Effectiveness of lifestyle intervention in subgroups of obese infertile women: a subgroup analysis of a RCT

Study question: Do age, ovulatory status, severity of obesity and body fat distribution affect the effectiveness of lifestyle intervention in obese infertile women? Summary answer: We did not identify a subgroup in which lifestyle intervention increased the healthy live birth rate however it did increase the natural conception rate in anovulatory obese infertile women. What is known already: Obese women are at increased risk of infertility and are less likely to conceive after infertility treatment. We previously demonstrated that a 6-month lifestyle intervention preceding infertility treatment did not increase the rate of healthy live births (vaginal live birth of a healthy singleton at term) within 24 months of follow-up as compared to prompt infertility treatment in obese infertile women. Natural conceptions occurred more frequently in women who received a 6-month lifestyle intervention preceding infertility treatment. Study design, size, duration: This is a secondary analysis of a multicenter RCT (randomized controlled trial), the LIFEstyle study. Between 2009 and 2012, 577 obese infertile women were randomly assigned to a 6-month lifestyle intervention followed by infertility treatment (intervention group) or to prompt infertility treatment (control group). Subgroups were predefined in the study protocol, based on frequently used cut-off values in the literature: age (≥ 36 or < 36 years), ovulatory status (anovulatory or ovulatory), BMI (≥ 35 or < 35 kg/ m) and waist-hip (WH) ratio (≥ 0.8 or < 0.8). Participants/materials, setting, methods: Data of 564 (98%) randomized women who completed follow-up were analyzed. We studied the effect of the intervention program in various subgroups on healthy live birth rate within 24 months, as well as the rate of overall live births (live births independent of gestational age, mode of delivery and health) and natural conceptions conceived within 24 months. Live birth rates included pregnancies resulting from both treatment dependent and natural conceptions. Logistic regression models with randomization group, subgroup and the interaction between randomization group and subgroup were used. Significant interaction was defined as a P-value <0.1. Main results and the role of chance: Neither maternal age, ovulatory status nor BMI had an impact on the healthy live birth rate within 24 months, nor did they influence the overall live birth rate within 24 months after randomization. WH ratio showed a significant interaction with the effect of lifestyle intervention on healthy live birth rate (P = 0.05), resulting in a lower