Table 2 Study findings on the relationship between weight change and dietary intakes.

Alvarez et al., 2017 [22]

1. Weight regain (%) at <50^th percentile

2. Weight regain (%) at >50^th percentile

-

-

Group 1 had lower fat intake (p < 0.05).

NS differences in energy intake, all other macronutrients, sweet foods, sugar, and alcohol.

Participants with less WR had lower fat intake.

NS differences reported for all other measured dietary intakes between comparison groups.

WR is defined by presentation of %WR>50^th percentile of the study cohort, which may affect generalizability of findings.

Amundsen et al., 2017 [23]

Comparison 1

1. Excess weight loss ≥50%

2. Excess weight loss <50%

Comparison 2

1. Weight regain ≤15%

2. Weight regain >15%

-

-

Comparison 1:

Group 1 had higher alcohol intake (p < 0.05). NS differences in energy, macronutrients, and all food groups.

Comparison 2:

NS differences in energy, macronutrients, all food groups, and alcohol intakes.

Participants with greater EWL had higher alcohol intake.

NS differences reported for all other measured dietary intakes between comparison groups.

Self-reported weight data.

Bobbioni-Harsch et al., 2002 [24]

Associations with excess weight loss and total weight loss

Positive association between EWL and energy intake (r2 = 0.13, p < 0.01).

Positive trend between TWL and energy intake (Coefficient 0.01, r2 = 0.56, p < 0.07).

NS associations between EWL or TWL and macronutrients.

-

-

EWL and TWL are associated with lower energy intake.

NS associations reported between EWL and TWL with all other measured dietary intakes.

-

Cadena-Obando et al., 2020 [25]

1. Excess weight loss ≥50%

2. Excess weight loss <50%

NS differences in energy and macronutrients.

-

-

NS differences reported for all measured dietary intakes between comparison groups.

Included only participants with complete medical files.

Chou et al., 2017 [26]

Associations with total weight loss

-

-

Positive trend between TWL and energy (r = 0.313, p = 0.052), NS associations with macronutrients.

Greater TWL is associated with higher energy intake.

NS associations reported between TWL with all other measured dietary intakes.

Study conducted in the Eastern context and may have less generalizability to the Western context.

1. Weight regain <25%

2. Weight regain ≥25%

Non-regainers had higher protein and fat (p < 0.05), but NS differences in energy or other macronutrient intakes.

Participants with less WR had higher protein and fat intakes.

NS differences reported for all other measured dietary intakes between comparison groups.

Dagan et al., 2017 [27]

1. Excess weight loss ≥60%

2. Excess weight loss <60%

NS differences in proportion of participants consuming ≥60 g/day protein.

-

-

Proportion of subjects having protein intakes of ≥60 g/day did not differ between participants with greater or less EWL.

Self-reported adherence to ≥60 g/day protein intakes.

Study was conducted in the Middle-Eastern context and may have less generalizability to the Western context.

da Silva et al., 2010 [28]

Odds of Weight regain ≥10%;

-

-

Inverse association between WR and better diet quality (OR 0.95, 95% CI 0.90–0.99, p < 0.05), NS associations with carbohydrate and fat intakes.

Lower odds of WR is associated with a better diet quality, but not with any other measured dietary intakes.

Diet quality measurement tool was designed for use in general population.

1. Weight regain <10%

2. Weight regain ≥10%

Group 1 had higher fruit (servings/day; p < 0.05) intake and better diet quality score (p < 0.01). NS differences in energy (kcal/day, kcal/kg ideal weight/day), carbohydrate (%), protein (g/day, g/kg ideal weight/day) and fat (%), vegetables, meats, beans, dairy, grains (p = 0.06), sugary and sweets, and fats and oils servings/day.

Participants with less WR had higher fruit intakes and better diet quality score.

NS differences reported for all other measured dietary intakes between comparison groups.

Dodsworth et al., 2012 [29]

Intervention: 6-month prescribed protein-enriched diet (960–1400 kcal/day, 40% carbohydrate, 30% protein, 25% fat).

Control: Usual care.

NS differences in BMI, EWL or TWL between groups.

-

-

Protein-enriched diet did not result in greater BMI, EWL, or TWL than having no prescribed diet.

Poor compliance at all timepoints which may affect interpretation of the effect of the prescribed diet.

Associations with BMI, excess weight loss and total weight loss

NS associations between BMI, EWL, or TWL, with protein intakes.

BMI, EWL or TWL were not associated with protein intakes.

Faria et al., 2009 [30]

Associations with average monthly weight loss

Moderate inverse associations with total energy intake (kcal/day) (r = 0.373, p < 0.01), carbohydrate (g/day) (r = −0.414, p < 0.01) and meal glycemic load (r = −0.364, p < 0.01).

Positive association with protein (%) (r = 0.305, p < 0.05) in bivariate analysis (unclear if included in final regression model). NS association with fat (g/day) in final model.

-

Greater average monthly weight loss is associated with lower energy, carbohydrate, meal glycemic load, and higher protein intakes.

NS associations reported between greater average monthly weight loss with all other measured dietary intakes.

53% of participants included in the data analysis were less than 12 months post-surgery, which may influence the comparability of findings to other studies where all participants were at least 12 months post-surgery.

Faria et al., 2009 [31]

1. <150 kcal of sweets/snacks between main meals. Highest excess weight loss (%) in cohort (p < 0.05 with 3. only).

2. ≥150 kcal of sweets between main meals.

3. ≥150 kcal of snacks between main meals.

-

Group with greatest EWL had lower energy; (p < 0.01) and carbohydrate (p < 0.01), but NS differences in fat intakes and; lower protein intakes than participants with least EWL(p < 0.01).

-

Participants with greater EWL had lower energy, carbohydrate, and protein intakes.

NS differences reported for all other measured dietary intakes between comparison groups.

Cohort-specific definition of significant EWL.

Differences in EWL reached significance (p < 0.05) between highest and lowest EWL groups only.

Faria et al., 2010 [32]

Intervention: 3-month prescribed high protein diet (1400 ± 180 kcal/day, 45% carbohydrate, 35% protein, 20% fat, 3 serves/day dairy).

Control: n/a

-

-

On average, participants had significant reduction of BMI (p < 0.001) and TBW (p < 0.001).

Intervention with prescribed high protein diet lowered the BMI and TBW of participants.

Findings may be confounded by the incentivized physical activity component.

No control group.

Reduced generalizability, as all participants were highly-motivated with demonstrated high adherence.

Freire et al., 2011 [33]

Associations with excess weight loss

-

-

NS associations between EWL and macronutrient composition (%) or any food groups.

EWL is not associated with macronutrient composition or any food group intakes.

-

1. Weight regain <2%

2. Weight regain ≥2%

Group 1 had lower energy (kcal/day) (p < 0.01), snacks and sweets (p < 0.05), and oils and fatty food (p < 0.01) servings/week, but NS differences in macronutrient composition, in fruit, vegetable, meats and eggs, beans, dairy, and grain servings/week.

Participants with less WR had lower energy, snacks and sweets, and oils and fatty food intakes.

NS differences reported for all other measured dietary intakes between comparison groups.

Gallé ́ et al., 2020 [34]

Intervention: 12-month exercise, motivational, and nutritional program.

Control: Received routine medical examinations only.

Intervention group had greater BMI loss (p < 0.01) than control.

Intervention group reported increased fruit or vegetables, and reduced sweets servings/day (p < 0.05 for all), while controls only reported reduced sweets servings/day (p < 0.05). There were NS changes in meat, fish, eggs, dairy, or grain intakes in both groups.

-

-

Lifestyle intervention resulted in greater BMI loss and increased fruit or vegetable, and reduced sweets intakes.

NS differences reported for all other measured dietary intakes between comparison groups.

Findings may be confounded by the physical activity component.

Intake differences between intervention and control not compared.

Iossa et al., 2020 [35]

Associations with excess weight loss, risk of excess weight loss <50%, risk of weight regain >25%

-

-

Moderate inverse association between EWL with energy (kcal/day; r = −0.54, p < 0.05) and fat (g/day; r = −0.35, p < 0.05) intakes, but NS associations with carbohydrate (g/day) and protein (g/day) intakes.

NS associations between risk of EWL < 50% to energy (kcal/day) and fat (g/day) intakes.

Positive association between risk of WR > 25% with energy ( > 1300 kcal/day; HR 4.2, 95% CI 5–16, p < 0.05), and fat (g/day; HR 4.2, 95% CI 6–11, p < 0.05) intakes.

Greater EWL is associated with lower energy and fat intakes, but not with any other measured dietary intakes.

Risk of EWL < 50% is not associated with energy and fat intakes.

Higher risk of WR is associated with higher energy and fat intakes.

Energy and fat intakes of Group 1< Group 3< Group 2, but significance was achieved between Groups 1 and 3 only. This may be attributable to the large difference in sample size between groups.

1. Excess weight loss >50% at first year + <25% regain of excess weight

2. Excess weight loss <50% at first year

3. >25% regain of excess weight loss in absence of surgical complications

NS differences in energy (kcal/day) and fat (g/day) intakes between Groups 1 and 2.

Group 1 had lower energy (kcal/day; p < 0.05) and fat (g/day; p < 0.05) intakes than Group 3.

Participants with greater EWL and less WR had lower energy and fat intakes than those with greater WR only.

Jassil et al., 2015 [36]

Intervention: 8-week exercise training, and group lifestyle and nutritional behavioural change sessions, at 3–6 months

Control: Standard post-surgical care and follow-ups

Intervention group had greater weight and BMI losses between 3–12 months (p < 0.05, p = 0.05), but not at 12 months.

Intervention group reported increased combined fruit and vegetable portions/day, and significantly reduced frequencies of ready meals (p < 0.05 for all). There were NS changes in deep-fried foods, crisps, cakes/ biscuits/ chocolate/ sweets, take-away meals, fizzy drinks, fruit juice, and liquid meals.

-

-

Lifestyle intervention resulted in greater weight and BMI losses, increased combined fruit and vegetable, and reduced sweets intake.

NS differences reported for all other measured dietary intakes between comparison groups.

Findings may be confounded by the physical activity component.

Intake differences between intervention and control not compared.

Kalarchian et al., 2016 [37]

Intervention: 4-month prescribed diet (4 serve/day vegetables, 2–4 serves/day meat, and 1–2 serves/day low glycaemic index grains) with home delivered, portion-controlled meals + usual care

Control:

Usual care

-

Intervention group had significantly greater weight loss at 16 months (p < 0.01) and 18 months (p = 0.05).

Group allocation had a significant main effect on weight loss from enrolment (F1,31 = 6.79, p = 0.01).

-

Intervention with prescribed diet resulted in greater weight loss.

Prescribed diet did not cover all food groups. Interpretation for potential effect of the prescribed diet was not possible.

Reduced generalizability, as home-delivered meals were provided as part of study.

Kanerva et al., 2017 [38]

Associations with initial weight loss between men and women, according to changes in intakes from 0–6 months.

At 10 years, greatest IWL were observed from macronutrient composition (%) of carbohydrate > fat (men only; p < 0.05), protein > carbohydrate (p < 0.05), and protein > fat (p < 0.01).

Greatest 10-year IWL was achieved by men and women with largest reduction of energy (kcal/day; p < 0.001) or carbohydrate (%; p < 0.05) intakes, men with largest reduction of fat (%; p < 0.001) intakes, women with largest increment of protein (%; p < 0.05) and least reduction of alcohol (%; p < 0.05) intakes, from 0–6 months.

-

-

Greatest 10-year IWL is associated with lower carbohydrate and fat, and higher protein intakes.

Participants with greatest 10-year IWL largely reduced their energy, carbohydrate, and fat (men only) intakes, and increased their protein (women only) intakes, from 0–6 months. Women with greatest 10-year IWL did not reduce their alcohol intakes from 0–6 months.

Absolute differences in IWL (%) between groups were <5% for all measures. These were described by the study as “non-clinically significant”.

It was not clear whether adjustment of baseline alcohol intakes were conducted.

Kruseman et al., 2010 [39]

1. Excess weight loss ≥50%

2. Excess weight loss <50%

NS differences in one-year intakes of energy (kcal/day), carbohydrate (%), protein (g/kg/day) and fat (%).

-

Group 1 had lower energy (kcal/day; p < 0.05) intakes but NS differences in carbohydrate (%), protein (g/kg/day), and fat (%) intakes.

Participants with greater EWL had lower energy intake at 8-years only.

NS differences reported for all other measured dietary intakes between comparison groups.

-

Lim et al., 2020 [40]

1. Excess weight loss ≥50%

2. Excess weight loss <50%;

At 6 months, Group 1 had lower energy (kcal/day; p < 0.01), carbohydrate (g/day only; p < 0.001) or fat (g/day only; p < 0.05), and higher protein (% only; p < 0.05) intakes.

At 12 months, Group 1 had lower energy (kcal/day; p < 0.001), carbohydrate (% and g/day; p < 0.001) or fat (% and g/day, p < 0.05), and higher protein (% and g/day, p < 0.05) intakes.

-

-

Participants with greater EWL had lower energy, carbohydrate and fat, and higher protein intakes.

Study was conducted in the Eastern context and may have less generalizability to the Western context.

Associations with excess weight loss within participants with excess weight loss ≥50%

There are moderate inverse associations between EWL with energy (kcal/day; r = −0.418, p < 0.01) or fat (%; r = −0.273, p < 0.001), but NS associations with carbohydrate (%) and protein (%).

In participants with greater EWL, greater EWL was associated with lower energy or fat intakes, but not with carbohydrate or protein intakes.

Associations with excess weight loss within participants with excess weight loss <50%

There are moderate inverse associations between EWL with carbohydrate (%; r = −0.3, p < 0.01) or fat (%; r = −0.266, p < 0.05), moderate positive associations with protein (%; r = 0.301, p < 0.01), and NS associations with energy (kcal/day).

In participants with less EWL, greater EWL was associated with lower carbohydrate or fat intakes, and with higher protein intakes, but not with energy intakes.

Associations with odds of excess weight loss ≥50%

There are small inverse associations between the odds of EWL ≥ 50% with carbohydrate (%; OR = 0.99, 95% CI 0.98–0.99, p < 0.001) or fat (%; OR = 0.96, 95% CI 0.93–0.98, p < 0.001), but NS associations with protein (%).

Greater odds of EWL ≥ 50% are associated with lower carbohydrate and fat intakes, but not with protein intakes.

12-month intake cut-offs for reaching excess weight loss ≥50%

Energy of <1523.0 kcal/day (AUC 0.912, 95%CI 0.872-0.953);

Carbohydrate of <49% and <172.5 g/day (AUC 0.714, 95%CI 0.637-0.792 / AUC 0.878, 95%CI 0.819-0.937);

Protein of >24.5% and >86.5 g/day

AUC 0.609, 95%CI 0.523–0.695 / AUC 0.618, 95%CI 0.531–0.705;

Fat of <28% and <52.5 g/day

(AUC 0.855, 95%CI 0.792-0.917 / AUC 0.781, 95%CI 0.709-0.853).

At 12-months, intake cut-offs for participants with EWL > 50% were <1523.0 kcal, 49% carbohydrates, >24.5% protein and <28% fats per day.

Lindroos et al., 1996 [41]

Associations with weight change

-

Positive association between weight change (kg) with polysaccharides (%) (t = 2.05, p < 0.05), protein (%) (t = 2.94, p < 0.01), or prepared meals (%) (t = 4.59, p < 0.001).

Inverse association between weight change (kg) with mono/disaccharides (%) (t = −3.16, p < 0.01), fat (%) (t = −2.11, p < 0.05) or sweet foods (%) (t = −3.61, p < 0.001).

NS associations with energy (kcal/day), sandwiches (%), and alcohol (%).

-

Positive weight change (weight gain) is associated with higher intakes of polysaccharides, protein or prepared meals.

Negative weight change (weight loss) is associated with higher intakes of mono/disaccharides, fat, and sweet foods.

NS associations reported between weight change with all other measured dietary intakes.

Direct comparison for carbohydrate intakes with other studies was not possible due to intakes being divided into mono/disaccharides and polysaccharides.

Total weight loss between participants with highest vs lowest quartiles of intakes

Greater TWL (p < 0.05) achieved from higher intakes of mono/disaccharides ( > 142 vs <72 g/day), fats ( > 96 vs <51 g/day), sweet foods ( > 2.83 vs <1.09MJ/day), and lower intakes of polysaccharides ( < 79 vs >132 g/day), protein ( < 56 vs >92 g/day), and prepared meals ( < 1.06 vs >2.08MJ/day).

Alcohol (0 g/day vs >6.7 g/day) and sandwiches ( < 0.9MJ/day vs >2.31MJ/day) did not result in differences in TWL (kg).

Greater weight losses are achieved by higher intakes of mono/disaccharides, fats or sweet foods, and lower intakes of polysaccharides, proteins and prepared meals.

NS differences reported for all other measured dietary intakes between comparison groups.

Masood et al., 2019 [42]

1. Weight regain <15%

2. Weight regain ≥15%

-

Higher proportion of weight loss maintainers consumed fat (p < 0.01), fruit (p < 0.001), and vegetable (p < 0.001) intakes of 3–5 exchanges/day.

NS differences in the proportion of participants eating carbohydrate of 1–5 exchanges/day, or ready-to-eat foods and fast foods of 0–1 times/week.

-

Participants who experienced less WR consumed fats, fruits, and vegetable intakes of 3–5 exchanges/day.

NS differences reported for all other measured dietary intakes between comparison groups.

Portion size of each exchange not specified.

Study was conducted in the Middle-Eastern context and may have less generalizability to the Western context.

Moizé, V. et al., 2013 [43]

Associations with excess weight loss

-

-

Inverse association with energy (kcal/day; r-value NR, p < 0.01), but NS association with macronutrient composition (%).

Greater EWL is associated with lower energy intake.

NS associations reported between EWL with all other measured dietary intakes.

The strength of association was not reported.

Novais et al., 2012 [44]

1. Excess weight loss ≥75%

2. Excess weight loss 50-75%

3. Excess weight loss <50%

-

-

NS differences in energy (kcal/day) and macronutrient composition (%).

Participants with greater EWL did not have different dietary intakes than those with less EWL.

-

Ortega et al., 2012 [45]

Associations with BMI loss.

-

-

Inverse associations with energy (kcal/day; B = −0.003, p < 0.01) intake. NS associations with macronutrient composition (%).

Greater BMI loss is associated with lower energy intakes.

NS associations reported between BMI loss with all other measured dietary intakes.

Self-reported weight data.

Palacio et al., 2020 [46]

Associations with odds of weight regain ≥15%

-

-

Positive associations with energy (kcal/day; OR 1.3, 95% CI 1.1–1.9, p < 0.05) but NS associations with carbohydrate (% and g/day) intakes.

Greater odds of WR is associated with higher energy intake.

NS associations reported between odds of WR with all other measured dietary intakes.

Weight data measured at 2 years but compared to diet data collected at 7-years.

1. Weight regain <15%

2. Weight regain ≥15%

Group 1 had lower 7-year energy (kcal/day; p < 0.001) and carbohydrate (% −p < 0.01, and g/day −p < 0.001) intakes.

Participants with less WR had lower energy and carbohydrate intakes.

Papalazarou et al., 2010 [47]

Intervention: 3-years person-centred lifestyle behaviour change program + usual care

Control: Usual care

Intervention group had greater EWL (p < 0.05) and lower TBW (p < 0.05), but NS differences in fruit, vegetable, and sweets servings/day than control.

Group allocation remained the only significant factor in weight losses (p-value NR).

-

Intervention group had greater EWL and lower TBW, higher fruit and vegetable, and lower sweet servings/day (p < 0.05 for all) than control.

Group allocation remained the only significant factor in weight losses (p-value NR).

Lifestyle intervention resulted in greater EWL, lower TBW, higher fruit and vegetable intakes, and lower sweet intakes.

Intervention group had higher physical activity levels than control at three years (p < 0.05). This may have confounded results at three-years.

Pinto et al., 2019 [48]

Associations with risk of obesity remission (BMI < 30)

Positive association with 0–3 months changes in protein (%; HR = 1.06, 95% CI 1.01–1.12, p < 0.05) intakes, but NS association with 0–3 months changes in carbohydrate (%) and fat (%).

-

-

Greater risk of obesity remission is associated with higher protein intakes.

NS associations reported between risk of obesity remission with all other measured dietary intakes.

-

Raftopoulos et al., 2011 [49]

Associations with BMI change and excess weight loss

Positive association with protein (g/kg/day) intakes (BMI change: B = 2.46, 95%CI 1.32–3.69, R2 = 0.041, p < 0.001; EWL: B = 8.28, 95%CI 3.65–12.92, R2 = 0.054, p < 0.01).

Positive association with consistent compliance to ≥1 g/kg/day protein intakes (BMI change: F = 5.097, p < 0.01; EWL: F = 4.415, p < 0.01).

-

-

Greater BMI change and EWL are associated with higher protein intakes and more consistent compliance to ≥1 g/kg/day protein.

12-month data was only available for 27% of participants with incomplete follow-up.

Reid et al., 2016 [50]

1. Total weight loss ≥38%

2. Total weight loss ≤30%

-

-

Group 1 had lower carbohydrate (g/day; F1,23 = 5.065, p < 0.05) and alcohol (g/day; F1,23 = 4.836, p < 0.05) intakes.

NS differences in energy (kcal/day), carbohydrate (%), protein (%, g/day, % of participants with intake of ≥60 g/day), and fat (% and g/day) intakes.

Participants with greater TWL had lower carbohydrate and alcohol intakes.

NS differences reported for all other measured dietary intakes between comparison groups.

Diet data were measured one-year after collection of weight data.

Ruiz-Lozano et al., 2016 [51]

1. Excess weight loss ≥50% at nadir and until last follow-up

2. Excess weight loss <50% at nadir and up to last follow-up

3. Excess weight loss ≥50% at nadir but <50% at last follow-up

NS difference in energy and macronutrient composition (% and g/day).

NS difference in energy and macronutrient composition (% and g/day).

Participants with greater EWL at nadir and at last-follow-up did not have different dietary intakes to those with less EWL.

No assessment of exact amount of WR. It is possible that significant amounts of WR were present in achievers despite maintaining >50%EWL at last follow-up.

Sarwer et al., 2012 [52]

Intervention: 4-month dietary counselling sessions + usual care

Control: Usual care

-

Intervention did not result in greater or less EWL, or higher or lower dietary intakes.

Did not result in any significant changes to enable determination of presence or absence of associations.

Schiavo et al., 2016 [53]

Intervention:

12-month prescribed protein-enriched diet

(1200 kcal/day, 37.3% carbohydrate, 47.7% protein, 15% fat)

Control:

12-month prescribed normal-protein diet

(1200 kcal/day, 61.7% carbohydrate, 23.3% protein, 15% fat)

NS differences in TBW between groups.

-

-

Protein-enriched diet did not result in lower TBW than normal-protein diet.

All participants were of male gender. Findings may be less comparable to all other included studies which had higher proportion of female participants.

Schiavo et al., 2018 [54]

1. Prescribed low-purine diet (890 kcal/day, 55% carbohydrate, 20% protein, 25% fat, emphasis on low-purine foods)

2. Prescribed normal-purine diet (890 kcal/day, 55% carbohydrate, 20% protein, 25% fat)

NS differences between groups in BMI and TBW.

-

-

Low-purine diet did not result in lower BMI or TBW than normal-purine diet.

Adherence assessment was based on prescribed energy and macronutrient, instead of purine intakes.

Taus et al., 2017 [55]

Intervention:

2-month prescribed ketogenic diet (800 kcal/day, 20% carbohydrate, 40% protein, 40% fat)

Control:

Prescribed usual care diet 800 kcal/day, 52% carbohydrate, 25% protein, 23% fat + Band calibration (Average 8cc)

-

Intervention had lower BMI, greater EWL, and lower TBW than control (p-values NR for all measures).

-

Ketogenic diet resulted in lower BMI and TBW and greater EWL than usual care diet.

At baseline, intervention group had lower BMI and TBW and greater EWL, than control. However, significance in weight differences between- and within- group, at baseline and at after intervention, were not described.

Wardé-Kamar et al., 2004 [56]

Associations with excess weight loss

-

-

Inverse associations with energy (kcal/day) (p-value NR). Predictors of 47% EWL were: Age, excess weight, pre-surgery weight, and energy (kcal/day) and fat (%) intakes (p < 0.001).

Greater EWL was associated with lower energy intakes. Energy and fat intakes were some of the predictors of 47% of EWL post-surgery.

NS associations reported between EWL with all other measured dietary intakes.

Self-reported weight data.

Interpretation of strength and direction of association (fat only) were not possible due to no reporting of respective statistics.

1. Excess weight loss ≥50%

2. Excess weight loss <50%

NS differences in energy (kcal/day) and macronutrient composition (%).

Participants with greater EWL did not have different dietary intakes to those with less EWL.

Yanos et al., 2015 [57]

Associations with total weight loss at nadir

-

-

NS associations between TWL at nadir, with having ≥60–80 g/day protein, ≥5 serves/day of fruit or vegetables, or avoiding sweets.

TWL at nadir is not associated with any measured intake variables.

Self-reported weight data.

Associations with risk of weight regain ≥20%

Avoidance of sweets was an independent predictor for risk of WR ≥ 20% (r2 = 0.22, p < 0.01; bivariate analysis: r = −0.28, p < 0.01).

NS associations with having ≥60–80 g/day protein, ≥5 serves/day of fruit, or ≥5 serves/day of vegetables.

Lower risk of WR ≥ 20% is associated with avoidance of sweets.

NS associations reported between risk of WR ≥ 20% with all other measured dietary intakes.