1550-7289/13/$http://dx.doi.org

*CorrespondUniversità deglitario, Via Giusti

E-mail: luca

Surgery for Obesity and Related Diseases ] (2013) 00–00

Original article

Long-term cardiovascular risk and coronary events in morbidly obesepatients treated with laparoscopic gastric banding

Luca Busetto, M.D.a,*, Fabio De Stefano, M.D.a, Sabrina Pigozzo, M.D.a, Gianni Segato, M.D.b,Maurizio De Luca, M.D.b, Franco Favretti, M.D.b

aDepartment of Medicine, University of Padova, Padova, ItalybDepartment of Surgery, Regional Hospital of Vicenza, Vicenza, Italy

Received January 31, 2013; accepted March 23, 2013

Abstract Background: Long-term studies reporting the effects of bariatric surgery on cardiovascular risk

– see/10.10

ence:Studniani.buset

factors and events are scarce. The aim of this study was to analyze reduction of multiple car-diovascular risk factors and rates of coronary events in morbidly obese patients treated with bariatricsurgery and with 410 years of follow-up.Methods: This was a prospective uncontrolled study with laparoscopic adjustable gastric banding.Cardiovascular risk factors (fasting glucose, total cholesterol, HDL-C, triglycerides, blood pressure)have been previously determined both at surgery and 12–18 months after in 650 patients treated withlaparoscopic adjustable gastric banding from 1993 to 1999. Cardiovascular risk status was againdetermined in 2010, and the rate of coronary events during long-term (410 years) follow-up wascollected.Results: A total of 318 patients (58 men and 260 women) were retrieved. Age at surgery was38.6 � 10.4 years. Body mass index was 46.7 � 7.2 kg/m2. Follow-up was 12.7 � 1.5 years.Weight loss was 17.6% � 15.7% of baseline weight at 12–18 months and 17.1% � 14.8% at 12.7years. A significant reduction in blood glucose, total cholesterol, triglycerides, and systolic anddiastolic blood pressure was observed at the short-term evaluation and confirmed in the long term.HDL-C was unchanged at 12–18 months and significantly increased at 12.7 years. Five coronaryevents (1.6%) were recorded during long-term follow-up. The rate of observed events was comparedwith the rate of events expected according to baseline 10-year probability of myocardial infarctioncalculated with the Prospective Cardiovascular Munster study (PROCAM) score. Observed rate(1.6%) was slightly lower than the expected rate (2.0% � 4.9%).Conclusions: Stable weight loss and significant improvement of cardiovascular risk profile wereobserved in morbidly obese patients 10 years after laparoscopic adjustable gastric banding. (SurgObes Relat Dis 2013;]:00-00.) r 2013 American Society for Metabolic and Bariatric Surgery. Allrights reserved.

Keywords: Cardiovascular risk; Obesity; Obesity surgery; Gastric banding

Obesity is often associated with multiple cardiovascularrisk factors, including impaired glucose tolerance and type 2diabetes, atherogenic dyslipidemia with high triglycerides,

front matter r 2013 American Society for Metabolic and16/j.soard.2013.03.010

Luca Busetto, M.D., Dipartimento di Medicina,i di Padova, Clinica Medica I, Policlinico Universi-2, 35128 Padova, Italy.to@unipd.it

low HDL-C levels, and hypertension [1]. Moreover, obesityis associated with increased cardiovascular morbidity andmortality in epidemiologic studies [2]. Even a modestweight loss in obese individuals with type 2 diabetes,hypertension, or dyslipidemia leads to improved glycemiccontrol, lower blood pressure, and a better lipid profile [3].However, the demonstration that these beneficial effects ofweight loss on the cardiovascular risk profile may ultimately

Bariatric Surgery. All rights reserved.

L. Busetto et al. / Surgery for Obesity and Related Diseases ] (2013) 00–002

translate to a reduction of cardiovascular events andmortality in obese patients has been prevented by theextreme difficult of obtaining and maintaining weight lossover the long term with lifestyle modifications and medicaltherapy [4].The superiority of bariatric surgery over conventional

medical therapy for the induction of a significant weightloss and a consensual improvement of multiple cardiovas-cular risk factors in patients with morbidly obesity has beenproved by 3 randomized controlled trials [5–7], and it issupported by a large number of short-term clinical obser-vations [8,9]. The long-term, prospective, controlled Swed-ish Obese Subjects (SOS) study confirmed that bariatricsurgery, compared with conventional medical therapy, wasassociated with a sustained and stable 10-year weight loss[10], a stable improvement of major cardiovascular riskfactors [10], a lower incidence of cardiovascular events[11], and a 30% reduction in total mortality [12]. However,apart from the SOS study, long-term studies reporting theeffects of bariatric surgery on cardiovascular risk factorsand events still remain scarce. In this study, we aimed toanalyze long-term reduction of multiple cardiovascular riskfactors and rates of coronary events in a cohort of morbidlyobese patients treated with restrictive bariatric surgery(laparoscopic adjustable gastric banding [LAGB]) at asingle institution and having at least 10 years of follow-up.

Methods

Study population

The levels of 5 metabolic cardiovascular risk factors(blood glucose, systolic and diastolic blood pressure, totalcholesterol, HDL-C, and triglycerides) have been previ-ously determined by our group both at surgery and 12–18months after in morbidly obese patients treated with LAGB[13]. Briefly, 830 consecutive patients (647 women and 183men) were selected according to standardized criteria [14]and underwent LAGB at our institution from September1993 to November 2000 [13]. LAGB was at that timeperformed with the use of the 9.75-cm LAP-BAND System(Allergan, Irvine, CA) applied with a perigastric technique[15]. For the metabolic study, we selected from the wholeseries those patients in whom a complete cardiovascular riskfactors profile was collected both before and 12–18 monthsafter surgery (mean follow-up time 15.3 � 2.1 months),when none of the patients reported weight changes 44 kgin the month preceding the postsurgical evaluation. Patientswith mild hypothyroidism with or without hormonal ther-apy and patients treated at any time during the study withdrugs known to affect lipid metabolism were excluded. Atotal of 650 morbidly obese patients (495 women and 155men) were then included in the study [13].Starting from January 2010 and ending in December

2010, an extensive search was made to retrieve and contact

all the patients participating in our first metabolic study.Patients who had dropped out from regular annual follow-up visits were attempted to be reached by phone. Vitalstatus was therefore determined as part of the routineclinical follow-up or by telephone interview. Patients stillalive were invited to participate in a further clinical visitwith cardiovascular risk factors determination. A total of331 patients (50.9%) were retrieved. Thirteen (2.0%)patients refused to have any contact with our team, and15 (2.3%) patients were determined to be dead. Therefore,318 patients participated to the study and cardiovascularrisk factors determination was obtainable in 303 patients.

Anthropometry

All anthropometric measurements were made with thepatients wearing only light clothes without shoes. Heightwas measured to the nearest .01 m using a wall-mountedstadiometer. Weight was determined to the nearest .05 kgusing a calibrated balance beam scale. Body mass index(BMI) was calculated as weight (kg) divided by the heightsquared (m2). Weight loss was reported as a percentage ofthe baseline weight.

Cardiovascular risk factors

Venous blood sampling was drawn early in the morning,after a 12-hour overnight fast. Blood samples were kept at41C for 1 hour and then centrifuged at 3000 rpm for 10minutes at 41C. Serum and plasma were kept at 41C andassayed within 3 hours. Fasting plasma glucose, totalcholesterol, HDL-C, and triglycerides were determined bypreviously described methods [16]. Systolic and diastolicblood pressure was measured in the sitting position, after a5-minute rest, with an appropriate size cuff.The presence of type 2 diabetes, hypercholesterolemia,

low HDL-C status, hypertriglyceridemia, and hypertensionwas determined at baseline, at the short-term evaluation(12–18 months after surgery), and at the long-term evalua-tion (410 years after surgery) with the use of the samediagnostic criteria [14]. Briefly, diabetes was defined asfasting plasma glucose level ≥126 mg/dL or use of anyantidiabetic drug [17]; hypercholesterolemia, as total cho-lesterol 4200 mg/dL [18]; low HDL-C status, as HDL-C o40 mg/dL in men and o50 mg/dL in women [18];hypertriglyceridemia, as triglycerides 4150 mg/dL (18);and hypertension, as blood pressure ≥140/90 mm Hg or useof any antihypertensive drug [19], regardless of diabeticstatus.The number of patients with 0, 1, 2, 3, or 4 factors of the

metabolic syndrome (obesity excluded) was also calculatedat baseline, at the short-term evaluation (12–18 months aftersurgery), and at the long-term evaluation (410 yearsafter surgery) with the use of the following diagnosticcriteria: fasting plasma glucose ≥110 mg/dL or use of any

Cardiovascular Risk After Gastric Banding / Surgery for Obesity and Related Diseases ] (2013) 00–00 3

antidiabetic drug; blood pressure ≥130/85 mm Hg or use ofany antihypertensive drug; fasting triglycerides ≥150 mg/dL; and HDL-C o40 mg/dL in men and o50 mg/dL inwomen [18].

Global cardiovascular risk and coronary events

The 10-year cardiovascular risk was calculated in eachpatient at baseline with the use of the Prospective Cardio-vascular Munster (PROCAM) Health Check Algorithm,based on 462 coronary events occurring among 18,460 menand 49 events occurring among 8,518 women aged 20–78years enrolled in the PROCAM study [20]. In this algo-rithm, 10-year risk of coronary events was estimated withthe use of the following parameters: age (20–75 years),gender, known diabetes mellitus or fasting blood glucoselevels ≥120 mg/dL, current nicotine consumption, positivefamily history of coronary heart disease (CHD) (a first-degree relative suffering a heart attack before the age of 60years), systolic blood pressure (100–225 mm Hg), LDL-C(75–250 mg/dL), HDL-C (25–75 mg/dL), and triglycerides(50–400 mg/dL). The algorithm is available online at theInternational Task Force for Prevention of Coronary HeartDisease website (http://www.chd-taskforce.com/procam_interactive.html).Rates of fatal and nonfatal coronary events during long-

term follow-up (410 years) were collected at the long-termevaluation on the basis of patients history and availableclinical documentation. Fatal coronary event was defined asany death possibly related to cardiovascular causes. Nonfatalcoronary events included any hospitalization for myocardialischemia or cardiac revascularization procedures.

Table 1Baseline characteristics in 318 morbidly obese patients treated withlaparoscopic adjustable gastric banding between September 1993 andNovember 2000

Mean � SD/Number (%)

Sex ¼ male 58/318 (18.2%)Age, yr 38.6 � 10.4Weight, kg 127. 5 � 23.8BMI, kg/m2 46.7 � 7.2Type 2 diabetes (%) 77/318 (24.2%)Hypercholesterolemia (%) 176/318 (55.4%)Hypertriglyceridemia (%) 115/318 (36.2%)Low HDL levels (%) 160/318 (50.3%)Hypertension (%) 114/318 (35.8%)OSAS/hypoventilation (%) 131/318 (41.2%)Heart failure (%) 6/318 (1.9%)Osteoarthritis (%) 197/318 (61.9%)

BMI ¼ body mass index; OSAS ¼ obstructive sleep apnea syndrome.

Statistical analysis

Results are expressed as means � standard deviation fornumerical variables and as frequencies for categoricalvariables. In all statistical analysis, P o .05 was consideredsignificant. Differences in weight and cardiovascular riskfactors levels between baseline status (before surgery), theshort-term evaluation (12–18 months after surgery), and thelong-term evaluation (410 years after surgery) were testedwith paired Student’s t test. Differences in the prevalence ofmetabolic abnormalities or in the proportion of patientshaving different components of the metabolic syndrome atbaseline status and at the short- and the long-term evalua-tion were tested with χ2 test or analysis of variance.Observed rates of coronary events were compared withestimated baseline 10-year cardiovascular risk (mean� standard deviations) in the global sample and in thefollowing predefined subgroups: men and women,patients 450 years old and younger patients, and patientswith and without type 2 diabetes. Statistical analysis wasperformed by using the Statistical Package for SocialSciences, version 18.0 (SPSS, Chicago, IL).

Results

Weight loss, complications, additional procedures, andtotal mortality

Baseline characteristics of 318 morbidly obese patientstreated with LAGB between September 1993 and Novem-ber 2000 are reported in Table 1. No patients in the studyhad previous myocardial infarction or known CHD. LAGBwas successfully positioned in all the patients by aperigastric surgical technique and without perioperativemortality. Conversion to open surgery was required in 12cases (3.8%), 9 for technical difficulties and 3 for intra-operative complications. Intraoperative complications wereobserved in 7 patients (2.1%): 3 cases of gastric bleeding, 3cases of minor gastric perforation, and 1 case of bandmalpositioning. The band was normally implanted in allthese patients, but an early band repositioning wasrequested in the case of malpositioning.Mean follow-up was 12.7 � 1.5 years, and all patients

had, as per protocol, at least 10 years of follow-up. Ten-yearweight loss, as a percentage of the baseline weight, isreported in an intention-to-treat analysis, independentlyfrom the banding status (original band still in place, bandrepositioned, band removed or switched to other bariatricprocedure), in Fig. 1 (upper panel). Maximal weight loss(21.4% � 9.6 % of baseline weight) was observed 2 yearsafter surgery. Thereafter, a gradual weight regain was notedand 10-year weight loss stabilized to 16.7% � 13.8% of thebaseline weight. Mean BMI was 46.7 � 7.2 kg/m2 atsurgery, 36.5 � 6.7 kg/m2 (P o .001) after 2 years, and38.6 � 7.9 kg/m2 (P o .001) after 10 years.A total of 136 late band-related complications were

observed during the long-term follow-up. The most frequentcomplication was pouch dilation, which was recognized in127 (39.9%) cases. Most of the cases of pouch dilation weretreated conservatively, with band deflation and subsequentgradual band reinflation, but 60 (18.0%) cases evolved in a

Fig. 1. (Upper panel) Ten-year weight loss, as a percentage of the baselineweight, in 318 morbidly obese patients treated with laparoscopic adjustablegastric banding between September 1993 and November 2000. (Lowerpanel) Ten-year weight loss, as a percentage of the baseline weight,according to banding status and redo surgery. The patients with the originalband still in place and without any redo surgery (closed squares) were usedas the reference group. The band was removed in 25 patients (8.2%) (opensquares) and was repositioned in 35 patients (11.0%) (closed circles),whereas 60 patients (18.0%) were switched to a secondary bariatricprocedures (open circles). Unpaired Student’s t test was performed.*P o .05; **P o .01; ***P o .001.

L. Busetto et al. / Surgery for Obesity and Related Diseases ] (2013) 00–004

permanent stomach slippage requiring redo surgery. Lessfrequent band-related complications were erosion (8 cases)and isolated band infection (1 case). A total of 116 patients(36.5%) required band-related redo surgery, with somepatients requiring more than one additional procedure.The band was repositioned in 35 patients (11.0%). Isolatedband removal was performed in 25 patients (8.2%), whereas60 patients (18.0%) were switched to a secondary bariatricprocedures (47 patients to a duodenal switch, 8 to sleevegastrectomy, and 5 to gastric bypass). Finally, 98 minorport-related procedures were performed during the follow-up (93 deconnected ports were substituted and 5 twistedport were repositioned).Weight loss according to banding status and redo-surgery

is also reported in Fig. 1 (lower panel). Patients wereclassified has not having or having band-related surgeryirrespective from the reasons of redo procedures (weightloss failure or occurrence of band-related complications).Compared with patients not having any band-relatedsurgery, patients having the band repositioned during

follow-up had a greater weight loss at 2 years (27.1% �11.5% versus 20.4% � 9.0% of baseline weight; P o .01),but a similar weight loss after 10 years (18.8% � 17.4%versus 15.0% � 11.5%). Patients having isolated bandremoval not followed by a further bariatric proceduretended to regain weight progressively and had a signifi-cantly lower weight loss at 10 years (8.4% � 12.0% versus15.0% � 11.5%; P o .05). On the contrary, patients hav-ing a secondary bariatric procedures, mainly represented byduodenal switch, had a significantly greater weight loss at10 years (24.8% � 15.8% versus 15.0% � 11.5%;P o .001).A total of 15 deaths in 318 patients were recorded during

follow-up, with a total 4.7% death rate. Seven deaths wererelated to cardiovascular causes (2 acute myocardial infarc-tion, 1 sudden death, 1 heart failure, 1 rupture of aorticaneurism, 1 aortic dissection), 3 deaths were caused bycancer (2 lymphomas and 1 gastric cancer), 2 deaths weretraumatic (1 car accident and 1 suicide), and 2 deaths wererelated to miscellaneous causes (1 cirrhosis and 1 pulmo-nary fibrosis). Finally, 1 patient died of acute pancreatitis inthe postoperative period after the secondary duodenalswitch.

Cardiovascular risk factors changes

Analysis of cardiovascular risk factors, performed in the303 patients having a metabolic evaluation at each of the 3evaluations (excluding the 15 patients who died), isreported in Table 2. Baseline evaluation was performedimmediately before surgery, short-term evaluation wasperformed 12–18 months after surgery, and long-termevaluation was performed at least 10 years after surgery.Weight and BMI dropped significantly at the short-termevaluation (P o .001) and remained stable in the long term.In the short term, significant reductions were observed infasting blood glucose (P o .001), total cholesterol(P o .001), triglycerides (P o .001), and both systolic(P o .001) and diastolic (P o .05) blood pressure levels.All these reductions were maintained even at the 10-yearevaluation. HDL-C levels, which were unchanged at theshort-term evaluation, were found to be significantlyincreased in the long term (P o .05).The prevalence of metabolic abnormalities at baseline, at

the short-term evaluation, and at the long-term evaluation isreported in Table 3. The prevalence of type 2 diabetes,hypercholesterolemia, and hypertriglyceridemia was signifi-cantly reduced at the short-term evaluation and remainedunchanged in the long term. The prevalence of low HDL-Clevels was unchanged at the short term, but it was found tobe significantly reduced in the long term. On the contrary,the prevalence of hypertension, which was found to besignificantly reduced at the short-term evaluation, returnedto baseline level at the long-term evaluation.

Table 2Anthropometry and cardiovascular risk factors in 303 morbidly obese patients treated with laparoscopic adjustable gastric banding between September 1993and November 2000

Baseline Short term Long term

Weight, kg 126.2 � 22.5 103.9 � 29.5* 103.8 � 22.7*BMI, kg/m2 46.5 � 6.9 38.2 � 9.2* 38.3 � 22.7*Glucose, mg/dL 104.9 � 33.9 94.4 � 18.5* 96.2 � 24.5*Total cholesterol, mg /dL 209.4 � 46.1 198.4 � 44.0* 185.4 � 2.0*LDL-C, mg/dL 132.4 � 41.6 128.6 � 37.5 123.5 � 37.9†

HDL-C, mg/dL 48.3 � 11.4 49.1 � 14.9 53.5 � 14.1‡

Triglycerides, mg/dL 146.7 � 90.6 111.5 � 64.3* 108.4 � 47.5*Systolic blood pressure, mm Hg 141.7 � 16.8 133.4 � 13.7* 130.6 � 7.3*Diastolic blood pressure, mm Hg 91.4 � 14.4 88.4 � 8.7† 86.8 � 7.5†

Baseline evaluation was performed immediately before surgery, at short-term evaluation (12–18 months after surgery), and at long-term evaluation (at least10 years after surgery) in all the patients. Paired Student’s t test was performed versus baseline values.BMI ¼ body mass index.*P o .001.†P o .05.‡P o .01.

Cardiovascular Risk After Gastric Banding / Surgery for Obesity and Related Diseases ] (2013) 00–00 5

The prevalence of metabolic abnormalities at the short-and at the long-term evaluation is reported according to co-morbidities status at baseline in Fig. 2. The rate of remissionof metabolic co-morbidities in patients having the co-morbidity at baseline was 57 of 72 patients (79.2%) at theshort-term and 54 of 72 patients (75.0%) at the long-termevaluation for type 2 diabetes, 51 of 104 patients (49.0%)and 35 of 104 patients (33.6%) for hypertension, 77 of 166patients (46.4%) and 84 of 166 patients (50.6%) forhypercholesterolemia, 66 of 111 patients (59.5%) and 83of 111 patients (74.8%) for hypertriglyceridemia, and 95 of153 patients (62.1%) at the short-term evaluation and 90 of153 patients (58.9%) at the long-term evaluation for lowHDL-C levels. The incidence of new cases of metabolicabnormalities in patients not having the co-morbidity atbaseline was 1 of 231 patients (.4%) at the short-termevaluation and 15 of 231 patients (6.5%) at the long-termevaluation for type 2 diabetes, 20 of 199 patients (10.0%)and 45 of 199 patients (22.6%) for hypertension, 27 of 137patients (19.7%) and 26 of 137 patients (19.0%) forhypercholesterolemia, 16 of 192 patients (8.3%) and 23 of192 patients (12.0%) for hypertriglyceridemia, and 74 of

Table 3Prevalence of metabolic abnormalities in 303 morbidly obese patients treated wNovember 2000

Baseline

Type 2 diabetes 72/303 (23.8%)Hypercholesterolemia 166/303 (54.8%)Hypertriglyceridemia 111/303 (36.7%)Low HDL-C 153/303 (50.7%)Hypertension 104/303 (35.3%)

Baseline evaluation was performed immediately before surgery, short-term e10 years after surgery. χ2 test was performed versus baseline values.

*P o .001.

150 patients (49.3%) at the short-term evaluation and 28 of150 patients (18.7%) at the long-term evaluation for lowHDL-C levels.The proportion of patients having obesity and 0, 1, 2,

or 42 of the other 4 components of metabolic syndrome(hyperglycemia, high blood pressure, hypertriglyceridemia,and low HDL levels) at baseline, at the short-term evalua-tion, and at the long-term evaluation is reported in Fig. 3.Compared with the baseline situation, the distribution ofpatients was significantly shifted toward less metabolicsyndrome factors at the short-term evaluation (P o .01)and remained unchanged over the long term (P o .01versus baseline). In particular, the proportion of patientshaving no factors of the metabolic syndrome was 47 of 303patients (15.6%) at baseline, increased to 93 of 303 patients(30.8%) at the short-term evaluation, and was 113 of 303patients (37.3%) at the long-term evaluation. On thecontrary, the proportion of patients having 42 factors ofthe metabolic syndrome in addition to obesity was 67 of303 patients (22.1%) at baseline, declined to 19 of 303patients (6.2%) at the short-term evaluation, and remainedto 27 of 303 patients (8.9%) at the long-term evaluation.

ith laparoscopic adjustable gastric banding between September 1993 and

Short term Long term

16/303 (5.3%)* 33/303 (10.9%)*116/303 (38.3%)* 108/303 (35.6%)*61/303 (20.1%)* 51/303 (16.9%)*132/303 (43.6%) 91/303 (30.1%)*73/303 (24.1%)* 113/303 (37.4%)

valuation 12–18 months after surgery, and long-term evaluation at least

Fig. 2. Prevalence of metabolic abnormalities at the short-term evaluation(12–18 months after surgery) (light gray bars) and at long-term evaluation(410 years after surgery) (dark gray bars) according to co-morbidity statusin 303 morbidly obese patients treated with laparoscopic adjustable gastricbanding between September 1993 and November 2000. Rates of remissionof metabolic co-morbidities in patients having the co-morbidity at baseline(upper panel) and the incidence of new cases of metabolic abnormalities inpatients not having the co-morbidity at baseline (lower panel) werereported. H-CT ¼ hypercholesterolemia; H-TG ¼ hypertriglyceridemia;HPT ¼ hypertension; L-HDL ¼ low HDL-cholesterol levels; T2 DM ¼type 2 diabetes.

Table 4Ten-year cardiovascular risk calculated before surgery with the use of thePROCAM Health Check Algorithm and observed rates of coronary eventsduring long-term follow-up (410 years) in 318 morbidly obese patientstreated with laparoscopic adjustable gastric banding between September1993 and November 2000

Estimated 10-yr RiskMean � SD

Observed Coronary EventsNo. cases/No. patients (%)

Total 2.0% � 4.9% 5/318 (1.6%)

Women 1.5% � 4.5% 3/260 (1.1%)Men 4.7% � 5.7% 2/58 (3.4%)

Age o50 yr 1.4% � 3.1% 3/281 (1.1%)Age 450 yr 7.2% � 10.6% 2/37 (5.4%)

Without type 2 diabetes 1.0% � 2.3% 3/241 (1.2%)With type 2 diabetes 4.7% � 7.9% 2/77 (2.6%)

PROCAM ¼ Prospective Cardiovascular Munster Study.

L. Busetto et al. / Surgery for Obesity and Related Diseases ] (2013) 00–006

Estimated cardiovascular risk and coronary events

The 10-year cardiovascular risk calculated before surgerywith the use of the PROCAM Health Check Algorithm in318 morbidly obese patients treated with LAGB betweenSeptember 1993 and November 2000 is reported in Table 4.Mean estimated baseline 10-year cardiovascular risk waslow in our sample (2.0% � 4.9%). Most of patients(67.1%) had a 10-year cardiovascular risk o1%, 22.7%

Fig. 3. Proportion of patients having obesity and 0, 1, 2, or 42 of the other4 components of metabolic syndrome (hyperglycemia, high blood pressure,hypertriglyceridemia, and low HDL levels) at baseline, at the short-termevaluation (12–18 months after surgery), and at long-term evaluation (410years after surgery) in 303 morbidly obese patients treated with laparoscopicadjustable gastric banding between September 1993 and November 2000.

of patients had a 10-year cardiovascular risk between 1%and 5%, 5.8% of patients had a 10-year cardiovascular riskbetween 5% and 10%, and only 4.4% of patients had a 10-year cardiovascular risk 410%. As expected, 10-yearcardiovascular risk was higher in men than in women(P o .001), higher in patients 450 years old than inyounger patients (P o .001), and higher in patients withtype 2 diabetes than in patients without diabetes at baseline(P o .001) (Table 4).Five coronary events were observed in 318 patients

during long-term (410 years) follow-up, including 3 fataland 2 nonfatal coronary events (Table 4). The observed rateof coronary events (1.6%) was slightly lower than theexpected risk (2.0% � 4.9%). The difference betweenobserved and expected rates was generally higher in thegroup of patients with a higher baseline events probability(men, patients 450 years old, those with diabetes) than inpatients with lower baseline risk (women, patients whowere younger, and patients without diabetes) (Table 4).However, giving the wide standard deviations observed inthe estimated baseline 10-year cardiovascular risk and thelow number of observed events, neither of these differenceswas statistically meaningful.

Discussion

In this study, we analyzed the long-term (410 years)changes of weight and major cardiovascular risk factors in alarge sample of morbidly obese patients treated with LAGBin which the short-term (12–18 months) weight andcardiovascular risk factors modifications have been previ-ously already determined [13]. We demonstrated that in thisselected group of patients most of the beneficial effects onweight and metabolic profile produced by bariatric proce-dure in the short term were maintained even in the longterm. This may ultimately translate to the prevention ofsome cardiovascular major event.

Cardiovascular Risk After Gastric Banding / Surgery for Obesity and Related Diseases ] (2013) 00–00 7

In the literature, an abundance of studies report thebeneficial effects on major cardiovascular risk factors fromany bariatric procedures in the short term, typically 1–3years [8–9]. However, few studies have been publishedreporting the sustainability of these effects on a longer term.Scopinaro et al. first reported on the reversal of all the majorcomponents of the metabolic syndrome in 312 obesepatients with type 2 diabetes who underwent biliopancreaticdiversion (BPD), with results being strictly maintained overa 10-year follow-up [21]. However, metabolic effects afterBPD, a predominantly malabsorptive procedure, may bepartly attributed to weight loss independent mechanisms,limiting the applicability of these results to more commonlyperformed bariatric techniques. In the SOS study, a largecohort of morbidly obese patients predominantly treatedwith restrictive bariatric procedures has been compared withan equivalent number of well-matched control patients notreceiving surgery. After a 10-year follow-up, SOS inves-tigators reported significantly reduced weight, glucose andtriglycerides levels, and increased HDL-C levels in thesurgically treated group [10]. Incidence rates of hyper-triglyceridemia and diabetes was markedly lower in thesurgically treated group than in the control group, and amore frequent recovery from hypertension, diabetes, hyper-triglyceridemia, and low HDL-C levels was also observed[10]. Our 10-year uncontrolled study substantially con-firmed the SOS results, demonstrating a sustained reductionof weight, glucose, triglycerides and blood pressure levels, asustained increase of HDL-C levels, and good incidenceand recovery rates for most of the metabolic abnormalitieseven in the longterm.In the SOS study, a 10-year weight loss corresponding to

16.1% of baseline weight was associated in the surgicalgroup to a 7% incidence of new type 2 diabetes cases, witha 70% reduction of incidence in comparison with thecontrol group [10]. In our sample, a 16.7% weight losswas associated to a 6.5% 10-year incidence of type 2diabetes. We therefore confirmed prevention of type 2diabetes as one of the most important effects of bariatricsurgery. In patients already having type 2 diabetes beforesurgery, both our results and the SOS data reporteda 470% rate of diabetes remission in the short term,consistent with several other short-term studies [9]. How-ever, whereas the rate of diabetes remission declined to 36%after 10-year in the SOS study, we observed a substantialstability of remission rate even in the long term, despitesimilar levels of weight loss and maintenance. Previousstudies reported that the probability of diabetes remissionafter bariatric surgery may depend not only on weight losslevels but also on weight independent factors, the mostimportant one probably being diabetes duration [22,23].Diabetes duration before surgery was neither recalled in

our database nor reported in the SOS study; therefore, wehave not means to control if the observed difference in thelong-term diabetes remission rates may depend on this or

other uncontrolled factors. Differences in the definition anddetermination of diabetes remission between these 2 verydifferent studies also should be taken into account.A sustained effect of weight loss on hypertension has

been observed in our study, with both systolic and diastolicblood pressure levels at 10 years significantly lower than atbaseline and a 50.8% rate of hypertension remission in thelong term. Our results substantially differed from results onhypertension observed in the surgical SOS group, in whichan initial decline in blood pressure levels was not main-tained in the long term and the 34% remission of hyper-tension observed at 2-year declined to 19% at 10 years [10].Long-term incidence of new cases of hypertension was alsolower in our study (22.6%) than in the SOS study (41%).One important difference between these 2 study populationswas age at surgery, which was around 10 years lower in ourstudy than in the SOS study (38.6 � 10.4 years versus47.0 � 5.6 years). This implies that after a 10-year follow-up, most of the patients included in our series were about 50years old, whereas most of the SOS population was close to60 years old. This age difference may be important inexplaining differences in prevalence and incidence ofhypertension between these 2 predominantly femalesamples.HDL-C levels showed a particular trend in our study,

with no changes in the short term and a subsequent lateincrease. Incidence of new low HDL-C cases peaked in theshort term and then gradually declined. No changes or amoderate reduction in HDL-C concentrations have beenpreviously reported in the phase of active weight loss [3]and has been attributed to differences in the actual energybalance of the obese patients included in the studies, with areduction in patients still losing weight and an increase afterweight stabilization [13].The improvements in major cardiovascular risk factors

and the reduction in the number of patients with a cluster ofmultiple risk factors (metabolic syndrome) observed in ourstudy should translate to a delay in the progression ofatherosclerosis and, ultimately, in the prevention of hardcardiovascular endpoints. A reduction in carotid intima-media thickness and an improvement of flow-mediateddilation, 2 structural markers of early atherosclerosis, havebeen recently reported in a small group of obese adultsexamined before and 18 months after bariatric surgery [24],providing a link between the weight loss-associatedimprovements of traditional and nontraditional risk factorsand the reduced long-term morbidity and mortality. Indeed,SOS investigators recently reported, with a median follow-up of 14.7 years, an incidence of fatal or nonfatalcardiovascular events 30% lower in the surgical group thanin the control group [11]. In our uncontrolled study, we didnot compare the rate of cardiovascular events observed inour surgical patients to an appropriate comparable controlgroup. Therefore, we tried to determine the predicted 10-year probability of having a cardiovascular event on the

L. Busetto et al. / Surgery for Obesity and Related Diseases ] (2013) 00–008

basis of preoperative characteristics and to compare theactual rate of cardiovascular events with the predicted 10-year risk. In contrast to other studies in which differentprediction algorithms have been used [25], and moreprobably as the consequence of the relatively young ageand the low prevalence of diabetes observed in ourpopulation, the mean baseline 10-year cardiovascular riskwas found to be low in our sample (2.0% � 4.9%), withonly 4.4% of patients having a 10-year cardiovascularrisk 410%. Accordingly, total number of fatal and nonfatalcardiovascular events observed in our study was low (5events in 318 patients, 1.6%), and definitely lower than therate observed in the surgical group of the SOS study (199events among 2010 patients, 9.9%) [11]. To find a reductionof cardiovascular events in a population with a lowcardiovascular risk would be extremely difficult, and thiswas the case in our study. We observed a trend toward alower than expected number of cardiovascular events bylimiting the analysis to the groups of patients with arelatively higher risk (men, older patients, and those withdiabetes), but the further reduction in the number of patientsand observations in these subgroup analyses still againlimited the reliability of our results.The most important limitations of our study were the

uncontrolled design and the low rate of follow-up in thelong term. We were able to retrieve and study at the 10-yearevaluation only half of the patients originally included inthe short-term study [13]. Baseline characteristics of thepatients included in the present report were not significantlydifferent from those of patients lost to follow-up (data notshown), but this did not eliminate the possibility that someselection bias may occur. The exclusion from our originalsample of patients treated with drugs known to affect lipidmetabolism may have limited our analysis to patients withless severe forms of dyslipidemia. Finally, different diag-nostic criteria for some of the metabolic abnormalities andfor the metabolic syndrome have been suggested in clinicalguidelines [26] and more accurate definitions of diabetesremission [27] have been proposed. The application ofdifferent diagnostic criteria or thresholds may have influ-enced the magnitude of our results, but it would be difficultfor the positive trends for all the major cardiovascular riskfactors observed in our study to be completely explained bymethodological factors.

Conclusion

In conclusion, our study confirmed that the improvementsin major cardiovascular risk factors frequently reportedshortly after bariatric surgery may be maintained for years,as long as weight loss is maintained. The effect that thesechanges may have on the progression of atherosclerosis andon the incidence of new cardiovascular events largelydepends on baseline patients characteristics, being moreclinically significant in patients with high cardiovascular risk

at baseline and low or negligible in patients with low risk. Amore precise quantification of the variation of this preventiveeffect according to patients preoperative characteristics maybe important for a more accurate definition of the risk/benefitratio of bariatric surgery in different patients groups.

Disclosures

The study was funded by the institutional research fundsof the University of Padova, Italy. L. Busetto declares tohave received support for travel to meetings and paymentfor lectures from Allergan Inc. F. De Stefano declares tohave meetings expenses covered by Allergan Inc. M. deLuca and F. Favretti declare to have received payment forlectures from Allergan Inc. No other authors declareconflicts of interest.

References

[1] National Institutes of Health. Obesity Education Initiative ExpertPanel on the Identification, Evaluation, and Treatment of Overweightand Obesity in Adults. Clinical guidelines on the identification,evaluation, and treatment of overweight and obesity in adults—theevidence report. Obes Res 1998;6(suppl 2):51S–209S.

[2] Whitlock G, Lewington S, et al. Body-mass index and cause-specificmortality in 900000 adults: collaborative analyses of 57 prospectivestudies. Lancet 2009;373:1083–96.

[3] Goldstein DJ. Beneficial health effects of modest weight loss. IntJ Obes 1991;16:397–415.

[4] Uusitupa M, Peltonen M, Lindström J, et al. Ten-year mortality andcardiovascular morbidity in the Finnish Diabetes Prevention Study—secondary analysis of the randomized trial. PLoS One 2009;4:e5656.

[5] Dixon JB, O’Brien PE, Playfair J, et al. Adjustable gastric bandingand conventional therapy for type 2 diabetes. A randomized con-trolled trial. JAMA 2008;299:316–23.

[6] Schauer PR, Kashyap SR, Wolski K, et al. Bariatric surgery versusintensive medical therapy in obese patients with diabetes. N EnglJ Med 2012;366:1567–76.

[7] Mingrone G, Panunzi S, De Gaetano A, et al. Bariatric surgery versusconventional medical therapy for type 2 diabetes. N Engl J Med2012;366:1577–85.

[8] Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery.A systematic review and meta-analysis. JAMA 2004;292:1724–37.

[9] Buchwald H, Estok R, Fahrbach K, et al. Weight and type 2 diabetesafter bariatric surgery: systematic review and meta-analysis. AmJ Med 2009;122:248–56.

[10] Sjöström L, Lindroos AK, Peltonen M, et al. Lifestyle, diabetes, andcardiovascular risk factors 10 years after bariatric surgery. N EnglJ Med 2004;351:2683–93.

[11] Sjöström L, Peltonen M, Jacobson P, et al. Bariatric surgery and long-term cardiovascular events. JAMA 2012;307:56–65.

[12] Sjöström L, Narbro K, Sjöström CD, et al. Effects of bariatric surgeryon mortality in Swedish obese subjects. N Engl J Med 2007;357:741–52.

[13] Busetto L, Sergi G, Enzi G, et al. Short term effects of weight loss onthe cardiovascular risk factors in morbidly obese patients. Obes Res2004;12:1256–63.

[14] Gastrointestinal surgery for severe obesity. National Institutes ofHealth Consensus Development Conference draft statement. ObesSurg 1991;1:257–66.

Cardiovascular Risk After Gastric Banding / Surgery for Obesity and Related Diseases ] (2013) 00–00 9

[15] Favretti F, Cadière GB, Segato G, et al. Laparoscopic adjustablesilicone gastric banding (LAP-BAND): how to avoid complications.Obes Surg 1997;7:352–8.

[16] Busetto L, Pisent C, Rinaldi D, et al. Variations of the lipid levels inmorbid obese patients operated with the LAP-BAND AdjustableGastric Banding System: the effects of different levels of weight loss.Obes Surg 2000;10:569–77.

[17] Report of the Expert Committee on the Diagnosis and Classificationof Diabetes Mellitus. Diabetes Care 2003;26:S5–S24.

[18] Third Report of the National Cholesterol Education Program (NCEP)Expert Panel on Detection, Evaluation, and Treatment of High BloodCholesterol in Adults (Adult Treatment Panel III). Final Report.Circulation 2002;106;3143–421.

[19] 1999 World Health Organization–International Society of Hyper-tension Guidelines for the Management of Hypertension: GuidelinesSubcommittee. J Hypertens 1999;19:151–83.

[20] Assmann G, Schulte H, Seedorf U. Assessing risk of myocardialinfarction and stroke: new data from the Prospective Cardiovas-cular Münster (PROCAM) study. Eur J Clin Invest 2007;37:925–32.

[21] Scopinaro N, Marinari GM, Camerini GB, Papadia FS, Adami GF.Specific effects of biliopancreatic diversion on the major components

of metabolic syndrome. A long-term follow-up study. Diabetes Care2005;28:2406–11.

[22] Schauer PR, Burguera B, Ikramuddin S, et al. Effect of laparoscopicRoux-en Y gastric bypass on type 2 diabetes mellitus. Ann Surg2003;238:467–84.

[23] Hall TC, Pellen MG, Sedman PC, Jain PK. Preoperative factorspredicting remission of type 2 diabetes mellitus after Roux-en-Ygastric bypass surgery for obesity. Obes Surg 2010;20:1245–50.

[24] Sturm W, Tschoner A, Engl J, et al. Effect of bariatric surgery on bothfunctional and structural measures of premature atherosclerosis. EurHeart J 2009;30:2038–43.

[25] Mackey RH, Belle SH, Courcoulas AP, et al. Distribution of 10-yearand lifetime predicted risk for cardiovascular disease prior to surgeryin the Longitudinal Assessment of Bariatric Surgery-2 Study. AmJ Cardiol 2012;110:1130–7.

[26] Alberti G, Zimmet PZ, Shaw J, Grundy SM. International DiabetesFederation 2006: The IDF Consensus Worldwide Definition of theMetabolic Syndrome. 2006. Available at: http://www.idf.org/webdata/docs/IDF_Meta_def_final.pdf. Accessed 17 July 2012.

[27] Buse JB, Caprio S, Cefalu WT, et al. How do we define cure ofdiabetes? Diabetes Care 2009;32:2133–5.