Weight-related behaviors and weight loss maintenance: a cross-sectional study in Cyprus

Weight loss maintenance has become the greatest challenge in the management of obesity. Despite the fact that many people manage to lose weight, the percentage of those who succeed in maintaining this loss remains low [1, 2]. There is currently no consensus on the definition of weight loss maintenance in adults. Some studies proposed a definition of intentional weight loss and maintenance for at least 6 months [3], while others proposed a definition of weight loss ≥5% of initial loss and maintenance for at least 1 year [4]. Achieving an intentional loss of ≥10% of maximum body weight and maintaining it for ≥1 year is the most widely used description of successful weight loss maintenance [5, 6], and based on it, ≥20% of at least overweight individuals who had previously lost weight manage in maintaining a significant amount of weight loss [5‐9].

Registries of individuals with successful long term weight loss have been developed worldwide to explore factors leading or related to successful maintenance of weight loss [6, 10‐13]. Notably, data within these registries include information on socioeconomic background, medical history of individuals as well as data related to individuals’ lifestyle such as sleep, physical activity, and dietary patterns. In particular, higher levels of physical activity, regular monitoring of diet and body weight were the most frequently reported factors [4, 14‐17]. In relation to the diet, several studies have examined specific nutrients, foods or dietary schemes as potential factors affecting weight loss maintenance; however, none of these studies has managed to identify a single nutrient, food or even a specific dietary scheme as the most effective [18].

Recent guidelines related to the management of overweight and obesity refer to the importance of incorporating behavioral techniques into various obesity treatment interventions and promote healthy weight-related behaviors [19, 20]. Findings from Weight Control Registries support that reducing energy intake by high frequency of self-weighing and regular meal frequency may promote weight loss maintenance [13]. In addition, it is apparent that individuals who maintain their weight report the use of more behavioral techniques [14, 15]. However, evidence supporting the use of specific weight-related behaviors that promote weight loss maintenance of overweight or obese individuals, who have previously lost significant weight, is scarce [13‐15]. Specifically, few behaviors were previously studied mostly on a piecemeal basis rather that in the context of an index that could potentially evaluate their additive, complementary or synergistic effect. In addition, comparisons between maintainers and regainers with significant prior weight loss, could enable researchers to identify important target behaviors for daily practice.

The current study aimed to explore differences in weight-related behaviors between maintainers and regainers enrolled in the MedWeight control registry in Cyprus. Moreover, another aim was to develop a weight-related behavior index predicting the likelihood of weight loss maintenance which could be used as a simple target tool by the Health Care Professionals to monitor weight loss maintenance.

Table 1 presents the characteristics of all the participants enrolled in the study, based on their weight maintenance status. In particular, 232 men and women enrolled in this study of which 145 were maintainers and 87 regainers. Concerning the total sample, 52.2% were women, 29.7% married and most of them employed (60.2%). Regainers were older than maintainers (37.3 ± 14.4 years vs 33.0 ± 12.2 years, P < 0.05) and they also had less years of education (13.3 ± 4.2 years vs 15.1 ± 3.2 years, P < 0.05). Moreover, initial weight loss was significantly higher among the maintainers compared to regainers (26.7 ± 16.0 kg vs 15.4 ± 7.6 kg, P < 0.05) and maintainers reported maintaining this loss for over 3 years. Although maintainers reported a lower weight and BMI, their max weight, max BMI and initial weight loss was higher than regainers (P < 0.05 for all comparisons).

For the development of WRBI index, the responses of the participants were grouped into two categories for each item of the index in order to simplify the final index scoring (Table 2). The Cronbach’s alpha score was considered acceptable (Cronbach’s alpha = 0.68). Results after the development of the WRBI index showed statistically significant differences between the responses of two groups regarding the following weight-related behaviors: eating episodes per day (P = 0.008) and frequency of eating home cooked meals (P = 0.004). There was a trend towards significance regarding frequency of eating out (P = 0.058). Although no statistically significant differences were observed regarding the other behaviors, their inclusion in the WRBI index was considered critical to provide a more comprehensive assessment of nutritional behavior, as described in other indexes or scores when assessing diet and other lifestyle factors [32, 33, 37]. Regarding the WRBI total score, statistically significant results were observed between the two groups (P = 0.022). Notably, results from logistic regression models indicated that for every point scored in WRBI total score the odds of maintaining weight loss increase at 30% (Model 1: P < 0.05, Odds ratio 1.306, 1.095–1.556 95% Confidence interval., Model 2: P < 0.05, OR 1.308, 95% CI: 1.097–1.560) and increase to 38% after adjusting for physical activity (Model 3: P < 0.05, OR 1.377, 95% CI: 1.114–1.701). Additional adjustment for baseline BMI did not change the results.

The performance of the WRBI was useful in discriminating maintainers among individuals, with an area under the receiver operating characteristic curve (AUC) of 0.589 (P = 0.023, 95% CI: 0.514–0.664.). The best cut-off value for the WRBI was ≥1.50 (Sens/1-Spec: 0.779/0.724) (Supplemental Figure 1).

The current study aimed to examine the association between eating behaviors and weight loss maintenance among adults who had previously lost weight and maintained it or regained it. The main findings indicate that maintaining weight loss is associated with having at least 6 eating episodes per day and almost always eating home cooked meals. Hence, the WRBI total score was positively associated with weight loss maintenance, indicating that the WRBI index could be a useful tool for monitoring weigh loss maintenance and/or as a guide for relevant nutrition counseling. Health care professionals need more practical tools to address weight management success. It was therefore considered wise to develop an index incorporating weight-related behavior that could potentially affect weight loss maintenance. A more holistic approach which incorporates interactions of eating behaviours, such as the WRBI index, may produce many benefits of clinical research and practice than any other single variable. To our knowledge this is the first study to produce such a weight-related behaviours index. Previous efforts regarding the development of similar tools were mostly based on foods or nutrients [38‐40], and not on behaviors.

In relation to specific behaviors, our results indicated that maintainers reported having more eating episodes per day than regainers. This is in line with the results of a previous study examining self-reported eating frequency of main meals and snacks consumed per day in weight loss maintainers, normal weight, and overweight individuals. Although there were no differences regarding participants’ main meals per day, maintainers and normal weight individuals consumed more snacks than overweight individuals [41]. However, results from the National Weight Control Registry in the USA, highlighted that maintenance of weight loss could also be achieved by having fewer eating episodes, a behaviour reported mostly by older people [42]. However, findings regarding the effect of meal frequency in weight loss maintenance seem to be conflicting and a possible reason could be the fact that different definitions of “a main meal” or “a snack” were used in various studies. For the purposes of the current study, we decided to choose having ≥6 eating episodes daily as the healthy behavior, due to reasons related with the inclusion of comparison groups and distinct definition of main meal/snack reported by previous study [41].

Although there is a great heterogeneity across studies examining the relationship between meal patterns and body weight mainly due to the method of assessment of main meals and snacks, most of them indicate a negative association between higher frequency of home cooked meals and body weight, in accordance with our study. As an example, in a cross sectional study of 11,936 participants, it was found that eating home cooked meals more frequently is associated with lower adiposity and better diet quality [43]. Previous studies indicated that eating home cooked meals could help individuals monitor their food and energy intake and that self-monitoring of food and energy intake was positively related to weight maintenance status [44, 45]. In any case, evidence across the above studies is a result of the cross-sectional nature and therefore, causation cannot be inferred. Further clarification is required through other prospective studies.

The strength of this study is that, to the best of our knowledge, the population used has not yet been examined related to weight loss maintenance. In addition, the population was consisted of both maintainers and regainers which allowed for direct comparisons of obesity related eating behaviors, as well as young volunteers, whereas in most relevant studies participants were middle-aged [15, 16, 46, 47]. Moreover, the use of two 24-h recalls as assessment tools allowed a more detailed dietary assessment. This study has also some limitations. In particular, the observational nature of this study indicates associations, but no causation can be drawn. Another limitation of the study is the fact that all data derived from questionnaires (e.g., weight measurement and dietary recalls) were self-reported, a fact which could potentially lead to misreporting and information bias. Yet, this design of contacting has been previously reported that might enhance response rates and improve sample representations, while maintain unbiased outcomes [48, 49]. Moreover, the tendency to underreport is a commonly observed behavior in studies of individuals with overweight or obesity, weight loss maintainers do not seem to underreport in a greater degree than individuals with overweight or obesity [50]. Also, no information was collected regarding the type of the weight loss program used by the participants, however it seems that despite the dietary method used for weight loss the results of weight maintenance do not appear to differ [51]. The small number of participants and the fact that they were not equally distributed in the two groups, resulting in a group of maintainers twice the size of the regainers’ group, are another limitation. However, it is unlikely that this inequality could have influenced our results as previously reported [52]. Lastly, the AUC demonstrated fair to poor discriminatory ability most likely due to the small number of participants. Therefore, further studies are required to assess the discriminative ability of the Weight-Related Behaviors Index in predicting weight regain vs. maintenance.

Adoption of healthy eating behaviors, such as eating more frequent home cooked meals and less eating away from home meals may influence weight loss maintenance. A Weight-Related Behaviors Index was designed to incorporate specific obesity-related eating behaviors previously shown to influence long term weight loss. The use of WRBI index by the Health Care Professionals could be either an assessment or a target tool to support individuals with overweight or obesity to maintain their weight after significant weight loss. Studies exploring the usability of this index in other setting or populations groups and its ability to discriminate between successful weight losers or maintainers and not are required in the upcoming future.