The American Heart Association and the American College of Cardiology have recommended physical activity and weight management for prevention of cardiovascular disease (CVD) events (1,2). Extensive data from cardiac rehabilitation studies have shown the possibility and benefits of improving fitness and losing weight in those with CVD (3). Jakicic et al. (4) showed that among the Look AHEAD cohort receiving a lifestyle intervention, those with a self-reported history of CVD improved cardiorespiratory fitness significantly less after 1 yr than did those without such history. However, beyond this early Look AHEAD report, no trials have assessed whether adults who are overweight or obese with diabetes and evidence or history of CVD can improve fitness that is similar to those without CVD. Moreover, it is unclear if similar weight loss can be achieved between individuals with and without a history of CVD, despite clinical recommendations indicating that weight loss may reduce CVD risk factors. Similarly, it is not known whether a preparticipation graded exercise tolerance test (GXT) that is abnormal by established criteria (5), but not severe enough to exclude participants from participation, will affect weight loss or improvement in cardiorespiratory fitness that occur during a lifestyle intervention program.
The Look AHEAD Study provides a unique opportunity to examine these questions. In Look AHEAD, overweight/obese participants were randomly assigned to an intensive lifestyle intervention (ILI) or a control group (diabetes support and education, or DSE) and followed over time. At 4 yr, the ILI group had greater weight loss and larger improvements in fitness than did DSE. However, the question of whether a history of CVD or an abnormal preparticipation GXT influences the change in weight or cardiorespiratory fitness achieved in response to a behavioral intervention focused on weight loss and physical activity has not previously been examined. These findings may provide insight into the characteristics of patients with type 2 diabetes that may influence response to a comprehensive behavioral weight management program. Moreover, these findings may be important for physicians and other health care providers when considering whether a lifestyle intervention focused on weight loss and physical activity is appropriate for individuals with a history of CVD or with the presence of abnormalities on a preparticipation GXT.
Baseline data on Look AHEAD participants (6) and study methods (7) have been published previously. Look AHEAD is registered at ClinicalTrials.gov (NCT00017953). Briefly, Look AHEAD enrolled 5145 participants aged 45 to 76 yr with type 2 diabetes and with body mass index ≥25 kg·m−2 (≥27 kg·m−2 if using insulin). Informed consent was obtained from each participant before screening. The study protocol was approved by the institutional review board of each center and the National Institute of Diabetes and Digestive and Kidney Diseases.
Assessing Cardiorespiratory Fitness
All participants underwent a symptom-limited maximal GXT before randomization to assess baseline fitness and rule out contraindications to moderate intensity exercise. Participants also underwent submaximal GXTs after 1 and 4 yr in the study. The maximal and submaximal GXTs have been described elsewhere (4,8). To proceed from the baseline GXT to randomization, the participant must have achieved at least 85% of predicted maximal heart rate (HRmax = 220 − age) and a calculated workload of at least 4 metabolic equivalents (METs). For participants taking a β-blocking medication, an RPE of at least 18 on the Borg 15-category scale was required rather than the heart rate criteria. The predetermined stopping point for each submaximal test conducted at 1 and 4 yr was achieving at least 80% of predicted maximal heart rate if not taking a β-blocking medication or RPE of at least 16 on the Borg scale if taking a β-blocker. Cardiorespiratory fitness was defined as the estimated MET level based on the treadmill work load (i.e., speed and grade). As previously described, to compute the change in cardiorespiratory fitness, MET level at the point of test termination for the submaximal GXT (80% of predicted maximal heart rate if not taking a β-blocking medication or RPE of at least 16 on the Borg scale if taking a β-blocker) was compared with the MET level attained at the same submaximal point from the baseline GXT as previously described (4,8) (see Slides, Supplemental Digital Content 1, which illustrate the of procedures for computing the change in fitness from baseline, using a maximal exercise test, to follow-up, using a submaximal exercise test, http://links.lww.com/TJACSM/A109). To evaluate heart rate and rhythm when determining an abnormal GXT, heart rate was determined from printed 12-lead electrocardiograms obtained during the last 10 s of every minute, immediately at test termination, and every 2 min during recovery.
Criteria for Abnormal GXT
An abnormal baseline GXT was defined by criteria associated with risk of acute exercise-related cardiovascular events or CVD mortality. Abnormal tests showed evidence of ischemia (ST-segment depression of ≥1.0 mm), ventricular arrhythmias (ventricular fibrillation, ventricular tachycardia, runs of at least three premature ventricular contractions, or exercise-induced bundle branch block), exercise-induced angina, abnormal hemodynamic response (heart rate recovery (HRR) <22 bpm 2 min after exercise), or impaired exercise capacity (IEC; highest achieved workload of <5 METs) (9). Participants with abnormal GXT results were excluded from the study or referred to their physician for evaluation after which some were subsequently cleared for study participation.
Defining History of CVD
At baseline, participants were asked about any history of myocardial infarction, stroke, transient ischemic attack, coronary artery bypass graft, coronary angioplasty/stent, carotid endarterectomy, lower extremity angioplasty, aortic aneurism repair, or heart failure. If a participant reported a history of any of these events or conditions, he/she was deemed to have a history of CVD. If any CVD events occurred within 3 months before screening, the participant was offered the opportunity to rescreen in 3 months or was excluded before randomization.
Assessment of Body Weight
Weight was measured annually in duplicate on a digital scale, and change was follow-up weight minus baseline weight.
The intervention has been described in detail elsewhere (6,10,11). Briefly, each ILI participant had contact weekly, in a combination of group and individual sessions, with a health coach for the first 6 months, three times per month for the next 6 months, and at least monthly through year 4. ILI participants were encouraged to adhere to a calorie-restricted diet (1200–1800 kcal·d−1) with <30% of calories from fat, <10% from saturated fat, and ≥15% from protein and had a weight loss goal of 10%. They were also encouraged to exercise at least 175 min·wk−1 at the intensity of brisk walking and were given support in problem solving, goal setting, and self-monitoring to achieve these objectives.
The DSE group was given three educational sessions per year focused on diet, physical activity, and social support, but they were not given weight loss, nutrition, or physical activity goals, nor were they given specific support for behavior changes (12).
The distributions of selected baseline characteristics and GXT abnormalities for participants grouped by intervention assignment were compared using χ2 tests and t-tests. Mixed-effects models were fitted to the longitudinal changes in weight and fitness using maximum likelihood (13). This approach uses all data collected through the 4 yr of follow-up and avoids the assumption that any data are missing completely at random (13). We used Wald tests to assess differences between participants grouped by intervention assignment with the presence of GXT abnormalities or history of CVD as predictors in models. Linear contrasts were used to estimate mean differences at year 4 to summarize effect sizes from these models. Participants were grouped by intervention assignment in all analyses (i.e., intention-to-treat), regardless of adherence.
This analysis includes the 5011 (97.4%) Look AHEAD participants for whom weight and fitness data were available at baseline and at least one annual follow-up visit. Baseline characteristics of the subset included in this analysis are shown in Table 1. All variables are balanced by intervention assignment. As shown in Table 1, 13.8% of patients reported a history of CVD and 19.3% experienced an abnormal baseline GXT test. ST depression and abnormal HRR were the most common abnormalities. At 4 yr, 2009 (78.2%) participants in ILI and 1919 (74.5%) in DSE completed GXTs.
TABLE 1 -
Baseline Description of Participants, Prevalence of Baseline GXT Abnormalities, and Self-Reported Prior CVDa
at Baseline for Participants Grouped by Intervention Assignment.
||DSE (n = 2497)
||ILI (n = 2514)
|Age (mean, SD), yr
|Sex, no. (% female)
|Race/ethnicity, no. (%)
| African American
| American Indian
| Not Latino
| African American
| American Indian
|BMI, mean (SD), kg·m−2
|HbA1c, mean (SD), (%)
|Insulin use, no. (%)
|ST depression, no. (%)
|Arrhythmia, no. (%)
|Angina, no. (%)
|HRR, no. (%)
|IEC, no. (%)
|Any GXT abnormality, no. (%)
|Self-reported prior CVD, no. (%)
|GXT abnormality and/or prior CVD, no. (%)
|GXT referral and/or prior CVD, no. (%)
|GXT referral, no. (%)
Data are limited to participants with at least one postrandomization annual weight measurement.
aDefined as self-report of myocardial infarction, stroke, transient ischemic attack, coronary artery bypass graft, angioplasty/stent, carotid endarterectomy, angioplasty of lower extremity, aortic aneurysm repair, or heart failure.
BMI, body mass index.
Compared with the DSE group, the ILI produced marked and statistically significant 4-yr weight losses, overall and after stratification by every category of GXT abnormality and history of CVD, after adjusting for age and sex (Fig. 1). After adjusting for age and sex, those with a self-reported history of CVD lost significantly less weight from baseline to year 4 in both DSE (−0.07% (weight gain) vs 0.73%, P = 0.01) and ILI (5.98% vs 6.69%, P = 0.02) compared with those without a history of CVD. In contrast, weight loss in either ILI or DSE did not differ between those participants who had any GXT abnormality and those who did not. Specifically within the ILI, mean weight loss was significantly greater in participants with arrhythmia (12.35% vs 6.57%, P = 0.008), angina (10.27% vs 6.56%, P = 0.01), or abnormal HRR (7.87% vs 6.48%, P = 0.03) compared with those without these abnormalities. However, the number of participants with each of these abnormalities was small, particularly ventricular arrhythmia (n = 6) and angina (n = 14). In DSE, none of the abnormalities observed on the GXT were associated with differences in 4-yr weight losses.
Change in Cardiorespiratory Fitness
The ILI had significantly greater improvements in fitness from baseline to year 4 than DSE, overall and within all subgroups. Relative to those without a history of CVD, self-reported history of CVD was associated with poorer 4-yr fitness changes in both the DSE (−0.12 METs vs 0.06 METs, P = 0.01) and ILI (0.43 METs vs 0.61 METs, P = 0.004) groups (Fig. 2). Those who had any GXT abnormality at baseline did not experience different fitness outcomes from those with no GXT abnormalities (Fig. 2). However, some specific GXT abnormalities at baseline did affect fitness change. In ILI, ventricular arrhythmia was associated with greater fitness improvement (1.47 METs vs 0.58 METs, P = 0.04). In both ILI and DSE, participants with IEC at baseline had smaller improvements in fitness than did those who did not have IEC (DSE: −0.17 METs with IEC vs 0.03 METs without IEC, P = 0.01; ILI: 0.25 METs with IEC vs 0.62 METs without IEC, P = 0.01).
Participants in ILI had significantly greater weight loss and improved fitness at 4 yr compared with those assigned to DSE, regardless of baseline history of CVD or GXT abnormalities, and this is consistent with the main findings from the Look AHEAD Study that demonstrated greater weight loss and improved fitness in ILI compared with DSE (10,11). However, the magnitude of the changes between those with and without a self-reported history of CVD, although statistically significant, was modest and may not be clinically meaningful. Moreover, magnitude of these changes between those with and without GXT abnormalities was not statistically or clinically meaningful. It has been suggested that a 3% difference in weight change (14) or a 2-MET difference in fitness (15) may be needed to affect CVD risk, which may support that these be considered when determining clinical meaningfulness of these observed differences. Physicians and other health care providers should expect similar responses in weight loss and fitness to an ILI in patients with overweight/obesity and with type 2 diabetes regardless of the presence of a history of CVD or abnormalities observed on a preparticipation GXT.
The underlying cause for this statistically significant, but clinically modest, difference in weight loss by CVD history is not clear. One hypothesis to explain this finding is that participants’ perception of the risk of participating in physical activity may be influenced by CVD history, which may have reduced energy expenditure and subsequent weight loss. For example, participants with a known history of CVD may have been reluctant to exercise as long, as often, or as vigorously as those without CVD. Jakicic et al. (16) found that at baseline, in the subset of Look AHEAD participants (n = 2240) whose physical activity was monitored by accelerometry, there was no significant difference between those with and without a history of CVD in the number of bouts of physical activity per day, minutes per bout, or number of bouts with intensities of at least 6 or 10 METs. However, it is possible that physical activity in response to the intervention differed between participants with and without CVD, and this warrants investigation.
The aforementioned analyses were repeated excluding participants who underwent bariatric surgery, were diagnosed with cancer or heart failure, underwent a coronary artery bypass graft, or died before their year 4 examination (data not shown). The results did not change materially, suggesting that these conditions, which are commonly associated with weight loss, were not the cause of weight loss seen in the trial.
Although weight loss in ILI and DSE did not differ between those who had an abnormal GXT and those who had a normal test, some specific abnormalities were associated with greater weight loss in ILI. After adjustment for age and sex, Look AHEAD participants ILI and who experienced ventricular arrhythmia, angina pectoris, or impaired HRR during their baseline stress test lost more weight than did their ILI counterparts without these conditions. These differences in weight loss were not associated with being referred for medical clearance on the basis of an abnormal GXT, as those referred did not lose more weight than did those not referred (P = 0.55, data not shown). However, the small number of participants with these abnormalities limits any conclusions that can be drawn from these findings and therefore should be interpreted with caution.
The association between angina pectoris on the baseline GXT and greater subsequent weight loss in the ILI group and a nonsignificant trend toward weight loss in the DSE group suggests a mechanism that is enhanced but not entirely created by the lifestyle intervention. This conclusion is supported by two observational studies that showed greater weight loss in those experiencing angina, either by self-report (17) or during a GXT (18), than in those without angina. In the latter study, those with angina and weight loss had worse clinical outcomes. Again, such interpretations are limited by the fact that only 14 participants in ILI and 14 in DSE had angina on the baseline GXT, and therefore, these findings warrant caution being interpreted for clinical meaningfulness.
Within each subgroup of GXT abnormality and among those with a history of CVD, the ILI was associated with improved fitness, relative to DSE. DSE participants with an abnormal baseline GXT did not experience fitness change differently from those with normal GXTs, except that baseline IEC or history of CVD was associated with declining fitness over 4 yr. An IEC at baseline or history of CVD was also associated with less improvement in fitness in the ILI group compared with ILI participants without these characteristics. Morris et al. (19) showed, in a large cohort referred for GXTs, that the expected rate of decline in fitness is greater in sedentary than in active men. Kokkinos et al. (20), in their review of the fitness studies in CVD, observed a strong, consistent, inverse association between both fitness and change in fitness over time and CVD outcomes, with this association being at least as strong as any other described (e.g., smoking, hyperlipidemia). IEC is closely correlated with physical activity level (21) in what is likely a bidirectional relationship. Those with IEC may have been more sedentary before the study, which could explain the more rapid fitness decline in the DSE group. A likely explanation for these fitness changes is that those with IEC had a greater burden of CVD at baseline, whether previously diagnosed or not, and experienced decline (DSE) or less improvement (ILI) in fitness over the observed period because of their underlying CVD. Thus, patients with overweight/obesity and type 2 diabetes should be monitored for potential declines in fitness and should also be encouraged to engage in safe and effective interventions, which may need to be medically monitored, to sustain or potentially enhance fitness.
A self-reported history of CVD was associated with a decrease in fitness in the DSE group and a lesser improvement in fitness in the ILI group; however, the ILI resulted in significant improvement in fitness among those with and without a history of CVD. As with IEC, this history of CVD may be a marker of subsequent clinical decline (22).
One potential incongruity merits attention. ILI participants with IEC at baseline lost more weight but improved fitness less than did those with normal exercise capacity; however, the modest differences may not be clinically meaningful and should be interpreted with caution. Given the strong correlation between weight loss and fitness gain, this divergence seems counterintuitive. In this analysis, the dietary aspect of the intervention, which has been shown by Wadden et al. (23) to have had a strong effect on weight loss, was not considered. Determining long-term outcomes in these participants is beyond the scope of this article; however, such evaluation may elucidate whether the weight loss seen in participants with IEC at baseline is a harbinger of overall clinical decline.
Strengths of this study include the size and diversity of the population included in the Look AHEAD trial and the high rate of participant retention. Look AHEAD also has also demonstrated the safety and effectiveness of the lifestyle intervention program in achieving both weight loss and fitness gain, which are commonly recommended for treating patients with type 2 diabetes to improve diabetes control and reduce its long-term complications.
This study also included limitations. Study participants included volunteers who met inclusion criteria expected to heighten adherence, which may result in these participants not reflecting the general population. Although retention was high, the analyses included only participants for whom follow-up data are available. Individuals at highest risk for CVD events may not be represented given the inclusion criteria for this study. Because the independent variables (abnormal GXT and history of CVD) are risk factors for CVD, the underrepresentation of those at highest risk for CVD events may make the results inaccurate with regard to both weight loss and fitness change. An additional limitation of this study may be the use of a submaximal GXT to monitor change in fitness across the intervention period, and future studies should consider using a symptom-limited maximal GXT at all assessment periods to evaluate change in fitness.
In this large cohort of participants with obesity and type 2 diabetes, those participants without a history of CVD achieved significant, but modestly greater, improvements in weight losses and fitness in both ILI and DSE than did those with a CVD history. Moreover, those with any GXT abnormality did not differ in weight loss or fitness changes from those with no abnormalities. However, in all subgroups, participants assigned to ILI had greater weight losses and improvements in fitness at year 4 than did those assigned to DSE regardless of CVD history or GXT abnormalities. Thus, overweight and obese adults with type 2 diabetes who have markers of CVD on a GXT or with a known history of CVD can lose weight and improve fitness with a lifestyle intervention. These findings provide insight into the characteristics of patients with type 2 diabetes that may influence response to a comprehensive behavioral weight management program. Moreover, these findings may be important for physicians and other health care providers, and suggest that a lifestyle intervention focused on weight loss and physical activity can be effective for treating overweight/obesity and improving fitness for individuals with a history of CVD or with the presence of abnormalities on a preparticipation GXT.
A complete list of the Look AHEAD Study Group is provided (see Document, Supplemental Digital Content 2, which provides a complete list of the Look AHEAD Study Group, http://links.lww.com/TJACSM/A110). Results of the present study do not constitute endorsement by the American College of Sports Medicine.
The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate manipulation. The following organizations have committed to make major contributions to Look AHEAD: FedEx Corporation; Health Management Resources; LifeScan, Inc., a Johnson & Johnson Company; OPTIFAST of Nestle HealthCare Nutrition, Inc.; Hoffmann-La Roche Inc.; Abbott Nutrition; and Slim-Fast Brand of Unilever North America. Dr. Jakicic is on the Scientific Advisory Board for WW (formerly Weight Watchers International, Inc.). No other potential conflicts of interest relevant to this article were reported.
The Look AHEAD study was funded by the National Institutes of Health (NIH) through cooperative agreements with the National Institute of Diabetes and Digestive and Kidney Diseases (DK-57136, DK-57149, DK-56990, DK-57177, DK-57171, DK-57151, DK-57182, DK-57131, DK-57002, DK-57078, DK-57154, DK-57178, DK-57219, DK-57008, DK-57135, and DK-56992). Additional funding was provided by the National Heart, Lung, and Blood Institute; National Institute of Nursing Research; National Center on Minority Health and Health Disparities; NIH Office of Research on Women’s Health; and Centers for Disease Control and Prevention. This research was supported in part by the Intramural Research Program of the National Institute of Diabetes and Digestive and Kidney Diseases. The Indian Health Service provided personnel, medical oversight, and use of facilities. The opinions expressed in this article are those of the authors and do not necessarily reflect the views of the Indian Health Service or other funding sources. Additional support was received from The Johns Hopkins Medical Institutions Bayview General Clinical Research Center (M01-RR-02719); the Massachusetts General Hospital Mallinckrodt General Clinical Research Center and the Massachusetts Institute of Technology General Clinical Research Center (M01-RR-01066); the Harvard Clinical and Translational Science Center (RR-025758-04); the University of Colorado Health Sciences Center General Clinical Research Center (M01-RR-00051) and Clinical Nutrition Research Unit (P30-DK-48520); the University of Tennessee at Memphis General Clinical Research Center (M01-RR-0021140); the University of Pittsburgh General Clinical Research Center (M01-RR-000056); the University of Pittsburgh Clinical and Translational Research Center funded by a Clinical and Translational Science Award (UL1-RR-024153) and NIH grant DK-046204; the Wake Forest Clinical and Translational Science Institute (UL1-TR-001420); the VA Puget Sound Health Care System Medical Research Service, Department of Veterans Affairs; and the Frederic C. Bartter General Clinical Research Center (M01-RR-01346).
1. Eckel RH, Jakicic JM, Ard J, et al. 2013 AHA/ACC guideline on lifestyle management to reduce cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation
. 2014;129(25 Suppl 2):S76–99.
2. Jensen MD, Ryan DH, Apovian CM, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Obesity Society. J Am Coll Cardiol
3. Anderson L, Taylor RS. Cardiac rehabilitation for people with heart disease: an overview of Cochrane systematic reviews. Cochrane Database Syst Rev
. 2014;12:CD011273(Art. No.: CD011273).
4. Jakicic JM, Jaramillo SA, Balasubramanyam A, et al. Effect of a lifestyle intervention on change in cardiorespiratory fitness in adults with type 2 diabetes: results from the Look AHEAD study. Int J Obes (Lond)
5. Smith SC, Blair SN, Bonow RO, et al. AHA/ADD guidlelines for preventing heart attack and death in patients with atherosclerotic cardiovascular disease. J Am Coll Cardiol
6. Wadden TA, West DS, et alLook AHEAD Research Group. The Look AHEAD Study: a description of the lifestyle intervention and the evidence supporting it. Obesity (Silver Spring)
7. Ryan DH, Espeland MA, Foster GD, et al. Look AHEAD (Action for Health in Diabetes): design and methods for a clinical trial of weight loss for the prevention of cardiovascular disease in type 2 diabetes. Control Clin Trials
8. Jakicic JM, Egan CE, Fabricatore AN, et al. Change in cardiorespiratory fitness and influence on diabetes control and CVD risk factors in adults with type 2 diabetes: 4-year results from the Look AHEAD trial. Diabetes Care
9. Curtis JM, Horton ES, Bahnson J, et al. Prevalence and predictors of abnormal cardiovascular responses to exercise testing among individuals with type 2 diabetes: the Look AHEAD (Action for Health in Diabetes) study. Diabetes Care
10. Pi-Sunyer X, Blackburn G, et alLook AHEAD Research Group. Reduction in weight and cardiovascular disease risk factors in individuals with type 2 diabetes: one-year results of the Look AHEAD trial. Diabetes Care
11. Wing RRLook AHEAD Research Group. Long-term effects of a lifestyle intervention on weight and cardiovascular risk factors in individuals with type 2 diabetes mellitus: four-year results of the Look AHEAD trial. Arch Intern Med
12. The Look AHEAD Reseach Group. The development and description of the comparison group in the Look AHEAD trial. Clin Trials
13. Little RJA, Rubin DR. Statistical Analysis With Missing Data
. 2nd ed. New York: John Wiley and Sons; 2002.
14. Stevens J, Truesdale KP, McClain JE, Cai J. The definition of weight maintenance. Int J Obes (Lond)
15. Gregg EW, Jakicic JM, et alLook AHEAD Research Group. Association of the magnitude of weight loss and changes in physical fitness with long-term cardiovascular disease outcomes in overweight or obese people with type 2 diabetes: a post-hoc analysis of the Look AHEAD randomised clinical trial. Lancet Diabetes Endocrinol
16. Jakicic JM, Gregg E, Knowler W, et al. Physical activity patterns of overweight and obese individuals with type 2 diabetes in the Look AHEAD study. Med Sci Sports Exerc
17. Wolinsky FD, Andresen EM, Malmstrom TK, Schootman M, Miller JP, Miller DK. Three-year measured weight change in the African American Health Study. J Aging Health
18. Myers J, Lata K, Chowdhury S, McAuley P, Jain N, Froelicher V. The obesity paradox and weight loss. Am J Med
19. Morris CK, Myers J, Froelicher VF, Kawaguchi T, Ueshima K, Hideg A. Nomogram based on metabolic equivalents and age for assessing aerobic exercise capacity in men. J Am Coll Cardiol
20. Kokkinos P, Faselis C, Myers J, Sui X, Zhang J, Blair SN. Age-specific exercise capacity threshold for mortality risk assessment in male veterans. Circulation
21. Lakoski SG, Barlow CE, Farrell SW, Berry JD, Morrow JR Jr., Haskell WL. Impact of body mass index, physical activity, and other clinical factors on cardiorespiratory fitness (from the Cooper Center Longitudinal Study). Am J Cardiol
22. Fogelholm M. Physical activity, fitness and fatness: relations to mortality, morbidity and disease risk factors. A systematic review. Obes Rev
23. Wadden TA, Neiberg RH, Wing RR, et al. Four-year weight losses in the Look AHEAD study: factors associated with long-term success. Obesity