No Evidence for Parent of Origin Influencing Premature Ovarian Failure in Fragile X Premutation Carriers

To the Editor:

We were interested to read, in the February edition of the Journal, an article by Hundscheid et al. (2000) that reported an imprinting effect on the fragile X premutation, such that paternally inherited premutations are more likely to give rise to premature ovarian failure (POF). We were impressed by the rigorous design of the study, which ensured that all women were interviewed personally and that adherence to strict criteria for the definition of POF was maintained. However, we were very surprised by the results of the study, in light of the results of our own study of a similar cohort of women from Wessex, United Kingdom.

We interviewed 352 women from families with the fragile X premutation about their reproductive and menstrual histories; of these women, 116 carried premutation-sized (51–200-repeat) alleles and were from 62 families (Murray et al., in press). These families have been investigated extensively during the past 15 years, and, therefore, we have almost complete ascertainment of “at-risk” individuals. All premutation and full-mutation carriers and their unaffected first- and second-degree relatives were invited to participate in the study and were interviewed either in person or via the telephone. Women who were using the contraceptive pill were excluded from the analysis, and those women who were still menstruating or who had undergone a hysterectomy were taken as censored values. We used Kaplan-Meier survival analysis to demonstrate that the premutation group had a mean age at menopause of 47.87 years, compared with 52.96 years for the full-mutation and normal groups combined (Murray et al., in press). It was apparent, from inspection of our data, that there were no parent-of-origin differences. In a separate study of women ascertained through the presence of POF, six pedigrees were illustrated (Murray et al. 1998); we could determine the origin of the premutation in only four women with POF (who were from two families), and the origin was maternal in each case.

However, the study by Hundscheid et al. (2000) prompted us to reevaluate our data and to analyze them in a similar fashion. Our population of 116 premutation carriers was comprised of 40 carriers with maternal transmissions, 51 carriers with paternal transmissions, and 25 carriers for whom we were not able to determine the origin of transmission. The results of survival analysis comparing maternal and paternal premutations demonstrated no significant shift in age at menopause between the two groups (χ²₁=0.0143; P=.91). POF was defined as spontaneous cessation of menses for >1 year, before age 40 years—a definition that is essentially the same as that in the study by Hundscheid et al. Similar to table 1 in the study by Hundscheid et al., table 1 in our study shows that we have only considered females that were of age ⩾40 years at the time of the interview; POF is not significantly more common in either group (two-tailed Fisher’s exact test; P=.669).

Table 1.

Origin of Premutation in Women of Age ⩾40 Years at the Time of the Interview

	No. of Premutationsof Origin
Classification	PIP	MIP	Total
Age at Menopause:
<40 years	2	5	7
⩾40 years	6	10	16
Proved ovarian functiona	4	3	7
Total	12	18	30

^aNot menopausal, with follicular-phase FSH level <40 U/liter.

The results of analyses of 116 female premutation carriers from families with fragile X do not provide any evidence with which to support the suggestion that there is imprinting of the FMR1 gene. We can provide no explanation for the discrepancy between our data and the material presented in the report by Hundscheid et al. (2000). In both studies, survival analysis was used to estimate the distribution of age at menopause in an uncensored cohort, since any method that excludes premenopausal subjects underestimates the mean. Survival analysis extracts full and unbiased information from all relationships to probands, which have different frequencies of paternally inherited fragile X premutations (PIP) and maternally inherited fragile X premutations (MIP). Daughters of normal transmitting males all have PIP, sisters of probands never have PIP, mothers of probands have a ratio of nearly 3:1, and so forth (Morton and Macpherson 1992). Neither the present study nor the study by Hundscheid et al. categorizes relationship, which presumably accounts for the observed difference in frequencies; however, this is irrelevant if survival analysis is used correctly. In both studies, all subjects were interviewed and hearsay evidence was rejected. In an unspecified proportion of cases, Hundscheid et al. obtained age at menopause from medical records, whereas we accepted the subject’s recall. Our definitions of POF, spontaneous menopause, unnatural menopause, menstrual history, and medication are indistinguishable from those of Hundscheid et al. We based our classification of MIP and PIP on several microsatellites in the FRAXA region, classifying 25 cases as being of unknown origin, according to conservative criteria. Hundscheid et al. did not specify whether markers were tested, how their classification was made, or how many subjects were unclassifiable. Regardless of whether this is consequential, the fact remains that we observed a significant difference between women with MIP and control individuals (log-rank χ²₁=8.52; P=.0035), whereas Hundscheid et al. did not. We are unable to explain this difference.

It would be very interesting to know whether other investigators find parent-of-origin differences in the frequency of POF in premutation carriers. As in all recent studies, the protocol should include interviews of all available female relatives, with rigorous definition of menopausal variables and mode of origin and with correct use of survival analysis. Only then will studies by different groups pass from debate to discovery.