Increased frequency of occult fragile X-associated primary ovarian insufficiency in infertile women with evidence of impaired ovarian function.

BACKGROUND The FMR1 premutation is associated with overt primary ovarian insufficiency (POI). However, its prevalence in women with occult POI (i.e. menstrual cycles, but impaired ovarian response) has not been examined. We hypothesized that both the FMR1 premutation and intermediate allele is more frequent in infertile women with occult POI than in controls, and that a repeat length cutoff might predict occult POI. METHODS All subjects were menstruating women <42 years old and with no family history of unexplained mental retardation, autism or fragile X syndrome. Cases had occult POI defined by elevated FSH or poor response to gonadotrophin therapy (n = 535). Control subjects (n = 521) had infertility from other causes or were oocyte donors. Prevalence of the FMR1 premutation and intermediate alleles was examined and allele length was compared between controls and women with occult POI. RESULTS The frequency of the premutation (7/535 versus 1/521; P< 0.05) and intermediate alleles (17/535 versus 7/521; P< 0.05) was higher in women with occult POI than in controls. The allele with the greatest number of CGG repeats was longer in women with occult POI compared with controls (32.7 ± 7.1 versus 31.6 ± 4.3; P < 0.01). A receiver operating characteristic curve examining repeat length as a test for occult POI had an area of 0.56 ± 0.02 (P < 0.01). A repeat cutoff of 45 had a specificity of 98%, but a sensitivity of only 5% to identify occult POI. The positive predictive value was only 21% for a fertility population that has ∼ 22% of its patients with occult POI. CONCLUSIONS The data suggest that FMR1 premutations and intermediate alleles are increased in women with occult POI. Thus, FMR1 testing should be performed in these women as some will have fragileX-associated POI. Although the FMR1 repeat lengths were longer in women with occult POI, the data do not support the use of a repeat length cutoff to predict occult POI.