This study shows that repetitive weight loss and regain during the career of elite athletes does not result in a massive body mass gain after retirement from sport competition. Our data also show high physical activity levels and low eating disorder scores in elite retired athletes. The BMI change after the retirement is influenced by the sport group but not by the number of diets performed during the career and the dieting or non-dieting behavior.
The studied population regrouped 136 French athletes of the main weight-class sports competing in striking (boxing) or grappling (judo and wrestling) combat sports and one endurance sport (rowing) with data corresponding to the sport career and after retirement.
The calculated correlation between the declared and the measured body mass suggests a good reliability of data from interviews. In addition, this result reflects the good ability for former athletes to assess their own body weight probably because they focused on it during their whole sport career.
The hypothesis stating that repeated weight loss may result in a mid-term body mass gain (and consequently a BMI increase) is not supported by our study. The BMI of retired athletes remains under the mean BMI of the general population matched by age group. The body mass gain between the end of the career and the day of the questionnaire was not different between the dieting and the non-dieting groups. Similar results were reported in other studies on former wrestlers in comparison with other athletes  and on the short-term effects (6-12-months) of diets in athletes who were not necessarily participating in weight class sports .
Our study brings a new approach through a mean follow-up of 22 years (since their age of 18 y and the time of the study) (mean age: 40.4 y ± 6.9). From the age of 18 years to 50 years, athletes gained 3.2 kg/m2 compared to 4.2 kg/m2 gained by the general population. It follows the BMI change in the population of developed countries with an initial growth until a peak at 55–60 years and a subsequent decline. Participants declared a mean 4.7 h/week of sport practice, which is higher than the 2 h-3 h/week of moderate physical activity recommended for general population . As recently reported, former elite athletes appear to maintain a high volume of physical activity  that could explain the limitation in the amount of body mass gain observed after retirement.
Our findings are in conflict with a previous study, reporting that weight gain post career in elite athletes having experienced weight cycling was higher than those observed in non-dieting athletes . In our study, non-dieting athletes presented a higher BMI due to the type of sport they represent (mostly heavyweight rowers), which leads to specific morphological particularities (they are heavier and taller than the other athletes). In this previous study, retired dieting athletes were reported with lower physical activity level than non-dieting athletes whereas such difference did not appear in our data. The control populations also differ between both studies: in the French population, 30% of individuals are overweight , in contrast with Saarni  who has a control group with 50% of overweight individuals. Moreover, in Saarni et al. participants competed from 1920 to 1965 whereas in the present study participants competed from 1978 to 2003. This could be explained by sport constraints as well as by athletes monitoring that differed between these two periods. Guidelines for nutrition behavior and lifestyle have markedly changed during the 20th century according to society and population changes that could have influenced the athletes weight management at the sport competition retirement .
The BMI trend is the same in dieting athletes, non-dieting athletes and in the general population. According to the declaration of athletes, we can stress on the minor amount of body mass lost during each diet. The elite athletes need only to adjust their mass up to a few kilos to enter their weight class. Wrestlers lost a mean of 7% of their body mass per diet, in contrast with the lightweight rowers who make the lightest diets with a mean loss of 4.5% of body mass probably due to the time of the weigh-in (2 h before competition). They can’t support a massive dehydration. Our results are in accordance with a study which reported a mean loss of weight of 6.9% for wrestlers . The BMI change statistically differs between the disciplines (despite similarities) and might depend on the sport-specific training developed by the athletes. In our population sample, wrestlers have the highest BMI that could be associated to their previously reported higher trunk strength when compared to judokas . Conversely, recall of body mass during the career showed that boxers had the lowest body mass and BMI as they principally belong to lightweight categories. Boxers show a higher body mass increase at the end of their career, without exceeding the values of the general population. Heavyweight rowers presented a high BMI, being taller (mean 1.91 m ± 0.05) and heavier (86.97 kg ± 5.65). According to the competition level, rowers anthropometrics may be linked to performance [38, 39].
Athletes competing in aesthetic or in combat sports have been assumed to be at greater risk for developing eating disorders that could influence BMI change post career . In this present study, no correlation is found between the score obtained at the EAT-26 questionnaire, the number of diets and the total body mass lost during the career. However, dieting athletes showed the higher scores in the subscale related to the food control. As observed in a larger population of athletes currently engaged in competition , our study does not show high levels of eating disorders in retired elite athletes (only 0.4% participants had an EAT-26 score above 20). Our results suggested that the athlete’s nutritional behavior during the sport career does not induce mid- and long-term eating disorders. During their career, athletes only control their body mass in order to avoid gaining fat mass before competitions without developing an eating disorder [36, 40]. Cereda et al.  described that body mass gain observed in obese participants may be due to a hyperphagic state following the diet whereas athletes appeared to maintain food intake control during their career as well as after retirement .
The main limitation of the present study could be that data were provided by phone interviews and some of them focused on past event. The comparison between self-reported and direct measurement of body weight supported the accuracy of the data we obtained from interviews. Moreover, concerning retrospective data, to our knowledge no adequate data base was previously built in French elite athletes population, inducing that no other method could be adopted to observe long-term effects of weight cycling.