Table 1. Summary of results.

Author Year Country	Bariatric procedure & modifiable behavior	Study design & population studied	Intervention/Group comparisons	Main results
Alvarez 2016 Chile	Sleeve gastrectomy D, B	Case control of 40 participants, 24 -55 months post-op, 80% female, avg 43 years old; pre-op BMI 35 kg/m^2. Self-reported food intake and standardized psychological questionnaires, and some guided.	2 groups: >50th %ile weight regain and <50th %ile weight regain, by 2 years post-op.	Group with higher vs lower weight regain: + 11.05 kg vs +3.55 kg. (p = .00). Gastric volume (p = .023), higher fat (p = .04) and energy intake (p = .09), and post-operative higher anxiety (p = .01) are associated with weight regain. Alcohol consumption (p = .81), fat as % total calories (p = .28), BED, depression, and exercise (p = .49) were not.
Bastos 2013 Brazil	RYGB D	Cross-sectional study. 64 patients, 53.4 months post-op, 89% female, avg 41 years old; pre-op BMI 49.5 kg/m^2. Interviews and questionnaires.	2 groups: > 50% EWL and < 50% EWL by 2 years post-op.	28% of patients experienced weight regain >15% of their lowest post-op weight. Working in food production (cafeteria, baker, snack bar, restaurant, grocery store) was correlated with weight regain (p = .003). Exercise (p = .075), alcohol intake (p = .572), and nutrition monitoring (p = .110) were not.
Faria 2010 Brazil	RYGB D	Intervention (non-randomized, non-blinded). 30 patients, 2–7 years post-op, 86% female, avg age 36 years old; pre-op BMI 43 kg/m^2.	3-month nutrition intervention.	Weight regain pre-intervention vs weight loss post- intervention: 8 kg vs 4.3 kg. 86% of patients had significantly lost weight and significantly reduced their BMI (p < .001) in patients with previous weight regain.
Yanos 2015 USA	RYGB D, B, E	Cross sectional. 97 patients, 8.76 years post-op, 77.3% female, 92.8% Caucasian, avg age 56.11 years; pre-op BMI: n/a. Assessed self- management behaviors, drug and alcohol use, food addiction, and physical activity.	N/A	Average of 26% weight regain from nadir weight. Significant weight regain (=20% weight regain from nadir weight) was associated with dietary adherence (p = .005-.018), physical activity modification (p = .002), nocturnal eating (p = .01), depression (p = .001), and problematic alcohol (p = .01).
Reid 2016 Canada	RYGB D	Cross sectional. 27 patients, 12.15 years post-op, 89% female, avg age 53 years old, pre-op BMI: n/a. Assessed dietary, vitamin compliance, physical activity, and follow-up.	2 groups: Weight maintainers: having lost ≥38% total body weight. Weight regainers: having lost ≤30% total body weight.	Weight loss of weight maintainers vs weight regainers: 44.4% vs 18.2%. People who regained their weight reported consuming more carbohydrates (p < .05) and alcohol (p < .05). No difference in frequency of vitamin supplementation and contact with a healthcare professional.
Himes 2015 USA	RYGB D, B	Intervention (non-randomized, non-blinded). 28 patients, avg 4 years post-op, 93% female, all Caucasian; avg age 53 years old; pre-op BMI: n/a.	6-week group CBT and DBT treatment.	Weight regain of 17 kg (37% of initial weight loss) prior to intervention. Weight loss of 1.6 kg (p ≤ .01) after the 6-week intervention. Patients reversed their pattern of weight regain (weight loss of 1.6 kg, p ≤ .01). Subjective binge eating (p ≤ .03), the number of daily snacks (3.9 to 2.7, p ≤ .01) and the number of eating episodes per day (6.7 to 5.5, p ≤ .01) significantly decreased during the intervention.
Mitchell 2016 USA	RYGB LAGB B	Cohort of 2022 patients, 3 years post-op, 78% female, avg age 47 years old; pre-op BMI n/a. Assessed dietary and lifestyle behaviors, including drug and alcohol abuse pre-op and post-op.	N/A	The factors associated with a 16% variability in weight at 3 years post RYGB are: weekly self-weighing, continuing to eat when feeling full more than once a week, and eating continuously during the day. Applying these positive habits post-op results in a 14% greater weight loss than those who don't apply them (p < .001).
McGrice 2015 USA	Sleeve gastrectomy RYGB LAGB D, B, E	Review: To review the challenges and solutions of interventions that improve long-term weight loss post bariatric surgery.	N/A	Loss of control eating, excessive energy intake, and a lack of exercise were associated with weight regain.
Lauti 2016 New Zealand	Sleeve gastrectomy S, E	Systematic review of 20–208 patients 2–9 years post-op age: n/a pre-op BMI 34.3-45.8 kg/m2. 5/21 studies addressed modifiable factors for weight regain Follow-up: 3 studies Lifestyle behavior: 2 studies	N/A	Rates of regain ranged from 5.7%at 2 years to 75.6%at 6 years. 2 studies addressing modifiable factors showed a regain of 20% EWL. Regular and frequent follow- up was associated with less weight regain, while maladaptive eating and lack of exercise was attributed to weight gain. Other non-modifiable factors were identified.
Karmali 2013 Canada	RYGB AGB VBG Sleeve gastrectomy Other D, B, S, E	Systematic review of 26–1845 patients 1–11.4 years post-op age: n/a pre-op BMI: n/a. 8/16 studies showed modifiable factors associated with weight regain.	N/A	Regain between 7.3–9% of EWL; increased BMI ~5.3 points; regain of =15% of total weight loss; regained ~22.6 lbs. Higher caloric intake, poor dietary quality, BED, grazing, depression, alcohol and substance abuse, lack of self-monitoring, little or no follow-up with team, and a lack of exercise were associated with weight regain. Other non-modifiable factors were identified.
Amundsen 2016 Norway	RYGB B, E	Case control. 40 patients, avg 5 years post-op, 82% female, avg 46 years old; pre-op BMI 44.1. Assessed dietary intake, eating behavior psychometrics, and physical activity through questionnaires.	2 groups: significant weight regain (>15%) vs normal weight regain (=15%).	Significant weight regain group: 43.7% total weight regain. Normal weight regain group: 6.8% total weight regain. Disinhibited eating (p = .015) and less exercise (p = .003) were associated with significant weight regain. Exercising for 567 min/week vs 287 min/week was associated with the group with normal weight regain.
Conceicao 2014a Portugal/USA	RYGB LAGB B	Cross sectional. 374 patients, 4 groups: pre- op, 6 months, 1 year, and 2 years post-op, 88.2% female; avg age 43.3 years old; avg pre- op BMI for all groups: 45 kg/m2. All patients assessed for maladaptive eating through self-reported questionnaires.	N/A	Higher weight regain in group with LAGB at year 1 and 2 post-op, amounting to 17.7% weight regain by year 2. Those with RYGB regained 5.5% body weight by year 2. Picking and nibbling was associated with weight regain (p < .000).
Nicolau 2015 Spain	RYGB Sleeve gastrectomy B	Cross sectional. 60 patients; 46.5 months post-op, 78.3% female, avg age 46.3 years old, pre-op BMI 48.3. Assessed dietary habits, grazing, depressive disorder and QOL via surveys and a semi-structured interview.	N/A	72% of people who grazed gained weight as opposed to people who did not graze (p < .000). Individuals who reported a grazing pattern were more prone to weight regain and achieved a lesser percentage of excess weight loss.
Pekkarinen 1994 Finland	VBG B	Cohort of 27 patients, avg 5.4 years post-op, 70.3% female, avg age 36; avg pre-op BMI: 50. Assessed dietary intake via a food record and disordered eating patterns through questionnaires further validated via a semi- structured interview.	2 groups: bingers vs non- bingers.	People who binged and people who did not binge had comparative results at 1 year post-op (55% EWL and 57% EWL, respectively), but at 2 years post-op, people who binged regained more weight than people who did not bingers (24% EWL and 50%EWL, respectively p = .04).
Conceicao 2014b Portugal	RYGB Sleeve gastrectomy LAGB B	Cross sectional. 168 patients, 3 groups: pre-op, short-term (11.4 months), and long-term (55.7 months), 88.1 % were women; and avg age 43.5 years; avg pre-op BMI 45.1 in long term group. Assessed disordered eating patterns, depression, and body image via interviews and self-reported measures.	3 groups: pre-op, <2 years, >2 years.	In long-term group (>2 years post-op) experiencing LOC, lowest BMI was 31.6 and BMI at time of study was 37.5 kg/m2. LOC eating was related with the highest BMIs, the least weight loss, most weight regain, and most psychological impairment in the long-term assessments, but not at short-term.
Rudolph 2013 Germany	RYGB LAGB VBG LGP (plication) Other S	Systematic review and meta-analysis. 13–144 patients, 0–53 months post-op, 12–100% female, 21–52.5 years old, pre-op BMI 42.6– 51.6 kg/m2. Various treatments observed: CBT, SGA (support group attendance), and physical activity.	N/A	3 studies observed weight up to 2–3 years post-op. Patients receiving behavioral management had greater weight loss than patients receiving usual care or no treatment. Papalazarou et al. (2010) showed that individual CBT given to VBG patients from the time of surgery and up to 3 years resulted in a maintenance of %EWL from 12 (76.4%) to 36 months (74.8%) post-op vs their controls (57.5% and 49.1%, respectively). Niazi et al. (2012) showed that group CBT to gastric plication patients showed 90.0 %EWL in treatment group vs 43.4%EWL in control group at 24 months post-op.
Gould 2007 USA	RYGB S	Cohort of 85 patients, 3–4 years post-op, =78% female in each group, avg age 43 years avg pre-op BMI: 50 kg/m2. Assessed for attendance to follow-up (f/u) visits.	3 groups: attended all visits for >3 years, attended all visits up to 1 year, and lost to follow-up before 1 year.	Patients who attended all scheduled follow-up appointments experienced greater long-term weight loss than those who did not (p < .05): 74% EWL in those who attended every f/u visit vs 61%EWL in those who attended f/u for first year post-op then lost to f/u vs 56% EWL in those who were lost to f/u before 1 year post-op.
Liebl 2016 USA	RYGB Sleeve gastrectomy LAGB S	Qualitative. 14 patients, 69 months post-op, 78% female, avg age 47 years, pre-op BMI: n/a, avg pre-op weight: 313 lbs. Semi-structured interviews	N/A	An average of 8% of EWL (22 lbs) was regained after surgery. To maintain weight loss, an individual must seek out and be surrounded by positive family and peer support influences. Positive support may provide the opportunity for an individual to place personal health needs as a priority, while negative influences must be identified by the individual and should be blocked out, avoided or contact may need to be eliminated.
Sarwar 2011 USA	RYGB LAGB B	Review observing potential threats posed with changes in dietary intake and eating behavior after bariatric surgery.	N/A	Clinical observations and some studies have suggested that suboptimal weight loss and other untoward outcomes (eg, nutrition, vomiting, and dumping) are often attributed to poor adherence to the postoperative diet and/or a migration to maladaptive eating behaviors. Sustained improvements in dietary intake and eating behaviors are critical to long-term success after bariatric surgery.
Silver 2006 USA	RYGB E	Cross sectional. 140 patients, 24.2 months post-op, 88.6% female, avg age 45 years pre-op BMI 49.8 kg/m2. Assessed behaviors, dietary behaviors, and physical activity via questionnaires.	N/A	A higher BMI at 2 years post-op was associated with less exercise (P=0.006). 82.9% continued to be physically active in an effort to lose or maintain weight, with 62.9% engaging in physical activity at least 3x/wk, with an average duration of 54.7 minutes.
Livhits 2010 USA	RYGB LAGB VBG DS E	Systematic review. 30–1585 patients, 18.9 months post-op, 'mostly' female, avg age 44 years, avg pre-op BMI 47.5.	N/A	6/13 studies observed exercise ≤16 months post- op. Meta-analysis suggests exercise results in a 4.2% greater degree of weight loss at 12 months, and greater weight loss persists out to 24 months. While Bond et al. (2004) report that patients who were active lost more weight than those who were sedentary 24 month after surgery, Larsen et al. (2006) found no evidence of a beneficial association at 24 months post-op.
Hsu 1998 USA	RYGB VBG B	Review.	N/A	Binge eating behavior and low metabolic energy expenditure, are associated with weight regain.

D = dietary, B = behavioral, avg = average, BMI = body mass index, BED = binge eating disorder, RYGB = Roux-en-Y gastric bypass, EWL = excess weight loss, E = exercise, LAGB = laparoscopic adjustable gastric band, S = support, VBG = vertical banded gastroplasty, QOL = quality of life, LGP = laparoscopic gastric plication, DS = duodenal switch