Our research identified a non-linear association between recreational activity time and the risk of female infertility, pinpointing an inflection point at 5.83 h/week (moderate intensity). Beyond this inflection point, as the duration of recreational activity extends, the risk of infertility correspondingly escalates (OR = 1.04, 95% CI: 1.02 to 1.06, P < 0.01). However, there was no similar association between work activity time, walking or bicycle time and infertility. In the results of the subgroup analysis, we observed that the association between recreational activity duration and infertility remained unaffected by these stratified variables, demonstrating its stability. In the study population that met our exclusion criteria, after adjusting for age, race, BMI, educational levels, marital status and smoking status, our finding suggested different associations between different types of physical activity and infertility.

Distinguishing our study from previous research, we specifically excluded female participants who were current alcohol consumers. This decision was based on the clear understanding that habitual alcohol consumption negatively impacts female reproductive function [25] and is typically avoided by women intending to conceive. Based on this exclusion criterion, our study identified a positive correlation between prolonged periods of recreational activity and the risk of infertility. This is consistent with the findings of a previous study, which demonstrated that high intensity and frequency of physical activity have a negative impact on female reproductive health [14]. However, other studies have discovered no significant link between physical activity and female infertility, [16,17,18,19] or have indicated that physical activity may actually act as a protective factor against infertility [15]. We believe that the divergence in research findings is likely due to the studies not considering the independent effects that different types of physical activity may have on the human body, as well as the lack of adjustment for certain confounding factors or the selection of appropriate inclusion criteria.

To the best of our knowledge, our study represents the first attempt to explore the relationship between various forms of physical activity and infertility. Our findings indicate that the relationship between various types of physical activity and infertility is not uniform. In our study, recreational activities had a more stable association with infertility than work activities, whereas traffic-related activities had no significant association with infertility. Two prior studies have similarly indicated that various types of physical activity exert distinct effects on the body, which supported the physical activity paradox [26, 27]. The variation observed might be attributed to self-determined motivation [28]. Recreational activity represent those chosen by individuals to engage in during their leisure time, whereas work activity are obligations that individuals must fulfill during their working hours. Consequently, recreational activity possess a more subjective nature compared to work activity. Moreover, distinct types of physical activity exhibit varying characteristics. Recreational activity predominantly involve high-intensity and short-duration exercises, whereas work activity tend to consist of prolonged periods of low-intensity and static tasks.

Engaging in high-intensity recreational activities for prolonged durations may result in infertility through various mechanisms. On the one hand, high-intensity physical activity may interact with additional psychosocial and metabolic stressors, prompting physiological stress responses. This can disrupt the pulsatile secretion of hypothalamic gonadotropin-releasing hormone (GnRH), which, via the hypothalamic-pituitary-ovarian (HPO) axis, impedes the production of estrogen and progesterone - pivotal hormones for ovulation and conception [29]. On the other hand, it can induce infertility by causing negative energy balance and impeding the necessary processes for ovulation [30].

Our study possesses several notable strengths. Firstly, we leveraged data from the NHANES database, which offers comprehensive coverage across all regions of the United States and ensures strong representativeness. Secondly, our investigation separately examined the relationship between various types of physical activity and infertility, uncovering a non-linear correlation between recreational activity time and infertility. Thirdly, by employing threshold effect analysis, we identified the inflection point of moderate intensity recreational activity time at 5.83 h/week, thereby offering valuable recommendations for the weekly exercise duration for women of childbearing age. Lastly, through subgroup analysis, we revealed that the relationship between recreational activity duration and infertility remained stable and unaffected by the stratified variables.

However, there are some limitations to our study. First, despite revealing a correlation between physical activity and infertility, establishing causation is not possible due to the cross-sectional nature of the study. Future prospective studies are required to investigate the causal relationship between the two factors. Second, our study is based on self-reported data, which includes information on infertility and physical activity. It is important to consider that self-reporting may introduce recall bias, as women might either overestimate or underestimate their exercise levels and misjudge their infertility status. Third, the NHANES dataset did not contain information on the precise length of infertility or the fertility status of their partners. Fourth, due to the lack of data on conditions such as polycystic ovary syndrome and endometriosis, which can have an impact on female fertility, within the NHANES database, we cannot exclude the influence of these potential factors on our results. Lastly, as the dataset originates from a nationwide survey in the United States, further validation is needed to confirm its generalizability across different racial groups.

In conclusion, our findings indicate a non-linear correlation between recreational activity duration and infertility, and the relationship between different types of physical activity and female infertility varies, which offering valuable insights for establishing healthy physical activity guidelines for women of childbearing age. However, because this study was a cross-sectional study, more prospective cohort studies are needed in the future to explore causality.