To our knowledge, this is the first study to investigate the relationship between TL, fecundability, and ART use. We found no significant associations between TL measures and fecundability or infertility in this large population-based study. However, in men, we found a higher likelihood of ART use among men with longer aTL, a pattern not found in women. Importantly, this increased likelihood of ART use in men with longer telomeres was not replicated when we used GRS for TL or genetically predicted aTL, suggesting the possibility of unmeasured confounding factors or bias influencing this outcome (34).
Our findings contrast with those of earlier studies reporting an association between shorter TL and infertility-related phenotypes in women, such as PCOS and endometriosis, as well as association between shorter TL and sperm quality in men (15, 16, 21). However, the fact that many studies in the literature find conflicting results may indicate that there is no true relationship between TL and infertility, which is in line with our main finding. Any potential association between TL and infertility in men needs to be validated.
Differences in TL measurement techniques may partly account for the inconsistent findings across studies. These methods range from polymerase chain reaction (qPCR)-based approaches to various fluorescent in situ hybridization (FISH) methods and Southern blotting (35–38). Additionally, telomere dynamics vary across different tissues, especially between germ cells and somatic cells (39). TL can be measured in different tissues and cells, including stromal cells, leukocytes, endometrial cells, tubal epithelial cells, granulosa cells or oocytes, and sperm cells (15). While telomeres shorten with age in most tissues, they remain stable in some, like testis and ovaries (40). Moreover, while there is considerable evidence linking oxidative stress and TL shortening in vitro, the impact of oxidative stress on TL shortening in vivo is less understood (41). The relatively small sample sizes in selected populations used in some studies may also lead to noise results. These methodological shortcomings could significantly affect the interpretation of the effect of TL on infertility-related phenotypes.
The observed modest association between longer TL and a higher likelihood of ART use in men remained stable, both when male factor infertility was considered as a main reason, when it was considered as a contributing reason for ART use, and when mutually adjusting for their partners’ aTL. However, this association was not replicated in analyses involving GRS for TL or genetically predicted aTL, indicating that the relationship might not be causal. Our results may also be biased by selection as men using ART was the smallest subsample in our study. This could induce associations of similar magnitude, indicating that the observed association could instead be influenced by lifestyle factors linked to TL, such as nutrition or smoking (42). Indeed, smoking is known to be an important external source of reactive oxygen species, which are predicted to negatively affect the triple guanine residues in telomeric repeats (43, 44).
Several studies have identified a positive correlation between offspring’s leukocyte TL and paternal age at conception, even after adjusting for offspring’s age (45–48). To examine whether potential effects of paternal age at conception (PAC) could explain the significant association between aTL and ART use in men, we calculated paternal residual aTL by regressing aTL on fathers’ age at birth when the women and men in our study population were born, and then repeating the analyses. The Pearson correlation between fathers’ ages at birth and aTL was 0.11 for women and 0.14 for men (Supplementary Figure S4). We found significant PAC effects, although adjusting for PAC in the analysis of aTL and ART use in men did not alter the results significantly, except for a slight strengthening of the association (Supplementary Figure S12, S13 and S14).
Studies have suggested that the TL in sperm cells increases with age, and it has been proposed that a more robust telomerase activity in these cells could be a mechanism for the observed PAC effect (48). Moreover, since older men have lower sperm counts, the amount of telomerase available per sperm cells is also greater in older men. Consequently, the offspring of older fathers tend to have longer telomeres in leukocyte cells (49). However, this significant PAC effect did not explain the association between longer telomeres and increased likelihood of using ART among men in our study. An alternative explanation might be related to common mechanistic patterns associated with male factor infertility. For instance, it has been proposed that male factor infertility, including low sperm count, may cluster within families (50, 51), potentially leading to longer telomeres in the sperm of fathers and the leukocytes of their offspring. This familial pattern could also contribute to the observed increased likelihood of ART use among men from such families. Further studies, with detailed analysis of male factor infertility across generations, are needed to elucidate the link between longer TL and the higher likelihood of ART use in men.
Key strengths of our study include our ability to investigate the associations using three measures of TL: aTL assessed through direct measurements by Southern blot, GRS constructed for TL from GWAS summary statistics, and genetically predicted aTL. Our measurements of TL were obtained using Southern Blot, which ensures higher accuracy compared to, for example, PCR-based methods. We also had detailed information on TTP, allowing us to estimate FRs. Moreover, we were able to investigate the risk of infertility among couples who conceived through sexual intercourse and those who conceived through ART. We also had detailed information on the reasons for using ART.
The limitations of our study include the fact that MoBa is a pregnancy cohort. Since all the participants were recruited based on having achieved a pregnancy, we were unable to investigate associations in women and men the most severe phenotypes, such as those who never conceived or experienced an early pregnancy loss. Given the lack of data on childless women and men, we were also unable to investigate the probability of using ART independent of treatment outcome. Recall bias may also influence our findings, given potential inaccuracies in participants’ recollections of their TTP. Lastly, MoBa comprises a selected and homogeneous group in general, representing women and men in a higher socioeconomic bracket compared to the population at large. This homogeneity may have contributed to TL being a poor biomarker of health-related factors in MoBa specifically (52).