24 weeks of Pilates-aerobic and educative training to improve body fat mass in elderly Serbian women

Dear editor

We read with interest the article by Ruiz-Montero et al, in which the authors used a before-and-after study design to examine changes in body composition (fat mass and lean body mass) related to an aerobic-Pilates program in elderly Serbian women.¹ The authors concluded that “a combined program of aerobic and Pilates, carried out under the supervision of an instructor, at least twice a week, produces health benefits in functionally independent women over the age of 60”. This conclusion is overly optimistic and not supported by the evidence provided. The authors used an uncontrolled study design so that, by definition, the observed changes in anthropometric endpoints can at best be interpreted as preliminary evidence of the effectiveness of the exercise program. Further, the evidence provided for the effectiveness of the exercise program is very weak. The authors attributed the changes in outcomes to the exercise program, but ignored the potential effect of the imposed study diet, health education, and other factors that may have independently affected the anthropometric outcomes. The endpoints chosen by the authors are surrogate outcomes, not patient relevant outcomes, so it is incorrect to refer to health benefits. With the exception of the waist-to-thigh ratio, imprecise thus not statistically significant changes were observed in muscle perimeters (ten outcomes), skinfold thickness (seven outcomes), and bone measurements (four outcomes). However, the authors inaccurately report “As for skinfold thickness, various results were obtained (P<0.05)”, and fail to address the problem of multiple testing in their discussion. In contrast, the observed change in fat mass may be of interest, but the clinical relevance of the magnitude of this finding is not discussed, nor is it put in context with the imposed study diet or other confounding factors. In conclusion, the authors seem to base their conclusions on an incorrect interpretation of their data, providing yet another example for the necessity of multidisciplinary collaborations, where at least health care specialists, biostatisticians, and methodologists team up to design, conduct, and report clinically meaningful research.