Des appareils connectés révèlent les liens entre les habitudes d’activité physique et la variation du poids

Veuillez noter qu’à l’exception du titre et du résumé, cet article est uniquement disponible en anglais.

• La santé digitale
• Unité de Recherche sur le Phénotypage Numérique Profond
• Activité Physique, Sport et Santé

Le D^r Guy Fagherazzi, directeur du Département de la santé de la population (DoPH) et chef de l’Unité de recherche Phénotypage numérique profond du LIH et le D^r Laurent Malisoux, chef du Groupe de recherche Activité physique, sport et santé du DoPH, ont exploité le potentiel des outils de santé numériques pour étudier la relation entre le niveau et la fréquence de l’activité physique et les variations du poids. L’étude, qui a été publiée le 15 avril dans la principale revue consacrée à la santé numérique, le « Journal of Medical Internet Research – mHealth and uHealth », est le fruit d’une collaboration avec l’Institut national de la santé et de la recherche médicale (Inserm) français et le fabricant français de dispositifs connectés de santé numérique Withings.

The benefits of physical activity on weight loss and on overall health status are well known, with 10,000 steps/day, three to four days a week, being a widely accepted indication of the level of exercise associated with beneficial outcomes. Nevertheless, robust evidence supporting the definition of a minimal optimal amount of physical activity resulting in weight loss is currently inconsistent. Moreover, most studies have only evaluated the effects of the total level of physical activity on weight change, without necessarily taking into account the complexity of an individual’s different physical activity patterns in terms of amount and frequency over time.

“In this context, our study aimed to elucidate the associations between physical activity expressed in terms of step count, and specifically its intensity and regularity, and weight change over a period of six months”, explains Dr Fagherazzi, corresponding author of the publication.  

At the heart of the research team’s approach were digital technologies, namely Withings consumer-based wearable activity trackers and connected digital scales. These allow individuals to objectively and continuously measure their physical activity levels and weight variations while minimising the bias typically associated with self-reported values from questionnaires, therefore constituting a valuable tool to monitor large populations in real life settings over extended time spans. However, connected devices and the generated data currently remain an untapped resource for clinical and epidemiological research purposes.

The scientists therefore analysed a variety of data, such as sex, age, number of steps per day, body weight, heart rate and blood pressure, from 26,935 Withings connected device users. They calculated the monthly average of daily step values as a proxy for physical activity levels and  derived the monthly coefficient of variation (CV) of daily step values to estimate the variability (i.e. the regularity) of physical activity. Through detailed statistical analyses, the team assessed the associations between these characteristics, their six-month changes and the six-month weight variations.

As expected, the results indicated that greater average physical activity levels correlate with more weight loss. Compared with individuals who were classified as sedentary (below 5,000 steps/day), low active (between 5,000 and 7,499 steps/day), somewhat active (between 7,500 and 9,999 steps/day) and active (above 10,000 steps/day) individuals reported a 0.21 kg, a 0.52 kg and a 1.17 kg greater decrease in weight, respectively. Moreover, each 1,000 steps/day increase in physical activity levels over the six-month follow-up was associated with a 0.26 kg weight reduction.

Interestingly, the findings revealed that a lower variability in the levels of physical activity, i.e. a greater regularity, was also associated with weight loss. Indeed, compared with users with a CV of their physical activity intensity greater than 63%, subjects with a CV ranging from 51 to 63%, from 40 to 51% and below 40% had a 0.19 kg, a 0.23 kg and a 0.33 kg greater decline in weight, respectively. Conversely, no correlation was found between the six-month change in the variability of physical activity levels and weight variations.

“In a nutshell, our study adds to the current body of knowledge that health benefits can already be observed below the 10,000 steps/day threshold. Naturally, the greater the intensity of the activity the more significant the ensuing weight loss, meaning that ‘some physical activity is good but more is better’ with respect to weight change. More importantly, we also showed that it is not only the intensity but the regularity of physical activity that plays a noteworthy role in short-term weight loss. Step count variability is therefore an interesting additional parameter to be considered as a digital biomarker candidate when assessing the impact of physical activity on health”, summarises Dr Fagherazzi.

“The strength of our work lies in the large study population, but also on the use of innovative digital tools that have the potential to disrupt the way clinical and epidemiological studies have traditionally been carried out. We expect our findings to set the basis for a more widespread adoption of wearable connected devices and machine-learning approaches for the development of truly personalised prevention programmes based on physical activity, tailored to an individual’s specific lifestyle, habits and living conditions”, adds Dr Malisoux.

“To this end, as shown by our cooperation with Withings, establishing close collaborations between academia and wearable device manufacturers is crucial in order to give researchers access to valuable and reliable real-time and real-life datasets to enhance population health”, concludes Dr Fagherazzi.

Looking forward, the study could be replicated by focusing also on different health parameters related to physical activity, including blood pressure, heart rate or fat mass percentage, which are easily measurable through connected devices and which represent key cardiometabolic risk factors to be taken into account in the contexts of the obesity epidemic and the digital revolution.