Abstract
Fragile X-associated primary ovarian insufficiency (FXPOI) occurs in about 20% of women who carry a premutation allele (55-200 CGG repeats). These women develop hypergonadotropic hypogonadism and have secondary amenorrhea before age 40. A non-linear association with repeat size and risk for FXPOI has been seen in multiple studies women with a premutation: those with a mid-range of repeats are at highest risk (∼70-100 CGG repeats). Importantly, not all carriers with 70-100 repeats experience FXPOI. We investigated whether AGG interruptions, adjusted for repeat size, impacted age at secondary amenorrhea. We have reproductive history information and AGG interruption data on 262 premutation women: 164 had an established age at amenorrhea (AAA) (for some, age at onset of FXPOI) or menopause, 16 had a surgery involving the reproductive system such as a hysterectomy, and 82 women were still cycling at the last interview. Reproductive status was determined using self-report reproductive questionnaires and interviews with a reproductive endocrinologist. For each of these 262 women, FMR1 repeat size and number of AGG interruptions were determined. We confirmed the association of repeat size with AAA or menopause among women with a premutation. As expected, both premutation repeat size and the quadratic form of repeat size (i.e., squared term) were significant in a survival analysis model predicting AAA (p < 0.0001 for both variables). When number of AGG interruptions was added to the model, this variable was not significant (p = 0.59). Finally, we used a regression model based on the 164 women with established AAA to estimate the proportion of variance in AAA explained by repeat size and its squared term. Both terms were again highly significant (p < 0.0001 for both), but together only explained 13% of the variation in AAA. The non-linear association between AAA and FMR1 repeat size has been described in several studies. We have determined that AGG interruption pattern does not contribute to this association. Because only 13% of the variation is described using repeat size, it is clear that further research of FXPOI is needed to identify other factors that affect the risk for FXPOI.