Association of lifestyle and non-lifestyle variables with fertility problems in a population-based cohort of women in Australia

Objectives: 

To assess the relative importance of a range of lifestyle and non-lifestyle variables in relation to the risk of fertility problems over time.

Study design: 

A prospective cohort study using data from the 1973-1978 birth cohort of women participating in the Australian Longitudinal Study of Women's Health.

Main outcome measures: 

Women who completed at least one survey between 2006 (used as baseline) and 2018 and who reported a known fertility status were included (n = 8475). Participants were aged 28-33 years in 2006. Exposure variables included age, body mass index (BMI), polycystic ovary syndrome (PCOS), physical activity, fruit and vegetable intake, smoking and alcohol, and level of anxiety. Women were asked to report their fertility status, categorised as "fertility problems" (with or without treatment) and "no fertility problems". A generalised estimating equation with log-binomial family was fitted, and the estimated probabilities along with 95% confidence intervals were obtained. The relative contribution of each variable to infertility was estimated by multimodel inference.

Results: 

The mean (SD) age and BMI of the 8475 women included was 33.2 (3.5) years and 25.9 (6.0) kg/m 2; of the sample, 24% had problems with fertility. Age, PCOS, anxiety, BMI, and the interaction of age x PCOS had the same relative contribution, of 8.5%, to the risk of fertility problems at any given point across the 12 years. When the contributions of age with any of the exposures (except PCOS) were explored together, the relative contribution was less than that of the respective individual exposures, indicating that their interaction weakly contributes to fertility problems.

Conclusions: 

BMI and anxiety contributed independently of age to infertility problems whereas the effect of PCOS was exacerbated with younger age. These findings highlight the importance of early pregnancy planning, particularly for women with PCOS, and support the need for preconception strategies targeting BMI and anxiety. However, temporal misalignment between current exposures and lifetime infertility reporting may limit causal interpretation.