pressure9? meansnearly thf; same thing as hypertensAonDad individuals with controlled highnJood pressure are frequentlyconsidered to have high blood pressure8\/en though, technicallY9 the bloodpressure levels are in the nornlal range.'I'he ternl "high blood pressureH is lesstechnical and more familiar toconsumers since blood pressuremeasurement is included in routinephysical examinations, and bloodpressure response is used to monitor thetreatment of hypertension. Becausesimple, uncomplica ted terminology isuseful for assuring that health claims areclear and understandable to consumers,FDA is proposing to require the use of ,the term "high blood pressure" and toallow for the optional addition of theterm "hypertension." The agencyrequests comments on theappropriateness of this proposed usage.

E. Additionallnformation

In § 101.74(d)(5), FDA is proposing toallow manufacturers to develop sodium/hypertension health claims that providefactual information about hypertension,including information contained in the"R~lationship"and HSignificance"statements included ias part of theregulation and estimates of the numberof people in the United States who areaffected with high blood pressure orhypertension. It is FDA's policy that oneof the purposes of health claims is toinform and educate the general public.Consequently, manufacturers should beallowed to include accurate, factualinformation in their health claims aboutthe prevalence and seriousness ofhypertension for the U.s. population.FDA is proposing to limit the additionalinformation allowed to that contained inthese statements because they arebased on FDA's revIew of the scientific

evidence concerning sodium andhypertension. By using an approxinlalec:stinll! te of prcv() lrnce" such as "one inthree" ndul~s, updc:ding ihj~) cstirnat(~ islikely to be !css of a problcn1 than if aIDO"[': prcci~;e c::tin1atc 1JJC~,(~ u~;cd.

rr~.t\ is ,~~ ,'" '''C',' -,.'~

§ 101.7~1[c) ;~~ ciairo. onsudluHl t'-ind The Llgcccy is,,--c,-a.," ,r,n: lhi3 rnodel t·,) (-l<~,si;;t

rn:1I1!.J.[acfurcr.s j:n ",""r""',,,

apprGpriat~~ cla~cl.

~l. Environm\;ctal [fi:.pgct

The age:tl,cy has detenninl~d und.::~r 21CFR 25,21(~:l) (11) that this nction is of atype that does not individually orcumulatively have a significant effect onthe t~um2n eflvironnlent. Therefore,neither an environrnental asscssn1.entnor an enviranrncntal hnpact statenlcntis required.

VI. Effe:L-:tive Date

FDA is proposing to make theseregulations effective 6 months after thepublication of a final rule based on thisproposal.

VII. Comments

Interested persons may, on or b~foreFebruary 25, 1992, submit to the DocketsManagement Branch (address above)written comment$ regarding thisproposal. Two copies of any commentsare to be submitted, except thatindividuals may submit one copy.Comments are to be identified with thedocket number found in brackets in theheading of this document. Receivedcomments may be seen in the officeabove between 9 a.m. and4 p.m.,Monday through Friday.

. VIII. Economic Impact

The food labeling reform initiative,taken as a whole, will have associatedcosts in excess of the $1"00 millionthreshold that defines a major rule.Therefore, in accordance with ExecutiveOrder 12291 and the RegulatoryFlexibility Act (Pub. L.'96-354), FDA hasdeveloped one comprehensiveregulatory impact analysis· (RIA) thatpresents the costs and benefits of all ofthe food labeling provisions taken 'together. The RIA is published,elsewhere in this issue ofthe FederalRegister. The agency requests commentson the RIA.

Appendix to the Pream~le~Con~umerHealth. Message Summc;lry~Sodium"andHigh Blood Pressure I

The following Appendix is; a proposed:consumer summary on sodium andhypertension. FDA solicits COffirnentson

this docurnent as explained in th(~

proposal on the general requircinpnts forhealth clahns published elsevvhere; inth:5 iSSU8 of th2 Federal Register.

Appendix~CoEsi,~r~H::rSUmnl,H"Y onSodhnn and fIigh Biood Prcs~,un=;

n1."'J-,"tnClr,,"~~' of the rcccn!

r~utrition e;,nd Educ:.\ ~inn ft.. ct :Jf

1990~ manufacturers cle;!rinformation on the about tLerr'"'llttr',1:-\,~·h>.." betwe~:i1 a nutrjent, such assodiulTI, a disease or health-·ndatedcondition, such DS hypertensiun. To

consumers frenA rnjded,aHovvs truthfuI

statements about diet and healthrelaHonships that are firmly supportedby the current scientific evidence. Thereis agreelnent that the scientific evidenceis strong enough to allo\v health claimsabout the relationship bet\veen sodiumin the diet and hypertension.

lvtany consumers have said thathealth claims on food labels could beuseful to them in n1aking improvementsin their diets. Ho\vever, label space isoften limited. Therefore, the labelstatement may refer to an attachedpcunphlet, or other adjacent labeling thatprovides additional information aboutthe health claims that appear on thelabel of the food product itself.

In addition to allowing heal th claimsabout the relationship het\AJeen sodiumand hypertension, FDA is allowinghealth claims about the relationshipbetween calcium and osteoporosis,saturated fat and cholesterol andcardiovascular disease, and fat 'andcancer. For information about theseother dietand health relationships, writeto: (to be supplied by manufacturer).

What is Hypertension?

Hypertension means high bloodpressure, a condition in which yourblood pressure goes up and stays abovea normal level. Blood pressure measuresthe force of blood again,st the artery\valls as the heart pumps blood throughthe'body. "

When you get your blood. pressurechecked,: you, are given two nUmbers.The firstnumber (systolic pre~;sure) isthe force of 'blood again,st th~ arterywalls when the heart beats. The secondnumber (diastolic pressure} is the forceon the artery ·walls when the heartrelaxes between beats. Currently,:people with systolic blood pressure of140 or :mor~ millimeters: of merc~ry (rnfl1,Hg) and/0rdiastolic blo:odpr~ss~r~of90 or Dlore ~mm Hg ,are Qo:n~idered ~o,

have high blood pressure.

60840 :Federal Register /·VoL 56, No. 229 IWednesdaYe November 27~ 199'1 I Proposed Rules

·\tVhy is ~rhere Concern Aboutl-Iypertension?

In the lJnited States, about one in. three adults has high blood pressure."{'he disease affects approximately 58.rniHion people and is a public healthconcern primarily because it is a majorrisk factor for death from coronary hear'tdisease and stroke. Risk of deathincreases steadily as blood pressureincreases. People with high bloodpressure levels are at greatest risk, andthe lower the blood pressure the lowerthe risk.

IIypertension occurs nlore frequentlyarnong persons with a family his tory ofhigh blood pressure, elderly men andwomen of all races, black men and~Nomen, and men a t an earlier age than·women. In the U.S., hypertension and its.rela ted risks increase with age. Lessthan 1 percent of people below age 18,about 23 percent of people between ages45 through 64, and about 38 percent ofpeople over 65 have hypertension.

Primarily because of increased publicu\vareness and treatment of the disease,hypertension has decreased somewhatin the U.S. population in recent years:nevertheless it remains a serious publicheal th concern.

What Is the Cause of I-Iypertension?

In most people with high bloodpressure the cause is unknown.Regulation of blood pressure by thebody is a complex process that is notconlpletely understood. Probably avariety of factors influence thedevelopment of hypertension in peoplewhose heredity makes them susceptibleto the disease.

Currently. scientists generally agreethat three major diet-related factorshave an effect on blood pressure­obesity or being overweight, excessivesodium in the diet. and excessivealcohol consumption.

The terms "salt" and Hsodium H oftenare used interchangeably, although salt(which is sodium chloride) is only partsodium. Salt is our most common sourceof dietary sodium.

Studies of populations around theworid provide the primary basis forassociating dietary sodium withhypertension. In populations that havediets low in sodium. high blood pressure ,is less common than in populations withdiets high in sodium. Scientists believethat dietary sodium is related tohypertension, and that diets which are~ower in sodiumwHl be asso.ciated withlower frequency of hypertension.

These studies also indicate, that inpopulations with diets low in sodium.blood pressure increases le,Ss rapidly ordoes not increase at aU with age. This

contras ts sharply wi th the bloodpressure increases with age that areseen in the U.S. Less salt in the diet maybe particularly appropriate for peoplowho are at increased risk for developinghypertension in later life, such as blacksand those with either a fanlily history ofhigh blood pressure or current highnormal blood pressure levels. The bloodpressure of some-but not all-peoplewill be lowered by decreasing dietarysodium. Persons whose blood pressureis decreased by lo\vering sodium areconsidered "salt sensitive." There is nopractical ·way to identify the "salt..sensitive" people in the population, topredict who might develop high bloodpressure, or to determine who 1Nillbenefit from reducing dietary sodium.Authorities currently recommend thatmost people use salt and sodium only inmoderation. Reduction in sodium willbenefi t those people whose bloodpressure rises with high salt intake. Noharmful effect is known to occur frommoderately reducing dietary sodium6

Do Most People Eat Too Much Salt andSodium?

Sodium is an essential nutrient that isrequired by the body. The NationalAcademy of Sciences has set a minimumsafe amount for adults of 500 milligrams(mg) per day under normal tempera tureand activity conditions. People who losea lot of sodium and water through sweatneed to drink extra water and in rarecases replace the lost salt. The Academyhas stated that there is no knovvnadvantage in consuming large amountsof sodium in excess of body needs. MostAmericans consume several times theminimum amount of sodium needede

The U.S. Public Health Service has seta national health goal for the public touse salt and sodium in moderation. Todo this, people are encouraged toprepare foods without adding salt1 toavoid salt at the table, and to make ahabit of purchasing foods that are low insodium or modified to lower sodiumcontent.

Vllhich Foods Are Sources of Sodium?

Sodium in the diet comes from manysources. Small amounts of sodium arefound naturally in many foods, so if youeat a variety of foods, you'll easily getthe minimum safe amount.

However, your salt intake canincrease dramatically depending on thechoices you make. Salt is added forflavoring and preserving duringprocessing of many foods, but productsare often available in a ulow sodium~'

version as well. Salt may also be addedduring cooking at home, or by yourselfat the table. In addition to table salt,manysubstan~sadded to foods, such

as baking soda, baking powder, sodiumnitrite~ and monosodium glutaolate(l\1SG), contain sodium.

A good way to learn about the amountof sodium in foods is to read nutritionlabels. ~1ost foods now have nutritionnnforrnation on their labels. The amount«Jf sodium in a serving of food is listed inmilligrams. FL1A has established HDaily""'Ii al-nes;· for several nutrients. including8odiuIIl, that are ilfnportant in diet andhealth relationships. l'he daily value isintended to help cc~nsurnersdett:C'nllnehovv a single serving of a foodcontributes to the total a'l,mount ofnutrient for the day. The daily value forsodi urn is 2.400 lng, based on a reportfroITt the National Acaderny of Science.'I"hereforeJ a food that contains 600 mgsod.iurn per serving would provide aboutone-quarter of the daily recom:mendedvalue for sodinol. WhE~n you add up thesodium frunl all the foods you ea t in aday, it shou.ld total less than 2.400 mg.

"\That Do Label Cla:hns }\hout Sodium.rvlean?

In addition to the a!TIount of sodiulnper serving on the nutrition label, youmay see other kinds of clairns aboutsodium on sonle food packages. 1"hereare t\VO kinds of label clainls-nutrientcontent claims and health claims.

Nutrient content claims nlay be :nadeabout the anlount of sodium the foodcontains. For example, a food thatcontains 35 mg sodium or less perserving may be labeled u very lowsodium." Foods that contain 5 mg or lessof sodium per serving may be labeledusodium free" or "no sodium," and foodsthat contain 140 mg sodium or less perserving may be labeled "low sodium." Areduced sodium claim on a food labelindicates that the sodium content hasbeen reduced by 50 percent or morecompared to the regular product.

Some foods that are low in sodiummay contain one or more nutrients thatmay increase the risk of a diet-relateddisease other than high blood pressure~

For example, a low sodium food couldbe high in saturated fat which has arela tionship to elevated bloodcholesterol and heart disease. A contentclaim about sodium cannot be made onsuch foods without indicating thepresence of the other nutrient. forexample, "Low, sodium: see nutritionlabel for saturated fat content:·

Health claims are those made aboutthe relationship between the nutrient.sodium, and the disease, hypertension~Health claims of this type may appearonly on foods tha t qualify as "lowsodium.H In addition. the food must notcontain any othernutrient that fTIA hasdetermined increases the risk of a diet..

II.. iOiIt ad

Federal Register I VoL 56, No. 229 I Wednesday, Novernbel' 27, 1991 I Proposed Rules 60841

rela.ted disease or health condition otherthan hypertensIon. For example. ahealth c}v.in1 could not appear on a Hlowsodiunl91 foou that contains a highHlnount of saturated fat, becausesaturated fat has H relationship to heartdisease.

tvlany foods are eligible to makesodium and hypertension claims. Forexample, at least some products in eachof the foBovving categories of foods canrnake such clairns: Fruits and vegetables;fruit juices and drinks; milk and dairyproducts; breakfas t cereals; cerealgrains (such as rice): pasta products[such as spaghetti); flours; legumes (peasHnd beans): nuts and seeds: andseafood.

Other Diet-Related R1Sk F~H:torR forIIypertension

In addition to sodium. there are atleast two other diet-related factors forhypertension over which a person hascontrol-body weight and alcoholconsumption. Increased body weight isrela ted to increased blood pressure, andblood pressure falls when weight isreduced. Weight loss is recommendedfor all overweight persons, particularlythose with hypertension. People whoregularly consume large amounts ofalcohol have higher blood pressure thanpeople who don't drink or who drinkonly in moderation. Authoritiesrecommend maintaining a healthy\veight and drinking alcoholic beveragesin moderation, if atalL

Facts to Keep in Mind

It's the total combination of foods thatyou eat regularly- both the kinds andthe amounts-that's important in termsof good nutrition. Eating particular foodsor one specific food isn't a magic keytha t will assure you have a more healthydiet.

Eating a healthy diet, in itself, doesnttguarantee good health. A healthy diet~

however, is an important part of ahealthy lifestyle that includes, forexample, regular physical exercise, notsmoking, .not drinking alcoholicbeverages to excess, and not abusingdrugs.

In addition to what you eat, manyfactors may affect your own chance ofdeveloping a particular disease. Amongthese are your heredity, yourenvironment, and the health care youreceive. Our knowledge about most diet~

heal th rela tionships is incornplete.andwill improve as scientific knowledgeincreases. However, enough is knowntoday about some of these relationshipsto encQur{:lge specific dietary practicestha tare believec1 to be beneficial.

IX. References

The following references have beenpLlced on dispt'ay in the DocketsManagernent Branch (address above]and may be seen by interested person~

between 9 a.m. and 4 p.rn., Mondaythrough Friday.

1. Ambard, 1... and E. Beaujard. "Causes dei.'hypertension Arterielle," ArchivesC;enerales de lvledicine, 1:520-533 (1904).

2. KirkendaH, \tV.M., W.E. Conner. F.i\l.iOoud. S.P. Rastogi, T.A. Anderson. and ~t

Fry, HThe Effect of Dietary Sodium Chlorideon Blood Pressure. Body Fluids. Electrolytes.R(~nal Function. and Serum Lipids ofNormotensive Man." Journal of Laboratoryand Clinicallvledicine, 87:4'18--434. 1976.

3. Select Committee on Nutrition andlluman Needs. "Dietary Goals for the UnitedStates," 2d ed.. U.S. Senate. 95th Cong.. 1stsess.. U.S. Government Printing Office.Washington. 1977.

4. World Health Organization. "Arteriall:lypertension." Report of a WHO ExpertCommittee, Technical Report Series 628,World Health Organization, Geneva. 1978.

5. Beaton. C.H., D.}. Milner, P. Corey, V.McGuire, M. Cousins, E. Siewart, M. deRamos, D. Hewitt, P.V. Grambsch, N. Kassim,and J.A. Little, "Sources of Variance in 24­hour Dietary Recall Da ta: Implica tions forNutrition Study Design and Interpretation:t

Anlerican Journal of Clinical Nl1trition~32:2456-2559,1979.

6. LSRO, UEvaluation of the J-Iealth Aspectoof Sodium Chloride and Potassium Chlorideas Food Ingredients," SCOCS-I02, FASEB,Bethesda, MD, 1.979.

7. Nlodern Nutrition In Health and Disease.6th ed., R.S. Goodhart, and M.E. Shils (eds.), .Lea and Febiger, Philadelphia, PAt pp. 362­377.833-835,1003-1004,1006-1020,1024, aod1031-1032, 1980.

8. Niveil, C.F. Jr., '7echnology of Sodium inProcessed Foods: General BacteriologicalPrinciples, With Emphasis on Canned Fruitsand Vegetables, and Dairy Foods." in("Sodium and Potassium in Foods and Drugs:·Conference Proceedings, White. P.L., and S.C.Crocco (eds.), American Medical Association.Chicago. IL, pp. 45-60, 1980.

9. "Nutrition and Your Health-DietaryGuidelines for Americans," Home andGarden Bulletin No. 232, USDA, DHHS, 1980.

10. Vetter, I.L., uTechnology of Sodium inBakery Products" in "Sodium and Potassiumin Foods and Drugs," ConferenceProceedings. White, P.L., and S.C. Crocco(eds.), American Medical Associa tion,Chicago, IL. pp. 61-65,1980.

11. Subcommittee on Investigations andOversight of the Committee on Science andTechnology, "Sodium in Food and I-ligh BloodPressure," Serial M,U.S. l-Iouse ofRepresentatives, 97th Cong., 1st sess., U.S.Government Printing Office. Washington, D~1981.

12. l-leimbach, J.T.,·"Public Understandingof Food Label Information, It Division ofConsumer Studies, FDA. 1982.

13. Marsh~A.C.• "Processes andForrnulations that :Affectthe Sodium Contentof Foods:· Food TechnololJY~ pp. 45-49,July1.983.

14. Porter, G.A., "Chronology of the Sc,ditHnHypothesis and Hypertension." Anno!s ofInterned Aledicine, 9B:72(}-72~J.19iH.

15. Resnick. LM.,J.ll Larngh. J.E. Sealey ..dnd MJ I. Alderman. "Divalent Cations inEssential Hypertension," Net\' Englandlournal of Medicine. 309:888-89'1.1983.

16. Sh~nk. F.R.. L. Larsen. F.E. Scarbrough,J.E. Vanderveen. and A.L. Forbes. "FDAPerspective on Sodium'" Food Technolugy.pp. 73-78, July 1983.

17. Joint National ConuniHee on Det.ection.\Evaluation, and Treahnent of High BloodPressure. HThe 1984 Report of the Joint.Na tional Commi ttee on Detection.Evaluation, and Treatment of lligh BloodPressun~.H Al'chie\-'es of Internol tvledicine,1-14:1045·-1057,1984.

18. Pennington, J.A.T.. D.B. ·Wilson. R.F.Newe!L B.F. Harland, R.D. Johnson, and ].E,Vanderveen, hSelected Minerals in FoodSurveys. 1974 to 1981/82," Journal of theArncrican Dietetic Associolion, 84:771-780f

1984.19. Resnick. L.M., R.K. Gupta. and .J.H.

Laragh, "Intracellular Free lvlagnesimn inErythrocytes of Essentiaillypertension:Rela Hon to Blood Pressure and SerumDivalent Cations," Proceedings of theNatJ'onal.Academy of Sciences of the llnitt?(JStates of America, 81~6511-6515,1984.

20. £-\rmstrong. L.E.. R.W. I-Iubbard. p.e,Szlyk, W.T. Matthew. and lV. Sils,"Voluntary Dehydration and ElectrolyteLosses During Prolonged Exercise in thel-Ieat," Aviation, Space, and E111/ironfnento!A1edici.ae, pp. 765-770, August 1985.

21. MacGregor, G.A., uSodium is Morehnportant than Calcium in Essentiall-Iypertension," Hypertension, 7:628-637, 1985

22. hNutrHion and Your Health-DietaryGuidelines for Americans," 2d ed., Home and(;arden Bulletin No. 232, USDA, DIU-IS, 1985.

23. Subcommittee on Definition andPrevalence of the 1984 Joint NationalComnlittee, Final Report. Hl-IypertensionPrevalence and the Status of Awareness i '

Treatrnent, and Control in the United Statcs/1

Jlypertension~ 7:457-468, 1985. .24. Weinsier, R.L.• and D. Norris, "Recent

Development-s in the Etiology and Treatn'1entof Hypertension: Dietary Calciun1, Fat, andMagnesium," Anlerican Journal of ClinicallVutrition~ 42:1331-1338. 1985.

25. Division of Consumer Studies, .6Trendsin Sodium Labeling of Supermarket Foods1978-1986-Food Label and Package Survey(FLAPS)," FDA. Center for Food Safety andi\pplied Nutrition, 1986.

26. Kaplan. N.~1., and R.B. ~1eese, HTht~

Calcium Deficiency lIypothesis ofllypertension: A Critique," Annals of lnternaAledic.ine, 105:947-955, 1986.

27. National Center for Health Statistics,Drizd. T., A.L. Dannenberg, and A. Engel.HBlood Pressure Levels in Persons 18-74Yea'rs of Age in 197~O, and Trends in BloodPressure From 1960 to 1980 in the UnitedStates:' Vital and Health Statistics. Datafrom the National Health and· NutritionExamination Survey, Series 11, No. 234.DI-Il-iSPub. No. (PIIS) 86-1684. DIIIIS,PublicHealth Service, (PHS), U.s. GovernmentPrinting Office, Washington, DC~ July 1966~

50842 Federal Register / Vol. 56, No. 229 I Wednesday, November 27 t 1991 / Proposed Rules

Japanese Populations," Journal of f-IunlonHypertension 9 3:315-321, 1989.

55, Lasaridis, A., C.A. Hatziioannou, P,GramaHcos, H. Dedusi, A, 50f03, C. Kcd~~is)

PJ. P:..rzoglou i S.B. Asvf:sta., P. iv1iSLdHdi~" P.Zembekalds, and A. Toul'kantords, "TheEffects of Daily Sodium Intake on theRelationship of Blood Pre~3Un\ RenD! Pr:~En1a

FlO'll! and l'·Jatriuresis in E8s.entL::dHypertensiDn," lournal of f{ypert.'en3/on, 7(SuppL) S18o-51~11, 19S9,

55. A,S., and J.T, Dt.:nrUU;]iCfL

PubHc EffortsDic:t," FDi\, Division of COnSlmr!erVVashington, DC, 1989.

57, LSRO, "Nu.trition idonHoring: theUnited States-An onNutrition rI,Jh}nitol'inj~?

the Dt~IHS, PHS, the USDitConsumer Services, Df1HS Publication Nr:'I ..

(PHS) 89-1255, PHS j VVashington, U.S.Government Printing Office? September

58. lVlancilha Carvalho y fl, R.G. BaruzzLP.F. Hov"rard, N'. PouHer? M.P, Alpers, L.J.Franco, L.F. Marcopito, V.J. Spooner, A.R.Dyer, P. EHiott, J. Stamler, and R. Stamler,"Blood Pressure in Four Remote Populationsin the INTERSALT Study," Hypertension9

14:238-246, 1989.59. f\1ancilha Carvalho, J.J., de R.Oliveira,

and R.J. Esposito, "Blood Pressure andElectrolyte Excretion in the YanomamoIndians, an Isolated PapulaHan," Journal ofHuman Hypertension, 3:309-314, 1989.

60. Marmot, M.G., and P. Elliott, "PublicHealth ~ieasures for Blood Pressure Controlin the VVhole Community," CDnical andExperimental Hypertension-Theory andPractice, All: (5 and 6), pp. 1171-1186, 1989.

61. ~vloore, T.J., "The Role of DietaryElectrolytes in Hypertension," Journal of theAmerican College ofNutrition, 8 (Sunpl),688-808, 1989.

62. NAS report, "Diet and Health­Implications for Reducing Chronic DiseaseRisk,n Committee on Diet and Health, Foodand Nutrition Board, Commission on LifeSciences, NRC, Nation,3} Academy Press,Washington, DC, pp. 3-40, 413-430, 549-561,1989. .

63. Subcommittee on the Tenth Edition ofthe RDA's, Food and Nutrition Board,Commission on Life Sciences, NRC,"Recommended Dietary Allowances,n 10thed., 'National Academy Press, Wasihington,DC,pp. 247-261, 1989.

64. Rose, G., and J. Stamler, on b.ehalf of theINTERSALT Cooperative Researc~'Group,

"The INTERSALT Study: Background,Methods an:d Main Results," Journal ofHuman Hypertension, 3:283-288, 1989.

65. Saito, K.,H. Sano, Y. Furuta, and H.Fukuzaki, "Effect of Oral Calcium on BloodPressure Response' in Salt-loaded BorderlineHypertensive Patients," Hypertension, 13(3):219-226, 1989.

66. Sowers, J.R., "Dietary Calcium Effects inSalt-sensitive Hypertension," ClinicalNutrition, pp. 158-163, 1989.

67. Sowers, J.R., P. Zemel, P. Standley, J.Kraniak, M~B. Zemel, "Role of Calcium inModulating Salt Sensitivity" in Salt andHypertension: Dietary Minerals, VolumeHomeostasis and Cardiovascular Regulation.Springer Verlag, New York, pp 169-175,1989.

Pedoe, "Urinary Electrolyte Excretion,tdcohol Consumption, and Blood Pressure inthe Scottish He3~t f-Iealth Study," BritishNledjcal Journal, 297:329-330, 1988.

42. Staesscn, J., C. J. Bu.lpitt, R. Fagard, J. V," "'.L,".",·";",."_ P. Lijnen, and A. Amery, "Salt

and Blood Pressure in the Generali-'Gj:Jt.u,atlon: l\. Controlled Intervention Trial in

T(yvvns," Journal of f1ypertension 9 6:955­97:1,19n8.

43. General's Rep8rt on "Nu.tritionand " Df-if-iS, PHS\ U.S. Gov£fnm.entPrinting Office, VVashington, DC, pp. 1-20,139-175,1988.

44. Australian National Health and r.,,1edlc21Research Council Dietary Salt StudyfAan2gernent ConuniHee, "Fall in BloodPressure w·ith fvIodest Reduction in DietarySeIt Intake in :fI.,,:ifld f-lypertension," Lancet~

pro 399-402, February 25, 1989.45. Australian. National Health and Medical

Research Council Dietary Salt Study~.Aanagem€ntCODlmittee, "Effects ofReplacing Sodium Intake in Subjects on a.LOl/v Sodimn Diet: A. Crossover Study,!!Clinical and Experimental Hypertensjon­Theory and Practice, Al1:1011-1022, 1989.

46. Bender, M., "Status of Nutrition andSodium Labeling on Processed Foods: 1988­Food Label and Package Survey (FLAPS),"Division of Consumer Studies, Center forFood Safety and Applied Nutrition, FD.A,1989.

47. Blumenthal, D., "The Health-Diet Link­Charting a Rising Awareness," FDAConsumer, pp. 23-27, October 1989.

48. Boegehold, M.A., and T.A. Kotchen,"Relative Contributions of Dietary Na + andCI- to Salt-Sensitive Hypertension,"Hypertension, 14:579-583, 1989.

49. Egan, B.M., and A.B. Weder,"Deleterious Effects of Short-Term DietaryNaCl Restriction in Man: Relationship to Salt­Sensitivity Status," presented at AmericanHeart Association's Scientific Sessions inNew Orleans, LA, November 15, 1989.

50. Elliott, P., "The INTERSALT Study: AnAddition to the Evidence on Salt and BloodPressure, and Some Implications," Journal ofHuman Hypertension, 3:289-298, 1989.

51. Elliott, P., A. Dyer, and R. Stamlert onbehalf of the INTERSALT CooperativeResearch Group, The INTERS.l\.LT Study:Results for 24 Hour Sodium and Potassiumby Age and Sex," Journal ofHumanHypertension, 3:323-330, 1989.

52. Elliott,P., M. Marmot, A. Dyer, J.Joossens, H. Kesteloot, R. Stamler, J. Staml~r,and G.Rose, "The INTERSALT Study: MainResults, Conclusions .and Some Implications,"Clinical and Experimental Hypertension­Theory andPractice, All:l025-1034, 1989.

53. Elliott, P., S. Rogers,' G. Scally,D.G.Beevers, M.J. Lichtenstein, G. Keenan,·R.Hornby, A. Evans, M.J. Shipley, and P.C.Elwood, "Sodium, Potassium, Body Mass,Alcohol, and.Blood Pressure in Three UnitedKingdom Centers (the INTERSALT Study),"European Journal of Clinical Nutrition, .44:637-645,1990.

54~ Hashimoto, T., Y. Fujita, H. Ueshima, S.Kagamimori, T. Kasamatsu, S. Morioka, K.Mikawa, Y. Naruse, H. Nakagawa, N. Hara,H.Yanagawa, and P. Elliott, "Urinary Sodiumand Potassium Excretion, Body Mass Ind~x,

Alcohol Intake and Blood Pressure in Three

28. Resnick, L.M., J.P. Nicholson, and JJ-LLaragh, uCalci~Hn ~1ctabolisin in EssentialHypertension: R~~];]tionship to Altered ReninSystem AcUv!i::/," FeJcralion Proceedings,45:2739-274.5,

29. Subccmrr:iUc'2 on Nonpb.armacologicTherapy of the 1984 Joint NaUonal Comrnitteeon Detection, EV:~!llatiGn, lind TreatInent of

Blood Pres:.;hre, Final Report,

, ". " A.pproache,~ to !?~~ TV

Control 01 Hl~:l Pressure, NI-J.tLnI, NLd,HFpertcnsionl a:444-467, 1986.

:3~. K.nocheL J: P':,~nd G: ~eed, '.'Disorders01 heal ReguJatlon In Clulleal D1sordersFlujd and l~!f;'ctrclvte jH'clabolism, 4th ed"H. Max\:velt C. R. 'Kleernan, and R. G. l\!arir:sfeds.), lvicGravv Hill, New' York, pp. 1197­1232,1987.

31. Kurtz, 'I'. II. .A. AJ-Bander, n, C.~',:orris, Jr.,. HGSr;;lt.Sensitive' Essex;.tialH.ypertension ;n Men---Is the Soanun IonI'l..lone Important?", lVeJ-tT England Journal ofJ..fedicine, 317:1843-1048, 1987.

32. McCarron., D...4.." and C. D. Morris, "TheCalciulTI Deficiency Hypothesis ofHypertension," Annals of Internall~1edjcjne,107:919-922, 1987.

33. l\Iiller, J. Z., M. H. Weinberger, S. A.Daugherty, N. S. Fineberg, J. C. Christian, andC. E. Grinl, "Heterogeneity of Blood PressureResponse to Dietary Sodium Restriction inNorn10tensive I\.dults," Journal of ChronicDisease, 40:245-250, 1987.

34. Sanchez-Castillo, C. P., W. J. Branch,and VI. P. T. Jalnes, "A Test of the Validity ofthe Lithium-marker Technique for MonitoringDietary Sources of Salt in Man," Journal ofClinical Science, 72:87-94, 1987.

35. Sanchez-Castillo, C. P., S.Warrender,and vV. P. T. James, "An Assessment of theSources of Dietary Salt in a BritishPopulation,n Journal of Clinical Sciences,72:95-102, 1987.

36. Gruchow, II. W., K. A. Sobocinski, andJ. Barboriak, "Calcium Intake and theRelationship ofDietary Sodium andPotassium to Blood Pressure," AmericanJournal o/Clinical Nutrition, 48:1463-1470,1988.

37. The INTERSALT Cooperative ResearchGroup, "Intersalt: An International Study ofElectrolyte Excretion and Blood Pressure:Results for 24-Hour Urinary Sodium andPotassium Excretion," British MedicalJournal, 297:319-328, 1988.

38. Jointr·Jational Committee ,on Detection,Evaluation, and Treatment of High BloodPressure, "The 1988 Report of the JointNational Committee on Detection,Evaluation, and Treatment of High BloodPressure," Archives of Internal Medicine,148:1023-1038, 1988.

39. Luft, F. C., and M.H. Weinberger,"Review of Salt Restriction and the Responseto Antihypertensive Drugs-Satellite

.Symposium' on Calcium .Antagonists,"Hypertension, 11 (Suppll) I-22~I-232,1988.

40. Masugi, F., T. Og-ibara, K.Hashizume, T.Hasagawa, K~ Sakafuchi, and Y.Kumahara,"Changes' in Plasma Lipids and Uric Acidwith Sodium Loading alld Sodium Depletionin Patients with Essential Hypertension,"JournalbfHuman Hypertension, ·1:293-29'8,1988.

41. Smith, W:.C.S., I. K. Crombie,R~ T.Tavendale, S. K. G\lllard.".and H. D. Tunstall-

I~ederal Rl~gister / Vol. 50, No. 229 / \Vednc!.Hlay, f'Jovculber 27. lH91 I Proposed Rules GOBi

68. Stamler. J., J.n. Nf~;don. n.N.\tVentworth. "'Ulood Pn:s~urc {Systolic andDiastolic) nnd Risk of Fatal Coronary IleartDisease." J-1vperti'llsion. l~l (Sunpl. !) 2---12,lBB9.

69. Stamler, J., G. Rose. R. Stamler, P.Elliott. A. Dyer. and M. fv1Brmot,&'INTERSALT Study Findings-Public llpaithand l\1edical Care IlnpHcations,"IJvpertension, 14:570-S77, 1989.

70. Stamler. R. o J. Stamler, F.e. (;osch, J.Civinelli, J. Fishman, P. McKee\'f~r. A.rv1cDonald, f\.R. Dyer. uPrifnary Prevention ofHypertension by nutritional-HygienicJ\lIeans-Final Report of 8 Randomized,Controlled Trial." Journal of the AmericanAfedical Association. 26:2:1 BOl-1807, 1989.

71. Takemori. K., S. rvlikami. S. Nihira, andN. Sasaki, Tohoku Journal of Experimental1\:ledicine, 158:269-281, 198B.

72. Alderman, f\.1.H., and B. Lamport~'ModerateSodium Restriction-Do theBenefits Justify the I-Iazards?", AmericanJournal of flypertension, 3:499-504, 1990.

73. Ernst, F. A., C.O. Enwonwu, and R. A.Francis, "Calcium Attenuates CardiovascularReactivity to Sodium and Stress in Blacks,"American Journal of Hypertension, 3:451-457,1990.

74. DJ-IHS, PHS, "Healthy People 2000­National Health Promotion and DiseasePrevention Objectives," DriHS Pub. No.(PI-IS) 91-50213, U.S. Government PrintingOffice, Washington, DC, 1990.

75. Kaplan, N.M., "New Evidence on theRole of Sodium in Hypertension-TheINTERSALT Study," American Journal of}{ypertension, 3:168-169,1990.

76. Koopman, H., C. Spreeuwenberg, R.F." \tVesterman, and A.J.M. Donker, "DietaryTreatment of Patients with Mild to ModerateHypertension in a General Practice: A PilotIntervention Study. (2) Beyond ThreeMonths," Journal of liunlan l/ypertension,4:372-374,1990.

77. Kurtz, T.W., and R.C. Morris. Jr.,("Sodium-Calcium Interactions and Salt­Sensitive Hypertension," American Journal ofJlypertension, 3:1525-1558, 1990.

78. Levy, A.S., and T. Guthrie, "Prevalenceand Distribution of Sodiurn AvoidanceDieting Behavior in the U.S.: 1982-1988:'Center for Food Safety and Applied Nutrition.1990.

79. Luft,' F.C., M.B. Zemel. J.A. Sowers, N.S.Fineberg, M.Il Weinberger, hSodiumBicarbonate and Sodium Chloride: Effects onBlood Pressure and Electrolyte Homeostasisnn Normal and Hypertensive fvfan," journal ofHypertension, 8:663-670, 1990.

80. Mtabaji, J.P., Y. Nara, and Y. Yamori,{'4The Cardiac Study in Tanzania: Salt Intakein the Causation and Treatment ofI-Iypertension," Journal of HumanJ/ypertension, 4:80-81, 1990.

81. National Center for }-{ealth Statistics,Adams, P.F., and V. Benson, "Current

: Estimates from the National Health InterviewSUl'vey.1989" in uVital and Health 'Sta tislics," Data from the National HealthSurvey, DI-~RIS Pub. No. (PHS) 90-1504.DJ-II-iS, -PHS. Centers for Disease Control.Series 10, No. 176,p~ 84, October 1990.

82~ National Center for Health Statistics,~~f\dvanceReport of Final MortaJityStc1tistics,' 1988." f"inal Data: froin the" National

Center for Ilpalth StatisUct. Mon!hly VitalStatistics Report. Viii IS, PI IS, Center~ forDisf'ase Control, vol. 39, No.7, Supplement. p.20, Novemher 28. 1!-l90

B~l. National Center for Health Statistics,"'Health, United States, 1989." DHHS Pub. No.(PI IS) 9()'-12~l2, DIll IS, PI IS, Centers forDisease ControL u.s. Government Printing()ffice. Washington, DC, pp. 25-27. 1990.

84. "Nutrition Recornntendations-TheRepoy·t of the Scientific Review Cornmitte~ It

Catalog NUITlber H49-42!1990E, CanadianGovernment Publishing Centre, Ottawa,Canada, 1990.

85. HNutrition and Your Health-DietaryC-uidelines for Americans," 3d ed., "Homeand Garden Bunetin No. 232," USDA, DHHS,1990.

86. Oparil. S.. ).M. vVyss. R. Yang, H. Jittand Y. Chen, "Dietary Ca2+ Prevents NaCl­sensitive Hypertension in Spontaneouslyf-Iypertensive Rats by a Sympatholytic~iechanism."American Journal offfypertension, 3(8):1795-1888,1990.

87. Passmore. J.C., and A.E. Jimenez,"'Separa te Hemodynamic Roles for Chlorideand Sodium in Deoxycorticosterone Acetate­salt Ilypertension, It Proceedings oftheSociety for Experimental Biology a,nd,Medicine, 194:283-288, 1990.

88. Resnick, L.M., "The Effects of Sodiumand Calcium in Clinical Hypertension:~tediatingRole:of Vitamin D Metabolism,"The Biology and Medicine of SignalTransductiofl in The Biology and Medicine ofSignal Transduction, Nishizuka, Y. et al.(eds.), Raven Press, New York, 1990.

89. Sharma, A.M., H.R. Arntz, A. Kribben,S. Schattenfroh, and A. Distler, "DietarySodium Restriction: Adverse Effect onPlasma Lipids," Klillische Wochenscbnfts,68:664-668,1990.

90. Swales, J.D., "Studies of Salt Intake inI-Iypertension-What Can EpidemiologyTeach US?,"American Journal ofJ/ypertension, 3:645-649, 1990.

91. Alderman, M.H., S. Madhavan, W.L.Ooi, H. Cohen, J.E. Sealey~ and J.H. Laragh;"Association of the Renin-Sodium Profilewith the Risk of Myocardial Infarction inPatients with Hypertension, It New EnglandJournal of A1edicine, 324:'1098-1104, 1991.

92. Boegehold, M.A.• and T.A. Kotchen,~~Importanceof Dietary Chloride for SaltSensitivity of Blood Pressure," HJpertension,17 (Suppl. I), pp. 1158-1161, 1991.

93. Crane, N .• memorandum of October 23.1991.

94. Cutler, J.A., D. Follrnann, P. Elliott. andII Suh, "An Overview of Randomized Trialsof Sodium Reduction and Blood Pressure,"llypertension, 17 (Supp!. I), 1-27-1-33, 1991.

95. Dustan, H.P.• "A Perspective on theSalt-Blood Pressure Relation.~' fJjpertensioJ1.17 (Suppl. I), 1-166-1-169, 1991.

96. Dzau, V.J., "Renin and MyocardialInfarction in Hypertension," New EnglandJournal of Medicine, 324:1128-1130,1991.

97. Elliott, P., uObservational Studies ofSalt and Blood Pressure." Hypertension~17(~uppL 1),1-3-1-8, 1991.

98. Eluler. P.J~. R.H. Grimm. Jr., J. Flack. andB. Laing, "Dietary Sodium Reduction for:Hypertension Prevention and Treatment,"J-Iypertensi'pn, 17 (Suppl. f), 1-182-1-189,1991.

99. Ernst, N.D., "Health Promotion Role's of.the: Federal Government and Food Industry' in

Nutrition and Blood Pi{~ssur~." / lypr:'rlt'E:':;/c'l7 (SuppL 1). 1-196---1--20U, 1991.

"100. Frost, C.D., M.R. Law, and N.l. V\/ald,"By How' Much Does Dietary Salt Ri:ductio~

Lower Blood Pressure? II-Analysis ofObservational Data Within Populations,'"British .Alle-dicolJournal, 302:81:1-818, 19~n,

101. l-iamet. P., E. Mongeau, J. Lamhert F.Bellavance, M. Daignault-Gelinas. tvL l.A~dOtJ

and L. vVhissell-Cambiotti, "Interactions/\rnong CakiuIT1, Sodium, and Alcohol lntd~

As Deterrninants of Blood Pressure."flypertfJl1sion. 17 (Suppl. f), 1-1SD-I--!S4.!HU

102. Hegsted, D.M., i'A Perspective onReducing Salt Intake." flypertension, 17(SuppI. I), 1-201-1-204. 1991.

103. Hypertension, 17 (SuppL IL 1-1-1-'2211;991.

104. Kuner, L,H.. HReseal'Ch and PolicyDirections-Salt and Blood Pressure,)'flypertension, 17 (Suppl. I). 1--211-1-21fi 1991

105. Kumanyika, S .. '''Behavioral AspectsIntervention Strategies to Reduce DietarySodiuIn/' Hypertension. 17 (Suppl. I), i-19o­"195,1991.

106. Law, M.R., C.D. Frost, and N.J. Wald,HBy How Much Does Dietary Salt Reductio!Lower Blood Pressure? I-Analysis ofObserva tionaI Da ta Among Popula Hons, H

British A1edical Journal. 302:811-815,1991.107. Law, M.R., C.D. Frost, and N.J. Wald,

I'By 1~low Much Does Dietary Salt ReductiolLower Blood Pressure? III-Analysis of Dat.From Trials of Salt Reduction,"BritishAJedicaJ Journal, 302:819--824, 1991.

108. LSRO, "Dietary Sodium Chloride andBlood Pressure, H FASEB, 1991.

109. Mascioli, S., R. Grimm, Jr., C. Launer,K, Svendsen, J. Flack, N. Gonzalez, P. ElmerJ. Neaton, "Sodium Chloride Raises BloodPressure in Normotensive Subjects, H

llypertension, 17 (Suppl. I), 1-21-1-26, 1991..110. McCarron, D.A., "A Consensus

Approach to Electrolytes and BloodPressure-Could We All Be Right?",lJypertension. 17 (SuppL I), 1-170-1-172, 199

111, Page, LB., HCommen tary onEpiderniological Studies and InterventionTrials, H l/ypertension, 17 (Suppl. I), 1-34-1-31991.

112. Peeker, M.S., and JJl Laragh. 64DietaISalt and Blood Pressure-A Perspective,"J-jypertension, 17 (Supp!. I), 1-97-1-99. 1991.

113. Prineas, R.f., "Clinical Interaction ofSalt and Weight Change on Blood PressureLevel." Hypertension, 17 (Suppl. I), 1-143-1­149,1991.

114. Stamler, R., ulmplications of theINTERSALT Study," Hypertension, 17 (SUP!I), 1-16-1-20, 1991.

115. Sullivan, J., "Session V,Environmental/Metabolic Modifiers ot Salt~

Blood Pressure Rela tion," Hypertension, 1'i(Suppl. I), 1-173-1-174.1991.

116. Sullivan, rM., "Salt Sensitivity:Definition, Conception, Methodology, andLong-term Issues," Hypertension, 17 (Suppl.1),1-173-1:-174,1991.

117. Armstrong. L.E., R.W. Hubbard, E.\V~Ask~w, and R.P. Francesconi. "Responses 0

Soldiers to 4 g and 8 g NaCI Diets Du~ing 10Days of Heat Acclimation," in press.

118. Francesconi, R., L. Armstrong, N. Lev,~. Moore, "~.Szlyk, W, ~atthew,\V.,Curtis,Jr.~ R. Hubbard,:E.W. Askew. .

60844 Feder~IRegister /VoL'56, No. 229 I vVednesd~y,'Nbveniber -~7, 1991 I,Proposed R~les

UEndocrinological Respons~~s to Dietary SaltRestriction During ,Heat Accli.nation," inpress.

119. Johnson, R.F., and D.}. ~1erullo,

jlReevaluation of Sodium Requirements for\\lork in the Heat: Sllbjective Reports of HeatIllness:' in press.

120. Folkow, B., i1Sa lt and Hypertension,"News in Physiological Sciences, 5:220-224,1990.

121. Dustan, H.P., and K.A. Kirk. "CororanLecture: The Case for or Against Salt inHypertension," Hypertension, 13:696-705,1989'.

122. !\.1acGregor, G.A., N.D. ~farkaudu, G.A.Sagnella,D.R. Singer. and F.P. Cappuccio.liDouble-blinde'd Study of Three SodiumIntakes and Long-term Effects of-Sodium

estriction in Essential Hypertension,"Lancet. pp. 1244-1247, Noven1ber 25,1989.

123. The TOHP Collaborative 'ResearchGroup, liResults of the Trials of f-IypertensionPrevention (TOl-IP)," abstract.

124. Hypertension Prevention TrialResearch Group, liThe HypertensionPrevention Trial: Three-year Effects ofDietary Changes on Blood Pressure,"Archives of In ternalMedjcjne. p. 150, 1990.

List of Subjects in 21 CFR Part 101

Food labeling,' Reporting andrecordkeeping requirements.

Therefore, under the Federal Food,Drug, and Cosmetic Act and under theauthority delegated to the Commissionerof Food and Drugs, it is proposed that 21CFR part 101 be amended as follows:

PART 101-FOOD LASE'LING

1. The authority citation for 21 CFRPart 101 is revised to read as follows:

Authority: Sees. 4, 5. 6 of the Fair Packagingand Labeling Act (15 U.S.C. 1453, 1454,1455) ;sec~ 201, 301~40~40a409,501, 50~ 50~701

of the Federal Food, Drug, and Cosmetic Act(21 U.S.C. 321, 331, 342, 343, 348, 351, 352. 355,371).

2. Section 101.74 is added to subpart Fto read as follows:

§ 101.74 Health claims: sodium and highblood pressure.

(a) Relationship betvveen sodium andhigh blood pressure.

As used here! hypertension, or highblood pressure, means systolic bloodpressure of grea ter than 140 millimetersof mercury (n1m Hg) ordiastolic bloodpressure ofgreater than 90 mm fIg.Normotension, or normal bloodpressure, -is a systolic blood 'pressurebelow 140 mm J-Ig and diastolic bloodpressure below 90mm Hg. Sodium isspecified here as the chelnical entity ormineral 'Isodium" and is distinguishedfrom sodium chloride or salt, which is 39

percent sodium by weight. The scientificevidence from epidemiological, clinical,and animal data establishes that highsodium intake is related to theprevalence of hypertension or highblood pressure and to the increase ofblood pressure with age, and that 10\vsodium intake is related to 'lowprevalence of hypertension or highblood pressure and to a low rise or noincrease of blood pressure with age.

(b) Significance of SOdjUfli bl affectinghigh blood pressure. High bloodpressure is a public health concernprimarily because it isa major riskfactor for rnortality from coronary heartdisease and stroke. There is acontinuum of mortality risk thatincreases as blood pressures rise.Individuals with high blood pressure areat greatest risk, and individuals withmoderately high, high normal, andnormal blood pressure are at steadilydecreasing risk. The 1990 "DietaryGuidelines for Americans" sta tes tha t:"In the United States, about one in threeadults has high blood pressure." Thescientific evidence from clinical dataindicates that reducing sodium intakelowers blood pressure and associatedrisks in SODle but not all hypertensiveindividuals; approximately 30 to 60percent respond to sodium reduction.There is some evidence that reducingsodium intake lowers blood pressurearid associated risks in many but not allnormotensive individuals as well:approximately 15 to 45 percent respondto sodium reduction. There are nopractical genetic markers to identifyresponsive individuals. The populationsat greatest risk for high blood pressure,and those most likely to benefit fromsodium reduction, include those withfamily histories of high blood pressure,the elderly of all genders and races,males because they develophypertension earlier in life, and blackmales and females. Sodium intake"alcohol consumption, and obesity areidentified risk factors for high bloodpressure. On a population-wide basis,the indications from epiden1iologicaland clinical data are that reducing theaverage sodium intake v~~ould have asmall but statistically significant effecton reducing the average blood pressure.Estimates suggest that reducing sodiumintake by 100 millimoles (mmol) per day(2,300 mg of sodium or approximatelyone rounded teaspoon of salt) wouldcorrespond to an average lowering ofblood pressure of approximately 2.2 mmHg systolic and 0.1 mm rIg diastolic.Because these are population-vvide

estlnlates, the magnitude of the-effect forsensitive individuals would be greater.Estimates suggest that, for the age rangefrom 25 to 55, ,a 100 mmol per day (2,300milligrams (mg) per day) lower lifetimeintake of sodiuol would correspond to areduction in mortality rates ofapproximately 16 percent for coronaryheart disease and ;23 percent for stroke.In order to reduce: sodium intake,indi~iduals can choose foods with lesssodilun and salt, reduce the an10unt ofsadiuol and salt used in foodpreparation and c'ooking, and reduce theamount of salt added at the table.

(c)Speclfic requirel11enls. A food labelor labeling may contain a sodium/hypertension health claim provided that:

(1) The health claim for a food or foodproduct meets all the generalrequirements of§101.14 for healthclaims.

(2) I'he health claim states that a 10\\1sodium diet is associa ted with or rela tedto lOvver blood pressure in some people.Alternatively, the health claim can statethata high 'sodium diet is associatedwith or' related to higher blood pressurein some people. '

(3) The health claim identifies thepopulations at gr~atest risk of ;developing high blood pI'essure! as beingthe elderly and those with fan1ilyhistories of high blood pressure andsta tes tha t other dietary risk factorsassociated with high blood pressureinclude alcohol 'consumption and excess\-veight.

(d) Optionalinfornlation. SodiuI11/hypertension in health clainls mayprovide- additional infotma tion:,

(1) The health clairn may state thatsodium is an essential nutrient ornecessary for good health, and tha t thetotal intake of sodium should be at least

'500 mg per day but notmore than 2,400mgper day.

(2J The health claim may state thatindividuals with high blood pressureshould consult their physicians formedical advice and treatment.

(3) In specifying the nutrient, thehealth claim Inay include the term I'salt"in addition to the ternl "sodium".

(4) In specifyi ng the disease, thehealth claim Inay include the tenn"hypertension" in addition to the termllhigh blood pressure".

(5) The health claim n1ay includeinfornlation from paragraphs (a) and (b)

Federal Register I Vol. 56, No. 229 I Wednesday, November 27, 1991 I Proposed Rules

of this section~ which include surnnlariesof the relationship betw(~(~n sodiunl Hndhigh blood pressure and of tll(~

s~gnificanceof sodiurn rcduciiu[l inaffecting high blood pf(~ssurCr

(elSainple heolth cJuinl. IIJigh ·bldodpressure is associated with Inanyfactors, includitlg a fanlily history of thedisease, gro\rv·in~ olJe~1 beingoverweight, drinking too Illuch alcohol.and diets high in sodiurn. 1\ low sodiullldiet is associa ted \!\'ith lower bloodpressure in some people.

Dcttccl: N()\'ernber 4. 10BI.

David A. Kessler,COl7!rnissiu!7er ofFoud und !)rlfg<,',

Louis W. Sullivan,Secrete;!,J/ ofJleoJth one! Jiu/nOll S'(.'rn'ces.

Note: The following tables will not appearin the annual Code of Fcdcrctl Regul,ttions.

6084~

TABLE 1.-INTERSALT STUDIES

Reference Study design and duration SUbjects Base diet Other factors Results 1 Assessment and comments

~

'I~en

Elliott (1989) IINTERSALT study design.(Ref. 50) Analysis of

normotensive SUbjects.

Elliott (1989) IINTERSALT study design.(Ref. 52) Main results and

implications for publichealth policy.

Elliott (1989) \INTERSALT study design.(Ref. 51) Analysis of all

INTERSALT data byage and sex.

Same as INTERSALT I Same as INTERSALT ..

~o

'"0ornro~

~c::::ro~

--.

<e..019'JZ?N~CO

.........

~ro8­coen0..PJ~

Zo<(0

3crro~

t'Y~

~coc.c~

JIoljCtlc..ct~e..~~

QC;~...CO~

~

Authors recommend modest In­crease in K and reductions inNa, obesity, and heavy alcoholdrinking and suggest these life­style changes could result indownward shift of populationblood pressure and prevalenceof hypertension.

Regression coefficient similar tofindings with the total INTERSALT popuiation. Total population: 2300 mg change in Na'ntake corresponded to 2,17~m Hg change in SSP.

Large scaleInternationalStandard assessmentsLittle data loss (approximately

3%). Study methods cons;stentwithin and across populations.Within center relationships andacrosa center changes in relationships with age were consistent. Across center relationshipsdepended on inclusion or exclusian 0" 4 population centers

SBP posi:ive and significantly re­lated to Na in 2 of 3 centers(p <0.05) (Belfast, SouthWales) and inconclusive inthird. DBP relationship inconclu­sive in 3 centers.

Within centers: Na significantly ra·lated to SSP and DBP. Across52 centers Na significantly re­lated to SSP, to DBP, and tochanges in SBP and DBP withage. Across 48 centers: Na sig­nificantly related to change inDBP with age and signi'ficantlyand negatively related to DBP.Estimated that 2300 mg lessNa per day corresponds tolower S8P (2.2 mm Hg), lowerDBP (0.1 mm Hg), and lowerchange of SSP (9.0 mm Hg)and DBP (4.5 mm Hg) with agefrom 25 to 45 years of age.

Na intake significantly related toblood pressure: 2300 mgchange in Na intake corre­sponded to 2.10 mm Hgchange in SSP after multipleregression analysis to adjust forconfounding factors.

Within centers: Associations stronger for womenAverage Na intake related to SSP than for men and for the older

and DBP in men and in women, age categories.to SSP for men and womencombined in 4 age categories,and to DBP in only the oldestage category. Across centersassociations influenced by 4centers with low Na intake.

The higher the center's median Authors suggest that small clinicalNa excretion, the steeper the changes could result in largeslope of blood pressure with benefits to a population, espe-age. dally in the cumulative effects

over a Iifet!me. Authors con·elude that the study gave "pow..erful qualitafive tests . . . poorQuantitative estimates of thesiz,3 of these reiationshlps "

Single Na measurement does notassess previous or habitual Naintake habits. Subjects on anti­hypertensive medication includ­ed thereby reducing the effectof Na on blood pressure. Urinecollection probably not com­plete in all subjects. Confound­ing factors: age, sex, bodymass index, alcohol intake, andpotassium intake. Significancesometimes lost after adjustmentfor confounding factors.

Multiple linear regression analy­ses in 52 centers, pooled,weighted, adjusted for potaSSi­um intake, alcohol intake, andbody mass index, and adjustedfor age or sex as appropriate.Corrected for individual variabili­ty of Na excretion ("reliability")by estimating the degree of re­gression dilution using datafrom 8 % of subjects who re­turned and provided second setof data.

Some subjects were on antihyper­tensive medication which wouldgive artificially low blood meas­urements and underestimateany effect of Na on blood pres­sure. Only simple adjustmentswere used to correct for biastoward zero. Some populationshad been subjected to healthcampaigns to reduce Na intakewhich wOl~ld underestimate cur­rent blood pressure effects dueto pievious habits.

43 subjects excluded (41 for in­complete urine collection and 2for pregnancy).

Normal dietsNa ranging from 730 mg

to 9,040 mg per 24hours.

Averages:Belfast: 3,470 mg Na per

24 hours.Birmingham: 3520 mg Na

per 24 hours.South Wales: 3500 mg

Na per 24 hours.

Normal diets 1 Multiple regression analysis ad-justed for a.ge, .sex, potassiumintake, alcohol intake, and bodymass index. Corrected for indi­vidual variability of Na excretion("reliability") (see below).

Normal diets ranging from5 mg Na per day to5,560 mg Na per day.

Same as INTERSALT ..

598 subjects:(299 men, 299 women)

10,079 subjects from 52population centers in 32countries.

(5,045 males, 5,034females).

(Goal of 200 at eachcenter, 25 in each of 8age and sex groups).

Normotensive subjectsfrom INTERSALTpopulation (SBP < 140mm Hg, DBP < 90 mmHg, not onantihypertensivemedication).

10,079 subjects:(5045 males, 5034

females).(Goal of 100 males and

100 females from eachof 52 centers, 25 ineach age category: 20­29, 30-39, 40-49, and50-59).

INTERSALT study design.Analysis of data fromthree United Kingdomcenters (Belfast,Birmingham, and SouthWales). Data collectionin 1985.

Cross sectional study.Sodium (Na) intakedetermined by single24-hour urine collectionblood pressuredetermined by averageof 2 seatedmeasurements.Repeated urinecollections for 8% ofsubjects to estimatewithin-individualvariability.

Elliott (1989) ..(Ref. 53) ..

INTERSALTCooperativeResearchGroup (1988).

(Ref. 37) .

Hashimoto (1989).. .1 iNTERSALT study design. 591 subjects:(Ref. 54) M l Analysis of data from (295 men~ 296 women}

~ three Japanese centers(urban Osaka, ruralTochigi, semi-ruraJToyama).

~

S;;

~

~roE-­c;t:J)

~Co;

~

Zo~ro3c­~

"""

--.

~~c.~,.,=-

<:c

Nt.Jc.:J

"t'...,o-0or;r;~Q..,

~C~I"J.)

~~

QQ;;.;-~--

--

N:~

~

~~~

Authors noted that when otrle(centers had low average bodyweights and alcohol intake, buthigh Na intake (2.760 to 4,830mg), the prevalence Of r"lyper·tension ranged from 8 to 19';'0.

2 years for planning, funding, andrecruitment of centers. Regionaltraining meetings in 1984- 1985.Field work and laboratory ana!­yses completed in 1987.

Authors recommend further re­ductions in Na, increases in K,and reductions in heavy drink~

ing,

Na excretion significantly relatedto blood pressure in individualsand to rise of blood pressurewith age.

Four populations had lowest aver~

age blood pressure comparedto other 48 centers (SBP of103 mm Hg vs 120 mm Hg,DBP of 63 mm Hg vs 74 mmHg). Four populations had little i

or no upward slope of bloodpressure with age. Hyperten·sion in 50/0 of Kenyan popula­tion and absent in other popula­tions. Four populations had lowaverage Na relative to other 48centers (1-3 grams NaCI vs 9grams NaCI).

Average SBP 96.0 mm Hg (range :from 78 to 128 mm Hg). Aver·age DBP 60.6 mm Hg (rangefrom 37 to 86 mm Hg). Lowaverage blood pressure. No hy­pertension. No increases ofblood pressure with age.

SSP negative and significantly re­lated to Na in 1 center(P<O.OOl). (Osaka), positiveand significantly related to Nain 1 center (p<O.05). (Toyama)and inconclusive in third DBPnegative and significantly relateed to Na in 1 center (p<0,01),(Osaka) and inconclusive inother two,

731 subjects:Yanomano Indians: 195

subjects.Xingu Indians: 198

subjects.Papua New Guinea: 162

subjectsKenya: 176 subjects

I 195 subjects:I (99 males, 96 females)

INTERSALT study design.Analysis of data fromYanomano. Indians, aseminomadic populationfrom the Brazilian,AmazQn rainforests.Data collected July1986.

Mancilla-Carvalho(1989).

(Ref. 59) ,\

Mancilla-CarvaJho INTERSALT study design.(1989). Analysis of data from

(Ret 58) four remote populationswith low 711 (NaC!)intake. (Yanomano andXingu Indians in Brazil.rural Kenya, and ruralPapua New Guinea),

Normal diets I9 subjects excluded for inc-om­Average Na of 4,300 mg I piete data. High within-individ~

per 24 hours ! ual variation. Public health cam~Averages: I paigns to reduce Na intake re-Osaka: 3,870 mg Na per I suited in declines in Na con-

24 hours. sumption, Blood Pressure, prev-Tochigi: 4,150 mg Na per I alence O.f hypertension, and

24 hours. stroke mortality. 16.7 normoten-Toyama: 4,890 mg Na per sives and 47.4 hypertensives

24 hours. I receiving medical treatment re-I ported reducing sait intake.Normal diets '''~ oo , •• , Average body weight low relativeMedian Na: 1 to other 48 centers. No or lowtanomano Indians: 5 mg average alcohol intake. Group

Na per 24 hours. ,: variability largest in Kenya pop-Xingu Indians: 130 mg Na ulation. Authors noted that

per 24 hours. I within-center association be-Papua New Guinea: 620 I tw.esn Na and blood pressure

mg Na per 24 hours. was unlikely due to small vari-Kenya: 1,180 rng Na per i ations in average Na and blood

24 hours. /' pressure. Adults were physicallyactive and healthy with no

I signs of malnutrition or proteinI deficiency.

! Normal diets " Ii Low average body mass index,I Average Na of 21 mg per j calcium intake, total fat and. 24 hours. Range from 1 I saturated fat. High K intake,

mg to 614 mg per 24 fiber. No alcohol intake. Almosthours. Diet of local no obesity. Relatively highcrops and game I physical activity and endurance.supplemented by wiid No physical signs of evidentfruits and irtsects. I malnutrition or protein deficien o

Banana and manioc I cy. Na not considered in corre-staple foods. Little salt, I lation analysis because nonerefined sugar, alcohol. !I normal distribution made it in-milk or dairy products in appropriate. May have eatendiet. I some food from investigators.

Rose (1989) INTERSALT study design. INTERSALT populations Same as INTERSALT l Planned study to be large enough(Ref. 64) Summary of I to observe an effect. Planned

background, methods. standardized methods to allowand main results. for appropriate pooling of re-Designed to apply I' suits. Planned analysis method-highly standardized ology to deal with confoundingmethods across varied factors. Planned random repeatpopulations, to examine I urine collections to estimatemajor confounding within-individual variability ("reli~factors. and to evaluate ! ability").the relationships in Iindividuals. I

i~-...J

60848 Federal Register I Vol. 56, No. 229 I Wednesday, November 27, 1991 I Proposed Rules

TABLE 2.-S0DIUM/HvPERTENSIONI STUDIES

inH 1'",li" I~~,-"ai~), butconfirmato~

stJdies in"responder" andj'nonresponde('populations ll::wenot been done,

Soundmethodology.

to 0.6(Sap) and 0.9(DBP) mmreduction incontrolpopulation(p<O.005) ..Results remainedsignif~cant aftermulti-variantanalysis to adjustfor age, weight,and initial bloodpressure,

Blood pressure oflow Na groupreduced anaverage 6.0(SBP) and 4.1(DBP) mm Hg inthe studypopulationrelative to 0.1(SSP) and 0.4(DBP) mm Hgreduction in thecontrolpopulation(p<O.001). Bloodpressure reducedan average 3.6(SSP) and 2.1(DBP) mm Hg inthe NaCI dietphase vs theplacebo dietphase.Differencesindependent oforder oftreatment.S.tatisticalsignificancegreater for DBPthan for SSP.

2 centers ..15 subjects

dropped outbetween two dietphases. 79subjectscompleted study.Confoundingfactors: age, sex,weight, initialblood pressure,center, order oftreatment

Low Na diet« 1840 mgNa per daymonitored by24-hr urinecollectionand dietarycounseling).

Base diet~lOt:~:::~~I---'-----;;~~~~~_. -- --'r - ASS~~~~~-~~~~~-I comments

--r-j-Lu-"W-N-a-·d:-r2~~:'Br:.=.....Blood pre;;sur~'~;-rl~und-~-j6-t.-hV-~dO!:~« 1840 mg 8 dropped low Nagroup Skewed sex .rat~o,

Na per day Large range reduced an but apiJfOxin1atelymonitored by of individual ave;';7,ge 6.124-hr urino variation SSP of (SSP) and 3.7 I to each groupcC;~8ction study and control (DBP) mrn Hg JI1 I Dk:;t~ry. f\jnand populations the restrictIon

approachedsame veJue ncarend of study D8Pof study andcontrolpopulationsapproached eachother butremained distinctat end of study.Confoundingfactors: age, sex,weight, initialblood pressure,center.

Intake of testmatedal

1840 mq Naper day,

1840 mg Naper day.

Studypopulation: 8(10mmol)slow releaseNaCI tabletsper day.

Controlpopulation~ 8placebotabletsps7day.

into tVJO

groups.

-- SUbiec~ ~reatmen.~ or,ts I Intervention

iStudy

popu!ation: 8(10 mmol)s10w re!r;aseNaC! l3b~ets

per day.Contro~

population: 8~!acBbo

tablets perday.

103 mildlyhypertehsivesubjects(DB?: 90­iOG mm

(36 rn~!e,

female).ago:

88 mildlyhypertensivesubjects(DBP: 90­100 mm Hg).

(73 male, 15female).

(Average age:58.6 years).

Randomizedinto twogroups.

Study designand duration

Double blind,placebocontrolled,cli,nica!interventiontria~.

6 weeks: run-inphase..

8 wee!.;s: dk~:1

phzse ..Subject;, secn

every 2weeks and24-hr urinE~

provided ateach visit.During dietphase, allsubjectsmonitoredandcounseled tokeep dletaryNa intake tobelow 1840mg Na perday. NaCI orplaceboadded to lowNa base diet

Double blind,placebocontrolled,clinicalinterventiontrial.Continuationof previousstudy andstudy designto includecrossoverphase.

6 weeks: run-inphase.

S weeks: dietphase 1.

8 weeks: dietphase 2.

DietarySaltStudyMan­agemen~

Commit~

tee(i989a).

(fief. 44} ....

Reference

Austra!ianNationalHealthandMedica;f1e~

seardlCouncii

AustralianNationa.HealthandMedicalRe­searchCouncilDietarySaltStudyMan­agementCommit­tee(198gb).

(Ref. 45).....

a

Federal Register I Vol. 50. No. 229 / VVednesduj Novernber 27. 1H91 I Proposed Rules

TABtE 2.-S0DIUM/HvPERTENStON S-rUD'Es-Continued

608~

Dustan("1989).

~AfJf.·12!) ..

Hyperten~

sionPreven­tionTriatRe4

searchGroup(1990).

(Ref. 124)

Study designand durat~on

CHntcalwnterventiontri~1.

Protocol .~: :3days:Controlphase. 4days: Nadepletionphase. 3days: Naloadingphase.

Protocol 2: 3days:Controlphase, 3days: Naloadingphase. 4days: Nadepletionphase. Nadeterminedby 24-hoururineexcretionbloodpressuredetermined4 times perday.

interventiontrial (paranetdesign) 3years. Nadeterminedby timed,ovemighlturinecollectionbloodpressuredeterminedas averageof 2·measure.mentstndividualsreceiveddtetarycou~seling

(once aweek for 10weeks. thenevery otherweek for 4weeks. thenevery othermonth forthe durationof the study).

Prutocol 1: 69normoten­SNas and 21hyperten­sives.

Protocot 2: :2-lnormolefl­sives and 19hyperten­sives.

·12normoten­sives and 9hyperten­sivesparticipatedin bothprotoco's.Hyperten4

sives wereeitheruntreated orhad notreceivedantihyperten·sivamedicationfor at 'east 1month.

841 subjects(OSP: 78-89mm Hg) (noantihyperten·sivemedicationor evidenceofcardiovascu.­lar disease)(65.3%mate)(82.2°/0white)(Averageage: 38.6years)Randomizedinto 5groups.

Treatmont or6nterventkm

Na dc--pletionphase:Furosemide(1 mg perkg) taken as2 divideddoses.

Na loadingphase:Isotonicsaline (3.8AmMNaCJper kg perday)suppUedintravenouslyover 4 tva

4 interventiongroups anda control.

5 groups: 1)Reducecalories, 2)Reduce N~3) Reducecalories andNa, 4)Reduce Naand increaseK~ 5) Control.

Intake of tostmateri31

Na ~oa~ng

phase: nomg Na porkg p8f day.

N/A n •••• ~.

Baso dlot

ControHe{1diets.

Contro! phdse:3450 mg Naper day(assumed asdiet sine,-'=)was notclear). Nadepletionand loadingphases: 210mg Na perday.

Group goal of50%reduction unaveragedietary NAIndividuaigoal of lessthan 1610mg/d.

Otllt)( faC~OiS

Only Na lO,lGinqpe/1ormedir.travenous~v

wttich introducesvariability. LowNa phase wasextreme: Only2~O mg per dav-

Muitipliers used toestimate 24-hourNa from timed,overnight. urinecollecttons. Ailgroups. includingthe control. hadreductions in Naexcretion and inblood pressureduring Jonow-up.Largest sustainedreduction in Naoccurred on thegro~p

encouraged toreduce both NAand calories.Interventionpopulations bloodpressure levelswere tower thanthose of thecontrolpopulation.

Protocol': MeanarteriaJ bloodpressure inhypertensives feUand rose with Na(116 to 104 to110 mm Hg).Mean arterialbiood pressure innormotensivesU'ema~ned stablethroughout (84 to83 to 81 mm Hg).

Protocol 2: Meanarterial bloodpressure in(hypertensivesand somenormotensivesrose and fell withNa (107 to 111to 98 mm Hg and83 to 87 to 82mm Hg).Separate analysisof subjects whoparticipated inboth protocolssuggested thatsequence wasnot important

Na reductionstatisticallysignificant at 6months (p =0.002) andmarginal at 3years (p =0.053) bloodpressurereductionsgenerally belowthose of controlpopulation, butchanges werenot statisticallysignificant.

Assessment amcomments

Short studyLow Na diet was

extremePopulationdiffered betwe'-O~

two protocols.Mea.ns ofadministering !N(differed betweerdiUerent IPhaS9~i

Authors note thatmaintainingdietary changesin free-livingpopulations isdifficult

.d0850 F~deral Registe,r tyoL56, No. 2~9'/ .Wednes:day, Nove~lber '27,1991 I Proposed Rules

TABLE 2.-S0DIUM/HvPERTENSION STUDIEs-Continued

Placebocontrolledcrossovertrial.

Phase 1: 4'days: low Nadiet. 7 days:low Na dietplus mineralwater withNaClorNaHC03.

Phase 2:,4days: low Nadiet. 7 days;low Nadietplus mineralwater withNaHC03 orNaC!. Allurinecollected.

Phase 1 ,andPhase 2conducted amonth apart.

Assessment andcomments

Small study.No untreated

control groupfollowedthro·ughout.

Small studyNot clear if' patients

were receiving orhad receivedantihypertensivemedication. LowDrenin patientsappeared torespond' better tochanges in Na.

mall sampl-e sizeNaCI and NaCH03

may differ in theireffects on bloodpressure.

_.. --

Subjects Treatment or Intake of test Base diet Other factors Resultsintervention material---- --_.~~-~.,._---

28 mild to Dietary N/A..................... Low Na diets 7 subjects dropped Average Namoderate counseling encouraged. out (4 for high decrease of 510hyperten- and DBP. 2 for .mg per 24 hrssives feedback suspected (p<O.05).(average monthly. angina, and 1, f,?f Ave.rage ,initial S8P: Dietary stroke). During decrease in SSP144.5 mm counseling 15t 3 months of' of 3.7 mm HgHg, average and Na and study. (p<0.05).initial DBP: blood participants had Average95.4 mm Hg}. pressure been, divided 'into decrease in DSP

measure- an intervention of 4.0 mm figmen1sended and a control (p <0;01). Na andat12 group. Control blood pressuremonths. group participants decrease during

recei'ied dietary first 6, monthscpunseling from then leveled off3'to 12"months. at lower ~alues.Intervention

, group, participantsreceived

t counseling fromO. to 12 months..

18 N/A..................... N/A..................... Controlled low .Blood pressure Average Na. rosehypertensive Na.diet: determined by from 1180 topatients. 1150 mg Na automated 4440 mg Na per

(10"males,8 per day. device. Patients day (p<0.OO1).-females). Controlled high on high Na diet Average supine

(Average age: Na diet: ' had an average blood pressure47.3 years). 4600 mg'Na increase in body rose 6.7 mm Hg

(Age range: per day. weight of ·1.3 kg (from 102.7 to30-64 years (p<0.01). 1.09.4,mm Hg)of age). "Responders" (8 (p<O.OOl).

patients with a Average standingchange in mean blood pressure

;supine blood rose 5.0 mmHgpressure >8 mm (from ·107.6 toHg). Low-renin 112.6 mm Hg)patients (6 (not significant).patients withplasma reninactivity <3 ng/mlper hr during Nadeprivation)."Responders"were virtually aUiaw-renin patients.

10 mildly NaHC03 1810 mg Na Controlled low Bfood pressure NaCI intake period: Shyp'&rtensive Phase: 3 per day. Nadiet determined by Blood pressuresubjects liters per day containing automated did not change in(blood of mineral 1380 mg Na device. NaHC03 hypertensive orpressure water per day. All minerai water normotensive>140/90 containing foods contained 12.69 subjects.mm Hg) and NaHC03 prepared mmo' CI per day. NaCH03 intake10 (26.2 mmolll and eaten at Base diet period: SBPnormoten- Na,33.03 research contained 60 decreased by 5sive subjects mmolll center. mmol CI per day. mm Hg in(blood HC03, and Same iood NaCI increased hypertensive ,pressure 4.23 mmol/l eaten for calcium excretion subjects<140/90 (CI). each meal whereas (p < 0,05). SSPmm Hg) (10 NaCI Phase: 3 every day NaHC03 did not did not change inmale, 10 liters per day during both normotensivefemale) (10 of mineral phClses' of subjects. DPS didblack, 10 water study. not change inwhite) containing hypertensive or(Aver~.ge NaCI (26.2 normotensiveage: 36 mmolll Na subjects.years) and 36.07Randomized mmoUi Cl).into twogroups.

I

Study designand duration

Clinicalinterventiontrial. Dietarycounselingandfeedback.fonow-·upfrom 3 to 18months Nadeterminedmonthly by24-hrexcretionbloodpressuredetermined'monthly.Participantsasked toreturn forevaluation a18 months.

Clinicalinterventiontrial.

2 days:adaptationperiod.

5 days: low Nadiet.

5 days: highNa diet.

Blood pressure(supine andstanding)determined3 times perday. Nadeterminedby 24-hrurineexcretion.

Reference

Koopman(1990).

(Ref. 76).....

Lasaridis(191)9).

(Ref. 55).....

Luft(1990).

(Ref. 79).....

Federal Register I Vol. 56, No. 229 I Wednesday, Novenlber 27, 1991 I Proposed Rules

TABLE 2.-S00lUM/HvPERTENSION STUDIEs-Continued

6085J

MacGre­gor(1989).

(Ref. 122)

Mascioli(1991).

(Ref. 109)

~)tviy d~si0n

and duration

Doublo blind.placebOcor1trol~ed.

crOsso~er

trial.2 months:

observaticn,4 weeks: lowNa dietphase, 11month: ~1sr

diet phase, 1month: 2nddiet phase, 1month: 3rddiet phase,12 months:follow up.Diet phasesincludedtotal dailydiets of1150,2300,and 4600mg Na.Bloodpressuredeterminedas averageof 5'measure­ments. Nadete'rminedas averageof 2 24-hrurinecollections.

Double blind,placebocontrolled,crossovertrial.

Phase 1: 4weeks: NaCIor placebocapsules.

Phase 2: 4weeks:placebo orNaGIcapsules.Phase 1preceded by2 weeks oftesting and 8weeks ofdietarycounselingto achievelow Na diet.2 weekwashoutperiod(placebocapsules)betweenphases. LowNa dietcontinuedthroughout.

Subjects

20 subjects I

with 'mildhypertension(DSP: 90­11~.

(11 men, 9women).

(15 whites, 5blacks).

(Average age:.57 years).'

(Age range:42-72 years).

48normoten­sive subjects.

(SSP<150mm Hg,D8P: 80-89).

(47 white, 1black).

(79% male) ...... ~.

(Average age:52 years)Randomizedinto twogroups.

Trcatm0r,t orintervention

1150 mg Naphase: 16placebotablets perday. 2300mg Naphase: 7 (10mmol) slowrelease NaCItablets plus9 placebotablets perday. 4600mg Naphase: 16(10 mmol)slow releaseNaCI tabletsper day.

Interventionphase: 6 (16meq) NaCIcapsules perday ControlPhase: 6placebocapsules perday.

Intake of te3tmaterial

1>:50 mg Naphase: 0 mg

: Na per day.2300 mg Na

: phase: 1610mg Na perday. 4600mg tJaphase: 3680mg Na perday.

2210 mg Naper day.

Base diet

low Na diet(690-1150mg Na perday)monitored by24-hr urinecollectionand dietarycounseling.

low Na diet«805 mgNa per timedovernight, 8~

hr urineexcretionassessedprior toPhase 1 by5consecutiveovernighturinecollectionsbelow 805mg Na).

Other factors

Excluded patientswith renal failure,

; ischaemic heartdisease,cerebrovasculardisease, andthose taking oralcontraceptives orother drugs.Weight increasedas Na increased,but change wasnot significant. 16to 20 (1 moved,3 medicated)controlled bloodpressure by sa.1trestriction alonefor the yearfollowing thestudy (Na of1420 mg per 24hr, SSP of 142mm Hg, DBP of87 mm Hg).

23% of initialparticipantsexcluded for highurine NA levels. 2subjects droppedout. Bloodpressuremeasured every2 weeks (twice ateach visit) 8-hrurine measuredat beginning ofPhase 1 and atend of eachphase. Subjectslost weight andblood pressuredropping duringdiet only phase.

Results

Na in :3.phas~ was1130, 2480,; and4370 mg per 24­hr. SSP in 3phases was 147,155, and 163 mmHg. DBP in 3phases was 91 ,95, :and 100'mmHg. Averagechange in bloodpressure fromlowest to highestNa intake was 16mill Hg S8P and9 mm Hg DBP(p<O.001).

Average SSP 3.6mm Hg higherduring NaGItreatment periodas compared withplacebo period(p<0.001).Average DBP 2.3mm Hg higherduring NaCItreatment periodas compared withplacebo period(p<0.005). 65%and 690/0 ofparticipantsexperienced anincrease of SSPand DBP,respectively,when on NaGIcapsules ascompared withplacebo capsules.

I

f. ssessment andcomments

prossuredifferences werenot affected, bythe order inwhich the Naintake wasaltered.Necessary tohave some salt­free products (i.e.salt-free broad) inorder to reachthe dietary Naintake of 690­1150 mg Na per-jay.

Sound methodologyEstimated 3-6 mm

Hg increase In

SSP and 2-4 mmHg increase inDSP associatedwith 2300 mgincrease in Naintake.

60852 Federal Register J V·oF.. 56~ No.. 229 I Wednesday, No.ve.m,ber 27tr 1991 I Propo.sed Rules:

TABLE' 2.-S0DtUMlHvPERTENSfON STUDiEs-Continued

Reference

-r-Study design, SubJ'ectsand duration

Treatment orintervention

Intake of testmaterial Base diet Other factors

·T-----~____r~-~·_--_·~_·"'"'

Results : Assessment and, comments

NIA. ~••••..••."...... N! I~" ••••••••• ,.......... Normal diets .

: Significant, difference

between twogroups within 4-5days (p<0.001).

Low Na group: Naof 1200 mg perday. Meanarterial bloodpressure fell ffom87 to 81 mm Hg~

, High Na group: Na:of 7750 mg perday. Meanarterial bloodpressure rose·from 86 to 89mm Hg.

Mtabaj~

(19-90).(Ret 80),...,

Smith(1988).

(Ref. 41).....

; CHnicalinterventiontrial.

3 days: control'phase.

7 days: dietphase.

One group onLow Na dietother onnormal dietplus Na asconsommesoup. Bloodpressuredetermineddaily. Nadeterminedby 24-hr Naexcretion.

Crosssectionalstudy. Partof Scottishheart healthstudy. Datacollectedfrom 1984-'1986. Naintakedeterminedby single 24­hr urinecollection.Bloodpressuredeterminedby averageof 2measure­ments.

'30normoten­siveTanzanianblack malesubiectsRandomizedinto twogroups.

7354 subjectsfrom 22districts inScotland(3754 males,3600females).

(Age range:40-59 years)Subjectschosen atrandom.

, High Na group: High Na group:.. 250mmol 5750 mg Na.

Nael perday asconsommesoup.

Control phase:Diet ofunspecifiedNa content.LowNagroup: 1150mg Na perday. High Nagroup:Normal dietofunspecifiedNa content.

Blood pressuredetermined byautomateddevice. Ages ofparticipants notspecified.

Single Nameasurementdoes not assessprevious orhabitual Naintake habits.74°10 responserate, 17,5°/0excluded(generally forfailure to provideurine). 1.6°10excluded due toantihypertensivemedication.Confoundingfa~tors: age t .

body mass index t !

ptlise rate~.

alcot:loLeOAsumption t

potassitWA: intake~

Control diet andnormal diet ofunknown Nacontent.

Single population:High Na dietphase wasexcessively highin NA (7750 mgper day),

Weak t positive Large studycorrelation population.between Na aAd: i Siflgle' communityblood pressure ill" (Scotland).both sexes. Nacorrelation with;SSP 0.025 formales and 0.055;for females. Nacorrelation withDBP 0.026 formales and (J,,052; :for females. Na;not independentfy'significant aftermultivariantanalysis.

Federal Register I Vol. 56, No. 229 / Wednesday, November 27, 1991 I Proposed Ruless=

TABLE 2.-S0DIUM/HvPERTENSION STUDIEs-Continued

6085:

---- .--.---_.- .._--------------- -------.

Reference~tudy designand duratinn

Subjects Treatment orintervention

Intake of testmaterial Base diet Other factors Results Assessment and!

comments

Staessen(1988).

(net 42) ...

Stamler,R.(1989).

(Ref. 70).....

Crosssectionalintclventiontrial. 5 years(1979-1985).flihiSS mediacampaign toavoid saltdirectedmainly atwomen in~:me town.Control townwasobserved.Urinary NAexcretionand bloodpressuredeterminedat beginningand end ofintervention..

Interventiontfial 5 years.

Interventionsubjects:Extensivedietary andlifestylecounselingvarying from2 visits perweek initiallyto 4 visitsper year.

Controlsubjects: 2visits peryear Bloodpressuremeasured 2times peryear Nafrom urinaryexcretionmeasured 1time p~;r

year.

Inter',/entiontown of12,000 andcontrol townof 9,000Belgianinhabj~ants,

2211subjectsexamined (5and 10%randomsample atbaseline incontrol andinterventiontown,respectively;doublad atfollow-up).Previousparticipantsexcludedfrom follow­up. Datafrom 1691subjectsanalyzed.(777 males,733 females)with andwithout 187subjects onantihyperten­sivemedication.

201'-. ...__ ....t}.,.... __:_....

I Iyt-/t:. ~~i I~rvl a

pronesubjects(high normalOBP: 85-89mm Hg) or(high norma!OBP: 80-84plus 10-49%overweight)and/or(rapid restingpulse rate>80 beatsper min).Randomizedinto twogroups (102interventionsubjects and99 controlsubjects).

Interventionto'~'\In (IT):Leaf!ets sentto all hcmes,postersdisplay0d,radio andnewspapefads run.Activesupport fromTownCounci~"

loca! healthofficials andpracbtioners,andinsuranceorganiza­tions. Locaubakers andrestaurantsasked toprepare low­salt foods.Women'sclubs, healtheducationcourses, andchildren'shomeworktargeted.Control town(CT): Saltnotmentionedas a healthhazard.

Intervention...... "'1"'. ~ \~val~. II

Reduce dailyNA intake,2) Reducealcoholintake, 3)Reduceoverweight,4) Increasemoderatephysicalactivity.

l'J,' A.......... Normal diets .

N/A.......•............. Goal of<1800 mgNa per day.13% ofinterventionsubjectsachieved Nainterventiongoal.Average Naintakereduced by24% ininterventiongroup (dropfrom 3980mg/day to3040 mg/day) vs 6%in controlgroup (dropfrom 4300mg/day to4060 mg/day)(p<O.OOl .

Data from teensexcluded (464subjects). Dataexcluded if urinevolume orcreatinineexcretion outsidelimits (50subjects) .. Bloodpressuredetermined a.saverage of 10measurementscollected at 2home visits 2-5weeks apart. Nadetermined by24-hr urineextraction.Baseline andfollow-up datataken on differentsubjects.

Multipleinterventions.Blood pressuremeasure atworksite and atofftce andworksitemeasurementsused forcomparison. 24­hr NA estimatedfrom timed, 8-hrNa excretion andmultipliers.Statisticallysignificantchanges in 3 of 4interventions: NAintake, alcoholintake, andweight reduction.Statisticallysignificantrelationship withblood pressure inone of 4interventions:weight reduction.

The trends in Na,SSP, and DBP inmen and worn,enwere similarbetween the twoto~vns except forNa in womenwhich wassignificantly lowerin the IT than inthe CT. Nachanges rangedfrom +180 mg infemales in CT to-410 mg infema~es in iTSSP changesranged from-4.4 mm Hg inmales in CT to-9.1 mm Hg infemales in IT.DBP changesranged from+1.8 mm Hg inmales in CT to- 2.8 in femalesto IT.

19 010 of controlsubjects and 9 0

/ 0

of interventionsubjectsdevelopedhypertensicn(DBP >90 mmHg ormedication). SBP:Reduction of 2.6mm Hg (from122.5 to 119.8mm Hg) ininterventiongroup vs 1.3 mmHg in controlgroup (from122.7 to 121.5mm Hg). DBP:Reduction of 1.3mm Hg (from82.5 to 81.2) ininterventiongroup vs 0.1 incontrol group(from 82.6 to82.5 mm Hg).Relationshipbetween Na andblood pressurenot independentlysignificant.

Large range ofvariation inresults affectsinterpretation(eg.: change inNa ill females incontrol town wa~

+ 180 ±~210

mg).l'Joindependentassessment otinformationavailable tosubjects incontrol townbet'Neen 1979and 1985regarding Na anhealth hazards,'nconclusive.

Appropriatestatistical toolsused. Lowdropout rate:87% participatirlat least 4 years

60854 Federal Register I Vol. 56, No. 229 I \Vednesday, Novelnber 27, 1991 / Proposed Rules

TABLE 2.-S0DIUM/HvPERTENSION STUDIEs-Continued

---- -_._---~-_._--... -~--_._--_.-

I Ii

,-----_. --_._--------------------_..

f I Study design I Subjects Treatment or Intake of test Base diet Other factors Results Assessment andRe,erence.

1a~~!~atio_~J intervention I material comments

Takemori ~oss 7,441 N/A...................../ N/A.....................

--f--. --------

Normal diet History of being Increase of 2300 Spot urine and(1989). I sectional Japanese

Iaveraging hypertensive or mg Na per day predictive

(Ref. 71 )..... study Na females from I 3720 mg Na on related to equations usedintake 88 urban I per day. antihypertensive increase of 4.5 to estimate 24-hrdetermined and 81 rural

Imedicine not mm Hg SSP Na adds to

by spot municipali- I considered.

I(urban: 4.1 mm uncertainty of

urine. Blood ties including

IConfounding Hg, rural: 4.9 mm results. Single

pressure all I factors: age, Hg) and an population.determined prefectures I height, weight, increase of 1.6by single in Japan. I I potassium. mm Hg DBPmeasure- (3933 urban

I I(urban: 1.2 mm

ment. Data subjects, Hg, rural: 2.0 mmcollected in 3508 rural I Hg).1985. subjects).

II

(3 age groups Ibetween 40 ! Iand 69 years

i I

of age). INone···················1 None..................Trials of Clinical 2182 subjects 7 interventions None ........................... 39% reduction in Authors suggestHyper- intervention with high (3 lifestyle Na at 18 months. that weight losstension trial. normal blood changes for SSP 1.5 mm Hg and Na restrictiorPreven- Investigation pressure. 18 months, lower at 18 are the mosttion of 7 (DBP: 80-89 4 nutrition months (p=O.05). promising(TOHP) nonpharma- mmHg). supplements DBP 0.8 mm Hg nonpharmacologi-Collabo- cological (Age range: for 6 lower at 18 cat interventionsrative interventions 30-54 years) months) and months (p=0.07). for hypertensionRe- in persons randomized 1 control). I control.search with high to 1 of 8 1) weight lossGroup normal blood groups. and(Ab- pressure. exercise, 2)stract) Checked at NA(1991). 6, 12, and restriction, 3)

(Ref. 123) .. 18 months stressinto study. manage-

ment,4)

I calciumsupplemen-

\ tation,5)II magnesium

supplemen-tation,6)potassium /

f·supplemen-tation, 7) fishoilsupplemen-tation,8)

i control.

TABLE 3.-S0DIUM/HYPERTENSION META...ANt~LYSES

enQCOQ1

~~

c.~...a.

~

~

.....c.oCO~

zo<ro3~ro...,

en~

Z~

NN:.c

<?-

~ro0..:Jroen0..~

"<.

4'c-0o

r.J'Jctc...:=dc::~en

--

--

~I~I. O'Ctn-S'...--

Variety of study methodologies (un­standardized).

Variety of study methodologies (un­standardized).

Variety of study methodologies (un­standardized). Authors suggest thatthe effect of moderate dietary saltreduction on mortality from strokeand ischemic heart disease wouldbe substantial. Authors concludedthat the effect of salt reduction onbiood pressure was larger than pre­V!OllS!y reported. Unclear exactlyhow authors controlled for con­founding factors in this study.

12,773 people from Europe,Asia, and the United States.

Deve:oped versus developing commu­nities.

Confounders: Potasz,ium, alcchol, andbody mass linked to Na. Thesevaried between, but not within. thetwo groups. Blacks excluded fromstudy because blood pressure inthese communit;es were i1igher thanfor communities with similar Naintake. The effect was measured inboth developing and developedcommunities.

Not stated ! Some of the trials Vi(;3re not random- Salt reduction lowered blood pressurei2ed which could introduce bias. in persons \;~fith high blood pressure

and in those wi til nQrma! blood pre:3~

sure, In people aged 50-·59. a re~

duction in daily Na !ntake of 50mrncl (about 3 9 salt}, aftsr a fewweeks, lowered SBP by ~n averageof 5 mm Hg and by 7 mm Hg intllose with high blood pressure DBPwas lowemd by about half thisamount

Reference Study design and duration I Subjects I Other factors Results j Assessment and comments-~.. ~~_.~~~.~l·· I

Cutler (1991) , Meta$analysis of 23 ra.ndom.. ITotal: 23 trials with 1536 sub.. Excluded trials with confounded de- Hypertensives: Net reductions in Na I Studies used a variety of different(Ret 94) "............. ized clinical tr!a!s published I jects. Data for. hypert.€.nSive signs. Excluded 2 trials for Na intake.ranged fiom 1290 to 2410 mg. Aver- I, methodologies (unstandardized). Se-

before January 'j 990. Ana- and normotensrve subjects outsido the usual ranges for Na. age pooled reductions in blood pres- 1 lected only randomized trials whichlyzed separately for hyper- analyzed separately and to- f\{1agnitude of results was reduced sure wem 4.9 ;Tim Hg (SSP) and 2.6 II €Iiminates selection bias. Authors in-tensives. normotensives. gather. slightly when invGrse variance mm Hg (DBP), dlGate there is evidence for a dose-and total subjects. weights were included. Ail results Normola,;sives: Net reductions in Na I response relationship between Na

statistically significam eXGcpt for ranged frcm 370 to 39: 0 mg. Aver~ and biood pressure.DBP in normotensives after includ~ age pco:ed reductiOils in blood pres-ing an adjustment for inverse var!- sure were 1. '? m;-n Hg (SSP) and 1.0anca weights. mm Hg (DBP).

Totals: Net reductions in Na rangedi from 370 to 3910 mg. Average

I pooled reductions In. blood pressure

Iwere 2.9 mm I-1g (SSP) and 1,6 mm

t Hg (DBP),Elliott (1991) · ···..·..·•···..1 Me.ta-ana!YSI~£,~.f: 1: Cb.se~a- 12,~03 subjects EXtCIU.'~;e,d stUd;~S that co~pa~:dchYP3r~ Reduct:~n 0,1 ~3~.:) n;9

n

N.:: rel~ted to : va~~e~y of study designs (unstand-(Ref. 97)........................... tronal stud,o:;,,:~ In 16 popula- (70 ...9 men~ 6136 women)........... .ens~\es W.tj1 normO.€i1SlilO;;,. eX- reduction In vDP OT ..,.7 ,11m Hg and I t.tdlzed methodology).

lJons. Only studies that pro- eluded studies that just reported sig- in DBP of 2.0 mm Hg. r~egressionvided 24-hr urine and blood niticance without quantific~tion. Data coefficients S0iY~ewhat larger inpressure data and published corrected for within-individual varia- VJomen than in men,quantitative r.;3gression or bility in Na ("reliability") using INmcorrelation estimates. TERSALT estimate. 2 studies of

only men. 1 study of only women, 2studies of only combined data for

,l v:,0men and men ccmbined. I .' r j r ,C::' Q

f..o,.e " , 2 mm Hg d8C,BD.Se .n blood P,€",,~ur..,

for every 100 rl-If~'ioi (.1'3CreaSe in 24hour NA int2ke. Weak efiect withinpopulations--reduced Na intake re-duced biood pressure sligritly. Un-clear how stud'-,' controlled for con-founding factci"s. One of three stud-ies suggestir.g that Na reduction re-duces SSP and the risk of mortalitydue to h'j'poiton3ion.

Blood pressuro varied according tointake. The cl"jange L1 blood pres­sure was reiativo to tt-Ie age andexisting blood pre~;sure, H appearedthat there was not a lower Nathreshold beIJ';'j ~Nhich no effect wasmeasured,

I

IFrost (1991)·.· 1 Metau~malYs.is of 14 published(Ref. 100)......................... studies of blood pressure

I and Na intake. 24 hour Na

I, excretion. One blood pres­

sure meaSl!F~lment 24 hourintake of NA adjustE1d due to I·I common underestiri1::l1ion.

ILaw (19~1a) \ Meta-ana'y~,js: , 147.000. ,people from 24 com-(Ref. 100)......................... 12 develop1ng rnunltles.

II. 12 developed communities .

Cross sectional studies

Ii

I

Law (1991b) , 1 Meta-analysis of 78 published(Ref. 107)......................... studies that recorded the

: e!1ect of salt restriction oni blood pressure.i Variable duration: Researchers

Isubdivided the data fromstu~ie~. tha~ reC~.i!ed .. bothsubjec,s with hi9n blood

I pressure and subjects \,-,Iith

I normal blood pressure to i

allow separate assessment II of the effect of sait restriceI tior. for each category.

60856 Federal Register / Vol. 56, No. 229 / Wednesday, November 27. 1991 I Proposed Rules

(FR Doc. 91-27168 Filed 1'1-2G-91; 8:45 an1]BILUNG CODE 4160-01-M

21 CFR Parts 5, 20, 100, 101, 105, and130

[Docket No. 91 N-0219]

RIN 0905-AD08

Regulatory Impact Analysis of theProposed Rules to Amend the Foodlabeling Regulations

AGENCY: Food and Drug Administration,I-IHS.ACTION: Regulatory in1pact analysisstatelnent.

SUMMARY: The Food and DrugAdministration (FDA) is publishingherein the regulatory impact analysis(RIA) tha tithas prepared underExecutive Order 12291 and theRegula tory Flexibility Act (Pub. L. 96­354) on the cos ts and benefits of thefood labeling regulations that 'FDA iscurrently proposing to amend. FDA isissuing these proposals (publishedelsewhere in this issue of the FederalRegister) in response to the NutritionLabeling and Education Act of 1990 (the1990 amendments) and as part of theSecretary of Health and ffumanServices' (the Secretary's) food labelingreform initiative. The agency hasprepared this comprehensive RIAdocument for these proposals because,when taken together, they constitute amajor rule.DATES: Written comments by February25,1992.ADDRESSES: Submit written commentsto the Dockets Management Branch(HFA-305), Food and DrugAdministration, Rm. 1-23, 12420Parklawn Dr., Rockville, MD 20857.Comments should be identified with thedocket number found in brackets in theheading of this document.FOR FURTHER INFORMATION CONTACT:Richard A. \Villiams, Jr., Center for FoodSafety and Applied Nutrition (HFF-303),Food and Drug Administration, 200 C St.SW., Washington, DC 20204, 202-485­0271.SUPPLEMENTARY INFORMATION: FDA ispublishing herein its RIA of theproposed rules to amend the foodlabeling regulations. This documentanalyzes both the cos ts and the benefits,including the impact on sn1all -businesses, of FDA's proposals(published elsewhere in this issue of theFederal Register) to reform the foodlabel in response to the 1990amendrnents and the Secretary's foodlabeling initiative. This analysis was

prepared by the Economics Section ofthe Office of Compliance in FDA'sCenter for Food Safety and AppliedNutrition (CFSAN).

The food labeling reform initiative.taken as a \vhole, will have associa tedcosts in excess of the $100 millionthreshold that defines a major rule.Therefore, in accordance with ExecutiveOrder 12291 and the Regula toryFlexibility Act (Pub. L. 96-354), FDA hasdeveloped one conlprehensive RIA thatpresents the costs and benefits of all ofthe food labeling proposals takentogether. FDA requests COffilnents on theRI.t\.

I. Introduction

The 1990 amendments amend theFederal Food, Drug, and Cosmetic Act(the act) to expand the coverage ofnutrition labeling to all food products(except Ineat and poultry), produce moreingredient labeling, regulate healthclaims, and standardize nutrient contentclaim definitions and serving sizes. The1990 amendments require that thenutrition information on both the foodlabel and on eating establishmentmenus be readily understandable by thepublic. These changes to the food labelare the most cOlnprehensive changes tobe proposed in 53 years. FDA hasproposed implementing regulations forthe 1990 amendments and estimated thecosts and benefits of the proposedchanges and regula tory options withinthe act. However, even before the 1990amendments were enacted FDAbelieved that the food label could beimproved and \vas engaged in proposinga series of similar regula tions.

In order to evaluate the need forFederal intervention, FDA examined themarket for food label information andfound that less than the optimal amountof nutrition information was beingproduced because consumers cannot,independently, determine the nutritionalquali ty of food, thus leading toinsufficient incentives for manufacturersto reveal the nutrient content of theirproducts or produce nutritious food.FDA undertook two studies to determinethe costs and benefits of these proposedregulations, by engaging a contractor,Research Triangle Institute (RTI). Thesestudies were done over a period of 3years under the direction of theEconOlnics Section of CFSAN.

A. Costs of the 1990 AnlendI17ents

rrhe cost study consisted of bothinterviews with food manufacturers anda mailed survey. l'he result was ageneric model which can be applied toany reguIation mandating a labelchange. Categories of costs includeadministra tive, analytical, printing,

inventory, and 1'eforn1ulution.Administrative costs are managenlentcosts \vhich are often high because ofthe prominence of the food label as anadvertising tool for packaged foods.Analytical costs are costs of testingproducts for nutrient con1position toconlply with labeling provisions.Printing costs are the costs of printingnew labels which may be either glue-onlabels or the food package itself. Thesecosts nlay include redesign costs whereextensive labeling changes arcundertaken. In the model, estimates ofprinting costs take into account normalfirm relabeling.

Inventory costs are the costs ofdisposal of existing labels where firmshave inventories that outlast thecompliance period, Le., the period oftime between issuance of a final ruleand its effective date. Inventories oflabels, both glue-on labels andpackages, range from only a few monthsto well over 10 years in the foodindustry. The last cost categoryreformulation includes the costs ofreformulating products and introducingnew ones in response to labelingregulations and market testing thoseproducts. No estimate of these costs isgiven because they depend on marketingdecisions and are impossible to predict.Moreover, they do not result directlyfrom these proposed rules. Regardless,FDA expects a substantial benefit to bederived from such reformulations, whichare likely to make foods more nutritious.In all cost categories, exceptadministrative costs, the costs ofrelabeling products produced andlabeled in foreign countries cannot beseparated from those produced andlabeled domestically. Thus,administrative costs considered aredomestic costs only, and printing,inventory, and analytical costs areconsidered multinational.

FDA estimates that about 17,000domestic food manufacturers and257,000 labels will be affected by theregulations promulgated in response tothe 1990 amendments. In addition,approximately 96,000 food service firmsmight be required to alter their nlenus ifthey are not in compliance with healthclaims or descriptors regulations. Themajority of the costs will occur in thefirst year. Recurring costs are assumedto continue 20 years into the future and

-are discounted back to the present at ara te of 5 percent.

The individual regula tions may bedivided into the following separablecategories: (1) Mandatory ingredientlabeling for standardized foods andcertified colors; (2) "voluntary" (seesection IILE. of this document) labeling