Br Heart J 1995;74:47-52

Insulin resistance syndrome in postmenopausalwomen with cardiological syndrome X

Ian F Godsland, David Crook, John C Stevenson, Peter Collins, Giuseppe M C Rosano,Belinda Lees, Mandeep Sidhu, Philip A Poole-Wilson

Wynn Institute forMetabolic Research,London andDepartment ofCardiac Medicine,National Heart andLund Institute,LondonI F GodslandD CrookJ C StevensonDepartment ofCardiac Medicine,National Heart andLung Institute,LondonP CollinsGM C RosanoP A Poole-WilsonWynn Institute forMetabolic Research,LondonB LeesM SidhuCorrespondence to:Dr I F Godsland, WynnInstitute for MetabolicResearch, 21 WellingtonRoad, London NW8 9SQ.Accepted for publication14 February 1995

AbstractObjective-To determine whether post-menopausal women with cardiologicalsyndrome X (chest pain and abnormalexercise electrocardiogram despite nor-mal coronary angiography) exhibit dis-turbances in the full range of proposedcomponents of the putative "insulinresistance syndrome".Patients and methods-20 post-menopausal women with syndrome Xand 20 healthy controls each underwentmeasurement of insulin resistance (byminimal model analysis of the intra-venous glucose tolerance test), lipid,lipoprotein, and apolipoprotein concen-trations, a range of haemostatic vari-ables, serum uric acid concentration, andcentrality of body fat distribution (bydual energy x ray absorptiometry).Results-Women with syndrome X hadhigher fasting triglyceride concentrationsthan controls (median: 1-60 v 1*02mmolIl, P < 0.05). Concentrations ofhighdensity lipoprotein cholesterol werelower (1.33 v 1'61 mmol/l, P < 0-05) aswere those of the high density lipoproteinapolipoproteins AI and All. Insulin and Cpeptide responses to the intravenous glu-cose tolerance test were higher (27.6 v19*8,uUlmllmin, P < 0 01; 101 v 72pmolmllmin, P < 0 05, respectively), andinsulin sensitivity was lower (1.89 v 3*09minI,uUIml, P < 0.05). There were, how-ever, no significant differences betweenother proposed components ofthe insulinresistance syndrome (blood pressure,glucose tolerance, proportion of centralbody fat, serum uric acid concentration,and plasminogen activator inhibitor-iactivity). Antithrombin III activity washigher in women with syndrome X (121 v113%, P < 0.01).Conclusions-Women with syndrome Xtend to be insulin resistant and have lipidand lipoprotein abnormalities, but do notexhibit all characteristics of the insulinresistance syndrome. Such variation incorrelated risk factors is consistent withunderlying heterogeneity in the insulinresistance syndrome and cardiologicalsyndrome X.

(Br HeartJ 1995;74:47-52)

Keywords: insulin resistance, postmenopausal women,syndrome X

The term "syndrome X" was introduced in1973 by Kemp' to acknowledge uncertaintiesover the aetiology of angina-like chest painaccompanied by angiographically normalcoronary arteries. Conventional diagnosticcriteria also include some further evidence ofmyocardial ischaemia, such as increasedmyocardial lactate production, reduced coro-nary sinus oxygen saturation, ST depressionon exercise electrocardiogram, and impairedvasodilator reserve.2 Inconsistencies in theexpression of these features, however, indicatea degree of heterogeneity in the condition.'

There is increasing evidence that cardiolog-ical syndrome X is accompanied by distur-bances in metabolic risk factors for vasculardisease, notably insulin resistance.45 Patientswith syndrome X might therefore exhibit theputative "insulin resistance syndrome" (itselfreferred to as syndrome X or metabolic syn-drome X).6 This syndrome has been invokedto account for reports of clustering amongincreased concentrations of insulin, glucose,and triglyceride, reduced concentrations ofhigh density lipoprotein (HDL), and raisedblood pressure.7 Experimental studies haveprovided evidence for biochemical and physio-logical links between insulin resistance andother risk factors for cardiovascular disease,thus supporting the hypothesis that insulinresistance is the abnormality that underliesthis syndrome of metabolic disturbance.89We have previously demonstrated these

disturbances in men with syndrome X,"0 butother studies have suggested less consistentmetabolic characteristics.45 There is alsouncertainty over the importance of impairedglucose tolerance as a component of theinsulin resistance syndrome, and controversyover the extent to which changes in bloodpressure are directly related to disturbances ininsulin metabolism.1' Furthermore, new can-didates for inclusion in the insulin resistancesyndrome have been proposed, includingincreased concentrations of uric acid andincreased activity of plasminogen activatorinhibitor-i (PAI-1).6 A central or android pat-tern of fat distribution may also be impor-tant.'2 These variables have yet to be exploredin detail in patients with cardiological syn-drome X, as have intercorrelations betweencandidate components of the insulin resis-tance syndrome. In particular, there is a lackof metabolic information concerning syn-drome X in women. Women comprise themajority of sufferers from the condition"3 andan involvement of the menopause is suggestedby their relatively high mean age. 4

47

Godsland, Crook, Stevenson, Collins, Rosano, Lees, Sidhu, Poole- Wilson

In the present study, we undertook todetermine whether insulin resistance syn-drome was present in 20 postmenopausalwomen with syndrome X by comparing themwith an age and weight matched group ofhealthy controls.

Patients and methodsPATIENTSWomen with syndrome X were recruited froma trial of transdermal oestrogen treatment.Inclusion in this trial required a history ofanginal chest pain, a treadmill or bicycle exer-cise test with >1 mm ST depression in anylead, and a coronary angiogram considerednormal-that is, not showing more than 30%stenosis, by two experienced observers. Leftventricular hypertrophy, mitral valve defect,or any other structural abnormality as diag-nosed by chest radiograph or echocardiogramconstituted exclusion criteria, as did any exist-ing electrocardiogram (ECG) abnormalitieslikely to interfere with interpretation of anECG during treadmill exercise testing, forexample, left bundle branch block. Womenwith blood pressure greater than or equal to160/100 mm Hg were also excluded. Allrecruits underwent evaluation for vasospasticangina (hyperventilation or ergonivine test)and all were negative. Four women hadreversible perfusion defects on thalium scan-ning. Postmenopausal status (natural or surgi-cal) was confirmed by amenorrhoea of at least6 months duration, a follicle stimulatinghormone level >40 lUll, and an estradiol con-centration <100 pmol/l. No hormone replace-ment treatment or cardiac medication (exceptglyceryl trinitrate as rescue medication) wasallowed for 4 weeks before baseline studies.Of the 26 women who entered the trial, threewere not white, two were grossly obese(>145% ideal body weight), and one had noinsulin measurements during her intravenousglucose tolerance test (IVGTT). Studies car-ried out before treatment in the remaining 20patients are evaluated in the present analysis.

Twenty healthy, non-hypertensive post-menopausal women attending a menopauseclinic were selected to provide a control groupof comparable age and weight to the patientgroup. Criteria for postmenopausal statuswere the same as for the women in the syn-drome X group, but cigarette smokers wereexcluded. No drugs likely to affect carbohy-drate or lipid metabolism were taken.

These studies were approved by the ethicscommittees of the Wynn Institute and theNational Heart and Lung Institute and eachparticipant gave written, informed consent.

PROCEDURESAll women were instructed to consume morethan 200 g/day of carbohydrate' for 3 daysbefore their tests. Tests began between 0900and 1000 after an overnight fast. Age, height,weight, time since menopause, and details ofalcohol consumption, smoking, family historyof diabetes and heart disease, and number ofpregnancies were recorded. Exercise habit was

defined as 'aerobic' for those women whoundertook, at least three times a week for 20min or more, exercise requiring a noticeableincrease in breathing rate, 'non-aerobic' forthose who described themselves as taking onlymoderate exercise, and 'none' for those whodescribed themselves as taking no exercise.Time since menopause for women who hadundergone a hysterectomy was taken as thetime since onset of menopausal symptoms.Numbers of close relatives (parents or sib-lings) with heart disease or diabetes wererecorded. Blood pressure was measured afterpatients had rested in a semirecumbent posi-tion for 15 min. -Indwelling cannulae wereinserted into an antecubital vein in each arm,one for glucose administration and the otherfor sampling. Blood samples were taken formeasurement of fasting plasma glucose,insulin, C peptide, serum lipid, and lipopro-tein concentrations and factors of the coagula-tion and fibrinolytic systems. In two cases nosamples were taken for coagulation and fibri-nolytic assays due to technical problems. Anadditional sample was taken after 10 min forreplicate measurement of fasting plasma glu-cose, insulin, and C peptide concentrations.An intravenous glucose injection (0 5 g glu-cose/kg body weight as a 50% (wt/vol) solu-tion of dextrose) was given over 3 min andsamples were taken at 3, 5, 7, 10, 15, 20, 30,45, 60, 75, 90, 120, 150, and 180 min.Regional distribution of fat and lean tissuewas measured by dual energy x ray absorp-tiometry using a whole body scanner (DPX:Lunar Radiation, Madison, Wisconsin).

LABORATORY PROCEDURESPlasma glucose was measured by a glucoseoxidase technique.'5 Plasma insulin and Cpeptide were measured using double antibodyradioimmunoassays (Guildhay, Guildford,UK). Serum triglycerides, total cholesterol,and uric acid were measured using enzymaticmethods. HDL cholesterol was measuredafter heparin manganese precipitation of otherlipoproteins'6 and HDL subfractions after fur-ther precipitation with dextran sulphate.'7Very low density lipoprotein (VLDL) choles-terol and triglyceride were measured afterpreparative ultracentrifugation at a solventdensity of 1 006 g/l, and low density lipopro-tein (LDL) cholesterol was estimated as thedifference between total cholesterol and thesum of HDL and VLDL cholesterol.Apolipoproteins AI, AII, and B were mea-sured by immunoturbidimetric assays,'8 andLp(a)' lipoprotein by an enzyme linkedimmunosorbent assay (Biopool AB, Umea,Sweden). Plasma fibrinogen, factors VIIa andX, protein C and protein S were measured bynephelometric techniques on an automatedcoagulation analyser (ACL 100;Instrumentation Laboratory, Milan, Italy).Antithrombin III (AT III) and plasminogenwere measured by enzymatic methods(Chromogenix, M6lndal, Sweden) on a dis-crete clinical analyser (Cobas Mira, RocheDiagnostics, Switzerland). Tissue plasmino-gen activator and PAI-1 were measured by

48

Insulin resistance syndrome in postmenopausal women with cardiological syndromeX

Table I Group characteristics in women with syndromeX and controls

nAge (years)Body mass index (kg/M2)Time since menopause (years)No of patients who underwent hysterectomy (%)*Alcohol consumption (Units/week)Cigarette smokers (%)

NeverPreviousCurrentt

No of patients taking exercise (%)NoneNon-aerobicAerobict

Family history of diabetes mellitus (%)tFamily history of heart disease (%)No of pregnancies (%)None1-2> 2

Women withsyndromeX

2058-0 (47-65)26-3 (22 4-30.4)7 5 (0-35)

402-0 (0-21)

453520

4050102560

205525

Controls

2057 0 (46-65)25-6 (22-6-27-7)6-5 (0-26)

153-5 (0-21)

70300

2535405

50

203050

Values are median (range) unless otherwise specified. *The greater proportion of women withsyndrome X who had a hysterectomy was of borderline significance (P = 0 08). tProportions ofcurrent cigarette smokers and women with a family history of diabetes were significantly higher inthe syndrome X group, and the proportion of women taking aerobic exercise significantly lower(P < 0 05) (Mann-Whitney U test).

manual enzymatic methods (Chromogenix)and fibrinopeptide A by radioimmunoassay(Byk-Sangtec, Dietzenbach, Germany).

DATA ANALYSISAndroid fat was estimated by dual energyx ray absorptiometry as the sum of fat mass inpredefined waist and subscapular areas, andgynoid fat as the sum of hip and thigh fat.19The ratio of fat in the android relative to thegynoid region was calculated, and proportionsof android and gynoid fat were expressed aspercentages of total body fat. Mean fastingplasma glucose, insulin and C peptide con-centrations were derived from the two basalmeasurements. Incremental glucose, insulin,and C peptide areas (the area between thefasting level and the IVGTT concentrationprofile) were calculated by the trapezium rule.Insulin sensitivity (Si) and glucose effective-ness were determined by minimal modelanalysis of the IVGTT glucose and insulinconcentration profiles.20 The relatively highglucose dose (05 g/kg) employed in the pre-sent studies allows for effective model identifi-cation,21 and insulin sensitivity estimates that

Table 2 Fasting serum lipid, lipoprotein and apolipoprotein concentrations in womenwith syndrome X and controls

Women with syndromeX Controls

n 20 20Triglycerides (mmol/l) 1-60 (0 68-4 60) 1-02 (0 57-3 50)*Total cholesterol (mmolIl) 5 99 (4 52-8-58) 5-76 (4-37-7-51)LDL cholesterol (mmol/l) 4-35 (2 70-6 46) 3-76 (2-76-5 48)HDL cholesterol (mmol/l) 1-33 (0 85-2 55) 1-61 (088-2 94)*HDL, cholesterol (mmol/l) 0 40 (0-07-1-40) 0 49 (0-10-1-69)HDL, cholesterol (mmol/l) 0 94 (0-68-1-25) 1-06 (0-78-1-33)VLDL cholesterol (mmol/l) 0 43 (0-08-1-71) 0-20 (0-08-1-62)VLDL triglyceride (mmol/l) 0-53 (0-17-3-24) 0-39 (0-15-2-23)VLDL cholesterol/triglyceride molar ratio 0-52 (0-32-0-81) 0 57 (0 40-0 87)Apolipoprotein AI (mg/dl) 132 (97-200) 149 (110-196)*Apolipoprotein AII (mg/dl) 38 (29-52) 43 (36-55)*Apolipoprotein B (mg/dl) 74 (48-108) 73 (43-98)Lp(a) lipoprotein (mg/dl) 8-9 (0-85) 7-6 (0-55)Values are median (range). *P < 0 05 v women with syndrome X (Mann-Whitney U test).LDL, low density lipoprotein; HDL, high density lipoprotein; VLDL, very low density lipoprotein.

correlate well with those derived from the gly-caemic clamp (r = 0 92).22 Insulin deliverycharacteristics were evaluated using the mini-mal model of post-hepatic insulin delivery,which provides measures of fractional insulinelimination rate (inversely related to theplasma insulin half-life), and measures of theresponsiveness of first and second phaseinsulin delivery to glucose (V01 and 02 respec-tively). A combined model of insulin and Cpeptide delivery2324 was used to quantify thefractional insulin elimination rate, a measureof the fraction of newly secreted insulin thatpasses out of the liver, and incremental pan-creatic insulin secretion during the IVGTT.Models describing insulin delivery have beenevaluated previously.212325

Differences between the groups wereassessed by the Mann-Whitney U test.Relations between individual components ofthe insulin resistance syndrome were exploredusing Spearnan's correlation. Statisticalanalyses employed BMDP statistical software(Los Angeles, California).

ResultsThe two groups did not differ significantly inage, body mass index, time since menopauseor alcohol consumption, nor in the percentageof women with a family history of heart dis-ease or in the number of pregnancies (table1). Four women with syndrome X were cur-rent smokers compared with none in the con-trol group. The percentages of women whohad undergone hysterectomy and who had afamily history of diabetes were higher inpatients with syndrome X and the proportionof women who took aerobic exercise waslower.

There were no significant differencesbetween women with syndrome X and con-trols in android to gynoid fat ratio (median:0 91 v 0 90, respectively), percentage ofandroid fat (32-0 v 32-6), percentage ofgynoid fat (35 5 v 34-5), diastolic or systolicblood pressure (70 v 72 and 130 v 120 mmHg), or serum uric acid concentrations (220 v218 ,mol/1). Serum triglycerides were 57%higher in patients with syndrome X than inthe controls and VLDL cholesterol andtriglyceride were appropriately increased(table 2). HDL cholesterol was 17% lower,and apolipoprotein AI and apolipoprotein AIIwere 11% lower than in the control group.The lower HDL cholesterol levels werereflected in lower values for HDL2 and HDL,cholesterol. Women with syndrome X hadnormal concentrations of total and LDL cho-lesterol, apolipoprotein B, and Lp(a) lipopro-tein.

There were no differences between womenwith syndrome X and controls in fastingplasma concentrations of glucose (median:5'29 v 5 10 mmol/l, respectively), insulin(0-065 v 0-065 pmol/ml) or C peptide (0 44 v0 39 pmol/ml). Insulin and C peptide areas ofthe IVGTT were respectively 39% and 40%higher in patients with syndrome X than incontrols and Si was 39% lower (table 3). The

49

Godsland, Crook, Stevenson, Collins, Rosano, Lees, Sidhu, Poole-Wilson

Table 3 Glucose, insulin, and C peptide responses to the intravenous glucose tolerancetest (IVG7T) and model derived variables in women with syndrome X and controls

Women with syndromeX Controls

n 20 20IVGTT incremental areas

Glucose (mmol.l-'.min) 528 (334-871) 550 (205-853)Insulin (pmol.ml-'.min) 27-6 (13-1-42-4) 19-8 (6 0-82-2)tC peptide (pmol.ml-'.min) 101 (50-190) 72 (35-201)*

Minimal model of glucose disappearanceSg (min-') 1-70 (0 10-3 22) 1-46 (0 50-6 05)Si (min-'.puU-'.ml) 1-89 (0-06-9-15) 3 09 (0-68-7 91)*

Post-hepatic insulin delivery modelni (min- ') 0-144 (0 040-0 257) 0-151 (0 022-0 240)0, (uU.ml-'.min.mg-'.ml) 4-64 (1-28-13-78) 3-36 (1 99-21-14)02 (U.ml-'.min-2.mg-'.ml) 10-9 (3-9-45 2) 9-2 (0-9-18 8)

Pancreatic insulin secretion modelF 0-98 (0-42-2-19) 0 94 (0-54-1-69)k- (min-') 0-080 (0 035-0-240) 0 110 (0-016-0-169)IS,,, (pmol.ml- '.min) 2-25 (1-36-4-84) 1-67 (0-21-6-45)

Values are median (range). *P < 0 05; tP < 0 01 (Mann-Whimey U test). Sg, glucose effective-ness; Si, insulin sensitivity; n,, fractional insulin elimination rate (post-hepatic insulin deliverymodel); 01, first phase insulin delivery to glucose; 02, second phase insulin delivery to glucose; F,fractional hepatic insulin throughput index; ISn,,, net increment in pancreatic insulin secretionduring the IVGTT; k, fractional insulin elimination rate (pancreatic insulin secretion model).

Table 4 Factors of the coagulation andfibrinolytic systems in women with syndromeXand controls

Women with syndromeX Controls

n 18 20Fibrinogen (mg/dl) 338 (236-417) 347 (222-518)Factor VII (%) 120 (82-169) 126 (82-156)FactorX(%) 119 (76-142) 112 (94-143)Protein C (%) 124 (98-148) 124 (71-169)Protein S (%) 118 (63-165) 104 (65-196)Antithrombin III(%) 121 (103-130) 113 (97-134)*Plasminogen (%) 108 (90-130) 114 (85-136)Plasminogen activator inhibitor-I (AU/ml) 8-2 (1-0-38-2) 5-2 (1-0-29-2)Tissue plasminogen activator (IU/ml) 1-35 (0 20-4 80) 1-05 (0 30-2 60)Fibrinopeptide A (ng/ml) 9.7 (3-7-79-6) 14-0 (3 1-89 0)

Values are median (range). *P < 0 01 (Mann-Whitney U test).

figure shows individual values for the Si index.Despite the significant difference in Si (P <005) there was still appreciable overlapbetween the two groups, and this was alsoapparent with the other metabolic variablesthat differed significantly (data not shown).One patient with syndrome X had unusuallyhigh Si. Exclusion of this patient's results didnot affect the significance of the difference.The higher second phase post-hepatic insulinresponsiveness ((p2) and incremental pancre-atic insulin secretion of patients with syn-

o drome X were of borderline significance (P =008). Glucose responses during the IVGTTdid not differ between the groups. AT III was7% higher in women with syndrome X than in

80

H_ a

2 - ca

Insulin sensitivity (S) inpostmenopausal womenwith cardiologicalsyndrome X (.) (n = 20)and in age and weightmatched healthy controls(:) (n = 20). The medianvalue is indicated by thehorizontal line in eachplot.

controls (table 4), otherwise there were nosignificant differences between coagulationand fibrinolytic factors.

Repeating these comparisons in analyses inwhich current smokers (n = 4), women whohad undergone hysterectomy (n = 11), orthose with a family history of diabetes (n = 6)were excluded did not affect the significanceof the differences seen. Exclusion of those tak-ing aerobic exercise (n = 10) reduced somesignificance levels, but in each of these com-parisons the magnitude of the differencebetween the groups remained the same. Forexample, median Si was between 36 and 44%lower, triglycerides between 48 and 75%higher, and HDL cholesterol between 13 and24% lower in the remaining women with syn-drome X.

Correlations between potential compo-nents of the insulin resistance syndrome wereexamined in the syndrome X and controlgroups separately. The magnitude of the cor-relations in each group were generally similar(results not shown), therefore these relationswere evaluated in the study group as a whole.There were significant associations amongpercentage of android fat, serum triglycerideand HDL cholesterol concentrations, andinsulin response to the IVGTT (table 5). Sialso correlated significantly with triglyceridesand HDL, although the association withproportion of android fat was of borderlinesignificance (r = -0-29, P = 006). PAI-1concentrations decreased with increasing Siand HDL cholesterol concentration and rosewith increasing percentage of android fat (r =0-31, P = 0 05). The only significant correlateof serum uric acid concentration was PAI-1activity. Systolic blood pressure correlatedwith IVGTT insulin response, but only withborderline significance (r = 0-32, P = 0-07).Otherwise there were no associations betweenblood pressure and components of the insulinresistance syndrome. Interrelationshipsbetween Si and several other variables wereexamined. There was a negative associationbetween Si and time since menopause (r =037, P < 0 05) while no such association wasseen with chronological age (r = 0 15, P >0 3). AT III levels correlated with insulinresponse to the IVGTT (r = 0 40, P < 001)and Si (r = 0-31, P = 006) and decreasing Siwas associated with increasing activity of fac-tor X. (r = - 0 33, P < 005).

Table 5 Univariate correlation coefficients (Spearman) among components of the insulin resistance syndrome for womenwith syndrome Xand controls (n = 40)

PlasminogenAndroid Insulin response Insulin activator Uricfat (%) Triglycerides HDL to the IVGTT sensitivity inhibitor-' acid

Triglycerides 0.43tHDL cholesterol - 0-33* 0-64tInsulin response to the IVGTT 0-36* 0-38* -0-31*Insulin sensitivity -0-29 -0-41t 0-38* -0-55tPlasminogen activator inhibitor-i 0-31 0-18 0-32* 0-24 -0-34*Uric acid 0-10 -0-17 0-05 0-16 -0-06 0-38*Systolic blood pressure -0 00 014 0-00 0-32 -0-00 0-28 0-03

*P < 0-05; tP < 0-01; tP < 0-001 (Spearman test). HDL, high density lipoprotein; IVGTT, intravenous glucose tolerance test.

10

9

8

7-

E6

,5aC-A

50

Insulin resistance syndrome in postmenopausal women with cardiological syndromeX

DiscussionThis study shows that postmenopausalwomen with syndrome X tend to have insulinresistance, hyperinsulinaemia, increasedtriglyceride concentrations, and reducedHDL cholesterol concentrations comparedwith those of healthy postmenopausal women.The two groups were of comparable age andbody mass index, although there were differ-ences in other characteristics. An increasedincidence of hysterectomy has previously beenfound in women with syndrome X'4 and thefewer women taking aerobic exercise is to beexpected. Cigarette smoking was an exclusioncriterion for the studies from which the con-trol group was drawn, so no inferences can bemade regarding the difference in smokinghabit between the two groups. The differencein family history of diabetes is unexpected andmay be worth examining in larger samples.Nevertheless, in subgroup analyses it wasapparent that the metabolic differencesbetween the groups were present regardless ofthese differences in group characteristics.

While three of the characteristics of theinsulin resistance syndrome, insulin resis-tance, high triglycerides, and low HDL cho-lesterol, were present in the cardiologicalsyndrome X group, blood pressure, glucosetolerance, PAI-1 activity, uric acid concentra-tion, and body fat distribution did not differfrom those of healthy controls. We have previ-ously observed similarly reduced Si, increasedtriglycerides, and reduced HDL cholesterol inmen with syndrome X, but in that study thepatients had significantly higher systolic bloodpressure.10 Other studies have examinedmixed groups of men and women. Wheremeasured, Si in those with syndrome X wasalso about 40% lower than in healthy con-trols.526 Glucose tolerance test studies have,with one exception,27 been consistent with thisfinding.428 Most studies show increasedtriglycerides and reduced HDL cholesterol,although these differences are generally lessprofound than the differences in insulinrelated variables. Where reported, most studiesalso show little difference in body fat distribu-tion or blood pressure. Small numbers insome of the studies described might underliethe inconsistent differences in lipid andlipoprotein concentrations reported, but theconsistent lack of a difference in blood pres-sure suggests that insulin resistance and raisedblood pressure may not be linked in thiscondition.

Intercorrelations between components ofthe insulin resistance syndrome have not beenexplored in previous studies of cardiologicalsyndrome X. We found significant associa-tions between insulin resistance, increasedtriglycerides, reduced HDL cholesterol, andproportion of android fat. The inclusion ofandroid fat in this cluster is noteworthy aswomen with syndrome X did not have anincreased proportion of android fat, suggest-ing that while android fat is an important cor-relate of metabolic disturbance it is not afactor in the differences seen between healthypostmenopausal women and those with syn-

drome X. Increased PAI- 1 activity was associ-ated with low HDL cholesterol, low Si and anincreased proportion of android fat, as wouldbe expected for a component of the insulinresistance syndrome. There were no associa-tions, however, between PAI-1 and triglyc-erides or the insulin response to the IVGTT.PAI-1 activity was higher in women with syn-drome X, but this was not statistically signifi-cant: possibly larger group sizes would haverendered this difference significant. Other fac-tors of the coagulation and fibrinolytic sys-tems have not previously been reported inwomen with syndrome X. The importance ofthe raised AT III activities in women withsyndrome X is difficult to assess: there seemsto be a biphasic association between AT IIIactivities and vascular disease, low and highactivities being associated with increasedrisk,29 but there are no reports of increasedthrombotic tendency in patients with cardio-logical syndrome X. The positive associationsamong AT III, insulin resistance, andincreased insulin levels suggest that increasedAT III activities might be a candidate for theinsulin resistance syndrome. Despite findingno significant differences in the activity offactor X between patients with syndrome Xand controls, we found a positive associationbetween Si and factor X activity. We are notaware of any previous reports linking insulinresistance with increased AT III or factor Xactivities.A heterogeneous manifestation of distur-

bances therefore seems to be associated withthe insulin resistance syndrome among thewomen who we studied, with a core of correla-tions among Si, triglycerides, and HDL cho-lesterol, probable involvement of android fatlevels and PAI-I activities, and no involve-ment of uric acid or blood pressure. Whetherthese differences and associations are suffi-cient for us to conclude that women with car-diological syndrome X manifest the insulinresistance syndrome is arguable. This uncer-tainty is emphasised by the considerable over-lap in individual values between the groups,even in those components of the insulin resis-tance syndrome that differed significantly. Aserious shortcoming in many discussions ofthe insulin resistance syndrome has been alack of diagnostic criteria: while correlationsand differences in median values may behighly suggestive with regard to groups as awhole, they are insufficient to confer a diagno-sis on an individual. Differentials betweeninsulin resistance as it relates to the glucoregu-latory, antilipolytic, and haemodynamiceffects of insulin may yet undermine the cen-tral position ascribed to insulin resistance inthe aetiology of correlated metabolic distur-bances. Furthermore, it is becoming increas-ingly apparent that insulin resistance can itselfbe induced by many of the disturbances thatare typical of the syndrome, making it difficultto justify a strict causal chain from insulinresistance to its various manifestations.9

In general, the prognosis of patients withcardiological syndrome X is good,'14 callinginto question a role for insulin resistance in

51

Godsland, Crook, Stevenson, Collins, Rosano, Lees, Sidhu, Poole-Wilson

the aetiology of coronary artery disease.Follow up studies of patients with syndromeX may have included a substantial proportionwho were not insulin resistant. Alternatively,the long-term effects of insulin resistanceassociated with vascular dysfunction may dif-fer from those of insulin resistance associatedwith more general, atherogenic metabolicderangement; moreover, a certain minimumcomplement of metabolic disturbances associ-ated with insulin resistance may be needed toinfluence long-term health. To resolve theseissues, future studies will need to pay particu-lar attention to potential heterogeneity in theinsulin resistance syndrome and cardiologicalsyndrome X.

We thank Dr Philip Sarrel for his help in initiating these studies,Drs David Lindsay, Nicholas Peters, David Lefroy, MichaelMarsh, and Sovra Whitcroft for clinical support, AnthonyProudler, Dr Carl Felton, Melek Worthington, and MaireCullinan for technical assistance, and Mary Miller and LeeMayne for nursing assistance. Financial support for this studywas provided by Ciba-Geigy Pharmaceuticals, Novo NordiskPharmaceuticals, the Heart Disease and Diabetes ResearchTrust, and the Rosen Foundation.

1 Kemp HG. Left ventricular function in patients with theanginal syndrome and normal coronary arteriograms.Am J Cardiol 1973;32:375-6.

2 Hutchinson SJ, Poole-Wilson PA, Henderson AH. Anginawith normal coronary arteries: a review. Q J Med 1988;72:677-88.

3 Cannon III RO, Camici PG, Epstein SE. Pathophysio-logical dilemma of syndrome X. Circulation 1992;85:883-92.

4 Dean JD, Jones CJH, Hutchinson SJ, Peters JR,Henderson AH. Hyperinsulinaemia and microvascularangina ("syndrome X"). Lancet 1991;337:456-7.

5 Botker HE, M0ller N, Ovesen P, Mengel A, Schmitz 0,0rskov H, et al. Insulin resistance in microvascularangina (syndrome X). Lancet 1993;342:136-40.

6 Reaven GM. Role of insulin resistance in human disease(syndrome X): an expanded definition. Annu Rev Med1993;44:121-3 1.

7 Moller D, Flier J. Insulin resistance-mechanisms, syn-dromes, and implications. N Engl J7 Med 1991;325:938-48.

8 Reaven G. Banting lecture 1988: role of insulin resistance inhuman disease. Diabetes 1988;37: 1595-607.

9 Frayn K. Insulin resistance and lipid metabolism. CurrentOpinion in Lipidology 1993;4:197-204.

10 Swan JW, Walton C, Godsland IF, Crook D, Oliver MF,Stevenson JC. Insulin resistance syndrome as a featureof cardiological syndrome X in non-obese men. Br HeartJ 1994;71:41-4.

11 Jarrett RJ. In defence of insulin: a critique of syndrome X.Lancet 1992;340:469-71.

12 Bjorntorp P. Metabolic implications of body fat distribu-tion. Diabetes Care 1991;14:1132-43.

13 Wenger N, Speroff L, Packard B. Cardiovascular healthand disease in women. NEnglJ7Med 1993;329:247-56.

14 Rosano GMC, Kaski JC, Collins P, Maseri A, Poole-Wilson PA. Clinical features and long-term follow-up of99 consecutive patients with syndrome X. [abstract] JfAm Coll Cardiol 1993;21:475.

15 Trinder P. Determination of blood glucose using an oxi-dase-peroxidase system with non-carcinogenic chro-mogen. 7 Clin Pathol 1969;22:158-61.

16 Warnick GR, Albers JJ. A comprehensive evaluation ofthe heparin-manganese precipitation procedure forestimating high density lipoprotein cholesterol. J LipidRes 1978;19:65-76.

17 Gidez LI, Miller GJ, Burnstein M, Slagle S, Eder HA.Separation and quantitation of subclasses of humanplasma high density lipoproteins by a simple precipita-tion procedure. J Lipid Res 1982;23: 1206-23.

18 Mount JN, Kearney EM, Rosseneu M, Slavin BM.Immunoturbidimetric assays for serum apoproteins Aland B using Cobas Bio centrifugal analyser. Jf Clin Pathol1988;41:471-4.

19 Ley C, Lees B, Stevenson J. Sex- and menopause-associ-ated changes in body fat distribution. Am Jf Clin Nutr1992;55:950-4.

20 Bergman RN, Ider YZ, Bowden CR, Cobelli C.Quantitative estimation of insulin sensitivity. Am JfPhysiol 1979;236:E667-77.

21 Walton C, Godsland IF, Proudler AJ, Felton C, Wynn V.Evaluation of four mathematical models of glucose andinsulin dynamics with analysis of the effects of age andobesity. Am J Physiol 1992;262:E755-62.

22 Swan J, Walton C, Godsland IF. Assessment of insulinsensitivity in man: a comparison of minimal model andeuglycaemic clamp derived measures in health and heartfailure. Clin Sci 1994;86:317-22.

23 Volund A, Polonsky K, Bergman R. Calculated pattern ofintraportal insulin appearance without independentassessment of C-peptide kinetics. Diabetes 1987;36:1195-202.

24 Watanabe R, V0lund A, Roy S, Bergman R. Prehepaticbeta-cell secretion during the intravenous glucose toler-ance test in humans: application of a combined model ofinsulin and C-peptide kinetics. J Clin Endocrinol Metab1989;69:790-7.

25 Godsland I, Felton C, Proudler A, Wynn V. Evaluation oftwo methods for estimating pancreatic insulin secretionby modelling analysis of the intravenous glucose toler-ance test. [abstract] Diabetic Med 199 1;8(suppl 1):23.

26 Fuh MM-T, Jeng C-Y, Young MM-S, Sheu WH-H, ChenY-DI, Reaven GM. Insulin resistance, glucose intoler-ance and hyperinsulinaemia in patients with microvascu-lar angina. Metabolism 1993;42:1090-2.

27 Saeed BT, Bartlett WA, Murray RG, Jones AF, NeattieJM. Insulin resistance in syndrome X [letter]. Lancet1993;342:555.

28 Chauhan A, Camici MC, Schofield PM. Insulin resistancein syndrome X [letter]. Lancet 1993;342:554.

29 Meade TW, Cooper J, Miller GJ, Howarth DJ, Stirling Y.Antithrombin III and arterial disease. Lancet 1991;338:850-1.

52