Treating morbid obesity with laparoscopic adjustable gastric banding

The American Journal of Surgery 194 (2007) 333–343

Clinical surgery–American

Treating morbid obesity with laparoscopic adjustable gastric banding Louis F. Martin, M.D., M.S.a,*, Gerard J. Smits, Ph.D.b, Robert J. Greenstein, M.D.c a

Weight Management Center, Louisiana State University Health Sciences Center, 533 Bolivar St, Rm 508, New Orleans, LA 70112, USA b Computer and Consulting Statisticians, Inc, 2528 Chapala St, Santa Barbara, CA 93105, USA c Department of Surgery, Veterans Affairs Medical Center, 130 Kingsbridge Rd, Bronx, NY 10468, USA Manuscript received August 21, 2006; revised manuscript March 29, 2007

Abstract Background: Morbid obesity results in multiple comorbidities and an increased mortality rate. The National Institutes of Health has stated that surgery is the most effective long-term therapy; therefore, we evaluated a laparoscopically implantable adjustable gastric band. Methods: We reviewed 2 multicenter prospective, open-label, single-arm surgical trials—trial A (3 years) and trial B (1 year)—with ongoing safety follow-up. These trials were conducted in United States community and university hospitals (trial A ⫽ 8 sites and trial B ⫽ 12 sites). Trial A comprised 292 subjects (mean ⫾ SD preoperative weight: 133 kg ⫾ 24.4), and trial B comprised 193 subjects (129 kg ⫾ 20.8). Intervention included placement of a constrictive, adjustable band around the upper stomach to limit food intake and induce weight loss. Main outcome measures were the primary efficacy end point of weight loss. Secondary end-points were change in quality-of-life, safety parameters, and complications, including band slippage, reoperation, and device explantation. Results: In the 2 trials, 485 devices were implanted (92% laparoscopically), and no deaths occurred. Of the patients in trial A, 206 (70.5%) completed the 3-year follow-up, and 142 (73.6%) of patients in trial B completed the 1-year follow-up. Weight-loss results, using the last value carried forward, for all 292 patients in trial A and all 193 patients in trial B demonstrated a change in mean body mass index (kg/m2) ⫾ SD from 47.4 ⫾ 7.0 to 39.0 ⫾ 7.3 in trial A and from 46.7 ⫾ 7.8 to 38.4 ⫾ 7.6 in trial B subjects at 1 year (P ⬍ .001 for both trials A and B), with minimal further change at 3 years (39.0 ⫾ 8.5) in trial A subjects. The percentage of initial body weight lost at 1 year was 17.7% ⫾ 9.4% for trial A subjects and 18.2% ⫾ 8.9% for trial B subjects, whereas the 3-year total for trial A subjects was 18.3% ⫾ 13.1%. At 1 year, 76% of patients in trial A and 66% of patients in trial B had complications, mostly related to upper gastrointestinal symptoms. By 9 years after surgery, 33% (96 of 292) of trial A subjects had their devices explanted because of complications or inadequate weight loss. Conclusions: These first-generation implantable adjustable gastric band results suggest that this is a viable bariatric surgery therapeutic option for the treatment of obesity. © 2007 Excerpta Medica Inc. All rights reserved. Keywords: Adjustable gastric banding; LAP-BAND; Clinical trials; Laparoscopy; Morbid (severe) obesity; Quality of life; Restrictive bariatric surgery; Safety; Weight loss

Obesity is prevalent [1] and is rapidly increasing in the United States [2] and other industrialized nations [3] as well as in developing countries [4]. This is causing United States [5,6] and international [7] governmental concern. Consequences of obesity include the development of comorbid diseases [8,9], a poorer [10] and shorter life [11], and substantial direct and indirect costs to our societies, eg, ⬎$99 billion in the United States in 1995 [12].

* Corresponding author. Tel.: ⫹1-504-568-4750; fax: ⫹1-504-568-4633. E-mail address: [email protected]

In the obese (ie, body mass index [BMI] ⱖ30 kg/m2), the United States Institute of Medicine (IOM) considers a loss of ⱖ10% initial body weight (IBW) for a period of 2 years clinically significant [13]. The United States Food and Drug Administration (USFDA) considers weight-control medications effective if treated subjects lose ⱖ5% IBW more than controls [14]. However, for the morbidly obese (BMI ⱖ40), nonoperative therapies [15–19] are not effective [6,20]. Only bariatric surgery [21–23] has documented satisfactory long-term weight loss in this population. Nevertheless, physicians and patients remain hesitant to consider bariatric surgery, partly because of the limited data

0002-9610/07/$ – see front matter © 2007 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2007.03.002

334

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

from controlled, randomized [6,24] or pair-matched [25] trials; the attendant surgical morbidity and mortality [26]; and the expense [27]. Accordingly, the development of effective, less-invasive, adjustable, reversible bariatric operations with minimal anatomic changes to the gastrointestinal (GI) tract are desirable. Laparoscopic adjustable gastric banding [28] may meet these criteria and is now the most common bariatric procedure performed outside the United States [29]. At present, ⬎100,000 of the devices used in the studies reported herein have been implanted worldwide. In the United States, prospective, multicenter, open-label clinical trials were performed on 2 adjustable gastric bands (trial A—1995, USFDA Investigational Device Exemption [IDE] number G-89034 and trial B—1998, USFDA IDE number G-950047). The purpose of this article is to present complete data from all sites of these 2 United States studies. Three-year efficacy and up to 9 years of safety data for trial A, together with 1-year safety and efficacy data for trial B, are presented. Methods Data for these trials were collected in compliance with USFDA regulations. They were compiled by the manufacturer, who was supervised by a non–sponsor-affiliated statistician who was also responsible for data analysis. The manufacturer’s USFDA premarket approval application used trial A data. The USFDA inspected half of the investigational sites in trial A as part of its Bioresearch Monitoring Program. Participants Institutional Review Board approval was obtained at each site before study commencement, and every patient signed an informed consent form before having a device implanted. Trial A was a 3-year, prospective, open-label, interventional study. Patients were enrolled at 8 bariatric surgical practices at community hospitals and academic medical centers. Prospective patients were considered for entry per medical history and physical examination. Standard preoperative diagnostic and laboratory tests were performed. Trial A inclusion criteria were age between 18 and 56 years and weight ⱖ100 lb (ⱖ45 kg) greater than ideal body weight [30]. Patients were required to have had multiple failed attempts to lose weight or to maintain lost weight during a period of at least 5 years using nonsurgical weight-loss methods. The company sponsoring the trial did not pay the expenses associated with the initial surgical intervention. Patients had to obtain permission from their medical insurance company for this elective procedure, pay for the expenses themselves, or qualify for indigent care. Exclusion criteria for both trials included previous gastric or small-bowel surgery, severe cardiopulmonary disorder, or any other medical condition making surgical placement of this device potentially life threatening. Subject accrual started June 1, 1995 and ended June 22, 1998. Trial B was a 1-year prospective, open-label, expandedaccess study conducted at 12 centers that had IRB approval before study commencement. All patients signed an informed consent form before device implantation. Inclusion criteria for trial B were a BMI of at least 35 to 40 kg/m2 and ⬎1 obesity-related comorbid condition, such as hyperten-

sion or diabetes. Inclusion criteria included age between 18 and 61 years. In addition, an attempt was made to recruit only investigators skilled in advanced laparoscopic abdominal surgical techniques. All participating surgeons had to have performed ⱖ25 laparoscopic Nissen fundoplications and/or Roux-en-Y gastric bypasses in the previous 12 months. An additional objective of trial B was to determine whether this cohort of experienced laparoscopic surgeons would have a lower band-slippage rate. Accrual for trial B started on March 1, 2000, and ended on June 21, 2001. Intervention The laparoscopic adjustable gastric banding system used in this study (Fig. 1) was a radio-opaque, silicone, elastomeric band with an inflatable balloon on its inner aspect and an integral lock (BioEnterics LAP-BAND System, INAMED, Santa Barbara, California). The lumen of the balloon is connected to a conventional access reservoir that is positioned in, or on, the rectus abdominus sheath. After surgery, the reservoir can be accessed transcutaneously using a Huber-tipped needle. By adding or withdrawing saline, the inner diameter of the band may be adjusted in a 10-minute outpatient procedure. The perigastric dissection technique, previously described [28,31] was used in all procedures during trial A for implantation of the adjustable gastric band system. The surgical techniques were still evolving while trial A was in progress. Several technical modifications were incorporated during trial B, especially the change of the posterior plane of dissection from the peri-gastric technique to the pars flacida technique [28]. The recommended band position was approximately 1 cm below the esophagogastric junction anteriorly and almost at the junction posteriorly. Multiple anterior gastrogastric serosal sutures were used to imbricate the stomach over the band as much as possible, and an additional posterior gastric suture was advocated if the band was not placed cephalad to the bursa omentalis [31].

Fig. 1. A drawing of the adjustable gastric band. The pouch calibration balloon, identified by the dotted line, was used in trial A patients. Many experienced surgeons throughout the world no longer calibrate the stoma size. The band is shown after the fundus has been circumvented but before its integral buckle being closed.

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

In trial A, investigators at 7 of the 8 sites planned a laparoscopic approach in all suitable patients. At the remaining site, the investigator elected to perform only open surgery. In trial B, all investigators scheduled laparoscopic operations where feasible. Reasons to schedule an open procedure included BMI ⬎60 kg/m2 (for some investigators) and concomitant incisional herniorraphy. Intraoperative indications for conversion to open procedure included uncontrolled hemorrhage, inability to safely perform the retrogastric dissection, or an enlarged liver preventing safe exposure. In both trials A and B, an upper gastrointestinal series was obtained the day after surgery and yearly thereafter for the duration of the protocol. This was done to confirm that there was no inadvertent gastrointestinal perforation at the time of surgery; to assess the size and shape of the gastric pouch; and to document the position of the band relative to the stomach and esophagus. Objectives The hypothesis being tested was that an adjustable gastric band could safely and effectively treat obesity. The primary efficacy end point was weight loss. Secondary efficacy end points were patient-reported changes in quality of life (QOL) variables. Safety parameters included death, perioperative morbidity, reoperation, device explantation, and withdrawal from the study. Because all patients were required to have failed previous nonsurgical weight-loss treatments before enrollment, they served as their own controls in the USFDAapproved IDE protocols. Outcomes In trial A, weight was recorded at entry (baseline) and at scheduled postoperative follow-up visits at 3 weeks and at 3, 6, 9, 12, 18, 24, 30, and 36 months after initial surgery. In trial B, weight was reported at baseline, 3 weeks, 6 months, and 12 months. Weight-related data are presented as changes in absolute weight (kg), absolute BMI (kg/m2), and %IBW lost, and percent excess BMI lost (%EBMIL) [32]. The parameter %EBMIL derives from National Institutes of Health guidelines defining excess weight as starting at BMI ⬎ 25 [33] and is calculated using the following formula: %EBMIL ⫽ 100 ⫺ (((postoperative BMI ⫺ 25) ⁄ (preoperative BMI ⫺ 25)) ⫻ 100). Self-assessed standardized quality-of-life questionnaires including the Beck Depression Inventory [34], the Multidimensional Body-Self Relations Questionnaire (MBSR) [35], and the Rand SF-36 Questionnaire [36]. These were completed at baseline and at 1 year (trials A and B) and 3 years (trial A). Safety parameters were evaluated and documented by each investigator. Complications were subdivided into death, surgery-related, device-related, total digestive, metabolic and nutritional, and other events. Device-related complications included the combined category of band slippage and pouch dilatation, band erosion, port leakage, infection, stoma obstruction, and esophageal dilatation.

335

Statistical analysis Based on previous experience with a first-generation adjustable silicone gastric band designed for open placement [37], sample size calculation indicated that 200 patients followed up for 36 months would be adequate to show weight loss efficacy, by IOM criteria, ⱖ10% IBW [13]. The study design for trial A assumed a 90% success rate for laparoscopic device implantation and patient attrition of 10% to 15% because of elective band explant or loss to follow-up (LTF). To compensate for this predicted trial A attrition rate, the objective was to enter 250 to 300 patients. A patient was defined as LTF if he or she did not attend the final on-study visit scheduled at 36 months within the protocol-stipulated ⫾ 90-day (3-month) “window.” However, in this report, we additionally present data for patients (n ⫽ 28) who missed the 36-month window but for whom data are available beyond 39 months. Accordingly, data for patients within the stipulated 36 ⫾ 3–month window are presented as “36-month data.” The inclusive cohort of all patients at or beyond 39 months is presented as “3-year” data. The 3-year database lock date for trial A for efficacy was November 30, 2000. Additional safety data were collected until December 15, 2003, with explant data collected until March 21, 2005. Most are available for ⬎5 because of an extension of the initial monitoring contract. The explanted medical devices (the band and reservoir) were returned to the sponsoring company for evaluation and to be made available to the FDA. As a consequence, explantation data are available for up to 9.5 years from the start of trial A. The 1-year database lock date for trial B for efficacy and safety was March 10, 2003. Explant data were collected until March 31, 2005. Once the FDA allowed the adjustable band to be marketed in June 2001, however, it was no longer necessary for study patients to return to the investigative sites to have a band replaced. Also, some subjects, especially in the B trial, could have had a band removed without the investigator knowing because no follow-up by the investigators was required after the end of the first postoperative year. Weight and QOL results are presented as mean ⫾ SD. Changes from baseline to all time points were analyzed using paired Student t test and by linear regression analysis when evaluating the relationship of weight loss to QOL changes. Statistical analyses were performed using PC SAS, version 8.2 (SAS Institute, Inc, Cary, NC). Weight loss was further stratified posthoc into groups of patients that lost ⬍10%, 10% to 20%, and ⬎20% IBW. Fisher’s exact test was used to compare 1-year complication rates between trials A and B because assignment to the 2 trials was not random. Therefore, the P values provided should only be considered as suggestive. The effect of race and sex were studied using posthoc analyses on combined data of trials A and B at year 1. In each model, baseline BMI was included as a covariate because we find this to be a good predictor of weight loss. Modeling was then performed using trial A data with follow-up for 3 years. Modeling was performed using PROC MIXED of SAS version 8.2 to examine the trend of weight change with time, in which covariance among repeated measures was modeled as a first-order autoregressive process. Only the main effects were included in the model. In

336

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

all of the analyses, the significance level was accepted as P ⱕ .05. We present outcome data in 3 ways. Summary data and statistics are based on available data, with no imputations for missing data, and assuming that data were missing at random and not predicted by previous values. We also present an LTF analysis, including all patients (trial A ⫽ 292 and trial B ⫽192), with baseline values carried forward if terminal values were not present. Finally, we show results substituting the last value carried forward (LVCF) for the missing terminal data. Results Patient characteristics The demographic characteristics of the 292 patients enrolled in trial A and 193 subjects in trial B were similar (Table 1). Patients in trial A had greater prevalence of comorbid conditions. (Table1) There were no significant demographic differences between patients LTF and the overall study population in trials A or B (data not presented). Subject disposition and participant flow Proportionally more patients had bands successfully implanted laparoscopically in trial B (98%), in which more experienced laparoscopic surgeons participated, than in trial A (89%) (Fig. 2). In trial A, 20 patients (7%) either had a planned laparotomy because of surgeon preference, a previous upper abdominal surgery, a BMI ⬎ 60 kg/m2 at entry, or a simultaneous laparotomy for ventral incisional hernia. A subgroup of 13 subjects (4%) in trial A was converted from scheduled laparoscopic operation (Fig. 2) because of anatomic features (n ⫽ 6), necessary concomitant surgery (n ⫽ 3), bleeding (n ⫽ 2), or technical difficulties (n ⫽ 2). There were no conversions for technical reasons in trial B, and fewer cases were initially scheduled as open procedures. At 36 ⫾ 3 months, 61% (178 of 292) of patients remained in trial A. We also present data on an additional 10% (28 of Table 1 Demographic data and baseline characteristics Characteristic

Trial A (N ⫽ 292)

Trial B (N ⫽ 193)

Age (y)* Weight (kg) BMI (kg/m2) Sex Male Female Race Caucasian African American Hispanic Asian Other Comorbidities (ⱖ5% in trial A) Hypertension Cholelithiasis Reflux and heartburn Diabetes (all) Asthma

38.8 (8.2) 133 (24.4) 47.4 (7.0)

41.5 (9.4) 129 (20.8) 46.6 (7.8)

45 (15) 247 (85)

35 (18) 158 (82)

236 (81) 44 (12) 12 (4) 0 (0) 0 (0)

179 (93) 6 (3) 5 (3) 1 (⬍1) 2 (1)

125 (43) 75 (26) 70 (24) 47 (16) 47 (16)

47 (24) 3 (1.6) 27 (14) 18 (9) 11 (6)

* Mean (SD); N (%).

Fig. 2. A flow diagram of the 485 patients entered in trials A and B. Patients in trial A were followed up for 3 years for efficacy and 9 years for safety. Subjects in Trial B were followed up for up to 1 year for both safety and efficacy.

292), for a total of 71% (206 of 292) (Tables 2 and 3 and Fig. 2). Thirty-three percent (96 of 292) of patients did not complete the study (Fig. 2). They either had removal of the adjustable band (18%; 51 of 292) or were LTF (15%; 45 of 292). LTF analysis found approximately one third less weight loss than the other approaches to handling missing data. Weight-loss outcomes At 1 year, mean weight loss among trial A and B patients was 23.6 ⫾ 13.2 kg and 23.7 ⫾ 12.5 kg, respectively. At 3 years, mean weight loss in trial A patients was 25.4 ⫾ 18.1 kg (all of these means are calculated using LVCF for the 485 patients initially entered in both studies). Additional weight loss data are presented in Tables 2 and 3 and Fig. 3, including BMI, %IBW, and %EBMIL for patients who remained in the study at 6 months, 12 months, 24 months, 36 months, “3-year,” final LTF, and final LVCF. In trials A and B at 3 years, 72% (148 of 206) ad 82% (116 of 142) of subjects, respectively, had lost ⱖ10% of their IBW (Table 3). Weight loss, whether expressed in terms of decrease in BMI, %IBW, or %EBMIL, was significant (P ⬍ .0001) at each time point compared with weight at study entry (baseline). A lower BMI at entry resulted in less absolute weight loss but a greater proportion of excess weight loss (Fig. 3). For each lower starting BMI unit, the data predicted .47 kg less weight loss (P ⬍ .0001). In contrast, there was an increase of .78 in %EBMIL for each lower starting BMI unit (P ⬍ .0001). These data predicted 7.8% greater %EBMIL for an individual who, at entry, weighed 10 BMI units less than another subject. The longitudinal modeling [38] of weight change with time using a first-order autoregressive process identified time from surgery, baseline weight, sex, age, investigational site, and race as predictors of weight loss, with time being

Visit

Study subjects

Change in weight (kg)

BMI

%IWL

%EBMIL

N

Mean ⫹ SD (median) range (95% confidence interval)

Mean ⫹ SD (median) range (95% confidence interval)

Mean ⫹ SD (median) range (95% confidence interval)

Mean ⫹ SD (median) range (95% confidence interval)

Trial A

Trial B

Trial A

Trial B

Trial A

Trial B

Trial A

Trial B

Trial A

Trial B

Baseline

292

193

—

—

—

—

—

233

138

12-Month*

233

142

24-Month*

189

—

⫺16.7 ⫾ 9.1 (⫺16.5) ⫺57.0-3.2 (⫺18.3–⫺15.2) ⫺23.7 ⫾ 12.5 (⫺21.8) ⫺66.4 –.2 (⫺25.8 –⫺21.6) —

—

3-Yeart

206

—

Final LTFtt

292

192

Final LVCFttt

292

192

30.2 ⫾ 16.2 (29.4) ⫺10.9 – 87.9 (28.1–32.2) 39.3 ⫾ 22.1 (36.4) ⫺9.8 –102.0 (36.4 – 42.1) 43.0 ⫾ 26.0 (38.6) ⫺20.7–117.1 (39.2– 46.7) 41.1 ⫾ 28.3 (38.8) ⫺11.2–134.4 (36.9 – 45.3) 39.8 ⫾ 29.0 (35.8) ⫺20.7–134.4 (35.8 – 43.8) 25.0 ⫾ 29.9 (16.5) ⫺11.2–134.4 (21.6 –28.5) 36.9 ⫾ 27.1 (33.7) ⫺20.7–134.4 (33.8 – 40.0)

29.3 ⫾ 14.6 (28.9) ⫺3.7–74.1 (26.8 –31.7) 41.2 ⫾ 21.6 (37.4) ⫺.8 –140.3 (37.6 – 44.8) —

178

13.5 ⫾ 6.8 (13.5) ⫺5.4 –31.6 (12.7–14.4) 17.7 ⫾ 9.4 (16.8) ⫺4.8 – 48.2 (16.5–18.9) 19.4 ⫾ 11.2 (18.0) ⫺7.5–53.8 (17.8 –21.0) 18.8 ⫾ 12. (17.7) ⫺4.7– 60.0 (16.9 –20.6) 18.3 ⫾ 13.1 (16.5) ⫺7.5– 60.0 (39.8 – 43.8) 11.4 ⫾ 13.4 (7.6) ⫺4.7– 60.0 (9.9 –13.0) 16.8 ⫾ 12.2 (14.9) ⫺7.5– 60.0 (15.4 –18.2)

12.9 ⫾ 6.3 (13.0) ⫺2.0 –33.5 (11.8 –13.9) 18.2 ⫾ 8.9 (18.0) ⫺.2– 49.1 (16.7–19.7) —

36-Month*

⫺18.1 ⫾ 9.7 (⫺17.5) ⫺56.4 – 8.7 (⫺16.8 –⫺56.4) ⫺23.6 ⫾ 13.2 (⫺21.5) ⫺66.6 –7.8 (⫺25.3–⫺21.9) ⫺25.9 ⫾ 15.7 (⫺23.9) ⫺75.3– 8.9 (⫺28.1–⫺23.6) ⫺25.4 ⫾ 18.1 (⫺22.7) ⫺114.5–⫺22.7 (⫺28.1–⫺22.8) ⫺25.2 ⫾ 18.2 (⫺22.3) ⫺114.5–⫺8.9 (⫺27.8 –⫺22.8) ⫺15.5 ⫾ 18.8 (⫺10.5) ⫺114.5–5.5 (⫺17.7–⫺13.3) ⫺22.8 ⫾ 17.7 (⫺19.5) ⫺114.5– 8.9 (⫺24.8 –⫺20.7)

46.7 ⫾ 7.8 (45.3) 32.9 –94.5 (45.5– 47.8) 40.4 ⫾ 7.2 (38.8) 28.5– 81.6 (39.2– 41.6) 38.4 ⫾ 7.6 (37.4) 19.6 –72.7 (37.1–39.7) —

—

6-Month*

47.4 ⫾ 7.0 (45.8) 36.6 –74.03 (46.6 – 48.2) 41.2 ⫾ 7.3 (40.0) 26.5– 67.8 (40.2– 42.1) 39.0 ⫾ 7.3 (38.3) 24.6 – 67.9 (38.0 – 40.0) 38.1 ⫾ 7.5 (37.9) 21.3– 67.3 (37.0 –39.1) 38.6 ⫾ 7.9 (39.0) 19.3– 63.6 (37.5–39.8) 39.0 ⫾ 8.5 (38.9) 19.3– 67.3 (37.8 – 40.2) 41.9 ⫾ 8.7 (41.5) 19.3–74.1 (40.9 – 42.9) 39.3 ⫾ 7.8 (41.5) 19.3– 67.3 (38.9 – 41.1)

— — ⫺18.4 ⫾ 14.5 (⫺17.8) ⫺66.4 –.2 (⫺20.5–⫺16.4) ⫺20.4 ⫾ 12.8 (⫺19.1) ⫺66.4 –.5 (⫺22.3–⫺18.6)

— — 40.0 ⫾ 7.9 (39.3) 19.6 –72.7 (38.8 – 41.1) 39.3 ⫾ 7.6 (38.5) 19.6 –72.7 (38.2– 40.4)

Note: All weight outcome parameters at every time point are significantly different from baseline values (P ⬍ 0.0001). * Data for subjects attending for visit within designated “window”. t Data for 3-year visit or last available visit, if still on study. tt Lost to follow-up. ttt Last visit carried forward.

— — 14.2 ⫾ 10.7 (14.5) ⫺.2– 49.1 (12.7–15.7) 15.8 ⫾ 9.3 (13.4) ⫺.5– 49.1 (14.4 –17.1)

— — 32.1 ⫾ 25.2 (31.8) ⫺.8 –140.3 (28.6 –35.7) 35.7 ⫾ 22.9 (34.0) ⫺1.5–140.3 (32.4 –39.0)

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

Table 2 Weight parameters include absolute weight lost in kg and BMI (kg/m2), %IWL, and %EBMIL by visit

337

338

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

Table 3 %IWL frequency by follow-up visit in categories of ⬍10%, 10% to 20%, and ⬎20% Visit

%IWL (N [%]) ⬍10%

3-mo* 6-mo* 12-mo* 24-mo* 36-mo* 3-y†

⬎20%

10% to 20%

Total

Trial A

Trial B

Trial A

Trial B

Trial A

Trial B

Trial A

Trial B

135 (55) 69 (30) 50 (22) 31 (16) 45 (25) 58 (28)

– 45 (33) 26 (18) – – –

106 (43) 126 (54) 94 (40) 76 (40) 57 (32) 63 (31)

– 79 (57) 61 (43) – – –

4 (2) 38 (16) 89 (38) 82 (44) 76 (43) 85 (41)

– 14 (10) 55 (39) – – –

245 (100) 233 (100) 233 (100) 189 (100) 178 (100) 206 (100)

– 138 (100) 142 (100) – – –

%IWL ⫽ percentage of initial weight lost. * Data for subjects attending for visit within designated “window.” † Data for 3-year visit or last available visit, if still in study.

the strongest predictor. The other strong predictor was baseline weight, with those with the greatest initial weight losing more kilograms with time. As expected, end points based on percentage of excess weight lost were inversely related to initial weight because of the effect of the larger denominator’s (initial weight) decreasing the percentage value of weight lost. Weight was lost most quickly initially and then leveled off after approximately 12 months after surgery. A log transform of time yielded an approximately linear trend in weight loss during the 3 years of follow-up examined, suggesting that weight loss slows as time elapses. Investigational site, age, race, and sex were weak or nonsignificant predictors of weight change. A posthoc subgroup analysis of 50 enrolled black patients (44 of 292 patients [15%] in trial A and 6 of 193 [3%] patients in trial B) was performed. For black patients, mean weight loss at 1 year was 19.4 ⫾ 10.23 kg (13.8 ⫾ 6.40 %IBW). For combined white, Hispanic, and Asian patients, the loss was 24.1 ⫾ 13.13 kg (18.4 ⫾ 9.37 %IBW; P ⬍ .01 for both). At 3 years, %EBMIL for black patients was 30.7% and for white patients was 43.3% (P ⱕ .01). In an additional posthoc subgroup analyses, weight loss was similar for men and women at 1 year in trials A and B. Men lost 23.7 ⫾ 12.79 kg, and women lost 23.1 ⫾ 13.76 kg, which was 18.5 ⫾ 9.28 %IBW and 14.9 ⫾ 8.24 %IBW, respectively, for A and B patients. At 3 years, men had lost 18.0% ⫾ 13.55% IBW and women 18.9% ⫾ 12.34% IBW.

Fig. 3. A composite graph addressing the effect of preoperative BMI on two parameters, %EBMIL and absolute weight loss (in Kg). There is a highly significant (P ⬍ .0001) negative relationship: Although the heavier patients lost more weight, those with a lower starting BMI lost a greater proportion of their excess weight.

Quality-of-life assessment Compared with baseline values, scores on the BDI significantly improved (P ⬍ .01) at 1 year, and improvement continued to 3 years (Table 4). MBSR responses indicated significant improvement in how patients evaluated their appearance (appearance evaluation P ⬍ .01 at 1 and P ⬍ .05 at 3 years). Their appearance orientation was normal before and did not change at 1 or 3 years. Appearance evaluation scores reflect how realistic their assessment of their appearance is to established norms. Appearance orientation is a reflection of the amount of care they take in grooming themselves. Composite scores and scores for each of the 8 subscales on the SF-36 Health Survey improved significantly at 1 year and remained significantly improved at 3 years (the closer the score is to an absolute of 100, the better the quality of life it represents). Complications Perioperative complications were significantly more common in trial A (43%) patients than in trial B (25%) patients (P ⬍ .05; Table 5). At 1 year, occurrences of band slippage and pouch dilatation, stomal obstruction, and nausea and vomiting were significantly less in trial B than trial A patients (Table 5). Complete trial A safety data were gathered during the formal 3-years postsurgical study period at all sites. (Table 5). The greatest proportion of patients experienced ⱖ1 complication, 76% (223 of 292) within the first year after implantation, 53% (141 of 265) of patients developed complications within the second year, and 46% (100 of 218) reported complications within the third year. Surgery-related gastrointestinal perforation in the second and third years after surgery in trial A patients (Table 5) occurred during reoperations. These reoperations were performed either to remove the device or to perform conversion to another bariatric procedure. Esophageal dilatation was reported in trial A patients, most commonly during the second and third years after surgery, and predominantly at 1 site (Table 6). Up to 1 year, there was 1 reported case of esophageal dilatation among trial B patients. In trial A, 18% (51 of 292) of patients had their bands explanted within 3 years of implantation. With a maximal follow-up of 9.5 years, this increased to 33% (96 of 292).

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

339

Table 4 QOL assessment by follow-up visit QOL parameter

Trial A baseline* (N ⫽ 282)

Trial B baseline* (N ⫽ 187)

Trial A 12 mo* (N ⫽ 198)

Trial B 12 mo* (N ⫽ 130)

Trial A 36 mo* (N ⫽ 145)

BDI MBSR appearance evaluation MBSR appearance orientation Rand SF-36 Health Survey Physical function Role limit: physical Role limit: emotional Energy and fatigue Emotional well-being Social functioning Pain General health Mental health composite Physical health composite

13.5 (8.49) 1.9 (.65) 3.8 (.61)

11.7 (7.76) 1.9 (.66) 3.7 (.64)

6.4 (7.25) 2.7 (.80) 3.9 (.59)

4.7 (5.36) 2.8 (.81) 3.8 (.61)

7.0 (8.30) 2.7 (.87) 3.8 (.66)

49.1 (24.52) 39.1 (40.55) 60.6 (40.26) 37.1 (20.99) 65.1 (18.63) 57.4 (27.30) 58.1 (25.41) 52.4 (23.02) 44.4 (11.04) 23.5 (8.46)

52.1 (23.07) 48.5 (39.76) 68.4 (38.73) 37.4 (20.70) 66.8 (18.04) 65.9 (26.10) 64.1 (24.31) 54.3 (21.67) 46.0 (10.56) 25.0 (7.59)

78.6 (24.88) 82.1 (31.46) 77.0 (36.83) 61.3 (20.90) 72.8 (19.45) 81.6 (24.17) 78.6 (22.50) 75.0 (18.92) 48.5 (11.85) 31.7 (7.84)

80.2 (24.46) 90.4 (24.59) 91.6 (21.05) 65.4 (19.99) 77.8 (15.98) 88.9 (17.51) 85.9 (17.99) 75.6 (17.33) 52.4 (7.78) 32.4 (6.56)

76.1 (25.93) 78.8 (35.93) 78.2 (34.55) 58.1 (21.60) 71.9 (20.15) 81.2 (24.28) 78.6 (24.67) 71.6 (19.65) 48.6 (10.14) 31.4 (7.60)

Note: All 12-month and 36-month values are significantly different (P ⬍ .01) from baseline except for MBSR appearance evaluation at 36 months where P ⬍ .05; comparing 36 with baseline values and for appointment orientation scores where there were no differences pre- and postoperatively. * N (%).

The most common reasons for band explantation in the first 5 years in trial A patients were band slippage and pouch dilatation (n ⫽ 15), stoma obstruction (n ⫽ 13), insufficient weight loss (n ⫽ 10), erosions (n ⫽ 3), and 10 other conditions that occurred in 1 patient each. The most common reason for band explantation in trial B at 1 year was for band slippage/pouch dilatation (n ⫽ 5), followed by stoma obstruction (n ⫽ 3). By the end of the third year, 36 (19%) of the patients in trial B had their bands removed, with the majority requesting conversion to gastric bypass for inadequate weight loss. By 5 years after surgery, band slippage and pouch dilatation occurred in 24% of the subjects (70 of 292) in trial A (Table 6). Of these patients, 21% (15 of 70) had their band removed. Stomal obstruction occurred in 14% (40 of 292) (Table 6) of patients, of whom 33% (13 of 40) had their band removed. There were 27 revision procedures by 5.5 years. Nine patients had removal of the band and a new band placed; the existing band was repositioned in 16 patients. Infection, Erosion, and Perforation: There were 4 intraoperative perforations. These were repaired, and band placement was then completed. Within the first postoperative month, 3 patients developed infections necessitating band removal. In the fourth patient, band erosion into the stomach was detected 11 months after a revision for a malpositioned band (16 months after initial placement). Two additional erosions of a band into the stomach, with no identifiable etiologic factor, were diagnosed approximately 3 years after initial band placement. In trials A and B combined, there were a total of 8 perireservoir infections. The entire system was removed in 3 patients. One of these patients had self-accessed his reservoir repeatedly and was converted to a gastric bypass. The other two patients had the system removed without further bariatric surgery. In 4 patients, the reservoir was removed, but the band was left around the stomach with its connecting tubing plugged. When the infection had cleared and the wound healed, a new reservoir was inserted, without further

sequelae. The eighth patient recovered with antibiotic treatment without removal of the reservoir. Superficial wound infections not involving the reservoir healed with local treatment and antibiotics without infecting the band system. Deaths: Two deaths occurred among patients enrolled in trial A. Neither death was technically related to the device. The first was a woman with a history of depression who had been evaluated and cleared for surgery by her psychiatrist. Sixteen months after implantation, she had an elective band explant because of reflux and dietary noncompliance. Ten days after a minor motor vehicle accident, with no significant injuries, she had uneventful band explantation. Six days after band removal, she was found dead in her home. The coroner’s autopsy report refers to injuries from the motor vehicle accident and concludes that death was caused by “mixed drug intoxication.” The second death involved a woman who had her device removed 43 months after implantation because of pouch dilatation, which was 36 months after an open revision because of band slippage. The open surgery entailed an uneventful conversion to a Roux-en-Y gastric bypass and removal of a large ovarian desmoid tumor. The patient died 24 hours later. The autopsy report recorded “multiple, fully formed pulmonary emboli bilaterally.” Comments This USFDA-monitored, open-label trial examining the safety and efficacy of an adjustable gastric band as a treatment for morbid obesity demonstrated that patients lost a mean of 18% of their initial body weight at 1 year and maintained this amount of weight loss to 3 years. Only approximately 20% of those in both trials lost ⱕ10% of their initial body weight at 1 year, a standard the IOM has suggested as being important in evaluating weight loss therapies caused by resolution of comorbid problems. Evidence-based medicine favors prospective, randomized, double-blinded, placebo-controlled studies. However, there are ethical concerns about placebo surgery [39]. In addition, comparative surgical studies [24,25,40,41] are dif-

340

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

Table 5 Number of subjects with ⱖ1 complication in each year Category/system

Trial A N (%) (0 to 1 y)

Trial A N (%) (1 to 2 y)

Trial A N (%) (2 to 3 y)

Trial B N (%) (0 to 1 y)

Total subjects enrolled Subjects with complications Surgery-related Gastrointestinal perforation Hepatic injury Improper band placement Wound infection Spleen injury Band specific Band slippage or pouch dilatation Stoma obstruction Esophageal dilatation Band disorder Band erosion Band leak Reservoir infection Port specific Site pain Displacement Disorder Leak Total digestive Nausea and vomiting Gastroesophageal reflux Constipation Diarrhea Abnormal stools Dysphagia Metabolic and nutritional Healing abnormal Dehydration Edema

292 (100) 223 (76)

265 (100) 141 (53)

218 (100) 100 (46)

193 (100) 127 (66)††

4 (1) 2 (1) 2 (1) 12 (4) 1 (⬍1)

1 (⬍1) 1 (⬍1) 0 (0) 0 (0) 0 (0)

1 (⬍1) 0 (0) 0 (0) 0 (0) 0 (0)

3 (2) 0 (0) 0 (0) 5 (3) 0 (0)

35 (12) 21 (7) 3 (1) 3 (1) 0 (0) 0 (0) 6 (2)

28 (11) 11 (4) 8 (3) 0 (0) 1 (⬍1) 0 (0) 0 (0)

18 (8) 11 (5) 10 (5) 0 (0) 2 (1)* 1 (⬍1) 0 (0)

13 (7) 4 (2)†† 1 (1) 0 (0) 1 (1) 0 (0) 2 (1)

14 (5) 12 (4) 3 (1) 0 (0)

7 (3) 3 (1) 1 (⬍1) 6 (2)

2 (1) 3 (1) 1 (⬍1) 2 (1)

39 (15) 41 (15) 1 ⬍ (1) 2 (1) 1 (⬍1) 8 (3)

21 (10) 36 (17) 3 (1) 1 (⬍1) 2 (1) 8 (4)

45 (23)† 25 (13) 10 (5) 4 (2)†† 0 (0)‡ 6 (3)

3 (1) 1 (⬍1) 0 (0)

1 (⬍1) 0 (0) 0 (0)

7 (4) 2 (1) 0 (0)†

111 (38) 48 (16) 22 (8) 20 (7) 15 (5) 12 (4) 20 (7) 7 (2) 7 (2)

14 (7) 10 (5) 2 (1) 2 (1)

* Includes 1 subject whose event occurred 41 days after the end of the 3-year period. † P ⬍ .0001 comparing decrease in 0- to 1-year incidence in trial B versus trial A. †† P ⬍ .05 comparing decrease in 0- to 1-year incidence in trial B versus trial A.

ficult to complete because patients who have decided they want a surgical or medical therapy usually are unwilling to be randomized to the opposite arm because of differences in efficacy and risk. Comparative surgical studies are also much more expensive [25] than medical studies, especially if no government funding mechanism is available for novel clinical trials. As a consequence, open-label studies, including the A and B trials presented here, are a common form of surgical investigation. Only a minority of prospective investigations monitored by institutional review boards are published [42], which is a cause for concern [43,44]. USFDA oversight of data collection is required for premarket approval submission of implantable devices. After gaining device approval, ongoing safety data submissions are usually required of investigators. This report includes safety data for ⬎9 years of follow-up, making it one of the longest safety evaluations of a bariatric surgical treatment. Our 3-year weight loss data were greater than results achieved with weight loss medications [16 –19] or with diet and exercise [15], which characteristically report weight loss at ⱕ1 year. Of subjects remaining in our study at 3

years, 28% lost ⬍10% IBW, and 72% had weight loss ⬎10% IBW (Table 3). Although other types of bariatric surgical procedures [23,40] have reported greater weight loss, these are associated with a higher mortality rate and greater metabolic and nutritional sequelae, concerns for prospective patients and many surgeons. Investigators still performing bariatric surgery at 4 of the 14 sites in this study, however, no longer offer the gastric band to their patients. They have concluded that the short- and long-term risk– benefit ratios of other bariatric procedures are preferable [45– 47]. The weight loss in our study was less than other largescale, long-term, international studies of adjustable gastric banding [28,31,48 –51]. This may be due to lower dietary fat intake in other countries, more smoking, easier access to medical care, a lower starting BMI, or less rigorous data accrual. We find that patients with a lower initial BMI lost proportionally more excess BMI (Fig. 3), possibly identifying a subpopulation more likely to benefit from gastric banding. The rate of complications in this study (Table 5) reflects the clinical importance of carefully monitoring patients after

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

341

Table 6 TRIAL A: complications up to 5.5 years and explantations up to 9 years as stratified by site Site no. (N)*

1 (18) 2 (25) 3 (28) 4 (20) 5 (36) 6 (36) 7 (75) 8 (54) Total⫽292

Complications

Reoperations

Band slippage or pouch dilatation N (%)§

Stoma obstruction N (%)§

Esophageal dilatation N (%)§

Band erosion N (%)§

Revisions N (%)§

Explants储 Total N (%)

Without conversion†† N (%)

With conversion†† N (%)

2 (11) 2 (8) 8 (29) 6 (30) 7 (19) 7 (19) 22 (29) 16 (30) 70 (24)

1 (6) 0 (0) 3 (11) 5 (25) 2 (6) 3 (8) 9 (12) 17 (31) 40 (14)

1 (6) 0 (0) 0 (0) 0 (0) 16 (44) 0 (0) 4 (5) 1 (2) 22 (8)

0 (0) 0 (0) 1 (4) 0 (0) 0 (0) 0 (0) 0 (0) 2 (4) 3 (1)

0 (0) 0 (0) 1 (4) 0 (0) 3 (8) 5 (14) 6 (8) 12 (22) 27 (9)

1 (6) 7 (28) 9 (32) 8 (40) 19 (53) 4 (11) 21 (28) 27 (50) 96 (33)

1 (6) 4 (16) 3 (12) 8 (40) 4 (11) 4 (11) 9 (12) 5 (9) 39 (13)†

0 (0) 3 (12) 4 (16) 0 (0) 15 (42) 0 (0) 12 (16) 22 (41) 57 (20)

* N ⫽ Number of subjects at site at baseline. † Includes 1 subject whose conversion status was unknown. †† Conversion to another bariatric procedure. § Percent of subjects with complication at individual site up to lock-out date of November 30, 2000, for the USFDA submission (3- to 5.5-year follow-up). 储 Values shown in these 3 explant columns were updated for all explants reported in trial A subjects up to December 15, 2003 (6- to 8.5-year follow-up).

bariatric surgery. For example, within 5 years of surgery, 98% of gastric bypass patients will have many of the complaints (such as abdominal pain, vomiting, and diarrhea) [52], that we included in this report as complications. A recent population-based study from Quebec Province in Canada reported that ⬎36% of patients (N ⫽ 1,035) who had a gastric bypass developed new digestive complaints within 5 years, but the obese control group also developed similar complaints in 25% of the 5,746 who were followed up [53]. Increasing international experience has resulted in fewer complications or reoperations and enhanced weight loss [28,31,50]. As a consequence, the investigators from these 3 centers now offer the gastric band exclusively, including revisional procedures, as obesity treatment for their patients because they are convinced it is the best bariatric procedure. We conclude that, although most prevalent within the first year, complications will probably occur as long as devices remain implanted. After device approval, ongoing safety data submissions are still required of investigators in FDA monitored studies. This report includes safety data for ⬎9 years of follow-up. Band repositioning or replacement was required for 9% of trial A patients by 5 years. Based on extensive experience in Belgium [28], Italy [49], and Australia [31,50], repositioning of a previously placed band is no longer recommended. All revisions should use a new band. Band explantation was undergone by 33% of the patients in trial A, with up to 9 years of follow-up (Table 6). This is similar to the conversion rate reported by Christou et al from failed vertical banded gastroplasty to gastric bypass in their study following up 194 patients for ⬎15 years in Quebec Province [53]. Medical insurance dictated which revisional procedure was performed in many of the cases in this study because before the band was approved by the FDA for general release in June 2001, many insurance companies would only approve conversion to a gastric bypass or vertical banded gastroplasty, not to another band, for these patients.

Explantation and revision rates are similar to the initial experience of other surgeons using adjustable gastric bands since the early 1990s [28,31,50,53,54]. Flum et al recently reported that the mortality rate for surgeons performing their first 19 procedures was 5 times that of more experienced surgeons, which is consistent with other learningcurve studies in general surgery [55,56]. However, more recent technical refinements and education of professionals and patients appears to have decreased the slippage and explantation rates. After intraoperative gastric perforation, our infection and erosion data suggest not proceeding with device implantation. Patients considering bariatric surgery have a different psychosocial profile [57] than those requesting less-invasive weight loss therapy. In this study, QOL parameters improved and were sustained during 3 years, despite reaching a weight loss plateau by 2 years, corroborating 2 year data of others [58]. All novel implantable devices have a rapid mechanical and surgical technique evolution [59]. The improved results in the trial B patients may be ascribed to a modified surgical technique, more experienced laparoscopic surgeons, and the learning curve. The intersite outcome differences presented may be an unavoidable consequence of multicenter studies. For example, 1 site, at which 7% (36 of 485) of patients of the total cohort had their device implanted [46], reported a 44% esophageal dilatation rate. This high rate was not observed at any other site in this study (Table 6) nor in any international series, including thousands of patients [28,31,48 – 51]. The frequency of this complication, however, may be underreported because routine radiological upper gastrointestinal series were not obtained at 3-year follow-up in the international studies, and not enough time may have evolved for the complication to have developed in the trial B patients. Esophageal dilitation may occur after undiagnosed preexisting esophageal dysmotility [60,61] and/or overly aggressive inflation of the band. Bands should not be

342

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343

implanted in patients with profound esophageal dysfunction. Esophageal dilatation may be prevented by refusing patient demands to overtighten the band to enhance weight loss. Timely removal of saline from an overinflated system will reverse esophageal dilatation in most patients who have a gastric band. Neither of our trials systematically evaluated the optimal manner of band adjustment, a subject that requires appropriate hypothesis-driven investigation. In addition, in our opinion, long-term, repetitive, barium swallow assessment of foregut anatomy and function and band position is indicated. The finding of our posthoc subgroup analysis for black patients is similar to a single-site analysis [46] in which African-American patients lost less weight than nonblack patients. This study was not designed to address this issue, and the number of subjects per race group is insufficient to provide confidence in this finding. Longitudinal modeling of weight loss suggested that baseline weight and time from surgery were the only strong predictors of weight loss in this study. Several factors could account for these differences other than race alone. These include dietary choices, socioeconomic status, educational background, and possible bias by treating professionals, although we did not collect these data prospectively. We conclude that appropriate prospective studies, with adequate subject accrual, are indicated to assess the effect of race on weight loss outcomes after gastric banding as well as other bariatric procedures. In summary, these results suggest that the laparoscopic adjustable gastric banding system may be a viable option for patients considering bariatric surgery, although safety beyond 8 years and efficacy beyond 3 years remain to be established. Physicians and potential patients should be aware that individual bariatric surgeons will offer different operations because of personal and professional preferences. Acknowledgments We are indebted to the individuals who participated in this study and to the investigators at each study site, including the following: Trial A—site 1: Lubomyr Kuzmak, M.D., Irvington, NJ; site 2: Robert McIntrye, M.D., Denver, CO; site 3: Alan Wittgrove, M.D., San Diego, CA; site 4: Cornelius Doherty, M.D., Iowa City, IA; site 5: Eric DeMaria, M.D., John Kellum Jr, M.D., and Harvey Surgeman, M.D., Richmond, VA; site 6: Kenneth MacDonald Jr, M.D., Greenville, NC; site 7: Robert Greenstein, M.D., and Michel Gagner, M.D., New York, NY; and site 8: Louis Martin, M.D., New Orleans, LA. Trial B—site 1: Joseph Amaral, M.D., Providence, RI; site 2: J. Ken Champion, M.D., Marietta, GA; site 3: Carlos Gracia, M.D., Pleasanton, CA; site 4: Kenneth MacDonald Jr, M.D., Greenville, NC; site 5: Lou-Ann Galibert, M.D., White Plains, NY; site 6: Louis Martin, M.D., New Orleans, LA; site 7: Philip Schauer, M.D., Pittsburgh, PA; site 8: Richard Rubenstein, M.D., East Patchogue, NY; site 9: J. Steven Scott, M.D., Wentzville, MO; site 10: Alan Wittgrove, M.D., San Diego, CA; site 11: Edward Phillips, M.D., and Scott Cunneen, M.D., Los Angeles, CA; and site 12: Michel Gagner, M.D, New York, NY. The following authors had responsibility for the named contributions — Study concept and design: Inamed Health

Corporation, Santa Barbara, CA; data acquisition: all 22 investigators at 16 sites; data analysis and interpretation: Gerard Smits, Ph.D., Louis Martin, M.D., and Robert Greenstein, M.D.; manuscript draft: Louis Martin, M.D., and Robert Greenstein, M.D.; critical revision of the manuscript for important intellectual content: Louis Martin, M.D., Gerard Smits, Ph.D., and Robert Greenstein, M.D.; statistical expertise: Gerard Smits, Ph.D.; study supervision: Inamed Health and USFDA; and administrative support: Louis Martin. The role of the study sponsor—These trials were entirely funded by the manufacturer of the adjustable gastric band, Inamed Health Corporation, Santa Barbara, CA. The company designed the study and had the design approved by the USFDA. The company’s clinical monitors collected and confirmed the data submitted at each site and also paid an independent company to review all data submitted at each site to have independent confirmation of the data sent to the statistician, Dr. Gerard Smits. They were allowed to review the manuscript after it was completed. Access to data—Gerard Smits, Ph.D., an independent contractor, managed all data for this study. Dr. Smits produced the data and statistics for the tables and figures initially requested by Drs. Martin and Greenstein. Many analyses could not be included in the manuscript because of space constraints. Dr. Smits advised Drs. Martin and Greenstein on additional data he thought should be reviewed and on indicated additional statistical analyses. The Inamed Health Corporation gave the authors compete access to all data without restrictions and obtained, collated, and updated datasets with additional information, from every site, that Drs. Martin and Greenstein requested. The authors had an independent statistician, Ronald Horswell, Ph.D., Louisiana State University School of Public Health, review the data and the manuscript. His suggestions for clarification were addressed to his satisfaction. These studies were entirely funded by the manufacturer of the adjustable gastric band, Allergan (Irvine, CA). References [1] Flegal KM, Carroll MD, Ogden CL, et al. Prevalence and trends in obesity among US adults, 1999 –2000. JAMA 2002;288:1723–7. [2] Prevalence of Overweight and Obesity Among Adults: United States, 1999 (initial results). National Health and Nutrition Examination Survey 1999 (NHANES 1999). Oct. 24, 2002. Available at: http://www.cdc. gov/nchs/products/pubs/pubd/hestats/obese/obse99.htm. Accessed January 13, 2004. [3] Doll S, Paccaud F, Bovet P, et al. Body mass index, abdominal adiposity and blood pressure: consistency of their association across developing and developed countries. Int J Obes Relat Metab Disord 2002;26:48 –57. [4] Friedrich MJ. Epidemic of obesity expands its spread to developing countries. JAMA 2002;287:1382– 6. [5] Detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel III). Third Report of the National Cholesterol Education Program (NCEP) Expert Panel: NHLBI of the NIH; May 2001. National Institutes of Health No. 01-3670. Bethesda, MD: National Institutes of Health. [6] National Institutes of Health. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults—the evidence report. Obes Res 1998;2(suppl 6):51S–209S. [7] Obesity: preventing and managing the global epidemic. Report of a WHO Consultation on Obesity. Report No. NUT/NCD/98.1. Geneva, Switzerland: World Health Organization; 1998.

L.F. Martin et al. / The American Journal of Surgery 194 (2007) 333–343 [8] Pi-Sunyer FX. Medical hazards of obesity. Ann Intern Med 1993; 119:655– 60. [9] Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of obesity, diabetes and obesity-related health risk factors. JAMA 2002;289:76 –9. [10] Stunkard AJ, Stinnett JL, Smoller JW. Psychological and social aspects of the surgical treatment of obesity. Am J Psychiatry 1986; 143:417–29. [11] Fontaine KR, Redden DT, Wang C, et al. Years of life lost due to obesity. JAMA 2003;289:187–93. [12] Wolf AM, Colditz GA. Current estimates of the economic cost of obesity in the United States. Obes Res 1998;6:97–106. [13] Committee to Develop Criteria for Evaluating the Outcomes of Approaches to Prevent and Treat Obesity, Institute of Medicine. Weighing the Options: Criteria for Evaluating Weight Management Programs. Washington DC: National Academy Press; 1995. [14] Guidance for the Clinical Evaluation of Weight Control Drugs (924-1996). Division of Metabolic and Endocrine Drug Products, Food and Drug Administration; Rockville, MD; 1996. [15] Wing RR, Venditti E, Jakicic JM, et al. Lifestyle intervention in overweight individuals with a family history of diabetes. Diabetes Care 1998;21:350 –9. [16] Lean ME. Sibutramine—a review of clinical efficacy. Int J Obes Relat Metab Disord 1997;21(suppl 1):S30 –S39. [17] Wadden TA, Berkowitz RI, Sarwer DB, et al. Benefits of lifestyle modification in the pharmacologic treatment of obesity: a randomized trial. Arch Intern Med 2001;161:218 –27. [18] Davidson MH, Hauptman J, DiGirolamo M, et al. Weight control and risk factor reduction in obese subjects treated for 2 years with orlistat: a randomized controlled trial. JAMA 1999;281:235– 42. [19] James WP, Astrup A, Finer N, et al. Effect of sibutramine on weight maintenance after weight loss: a randomized trial. STORM Study Group. Sibutramine Trial of Obesity Reduction and Maintenance. Lancet 2000;356:2119 –25. [20] Gastrointestinal surgery for severe obesity: National Institutes of Health Consensus Development Conference Statement. Am J Clin Nutr 1992;55(2 suppl):615S–9S. [21] Mason EE, Ito C. Gastric bypass. Ann Surg 1969;170:329 –36. [22] Mason EE. Vertical banded gastroplasty for obesity. Arch Surg 1982; 117:701– 6. [23] Hess DS, Hess DW. Biliopancreatic diversion with a duodenal switch. Obes Surg 1998;8:267– 82. [24] Andersen T, Backer OG, Stokholm KH, et al. Randomized trial of diet and gastroplasty compared with diet alone in morbid obesity. N Engl J Med 1984;310:352– 6. [25] Sjostrom L, Larsson B, Backman L, et al. Swedish obese subjects (SOS). Recruitment for an intervention study and a selected description of the obese state. Int J Obes Relat Metab Disord 1992;16:465–79. [26] Mason EE, Renquist KE, Jiang D. Perioperative risks and safety of surgery for severe obesity. Am J Clin Nutr 1992;55(2 suppl):573S– 6S. [27] Martin LF, White S, Lindstrom W Jr. Cost-benefit analysis for the treatment of severe obesity. World J Surg 1998;22:1008 –17. [28] Belachew M, Legrand MJ, Vincent V. History of Lap-Band: from dream to reality. Obes Surg 2001;11:297–302. [29] Scopinaro N. The IFSO and obesity surgery throughout the world. International Federation for the Surgery of Obesity. Obes Surg 1998; 8:3– 8. [30] Build Study. Chicago, Illinois: Society of Actuaries and Association of Life Insurance Medical Directors of America Statistical Bulletin; Jan-Jun 1983. [31] O’Brien PE, Brown WA, Smith A, et al. Prospective study of a laparoscopically placed, adjustable gastric band in the treatment of morbid obesity. Br J Surg 1999;86:113– 8. [32] Deitel M, Greenstein RJ. Recommendations for reporting weight loss [editorial]. Obes Surg 2003;13:159 – 60. [33] Kuczmarski RJ, Flegal KM. Criteria for definition of overweight in transition: background and recommendations for the United States. Am J Clin Nutr 2000;72:1074 – 81. [34] Beck AT, Steer RT. Manual for the Beck Depression Inventory. San Antonio, TX: Psychological Corporation; 1993. [35] Cash TF. Body-image attitudes among obese enrollees in a commercial weight-loss program. Percept Mot Skills 1993;77:1099 –1103.

343

[36] Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30:473– 83. [37] Kuzmak LI. A review of seven years’ experience with silicone gastric banding. Obes Surg 1991;1:403– 8. [38] Diggle PJ, Heagerty P, Liang K-Y, et al. Analysis of Longitudinal Data. 2nd ed. [Vol. 1]. New York, NY: Oxford University Press; 2002. [39] Horng S, Miller FG. Is placebo surgery unethical? N Engl J Med 2002;347:137–9. [40] Sugerman HJ, Starkey JV, Birkenhauer R. A randomized prospective trial of gastric-bypass versus vertical-banded-gastroplasty for morbidobesity and their effects on sweet-eaters versus non-sweet-eaters. Ann Surg 1987;205:613–22. [41] Moseley JB, O’Malley K, Petersen NJ, et al. A controlled trial of arthroscopic surgery for osteoarthritis of the knee. N Engl J Med 2002;347:81– 8. [42] Pich J, Carne X, Arnaiz JA, et al. Role of a research ethics committee in follow-up and publication of results. Lancet 2003;361:1015– 6. [43] Antes G, Chalmers I. Under-reporting of clinical trials is unethical. Lancet 2003;361:978 –9. [44] Mann H. Research ethics committees and public dissemination of clinical trial results. Lancet 2002;360:406 – 8. [45] Doherty C, Maher JW, Heitshusen DS. Long-term data indicate a progressive loss in efficacy of adjustable silicone gastric banding for the surgical treatment of morbid obesity. Surgery 2002;132:724 – 8. [46] DeMaria EJ, Sugerman HJ, Meador JG, et al. High failure rate after laparoscopic adjustable silicone gastric banding for treatment of morbid obesity. Ann Surg 2001;233:809 –18. [47] Kellum JM. Gastric banding. Ann Surg 2003;237:17– 8. [48] Forsell P, Hallerback B, Glise H, et al. Complications following Swedish adjustable gastric banding: a long-term follow-up. Obes Surg 1999;9:11– 6. [49] Angrisani L, Alkilani M, Basso N, et al. Laparoscopic Italian experience with the Lap-Band. Obes Surg 2001;11:307–10. [50] Allen JW, Coleman MG, Fielding GA. Lessons learned from laparoscopic gastric banding for morbid obesity. Am J Surg 2001;182: 10 – 4. [51] Dargent J. Laparoscopic adjustable gastric banding: lessons from the first 500 patients in a single institution. Obes Surg 1999;9:446 –52. [52] Pories WJ, MacDonald KG Jr, Morgan EJ, et al. Surgical treatment of obesity and its effect on diabetes: 10-y follow-up. Am J Clin Nutr 1992;55(2 suppl):582S–585S. [53] Christou NV, Sampalis JS, Liberman M, et al. Surgery decreases long-term mortality, morbidity, and health care use in morbidly obese patients. Ann Surg 2004;240:416 –24. [54] Flum DR, Dellinger EP. Impact of gastric bypass operation on survival: a population-based analysis. J Am Coll Surg 2004;199: 543–51. [55] Flum DR, Koepsell T, Heagerty P, et al. The nationwide frequency of major adverse outcomes in antireflux surgery and the role of surgeon experience, 1992–1997. J Am Coll Surg 2002;195:611– 8. [56] Flum DR, Koepsell T, Heagerty P, et al. Common bile duct injury during laparoscopic cholecystectomy and the use of intraoperative cholangiography: adverse outcome or preventable error? Arch Surg 2001;136:1287–92. [57] Martin LF. The biopsychosocial characteristics of people seeking treatment for obesity. Obes Surg 1999;9:235– 43. [58] Dixon JB, Dixon ME, O’Brien PE. Quality of life after Lap-Band placement: influence of time, weight loss, and comorbidities. Obes Res 2001;9:713–21. [59] Peppelenbosch N, Buth J, Harris PL, et al. Diameter of abdominal aortic aneurysm and outcome of endovascular aneurysm repair: does size matter? A report from EUROSTAR. J Vasc Surg 2004;39:288 –97. [60] Greenstein RJ, Nissan A, Jaffin BW. Esophageal anatomy and function in laparoscopic gastric restrictive bariatric surgery: implications for patient selection. Obes Surg 1998;8:199 –206. [61] Jaffin BW, Knoepflmacher P, Greenstein RJ. High prevalence of asymptomatic esophageal motility disorders among morbidly obese patients. Obes Surg 1999;9:390 –5. The results of the US Food and Drug Administration monitored premarket approval trial of the first adjustable gastric band approved for use in the United States as a surgical treatment for morbid obesity and its co-morbid medical conditions are presented.