日本体育・スポーツ・健康学会第73回大会

講演情報

ポスター発表(専門領域別)

専門領域別 » 測定評価

測定評価(偶数演題)/ポスター発表

2023年9月1日(金) 13:20 〜 13:50 RY208 (良心館2階RY208番教室)

[08測-ポ-02] Physical activity volume and intensity may not be associated with vaso-occlusive crises in sickle cell anemia patients

A new approach in sickle cell anemia patients physical activity management ?

*Gallou Guyot Matthieu1,2、Miyachi Motohiko3 (1. Japan Society for the Promotion of Science、2. Department of Human-Environmental Science, Ochanomizu University, Tokyo, Japan、3. Faculty of Sport Sciences, Waseda University, Tokorozawa, Saitama, Japan)


Patients with sickle cell anemia (SCA) synthetize an abnormal hemoglobin, which is responsible for several severe clinical symptoms, including painful vaso-occlusive crises. Because physical exercising is considered a potential trigger of vaso-occlusive events, SCA patients commonly adopt sedentary lifestyles, preventing the development of vascular. In the present study, we assessed the time relationship between physical activity (PA) and the occurrence of painful events in SCA patients.

SCA patients were recruited in Dakar, Senegal. They were asked to wear a PA monitor (Fitbit Alta) and to fill a symptoms diary for 5-15 weeks. One-hundred and seventeen data features were extracted in both the time and frequency domains from the min-per-min data (step-count, energy expenditure, intensities) and 3 features were extracted from the pain diary. We evaluated their association using a 4-day sliding window. The extracted data were split at an 8:2 ratio and assigned to training or test in a random forest models made of 20 decision trees (sklearn library, Python). The analysis was repeated 5 times with different training-test splits, and (1) with or (2) without the inclusion of the pain diary features within the descriptors.

We recruited 76 patients (29 +/- 8 years old), for a total of 5553 days of observations. The patients performed 10000 steps per day (10377 +/- 3296) and an average of 210 +/- 96 min per day of 3 MET or more PA. The testing mean balanced accuracy for the two models were (1) 58+/-1.5 % and (2) 51+/-0.6 %.

Despite the use of an exhaustive panel of PA data features and the very high PA level of the studied population, the random forest models that used PA features only were not able to predict painful days. Our results indicate no time relationship between PA done in the course of the daily life and pain days, advocating for a less conservative approach regarding PA recommendation in SCA patients.