The Effect of Rate of Weight Loss on Long-Term Weight
Regain in Adults with Overweight and Obesity
Roel G. Vink, Nadia J. T. Roumans, Laura A. J. Arkenbosch, Edwin C. M. Mariman, and Marleen A. van Baak
Objective: To investigate the effect of rate of weight loss, with similar total weight loss, on weight regain
in individuals with overweight and obesity.
Methods: Fifty-seven participants (BMI: 28-35 kg/m
2) underwent a dietary intervention (DI). They were
randomized to a low-calorie diet (LCD; 1250 kcal/day) for 12 weeks (slow weight loss) or a
very-low-calorie diet (VLCD; 500 kcal/day) for 5 weeks (rapid weight loss) (weight loss (WL) period) followed by a
4-week weight-stable (WS) period and 9 months follow-up. Body weight and body composition (BodPod)
were determined at study start and after each period.
Results: Weight change was similar in both groups after WL (LCD: 28.2 kg and VLCD: 29.0 kg,
P 5 0.24). Weight regain after follow-up was not significantly different between groups (LCD: 4.2 kg and
VLCD: 4.5 kg, P 5 0.73). Percentage fat-free mass loss (%FFML) was higher in the VLCD-group
com-pared to the LCD-group after DI (8.8% and 1.3%, respectively, P 5 0.034) and was associated with
weight regain during follow-up in the whole group (r 5 0.325, P 5 0.018).
Conclusions: The present study showed that, with similar total weight loss, rate of weight loss did not
affect weight regain. However, %FFML after DI was associated with weight regain.
Obesity (2016) 24, 321-327. doi:10.1002/oby.21346
Introduction
The number of people with overweight and obesity worldwide has risen from 857 million in 1980 to 2.1 billion in 2013 (1). Despite the adverse health problems associated with obesity (2), not a single country in the world had a significant decrease in obesity prevalence in the past three decades (1). While a dietary intervention (DI) can achieve significant weight loss, the greatest challenge is the seem-ingly inevitable weight regain in the following years. One year after weight loss,20% of individuals were able to remain weight stable, when weight stable was defined as an intentional weight loss of 10% maintained for at least 1 year (3,4). In a more recent study, participants regained on average70% of their lost weight over 2 years following diet-induced weight loss (5). Insight in the etiology of weight regain and long-term weight management are therefore strongly needed.
A widely discussed topic in the prevention of weight regain is the rate of weight loss. The concept that rapid weight loss increases long-term weight regain compared to a more gradual weight loss approach is a belief held by the general public (6,7), and dietary guidelines in several countries recommend the latter approach in obesity management (8,9). Rapid weight loss diets create a larger energy deficit and contain lower absolute amounts of protein
compared to a more gradual weight loss approach, which increases the risk of muscle mass loss. Muscle mass is a key contributor to resting energy expenditure (10,11) and could play a role in long-term weight management. However, the belief that a more gradual weight loss approach is preferred over rapid weight loss in terms of long-term weight control is not supported by scientific evidence (6,7). The discussion is complicated by the fact that most studies on rate of weight loss do not dissociate between the effects of the rate of weight loss and the total weight loss. At the start of our study only one other study had investigated the effect of two different rates of weight loss, with similar total weight loss, on weight regain. The results from that study showed that an 8-week very-low-calorie diet (VLCD) or a 17-week low-calorie diet (LCD) did not result in significant differences in weight regain after a 1-year weight maintenance diet or after a subsequent 1-year follow-up period (12). However, it should be noted that both weight loss diets were supplemented with the anorectic compounds caffeine and ephedrine, and that participants were randomized to two different weight maintenance diets. Thus, whether—under con-ditions of similar weight loss—a more gradual weight loss approach is to be preferred over rapid weight loss with respect to successful weight loss maintenance is so far not supported by strong scientific evidence.
Department of Human Biology, NUTRIM School of Nutrition and Translational Research in Metabolism, Maastricht University Medical Centre1,
Maastricht, The Netherlands. Correspondence: Roel G. Vink (r.vink@maastrichtuniversity.nl)
Funding agencies: Netherlands Organisation for Scientific Research TOP, grant number: 200500001. Disclosure: The authors declare no conflict of interest.
We therefore performed a study where weight regain was studied after two different rates of diet-induced weight loss with similar total weight loss in the absence of dietary advice during the follow-up period.
Methods
Subjects
Sixty-one individuals with overweight and obesity (BMI 28-35 kg/ m2) were recruited by advertisement via local media. Exclusion cri-teria were smoking, cardiovascular disease, type 2 diabetes mellitus, liver or kidney disease, use of medication that influences body weight regulation, pregnancy, marked alcohol consumption (>21 alcoholic units per week for men and >14 alcoholic units per week for women), elevated fasting glucose (>6.1 mmol/L), total choles-terol (>7.0 mmol/L) or triacylglycerol (>3.0 mmol/L) concentra-tions, or blood pressure (>160/100 mmHg). Furthermore, partici-pants had to remain weight stable (weight change <3.0 kg) 2 months prior to the start of the study. Participants were all Cauca-sians. All subjects gave their written informed consent before partic-ipation in the study. The study was performed according to the Dec-laration of Helsinki and was approved by the Medical Ethics Committee of Maastricht University Medical Centre.
Experimental protocol
The participants in our study followed a DI program that was divided in three periods: a 12-week LCD-period or 5-week VLCD-period (weight loss VLCD-period, WL), a 4-week weight-stable VLCD-period (WS), and a 9-month follow-up period. The WL-period and WS-period taken together was named the DI-WS-period (Figure 1). Partici-pants were randomly assigned to either the LCD (slow weight loss) or VLCD (rapid weight loss) group. Both interventions aimed at a weight loss of10%. In the slow weight loss program, participants underwent a 12-week LCD providing 1,250 kcal/day designed by the dietician. The LCD consisted of one meal that was replaced by meal replacements (Modifast; Nutrition et Sante Benelux, Breda, The Netherlands), two meals that the participants prepared them-selves based on meal plans designed by our dietician, and three in-between meal snacks. In the rapid weight loss program, participants underwent a 5-week VLCD in which three meals per day were replaced by meal replacements, providing 500 kcal/day. During this period participants were advised by our dietician (five meetings) and were allowed to consume two 100 mL instant broth drinks per day containing a high amount of sodium and 7 kcal each and an
unrestricted amount of low-calorie vegetables. Both groups subse-quently underwent a 4-week WS-period with a diet based on the energy requirement of the participants. This allowed us to investi-gate the effect of weight loss, without the interfering effect of a neg-ative energy balance. The study dietician provided dietary advice according to the Dutch national guidelines (13) to both groups, to assist in remaining weight stable throughout the WS-period (four meetings) and to assist in weight loss during the WL-period in the LCD-group (five meetings). Finally, body weight and blood pressure were measured monthly at the Maastricht University Medical Centre (MUMC) for 9 months during follow-up. However, dietary advice was no longer given to mimic non-restricted free-living conditions. At the start of the study and at the end of each period body compo-sition was determined, physical activity (PA) questionnaires were completed, and weight, height, and waist and hip circumference were measured. Furthermore, at each visit adverse events were monitored, and body weight and blood pressure were determined. The researchers, study participants, and dietician were not blinded to the intervention. This trial is registered with www.clinicaltrials.gov as NCT01559415.
Diet composition
The VLCD contained: 52 g protein (43 En%), 52 g carbohydrate (43 En%), and 8 g fat (14 En%). The LCD contained: 90 g protein (29 En%), 150 g carbohydrate (48 En%), and 32 g fat (23 En%). Estimated protein intake was 0.56 g/kg in the VLCD-group and 0.97 g/kg in the LCD-group.
Anthropometric measurements
Participants were weighed on the same scale (Seca model 861, Hamburg, Germany) accurate to the nearest 0.1 kg in light clothing after an overnight fast of at least 10 h. Blood pressure was measured while participants were seated in a chair with a digital automatic blood pressure monitor (Intellisense, Omron Model M6 comfort). We measured waist circumference above the umbilicus and hip cir-cumference at the widest part of the buttocks, both to the nearest 0.5 cm. Body volume was determined with air-displacement plethys-mography (ADP) using the Bod Pod device (Cosmed, Italy, Rome) according to the manufacturer’s instructions and as described by Dempster and Aitkens (14). The thoracic gas volume was predicted using the equations incorporated in the Bod Pod software. Body density, as calculated by the Bod Pod, was used to calculate body
composition according to the two-compartment model as described by Siri (15).
Questionnaires on PA and weight cycles
Physical activity score (PA-score) was determined with the Baecke ques-tionnaire for habitual PA (16). The PA-score determines habitual PA in the preceding period, whereas the PA-score at study start was a measure of the year before study participation.
Number of weight cycles was determined at study start with a tionnaire that we designed. The questionnaire consisted of one ques-tion: write down the number of weight cycles that you have experi-enced in your life, when a weight cycle is defined as losing and subsequently regaining at least 5 kg of body weight.
Calculations
The primary outcome of this study was absolute body weight change during follow-up (weight at follow-up minus weight at the end of DI). Percentage weight regain was calculated by dividing the weight change during follow-up by the change in weight from study start to end of DI (study start minus end of DI), multiplied by 100. Percent-age fat-free mass loss (%FFML) and %FFM gain were calculated by dividing delta FFM by delta body weight (%FFML: pre 2 post, %FFM gain: post 2 pre) over a certain time period, multiplied by 100.
Statistics
Data are presented as mean 6 SEM. Comparisons of variables between time points within groups were made with the
paired-sample t test. Between-group comparisons were made with the independent-samples t test. Correlations were calculated with Pearson (normally distributed) and Spearman’s rank (non-normally distributed, variable: “weight cycles”) correlation coefficients. Sta-tistical calculations were performed with SPSS for Macintosh, Version 21 (Chicago, IL). P < 0.05 was considered statistically significant.
Only 9.8% of our participants were drop-outs and these were explained by circumstances deemed unrelated to the intervention (cancer development, medication use, and personal circumstances). Furthermore, because of the near-complete absence of missing val-ues the data were considered missing at random.
Results
Four participants withdrew from the study during the DI, two because of cancer (VLCD) and two because of personal circum-stances (one in each group). Characteristics of the remaining 57 participants are displayed in Table 1. No significant differences were observed between diet groups at the start of the study (Table 1).
Two participants started to use prescription medication that could influence body weight during follow-up (one in each group) and were excluded from follow-up analysis.
Diet-induced changes in anthropometry, blood
pressure, and PA-score
The WL-period induced comparable body weight changes in the slow weight loss group (LCD-group) and rapid weight loss group (VLCD-group) (28.2 6 0.5 kg vs. 29.0 6 0.4 kg, respectively, P 5 0.24, Table 2), with an average weight loss of 0.7 kg/week and 1.8 kg/week, respectively. In the subsequent 4 weeks of the WS-period, body weight did not change significantly in either group (Table 2). Furthermore, only waist circumference was significantly different between groups over the 4-week WS-period (LCD: 21.0 6 0.5 cm vs. VLCD: 10.8 6 0.6 cm, P 5 0.037). Systolic blood pressure decreased significantly after rapid weight loss but not after slow weight loss at the end of both WL and DI (Table 2), but these changes were not significantly different between groups. Fat percentage decreased significantly in both diet groups (Table 2) but remained unchanged between groups. FFM loss is often expressed as the percentage of weight lost as fat-free mass (%FFML). %FFML was higher in the VLCD-group compared to the LCD-group after WL (17.5% vs. 7.0%, P 5 0.003) and after DI (8.8% vs. 1.3%, P 5 0.034) (Figure 2A). At the end of DI, the VLCD-group had a 0.6 kg higher FFM loss than the LCD-group. %FFML was higher after WL compared to after DI within groups (LCD: 7.0% vs. 1.3%, P 5 0.030; VLCD: 17.5% vs. 8.8%, P < 0.001, Figure 2A). The %FFML was not significantly different between men and women after WL and after DI, both within the LCD- and VLCD-group (Figure 2B). Furthermore, no significant difference was observed in %FFML after DI compared to %FFM gain during follow-up in the whole group (6.52 6 1.95% vs. 4.40 6 5.21%, respectively,P 5 0.664).
TABLE 1Baseline characteristics of the study population Study start LCD (n 5 29) VLCD (n 5 28) Sex (male/female) 14/15 13/15 Age (years) 51.8 6 1.9 50.7 6 1.5 Weight (kg) 92.4 6 1.9 92.6 6 1.8 BMI (kg/m2) 31.3 6 0.5 31.0 6 0.4 Hip circumference (cm) 110.8 6 1.3 111.1 6 1.1 Waist circumference (cm) 102.6 6 2.0 101.9 6 1.5 Waist/hip ratio 0.93 6 0.1 0.92 6 0.1 Body fat (%) 39.9 6 1.8 39.7 6 1.5 Body fat (kg) 36.6 6 1.7 36.2 6 1.2 FFM (kg) 55.4 6 2.2 55.9 6 2.2
Systolic blood pressure (mmHg)
126.8 6 2.4 123.3 6 2.7 Diastolic blood pressure
(mmHg)
85.9 6 1.9 84.6 6 1.9
PA-score 9.06 6 0.18 9.05 6 0.21
Weight cycles 2.3 6 0.3 2.0 6 0.4
Values are mean 6 SEM.
Changes in anthropometry, blood pressure, and
PA-score during follow-up
Forty-one out of fifty-five participants (75%) regained >2 kg body weight during the 9-month follow-up period. Mean weight regain was 4.2 kg in the LCD-group and 4.5 kg in the VLCD-group and was not different between groups (P 5 0.73, Table 2). On average, participants on both diets regained more than 50% of their lost weight within 9 months (LCD: 58.6%, VLCD: 54.7%). Variation in weight loss during the DI-period explained less than 1% of the vari-ation in weight regain after follow-up in the whole group and was therefore not used in further analyses. Weight regain during follow-up was also not associated with gender, age, or BMI at start of the study (whole group, data not shown).
Interestingly, %FFML at the end of DI was positively correlated with weight regain during follow-up in the whole group (r 5 0.325, P 5 0.018, Figure 3A). This correlation remained significant after adjusting for body fat percentage at study start (r 5 0.384, P 5 0.011). Thus, despite this positive correlation and a significant difference in %FFML between groups after DI, this did not translate into significant differences in weight regain between groups. The PA-score after follow-up, which measures habitual PA during follow-up, was negatively correlated with weight regain in the whole group (r 5 20.330, P 5 0.014, Figure 3B) and was mainly explained by a stronger correlation in the slow weight loss group compared to the rapid weight loss group (r 5 20.586, P 5 0.001 vs. r 5 20.094,
TABLE 2Changes in characteristics at the end of WL and DI compared to study start, and at the end of follow-up compared to the end of DI End of WL vs. study start End of DI vs. study start End of follow-up vs. end of DI LCD (n 5 29) VLCD (n 5 28) LCD (n 5 29) VLCD (n 5 28) LCD (n 5 28) VLCD (n 5 27) Weight (kg) 28.2 6 0.5*** 29.0 6 0.4*** 28.4 6 0.5*** 29.3 6 0.5*** 14.2 6 0.6*** 14.5 6 0.7*** BMI (kg/m2) 22.8 6 0.2*** 23.0 6 0.1*** 22.9 6 0.2*** 23.1 6 0.1*** 11.4 6 0.2*** 11.5 6 0.2*** Hip circumference (cm) 24.9 6 0.6*** 25.9 6 0.6*** 26.2 6 0.7*** 26.3 6 0.6*** 2.0 6 0.7** 1.0 6 0.9 Waist circumference (cm) 27.3 6 0.8*** 27.7 6 0.6*** 28.3 6 1.0*** 26.9 6 0.7*** 13.8 6 0.8*** 13.0 6 0.7*** Waist/hip ratio 20.03 6 0.01** 20.02 6 0.01*** 20.02 6 0.01** 20.01 6 0.01 10.02 6 0.01* 10.02 6 0.01* Body fat (%) 25.4 6 0.5*** 24.7 6 0.4*** 26.1 6 0.5*** 25.9 6 0.5*** 12.9 6 0.5*** 12.8 6 0.6*** Body fat (kg) 27.6 6 0.5*** 27.4 6 0.4*** 28.3 6 0.6*** 28.6 6 0.5*** 13.9 6 0.6*** 14.0 6 0.7*** FFM (kg) 20.6 6 0.2* 21.6 6 0.2***,‡ 20.2 6 0.2 20.8 6 0.2**,† 10.4 6 0.3 10.5 6 0.2* Systolic blood pressure (mmHg) 23.1 6 2.1 27.5 6 1.8*** 22.7 6 2.1 26.1 6 1.7*** 13.0 6 2.1 4.7 6 1.6** Diastolic blood pressure (mmHg) 24.8 6 1.9* 26.8 6 1.2*** 24.9 6 1.5* 26.3 6 1.3*** 12.1 6 1.0* 14.8 6 0.9** PA-score 10.27 6 0.13 20.11 6 0.15 10.18 6 0.13 20.15 6 0.13 20.3 6 0.1* 20.3 6 0.1* Values are mean 6 SEM.
b
*P < 0.05, **P < 0.01, ***P < 0.001, paired sample t-test, within-diet change from end of WL vs. study start, end of DI vs. study start, and end of follow-up vs. end of DI.
†
P < 0.05,‡P < 0.01, change from start of study between diets. No significant differences were observed between the LCD- and VLCD-group at the end of follow-up com-pared to the end of DI.
c
LCD: low-calorie diet; VLCD, very-low-calorie diet; BMI, body mass index; FFM, fat-free mass; PA-score, physical activity score; WL, weight loss; DI, dietary intervention.
P 5 0.642, respectively). Furthermore, the PA-score after follow-up was significantly lower compared to after DI (Table 2) and the other two time points (data not shown).
Effects of weight cycling
Number of weight cycles was significantly higher in women com-pared to men (2.7 6 0.3 vs. 1.4 6 0.2, respectively, P 5 0.001). Number of weight cycles did not correlate with weight regain after follow-up (whole group: Spearman’s rho 5 0.026,P 5 0.859).
Discussion
The present study showed that, with similar total weight loss, the rate of weight loss did not affect weight regain. Participants regained on average more than 50% of their lost weight within the 9-month follow-up period in both groups. This finding directly con-tradicts the current dietary guidelines of several countries, which recommend a more gradual weight loss approach for prevention of weight regain (8,9). While VLCDs are often seen as a temporary means to lose weight, LCDs are thought to mimic a healthy lifestyle
more closely, which could improve successful long-term weight management. Regardless, VLCDs remain a popular strategy for weight loss in the general population because they are easy to use and produce quick weight loss results. In accordance with our results, one study showed that an 8-week VLCD or a 17-week LCD did not result in significant differences in weight regain after a 1-year weight maintenance diet or after a subsequent 1-1-year follow-up period (12). More recently Purcell et al. (5) compared rapid (12-week VLCD) with gradual weight loss (36-(12-week LCD) diets, which resulted in similar total weight loss, on 2-year weight regain in a large group of individuals with obesity. The results also showed that the rate of weight loss did not affect the proportion of weight regained. Additionally, even when the initial weight loss is greater, as is more common with VLCDs compared to LCDs, studies have shown that this was correlated with improved weight loss mainte-nance (17-20). Therefore, the current scientific evidence does not support a gradual weight loss approach over rapid weight loss in the prevention of weight regain.
Diet-induced weight loss is often accompanied by loss of FFM, which varies based on the type of weight loss intervention. Chaston et al. (21) previously stated that a comparison of LCDs and VLCDs gave clear evidence that the degree of caloric restriction affects %FFML. However, this systematic review used studies that differed in total weight loss and did not include studies in which participants were randomly assigned to a LCD or VLCD. Our study is, to our knowl-edge, the first to directly confirm with a randomized approach with similar weight loss that a VLCD induced greater loss of FFM com-pared to a LCD. Nevertheless, the clinical relevance of this finding remains to be established since the difference in FFM loss between groups was relatively small (0.6 kg) and did not result in a difference in weight regain between groups. Interestingly, Purcell et al. (5) did not observe a difference in FFM loss between rapid and gradual weight loss groups in participants that lost 12.5% of body weight. Bioelectrical impedance was used to assess FFM in the latter study while the present study used ADP (Bod Pod), and the difference in techniques might explain this discrepancy. Furthermore, %FFML was not significantly different between men and women in this study, in contrast to a recent study that showed that FFM loss was greater in men with overweight and obesity compared to women (22).
A remarkable finding was that the %FFML after DI was lower than after the WL-period. Heymsfield et al. already stated that %FFML can vary with the amount of energy intake and the different phases of a diet (23). These phases can induce changes not only in the mus-cle mass, but also in the hydration of FFM. Indeed, within 4 days after starting a VLCD it was shown that the average glycogen loss was 0.4 kg (24), which can account for 1.6-2.0 kg of FFM because glycogen is stored with three to four parts of water (25). Additional water losses may have occurred during weight loss because increased ketone-body excretion will lead to extra sodium and potassium losses together with water (26,27). Unfortunately we are unable to distin-guish between the water and muscle content of FFM with the current technique used, ADP. However, we believe that the loss of glycogen and water at the end of weight loss is restored with the initiation of a normal diet in the following weeks. Therefore, the FFM measured at the end of DI reflects, in our view, the actual muscle mass changes more accurately than at the end of the WL-period.
Although the rate of weight loss and weight cycling were not corre-lated with weight regain, the PA-score after follow-up and the
%FFML during DI were correlated with weight regain. The reduced level of PA during follow-up could have increased weight regain via a lowered energy expenditure. In accordance, the level of PA was previously shown to be related to long-term weight maintenance (3,28). However, the drop in PA level during follow-up was surpris-ing and difficult to explain. It might have occurred due to reduced motivation in some participants in response to the regained weight. Muscle mass is a key contributor to resting energy expenditure (10,11), and loss of muscle mass, and possibly organ mass, can reduce total energy expenditure and potentially increase weight regain. Alternatively, a compensatory increase in food intake during follow-up could have occurred in the current study, since a reanaly-sis of the Minnesota Experiment showed that the degree of FFM depletion induced by starvation was a predictor of subsequent hyper-phagia upon ad libitum food intake (29). While %FFML after DI was correlated with weight regain, the increased %FFML in the VLCD-group did, however, not result in differences in weight regain between groups. The difference in %FFML between the groups might have been too small to cause differences in weight regain between groups. Also, it can be appreciated that weight regain after weight loss is a multifactorial problem and that changes in FFM alone are insufficient to explain the observed results. Finally, weight cycling has been associated with an increased risk for all-cause and cardiovascular mortality (30-32), but controversy exists regarding whether weight cycling promotes obesity (33). This study did not find an association between the number of weight cycles and weight regain. This is consistent with increasing evidence that weight cycling does not predispose people with overweight and obesity to regain more body fat than what had been lost (34).
A strength of this study was a low drop-out rate and the fact that these drop-outs were mostly unrelated to the intervention. Also, the WS-period allowed for a separation of the effect of a negative energy balance and the effect of weight loss. The importance of the use of a WS-period was stipulated by the strong changes in %FFML during this period in the absence of important changes in body weight. In our opinion, the absence of dietary advice in the follow-up period reflects common practice, because most individuals with obesity do not receive dietary counseling after a weight loss attempt. A weakness of this study was the use of the Bod Pod to assess body composition. Although it is a very cost-efficient method, it does not always accurately reflect changes in body composition associated with weight loss due to changes in hydration of FFM (35). Further-more, a questionnaire was used to assess the level of PA while more accurate alternatives are available.
Conclusion
This study showed that, with similar total weight loss, the rate of weight loss did not affect long-term weight regain in individuals with overweight and obesity. We identified that a VLCD induced greater loss of FFM compared to a LCD but did not result in differ-ences in weight regain between groups. With respect to measure-ments of FFM loss with ADP it is important that, especially with respect to VLCDs, they are performed under conditions of energy balance. Furthermore, the diet-induced loss of FFM and lowered habitual PA during follow-up could have decreased energy expendi-ture and thereby increased weight regain. These data reveal that the
role of rate of weight loss in dietary recommendations and in the etiology of weight regain should be critically reviewed.O
Acknowledgments
Authors thank Imco Janssen, Helena Schaap, and Christianne Pijls for their assistance on test days. Also, authors thank the study par-ticipants for their contribution to the trial.
VC 2016 The Obesity Society
References
1. Ng M, Fleming T, Robinson M, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the Global Burden of Disease Study 2013.Lancet 2014;384:766-781. 2. Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment of burden of
disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.Lancet 2012;380:2224-2260.
3. Wing RR, Phelan S. Long-term weight loss maintenance.Am J Clin Nutr 2005;82: 222S-225S.
4. McGuire MT, Wing RR, Hill JO. The prevalence of weight loss maintenance among American adults.Int J Obes Relat Metab Disord 1999;23:1314-1319. 5. Purcell K, Sumithran P, Prendergast LA, Bouniu CJ, Delbridge E, Proietto J. The
effect of rate of weight loss on long-term weight management: a randomised controlled trial.Lancet Diabetes Endocrinol 2014;2:954-962.
6. Casazza K, Brown A, Astrup A, et al. Weighing the evidence of common beliefs in obesity research.Crit Rev Food Sci Nutr 2014;55:2014-2053.
7. Casazza K, Fontaine KR, Astrup A, et al. Myths, presumptions, and facts about obesity.N Engl J Med 2013;368:446-454.
8. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults.Am J Clin Nutr 1998;68:899-917.
9. NICE. Obesity: Identification, Assessment and Management of Overweight and Obesity in Children, Young People and Adults.Partial update of CG43. 2014. NICE Clinical Guidelines: http://www.nice.org.uk/guidance/CG189, accessed on August 19, 2015.
10. Sparti A, DeLany JP, de la Bretonne JA, Sander GE, Bray GA. Relationship between resting metabolic rate and the composition of the fat-free mass. Metabolism 1997;46:1225-1230.
11. Deriaz O, Fournier G, Tremblay A, Despres JP, Bouchard C. Lean-body-mass composition and resting energy expenditure before and after long-term overfeeding. Am J Clin Nutr 1992;56:840-847.
12. Toubro S, Astrup A. Randomised comparison of diets for maintaining obese subjects’ weight after major weight loss: ad lib, low fat, high carbohydrate diet v fixed energy intake.BMJ 1997;314:29-34.
13. Richtlijnen voedselkeuze. Available at: http://www.voedingscentrum.nl/ professionals/schijf-van-vijf/Richtlijnen.aspx, 2011, accessed on August 19, 2015. 14. Dempster P, Aitkens S. A new air displacement method for the determination of
human body composition.Med Sci Sports Exerc 1995;27:1692-1697.
15. Siri WE. The gross composition of the body.Adv Biol Med Phys 1956;4:239-280. 16. Baecke JA, Burema J, Frijters JE. A short questionnaire for the measurement of
habitual physical activity in epidemiological studies.Am J Clin Nutr 1982;36:936-942.
17. Astrup A, Rossner S. Lessons from obesity management programmes: greater initial weight loss improves long-term maintenance.Obes Rev 2000;1:17-19.
18. Unick JL, Hogan PE, Neiberg RH, et al. Evaluation of early weight loss thresholds for identifying nonresponders to an intensive lifestyle intervention.Obesity (Silver Spring) 2014;22:1608-1616.
19. Wing RR, Hamman RF, Bray GA, et al. Achieving weight and activity goals among diabetes prevention program lifestyle participants.Obes Res 2004;12:1426-1434.
20. Nackers LM, Ross KM, Perri MG. The association between rate of initial weight loss and long-term success in obesity treatment: does slow and steady win the race? Int J Behav Med 2010;17:161-167.
21. Chaston TB, Dixon JB, O’Brien PE. Changes in fat-free mass during significant weight loss: a systematic review.Int J Obes (Lond) 2007;31:743-750.
22. Millward DJ, Truby H, Fox KR, Livingstone MB, Macdonald IA, Tothill P. Sex differences in the composition of weight gain and loss in overweight and obese adults.Br J Nutr 2014;111:933-943.
24. Kreitzman SN, Coxon AY, Szaz KF. Glycogen storage: illusions of easy weight loss, excessive weight regain, and distortions in estimates of body composition.Am J Clin Nutr 1992;56:292S-293S.
25. Olsson KE, Saltin B. Variation in total body water with muscle glycogen changes in man.Acta Physiol Scand 1970;80:11-18.
26. Henry RR, Wiest-Kent TA, Scheaffer L, Kolterman OG, Olefsky JM. Metabolic consequences of very-low-calorie diet therapy in obese non-insulin-dependent diabetic and nondiabetic subjects.Diabetes 1986;35:155-164.
27. Saris WH. Very-low-calorie diets and sustained weight loss. Obes Res 2001;9 (Suppl 4):295S-301S.
28. Elfhag K, Rossner S. Who succeeds in maintaining weight loss? A conceptual review of factors associated with weight loss maintenance and weight regain.Obes Rev 2005;6:67-85.
29. Dulloo AG, Jacquet J, Girardier L. Poststarvation hyperphagia and body fat overshooting in humans: a role for feedback signals from lean and fat tissues.Am J Clin Nutr 1997;65:717-723.
30. Lissner L, Odell PM, D’Agostino RB, et al. Variability of body weight and health outcomes in the Framingham population.N Engl J Med 1991;324:1839-1844. 31. Diaz VA, Mainous AG III, Everett CJ. The association between weight fluctuation
and mortality: results from a population-based cohort study.J Community Health 2005;30:153-165.
32. Peters ET, Seidell JC, Menotti A, et al. Changes in body weight in relation to mortality in 6441 European middle-aged men: the Seven Countries Study. Int J Obes Relat Metab Disord 1995;19:862-868.
33. Montani JP, Viecelli AK, Prevot A, Dulloo AG. Weight cycling during growth and beyond as a risk factor for later cardiovascular diseases: the ‘repeated overshoot’ theory.Int J Obes (Lond) 2006;30 (Suppl 4):S58-S66.
34. Dulloo AG, Jacquet J, Montani JP, Schutz Y. How dieting makes the lean fatter: from a perspective of body composition autoregulation through adipostats and proteinstats awaiting discovery.Obes Rev 2015;16(Suppl 1):25-35.
