Change in Pain and Physical Function Following Bariatric Surgery for Severe Obesity

Importance  The variability and durability of improvements in pain and physical function following Roux-en-Y gastric bypass (RYGB) or laparoscopic adjustable gastric banding (LAGB) are not well described.

Objectives  To report changes in pain and physical function in the first 3 years following bariatric surgery, and to identify factors associated with improvement.

Design, Setting, and Participants  The Longitudinal Assessment of Bariatric Surgery-2 is an observational cohort study at 10 US hospitals. Adults with severe obesity undergoing bariatric surgery were recruited between February 2005 and February 2009. Research assessments were conducted prior to surgery and annually thereafter. Three-year follow-up through October 2012 is reported.

Exposures  Bariatric surgery as clinical care.

Main Outcomes and Measures  Primary outcomes were clinically meaningful presurgery to postsurgery improvements in pain and function using scores from the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) (ie, improvement of ≥5 points on the norm-based score [range, 0-100]) and 400-meter walk time (ie, improvement of ≥24 seconds) using established thresholds. The secondary outcome was clinically meaningful improvement using the Western Ontario McMaster Osteoarthritis Index (ie, improvement of ≥9.7 pain points and ≥9.3 function points on the transformed score [range, 0-100]).

Results  Of 2458 participants, 2221 completed baseline and follow-up assessments (1743 [78.5%] were women; median age was 47 years; median body mass index [BMI] was 45.9; 70.4% underwent RYGB; 25.0% underwent LAGB). At year 1, clinically meaningful improvements were shown in 57.6% (95% CI, 55.3%-59.9%) of participants for bodily pain, 76.5% (95% CI, 74.6%-78.5%) for physical function, and 59.5% (95% CI, 56.4%-62.7%) for walk time. Additionally, among participants with severe knee or disability (633), or hip pain or disability (500) at baseline, approximately three-fourths experienced joint-specific improvements in knee pain (77.1% [95% CI, 73.5%-80.7%]) and in hip function (79.2% [95% CI, 75.3%-83.1%]). Between year 1 and year 3, rates of improvement significantly decreased to 48.6% (95% CI, 46.0%-51.1%) for bodily pain and to 70.2% (95% CI, 67.8%-72.5%) for physical function, but improvement rates for walk time, knee and hip pain, and knee and hip function did not (P for all ≥.05). Younger age, male sex, higher income, lower BMI, and fewer depressive symptoms presurgery; no diabetes and no venous edema with ulcerations postsurgery (either no history or remission); and presurgery-to-postsurgery reductions in weight and depressive symptoms were associated with presurgery-to-postsurgery improvements in multiple outcomes at years 1, 2, and 3.

Conclusions and Relevance  Among a cohort of participants with severe obesity undergoing bariatric surgery, a large percentage experienced improvement, compared with baseline, in pain, physical function, and walk time over 3 years, but the percentage with improvement in pain and physical function decreased between year 1 and year 3.

Trial Registration  clinicaltrials.gov Identifier: NCT00465829

Severe obesity is associated with significant joint pain and impaired physical function (ability to bend, lift, carry, push, and walk).^1,2 Excess weight bearing can lead to joint damage and pain, resulting in activity restriction and walking limitations.³ Obesity can also contribute to pain and physical limitations through factors such as impaired cardiorespiratory function,⁴ systematic inflammation,⁵ reduced flexibility of movement,⁶ low strength per body mass,⁵ and depression.⁷

Bariatric surgery is effective at achieving and maintaining weight loss, and inducing remission or reducing severity of many comorbidities such as type 2 diabetes mellitus, hypertension, dyslipidemia, and depression.^8,9 Although evidence of improvements in pain and physical function following bariatric surgery is increasing, the variability and durability of improvement have not been well described—with most studies limited by small sample size and follow-up of 1 year or less or by the study of obsolete surgical procedures.^10-12

This report examines pain and physical function outcomes in a large multisite cohort study with annual follow-up. The aim of the study was to evaluate changes in bodily and joint-specific pain and physical function, including perceived and objectively measured walking capacity in the first 3 years following bariatric surgery, and to identify factors associated with presurgery-to-postsurgery improvements.

The Longitudinal Assessment of Bariatric Surgery-2 (LABS-2) study is an observational study of 2458 adults who underwent an initial bariatric surgical procedure between March 14, 2006, and April 24, 2009, at 1 of 10 hospitals at 6 US clinical centers.^13,14 The institutional review boards at each center and the data coordinating center approved the protocol and all participants provided written informed consent.

LABS-2 had a target sample size of 2400 participants, based on anticipated loss to follow-up of 17% to 25% and the desire to detect small effect sizes for continuous outcomes (requires ≤1800 participants) and odds ratios (ORs) of at least 2.0 for categorical outcomes with 90% power (requires ≤2000 participants) for discrete outcomes in most circumstances.

Assessments were conducted by trained personnel within 30 days prior to scheduled surgery and annually following surgery. To be included, participants had to complete the baseline and at least 1 follow-up assessment within the first 3 postsurgery years (2221 participants; 91% with baseline data), with data collection ending October 2012 (Figure 1).

The Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) is a generic measure of functional health and well-being with proven validity, reliability, and sensitivity to change.¹⁵ Two domain scores were examined: bodily pain, composed of 2 items that assess the magnitude of bodily pain and how much it interferes with activities; and physical function, composed of 10 items that assess whether health limits various activities. Norm-based methods were used to transform the scores to a mean (SD) of 50 (10) in the general US population.¹⁶ Higher SF-36 scores indicate less pain or better function. An increase of at least 5 points represents a clinically important improvement.¹⁷ The 3 items specific to walking limitations were also examined.

The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) measures symptoms of hip and knee osteoarthritis. It has demonstrated reliability and validity and is sensitive to preintervention-to-postintervention changes.¹⁸ This analysis used 2 scores per joint: pain, composed of 5 items that assess pain level during various activities; and function, composed of 17 items that assess difficulty performing various activities. Scores were transformed to a scale (0-100; lower scores indicate less pain and better function).¹⁹ A decrease of at least 9.7 pain points and at least 9.3 function points represents a clinically important improvement.²⁰

Participants reported history of back, hip, knee, and ankle surgery; use of pain medication (prescription or over-the-counter) for back, hip(s), knee(s), or ankle(s), in the past week; the effects and level of dissatisfaction or satisfaction with back or leg pain; use of a mobility aid; and severe walking limitation (inability to walk 61 m unaided).¹³

The 400-m Long-Distance Corridor Walk (LDCW) was used to objectively measure walking capacity.^21,22 To minimize risk, participants were instructed to walk the 400-m course at their usual pace and to wear a heart rate monitor. After completion, participants were asked if they experienced back, hip, knee, or foot pain during the LDCW. Testing could be terminated prior to completion for safety reasons. Participants were ineligible to attempt the LDCW if they had a contraindication to exertion, had any of several cardiovascular risk factors (eg, hospitalized for myocardial infarction, underwent angioplasty, or saw a clinician for new or worsening chest pain in the past 3 months), needed a mobility aid other than a straight cane, or reported feeling unsafe. Participants were categorized as having a mobility deficit if they met the LDCW exclusion criteria or stopping criteria or exceeded 7 minutes to walk 400 m (which equates to a cardiorespiratory fitness level of <12 mL oxygen/kg/min, the minimal level deemed necessary to safely cross a traffic intersection).²³ A decrease of at least 24 seconds indicates a substantial improvement.²⁴

Heart rate, measured after a minimum of 5 minutes of seated rest, was used as a proxy for cardiovascular fitness.

Primary end points were clinically meaningful presurgery-to-postsurgery improvements in pain and function scores from the SF-36 Health Survey and 400- m walk time. Secondary end points were clinically meaningful improvements in the WOMAC and remission of mobility deficit.

Anthropometric measurements followed standardized protocols.⁸ Sociodemographics and smoking status were self-reported. Race was set to missing for participants who did not self-report as one or more of the following: white/Caucasian, black/African American, Asian, American Indian/Alaska Native, Native Hawaiian/other Pacific Islander. β-Blocker use was determined from the therapeutic class of self-reported prescribed medications. Diabetes and history of stroke, cardiovascular disease (CVD), asthma, sleep apnea, and venous edema with ulcerations were determined using laboratory values (eg, hemoglobin A_1c), physical examination measures (eg, blood pressure), patient-reported medication use, comorbidity diagnoses from clinicians, and medical records review using standard definitions.¹⁴ Due to lack of sleep studies following surgery, only baseline sleep apnea status was used in this analysis. Depressive symptoms in the past week were assessed using the Beck Depression Inventory (BDI) version 1.²⁵ Higher BDI scores (range, 0-63) indicate greater depressive symptomatology.

Analyses were conducted using SAS version 9.3. Potential selection bias was examined by comparing preoperative characteristics and presurgery-to-postsurgery percent weight loss of LABS-2 participants in the analysis sample (n = 2221) to those excluded (n = 237) using the Pearson χ² test for categorical variables and the Wilcoxon rank-sum test for continuous variables. Baseline characteristics were summarized with frequencies and percentages for categorical data and median plus interquartile (IQR) ranges for continuous data.

Longitudinal analyses was performed with mixed models (eAppendix 1 [Supplement]) using all available data, with control for age and site, which were associated with missing follow-up data (eTable 1 [Supplement]). Sensitivity analyses, performed to examine the robustness of results with respect to the missing at random assumption, indicated that missing follow-up data was not related to outcomes (eAppendices 1-2, eTable 1 [Supplement]).

Mixed models were used to estimate and test changes in pain and physical function over time. Poisson mixed models with robust error variance were used for binary measures (eg, mobility deficit); mixed-effects multinomial logistic regression models for nominal categorical measures (eg, LDCW status); mixed-effects ordinal logistic regression models for ordinal measures (eg, degree to which pain interfered with work); and linear mixed models for continuous measures (eg, SF-36 score). Analysis of WOMAC scores was limited to participants with symptoms indicative of osteoarthritis (severe or extreme rating on ≥1 item in the relevant joint²⁶) at baseline. Analysis of pain during the LDCW excluded participants who did not start, and LDCW completion time excluded those who did not complete the LDCW. Analysis of heart rate excluded those who reported taking β-blockers. Pairwise comparisons were made between baseline and each follow-up and between years 1 and 3. The 4 comparisons were tested using the t statistic with P values adjusted to control for overall type I error (eAppendix 3 [Supplement]).²⁷

The proportions of participants with clinically meaningful presurgery-to-postsurgery improvements based on SF-36 scores (among the total sample), WOMAC scores, and the LDCW time (among subgroups previously described) were calculated using established thresholds^17,20,24 with Poisson mixed models with robust error variance. An additional 12 to 23 patients whose baseline WOMAC scores did not meet the thresholds for clinically important improvement were excluded. Postsurgery remission of mobility deficit (among patients with a presurgery mobility deficit) was also calculated. Year-1 and year-3 proportions were compared using t statistic (eAppendix 4 [Supplement]).

Modeled percentages or means and 95% CIs, are reported. Summary statistics of observed data are reported online (eTables 2-3 [Supplement]).

Poisson mixed models with robust error variance were used to identify factors related to clinically important presurgery-to-postsurgery improvements in SF-36 and WOMAC scores, LDCW time, and remission of mobility deficit in years 1, 2, and 3. Sex, race, baseline age, household income, body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), depressive symptoms, surgical procedure, presurgery-to-postsurgery smoking status, change in depressive symptoms, and percent weight change were included in all models as independent variables, with control for site. Baseline and presurgery-to-postsurgery change in bodily pain were also included as independent variables in models of physical function and LDCW outcomes. Baseline sleep apnea status, history of stroke, and postsurgery status of diabetes, asthma, venous edema with ulcerations, and CVD symptoms with consideration for baseline status (ie, current vs remitted/no symptoms in past 12 months, and no history) were also considered as independent variables and retained if statistically significant (P < .05). Adjusted relative risks (RRs), 95% CIs, and P values are reported. All reported P values are 2-sided; P <.05 was considered to be statistically significant.

This report includes 2221 of 2428 study participants (91%) with baseline data (Figure 1; Table 1). Pain and function data were obtained in 2042 (84%) participants at year 1, 1794 (74%) at year 2, and 1724 (72%) at year 3. Participants who were excluded (n = 237) vs included (n = 2221) from the analysis sample were younger (median age, 41 vs 47 years) and had a lower household income, a higher proportion of current smokers (22% vs 12%), and a lower proportion of sleep apnea (43% vs 54%) (P value for all <.05). There was also a significant difference in site representation between participants who were excluded vs included. There were no other statistically significant differences in baseline characteristics, or in percent of weight loss from baseline at year 1, year 2, or year 3, controlling for baseline characteristics that differed by group (P value for all ≥.05).

RYGB was the most common surgical procedure (70.4%); one-fourth (25.0%) of patients underwent LAGB; and less than 5% underwent another procedure (sleeve gastrectomy, banded gastric bypass, or biliopancreatic diversion with duodenal switch). Median (IQR) percentage weight loss of baseline weight at year 1 was 30.5% (21.3%-37.5%), at year 2 it was 30.5% (21.3%-38.5%), and at year 3 it was 28.2% (19.8%-36.4%) overall. Following RYGB, percentage weight loss at year 1 was 34.1% (28.7%-39.2%), at year 2 it was 34.1% (27.7%-40.3%), and at year 3 it was 31.5% (24.8%-8.4%). Following LAGB, percentage weight loss at 1 year was 14.0% (9.7%-19.7%), at year 2 it was 16.1% (9.4%-23.0%), and at year 3 it was 16.2% (8.1%-23.1%).

Following surgery, SF-36 scores were significantly higher for bodily pain (baseline, 39.9 [95% CI, 39.5-40.3] to 44.8 [95% CI, 44.3-45.3] at year 3) and for physical function (baseline, 36.5 [95% CI, 36.1-37.0] to 47.8 [95% CI, 47.4-48.3] at year 3). WOMAC scores were significantly lower for knee pain (baseline, 46.5 [95% CI, 44.9-48.1] to 26.2 (95% CI, 24.1-28.2] at year 3), hip pain (baseline, 47.4 [95% CI, 45.6-49.2] to 25.7 [95% CI, 23.4-28.0] at year 3), knee function (baseline, 48.6 [95% CI, 47.2-50.0] to 24.6 [95% CI, 22.7-26.6] at year 3), and hip function (baseline, 46.7 [95% CI, 45.0-48.3] to 22.2 [95% CI, 20.0-24.4] at year 3), indicating improvements in the entire sample and specifically among participants with severe knee or hip pain or disability at baseline (Table 2). The degree to which back or leg pain interfered with work and level of dissatisfaction with current back or leg pain symptoms also improved postsurgery (Figure 2; eTable 4 [Supplement]). Likewise, at follow-up, a smaller percentage of participants reported medication use for leg or back pain, leg or back pain during the LDCW, back or leg pain that prevented going to work or school, health limitations that impeded walking and mobility aid use, and were measured to have a mobility deficit, reflecting a higher percentage of participants able to complete the LDCW and faster LDCW completion time. Resting heart rate also improved (Table 2). In contrast, the prevalence of severe walking limitations at year 3 was not significantly different from baseline.

At year 1, the majority of participants had clinically meaningful presurgery-to-postsurgery improvements for bodily pain (57.6% [95% CI, 55.3%-59.9%]), physical function (76.5% [95% CI, 74.6%-78.5%]), and walk time (59.5% [95% CI, 56.4%-62.7%]). Additionally, the majority of participants with severe knee or hip pain or disability at baseline experienced joint-specific improvements (for knee pain, 77.1% [95% CI, 73.5%-80.7%]; hip pain, 74.1% [95% CI, 69.7%-78.4%]; hip physical function, 79.2% [95% CI, 75.3%-83.1%]), and the majority of participants with a mobility deficit at baseline experienced remission by year 1 (55.6% [95% CI, 52.0%-59.3%]) (Table 3).

Rates of improvement in LDCW time, knee and hip pain, knee and hip function, and mobility deficit remission did not significantly differ between year 1 and year 3 (Table 3). However, by year 3, rates were significantly lower vs year 1 for bodily pain (48.6% [95% CI, 46.0%-51.1%]) and physical function (70.2% [95% CI, 67.8%-72.5%]) (Table 3). Likewise, the prevalence of medication use for pain, back pain during the LDCW, health limitations to walking, and mobility aid use postsurgery increased during follow-up (Table 2). Despite these postsurgery deteriorations in improvement, year-3 status was significantly better than baseline status.

Associations between sociodemographics and clinical characteristics with clinically meaningful improvements in bodily pain (SF-36 score), physical function (SF-36 score), walking capacity (LDCW completion time), and knee pain (WOMAC score) are shown in Table 4. Associations with additional clinically meaningful improvements in hip pain, knee function and hip function (WOMAC scores), and no longer having a mobility deficit are provided in eTable 5 (in the Supplement). Younger age, higher household income, and fewer depressive symptoms presurgery, greater presurgery-to-postsurgery percent weight loss, decline in depressive symptoms, and remission or no history of diabetes were associated with greater likelihood of clinically meaningful improvements in most measures of pain and physical function. Additionally, men, as compared with women, had greater likelihood of improvement in joint-specific and bodily pain, while lower BMI and less bodily pain presurgery, greater presurgery-to-postsurgery decline in bodily pain, and remission or no history of venous edema with ulcerations were associated with greater likelihood of improvement in multiple measures of physical function. More pain presurgery was associated with higher likelihood of improvement in pain, and worse function presurgery was associated with higher likelihood of improvement in function. Type of surgical procedure (RYGB, LAGB, and other) was not related to any outcomes. The numbers of participants with clinically meaningful improvements by year, outcome, and categorical variables in Table 4 and eTable 5 (Supplement) are reported in eTable 6 and eTable7 (Supplement).

The incidence of past-year hip, knee, or ankle surgery in year 1 was 3.7% (95% CI, 2.9%-4.6%), for year 2 it was 4.9% (95% CI, 2.9%-4.6%), and for year 3 it was 4.6% (95% CI, 3.6%-5.6%); the majority of which were knee surgeries (eTable 8 [Supplement]). Past-year back surgery incidence ranged from a year-1 level of 1.5% (95% CI, 0.9%-2.0%) to a year-3 level of 2.3% (95% CI, 1.5%-3.1%).

The primary findings of this study are that through 3 years of follow-up: (1) approximately 50% to 70% of adults with severe obesity who underwent bariatric surgery experienced clinically significant improvements in perceived bodily pain and physical function and in objectively measured walking capacity; and (2) approximately three-fourths of participants with severe knee and hip pain or disability at baseline experienced clinically significant improvements in symptoms indicative of osteoarthritis. Additionally, results of this study suggest that, while response to surgery was variable, there were several presurgery factors and postsurgery changes that were consistently associated with improvements in pain and physical function following bariatric surgery.

Changes in SF-36 and WOMAC pain scores observed in this study indicate that following bariatric surgery, the majority of patients initially experience clinically meaningful improvements in bodily and joint-specific pain, although the percentage with improvement in pain and function, as measured by the SF-36, decreased from year 1 to year 3. Additionally, the findings that following surgery, a smaller percentage of patients had back or leg pain that prevented them from going to work or school and that the degree to which back or leg pain interfered with work improved suggest that bariatric surgery may lead to improvements in work productivity and related costs. However, when considering the clinical implications of bariatric surgery on pain, it is important to note that bariatric surgery patients, as a group, continue to have more pain following surgery than the general US population, as indicated by the mean standardized SF-36 bodily pain score.¹⁶ Additionally, at year 3, approximately 1 in 3 participants took pain medication within the prior week for back pain or leg pain and were dissatisfied with their level of back or leg pain.

This study revealed substantial improvements in joint-specific and general measures of function. The mean standardized SF-36 physical function score improved approximately 1 standard deviation following surgery, such that it was close to that of the general US population at year 3 (48 vs 50).¹⁶ Improvements in resting heart rate and LDCW walking time and the related improvement (decrease) in the proportion of patients with an objectively defined mobility deficit are also noteworthy—as walking capacity is a strong predictor of incident mobility limitations and all-cause mortality. Still, at year 3, approximately one-fourth of patients self-reported limitations with walking several blocks and exhibited evidence of an objectively measured mobility deficit, indicating that a sizable portion of postsurgery patients may have walking limitations that hinder ability to follow physical activity recommendations for weight loss maintenance.²⁸

The findings from this study reinforce shorter-term results from studies that have reported significant improvements in SF-36 bodily pain and physical function scores,^29-32 WOMAC scores,^29,33,34 walking capacity (as measured by the LDCW or the 6-minute walk test),^30,31,35,36 resting heart rate,^35,36 or other measures of pain and function^10-12 in the first 3 to 12 months following RYGB or LAGB. Few studies have reported on pain and function from longer-term follow-up of these procedures. In a cross-sectional analysis, Sanchez-Santos et al³⁷ reported that a smaller percentage of adults at least 5 years after RYGB (22% [50]) had difficulty with mobility when compared with controls matched on presurgery characteristics (55%; P < .01). Likewise, Raoof et al³⁸ reported that mean (SD) scores for bodily pain and physical function on the SF-36 were better 12 (3) years post-RYGB when compared with those of morbidly obese controls who were awaiting surgery (matched on age and sex). Another study,³⁹ which provided follow-up for 145 LAGB patients for 3 to 8 years, reported that fewer patients had significant knee pain at follow-up vs baseline (38% vs 47%; P < .01). The authors speculated that the majority of patients may not have experienced a clinical improvement because of irreversible degenerative joint damage.

In the Swedish Obesity Study, which primarily studied vertical banded gastroplasty, a smaller percentage of male and female surgical patients reported pain in most body parts (neck, back, hip, knee, and ankle) 2 years following surgery vs nonsurgical controls. However, most comparisons were not statistically significant at 6-year follow-up, despite substantial (albeit smaller) differences in 6-year weight change (mean difference >20 kg in both men and women).⁴⁰ Given that we detected deteriorations in some measures of pain and function between year 1 and year 3, evaluating longer-term follow-up will be important for elucidating the durability of surgery-induced improvements and examining how weight regain, which becomes common 2 or more years following RYGB and LAGB,⁸ and other factors (such as physical activity participation) may explain deteriorations.

This study identifies several baseline characteristics such as younger age, higher household income, fewer depressive symptoms, and no history of diabetes related to improvements in pain and physical function following surgery, as well as presurgery-to-postsurgery changes associated with improvements in pain and function. Similar to 2 previous studies, degree of weight loss was consistently related to improvement,^29,40 while surgical procedure, independent of weight loss, was not. However, contrary to a previous study,²⁹ which did not find an association between resolution of comorbid medical conditions and improvements in SF-36 or WOMAC scores 6 to 12 months following RYGB (N = 48), our data suggest that change in several comorbidities was associated with changes in pain and function. Specifically, patients who had remission of diabetes, remission of venous edema with ulcerations, and no symptoms of cardiovascular disease in the past 12 months had greater improvements in either pain, function, or both than patients who continued to have each condition and/or symptoms, respectively. Also, improvement in depressive symptoms was related to 7 of 8 outcomes. Although this association may be bidirectional, the finding that having fewer presurgery depressive symptoms was associated with improvements in pain suggests amelioration of depressive symptoms may contribute to postsurgery pain perception. Less bodily pain presurgery and presurgery-to-postsurgery decline in bodily pain were associated with improvements in physical function and walking capacity, after controlling for factors related to both pain and function. Thus, effective pain management may help postsurgical patients improve their physical function and walking capacity.

Limitations of this study include the lack of a nonsurgical control group, precluding us from establishing that surgery caused observed changes in pain and physical function. Additionally, we do not know whether knee or hip pain reflected osteoarthritis pain or widespread chronic pain, and we did not assess abdominal pain, which may affect perceived joint or bodily pain, or the number of years of obesity and severe obesity, which might affect likelihood of improvement. Missing follow-up data are also a concern because they can affect statistical power or bias the findings. However, the initial sample size and retention rate were adequate to ensure sufficient power for the analyses; all longitudinal analyses controlled for baseline factors (age and site) related to missing follow-up data; and the sensitivity analysis showed that those missing pain or function outcomes at 1, 2, or 3 years vs those not missing these data had similar improvement rates at the other follow-up time points, indicating the missing data have a minimal affect on the results. The current study’s large geographically diverse sample, inclusion of multiple validated measures of pain and physical function, longitudinal design, and follow-up through 3 years make it one of the most informative studies of pain and function following RYGB and LAGB to date.

Among a cohort of patients with severe obesity undergoing bariatric surgery, a large percentage experienced improvement compared with baseline in pain, physical function, and walk time over 3 years. However, the percentage with improvement in pain and physical function decreased between year 1 and year 3 following surgery.

Corresponding Author: Wendy C. King, PhD, Epidemiology Data Center, Room 105, 130 DeSoto St, University of Pittsburgh, Graduate School of Public Health, Pittsburgh, PA 15261 (kingw@edc.pitt.edu).

Author Contributions: Drs King and Chen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: King, Mitchell, Pories, Wolfe.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: King.

Critical revision of the manuscript for important intellectual content: King, Chen, Belle, Courcoulas, Dakin, Elder, Flum, Hinojosa, Mitchell, Pories, Wolfe, Yanovski.

Statistical analysis: King, Chen, Wolfe.

Obtained funding: Belle, Flum, Mitchell, Pories, Wolfe.

Administrative, technical, or material support: Belle, Courcoulas, Dakin, Flum, Wolfe.

Study supervision: Wolfe, Yanovski.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Coucourlas reports receipt of research grants from Covidien, Ethicon, Nutrisystem, and PCORI; and consultant fees from Apollo Endosurgery. Dr Dakin reports receipt of consultant fees from Covidien. Dr Pories reports receipt of research grants from Johnson & Johnson Janssen Pharmaceuticals. Dr Wolfe reports receipt of consultant fees from EnteroMedics. Dr Flum reports having had an advisor role with Pacira Pharmaceuticals, providing expert testimony for Surgical Consulting LLC, and receiving travel expenses from PCORI. The other authors report no disclosures.

Funding/Support: LABS-2 was a cooperative agreement funded by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). Grant numbers: Data Coordinating Center–U01 DK066557; Columbia University Medical Center–U01-DK66667 (in collaboration with Cornell University Medical Center CTRC, grant UL1-RR024996); University of Washington–U01-DK66568 (in collaboration with CTRC, grant M01RR-00037); Neuropsychiatric Research Institute–U01-DK66471; East Carolina University–U01-DK66526; University of Pittsburgh Medical Center–U01-DK66585 (in collaboration with CTRC, grant UL1-RR024153); Oregon Health & Science University–U01-DK66555.

Role of the Sponsor: The NIDDK scientists contributed to the design and conduct of the study, which included collection and management of data. The project scientist from the NIDDK served as a member of the steering committee, along with the principal investigator from each clinical site and the data coordinating center. The data coordinating center housed all data during the study and performed data analyses according to a prespecified plan developed by the data coordinating center biostatistician and approved by the steering committee and independent data and safety monitoring board. The decision to publish was made by the LABS-2 steering committee, with no restrictions imposed by the sponsor. As a coauthor, an NIDDK scientist contributed to the interpretation of the data and preparation, review, or approval of the manuscript.