Background Physical activity may improve glucose homeostasis and reduce the risk of adverse outcomes associated with obese pregnancy. An understanding of physical activity determinants in obese pregnant women is needed to inform interventions.

Aim To identify predictors of physical activity in obese pregnant women in the UK Pregnancy Better Eating and Activity (UPBEAT) pilot trial.

Methods 102 participants in UPBEAT, recruited at 15-18 weeks' gestation, were included. Women were randomly allocated to receive a combined lifestyle intervention, which included a pedometer-based goal oriented walking intervention, or usual care. Physical activity was assessed at baseline, 28 and 35 weeks' gestation by accelerometer. Linear regression was used to identify factors associated with combined moderate and vigorous physical activity (MVPA).

Results Data were available for 102, 69 and 53 women at baseline, 28 and 35 weeks' gestation respectively. Physical activity declined with gestation from median 37 to 23 minutes per day from baseline to 35 weeks (p<0.001). Age and parity were negatively associated with activity at baseline; higher socio-economic position (Index of Material Deprivation) was positively associated with activity. Baseline activity was the strongest predictor of activity at 28 and 35 weeks' gestation. Intervention status was not a significant predictor of activity at 28 or 35 weeks' gestation.

Conclusion Younger age, nulliparity and higher socio-economic position were associated with higher physical activity at baseline. Baseline activity, reflecting prepregnancy activity, was the strongest predictor of remaining active during pregnancy. Women should be encouraged to participate in physical activity as early as possible in pregnancy.