Do mothers of dizygotic twins have earlier menopause? A role for fragile X?

Dizygotic (DZ) twinning appears to run in families and the underlying trait, multiple ovulation, appears to be genetically influenced [Lewis et al., 1996; Meule-mans et al., 1996]. Multiple ovulation is considerably more frequent in mothers of DZ twins than in controls [Martin et al., 1991b] although its hormonal control is far from understood [Martin et al., 1991a]. Whatever the cause, if the supply of ovarian follicles is finite, then it might be expected that a woman predisposed to multiple ovulation would more quickly exhaust her supply. If this in turn is a precipitator of menopause [te Velde, 1993], one might expect mothers of DZ twins to have earlier menopause than controls. One predisposing factor to DZ twinning may be het-erozygosity for expansion of the CGG repeat number in the fragile X (FRAXA) locus into the premutation range (50–200 repeats) [Fu et al., 1991]. Previously we have shown that such women have a significantly higher DZ twinning rate than control women and have speculated that they should have earlier menopause [Turner et al., 1994]. Now Conway et al. [1995] have found 2 out of 9 women with premature ovarian failure with FRAXA premutations. This has prompted us to investigate: (i) the frequency of early menopause in mothers of DZ twins (MODZTs), and (ii) whether FRAXA premuta-tions contribute to any tendency to earlier menopause among MODZTs. Health and Lifestyle questionnaires were sent to mothers of 3,220 adult twin pairs enrolled on the Australian NHMRC Twin Registry; 65% responded (ages 38–96, mean 58) and 59% answered " Have you reached menopause ('natural change of life')? " and " How old were you when your periods stopped? " (Table I). They also answered questions about hysterectomy and its temporal relation to natural menopause. While it is difficult to evaluate the validity of self-reported age at menopause, the consistency of this reporting over time appears to be good; for 357 postmenopausal twins who reported age at natural menopause in 1980 and again in 1988, the 8-year test-retest correlation was 0.80 Ϯ 0.03, and in an overlapping sample of 85 women, the 2-year consistency was 0.92 Ϯ 0.04. From Kaplan-Meier survival analysis, the mean age at menopause was 49.9 Ϯ .2 in MODZTs and 50.9 Ϯ .2 in MOMZTs (P ϭ 0.0013). Smoking (pack-years; P Ͻ .0001), education (P ϭ 0.034) and body mass index (P ϭ 0.031) were significant covariates, as found by other [Willett …