Heritability of Age at Natural Menopause in the Framingham Heart Study

Document Sample
Heritability of Age at Natural Menopause in the Framingham Heart Study Powered By Docstoc
					    Heritability of Age at Natural Menopause in the Framingham Heart Study
           Joanne M. Murabito, Qiong Yang, Caroline Fox, Peter W. F. Wilson and L. Adrienne Cupples

J. Clin. Endocrinol. Metab. 2005 90:3427-3430 originally published online Mar 15, 2005; , doi: 10.1210/jc.2005-0181

To subscribe to Journal of Clinical Endocrinology & Metabolism or any of the other journals published by The Endocrine
                            Society please go to:

                      Copyright © The Endocrine Society. All rights reserved. Print ISSN: 0021-972X. Online
0021-972X/05/$15.00/0                                                                     The Journal of Clinical Endocrinology & Metabolism 90(6):3427–3430
Printed in U.S.A.                                                                                                  Copyright © 2005 by The Endocrine Society
                                                                                                                                     doi: 10.1210/jc.2005-0181

Heritability of Age at Natural Menopause in the
Framingham Heart Study
Joanne M. Murabito, Qiong Yang, Caroline Fox, Peter W. F. Wilson, and L. Adrienne Cupples
National Heart, Lung, and Blood Institute’s Framingham Heart Study (J.M.M., Q.Y., C.F., L.A.C.), Framingham,
Massachusetts 01702; Section of General Internal Medicine (J.M.M.), Boston University School of Medicine, Boston,
Massachusetts 02118; Departments of Biostatistics and Neurology (Q.Y., L.A.C.), Boston University Schools of Public Health
and Medicine, Boston, Massachusetts 02118; National Heart, Lung, and Blood Institute (C.F.), Bethesda, Maryland 20892;
Department of Endocrinology, Diabetes, and Hypertension (C.F.), Brigham and Women’s Hospital and Harvard Medical
School, Boston, Massachusetts 02115; and Departments of Endocrinology, Diabetes, and Medical Genetics (P.W.F.W.),
Medical University of South Carolina, Charleston, South Carolina 29425

Background: Twin registries and family history studies provide            Analysis Routines (SOLAR) computer package. Covariates in the
evidence that genetic factors contribute to the onset of menopause, but   multivariable models included generation, number of cigarettes
heritability estimates in population-based samples are limited. We        smoked, body mass index, and parity.
sought to estimate heritability of age at natural menopause in women
participating in the multigenerational Framingham Heart Study, a          Results: The mean age at natural menopause was 49.1 and 49.4 yr
community-based epidemiological study.                                    in original cohort and offspring women, respectively. The multivari-
                                                                          able-adjusted correlation coefficients for mother-daughter, sister-
Methods: A total of 1500 original cohort and 932 offspring cohort         sister, and aunt-niece pairs were 0.21, 0.22, and 0.12, respectively.
women from 1296 extended families reported a natural menopause            The crude and multivariable-adjusted heritability estimates for age
defined as the natural cessation of menses for 1 yr or more. Corre-       at natural menopause were 0.49 (0.37, 0.61) and 0.52 (0.35, 0.69).
lation coefficients were calculated using family correlations in Sta-
tistical Applications for Genetic Epidemiology for mother-daughter,       Conclusions: Our data suggest that at least 50% of the interindi-
sister-sister, and aunt-niece pairs. Heritability was estimated using     vidual variability in menopausal age appears to be attributable to
variance-components methods in the Sequential Oligogenic Linkage          genetic effects. (J Clin Endocrinol Metab 90: 3427–3430, 2005)

M       ENOPAUSAL AGE INFLUENCES risk for death and
         many serious illnesses in women. For each year that
menopause is delayed, cardiovascular disease mortality has
                                                                          (6). In a population-based sample of women from The Neth-
                                                                          erlands participating in a breast cancer screening project,
                                                                          heritability estimates for menopausal age were even higher
been estimated to fall by 2% (1). Menopausal age varies                   (0.71– 0.72 for twin sisters and 0.85– 0.87 for singleton sisters)
widely, between 40 and 60 yr, and appears to be modulated                 (7). Recent work using the same Netherlands population
by both environmental and genetic factors (2–7). Smoking is               sample examining mother-daughter pairs found heritability
one of the most important environmental factors influencing               of age at natural menopause to be 0.44 (13). However, her-
menopausal age with smokers experiencing menopause on                     itability estimates of menopausal age have not been exam-
average 0.8 –2.0 yr earlier than nonsmokers (8 –11). However,             ined in unselected population-based samples in the United
it has been suggested that only a small proportion of the large           States. We hypothesized that age at natural menopause is
variation in age at natural menopause can be explained by                 heritable, and we sought to test this in the population-based
environmental factors (12).                                               sample of women participating in the Framingham Study
   Family history studies and heritability estimates of meno-             cohorts. Furthermore, we sought to examine the effect of
pausal age provide support for genetic factors contributing               important environmental factors on heritability estimates in
significantly to the onset of menopause. Women reporting a                this sample of women.
mother with an early age at menopause had a 6-fold increase
in odds of a premature or early menopause (3, 4). Volunteer                               Subjects and Methods
twin registries have reported substantial heritability esti-
                                                                          Study sample and definitions
mates for age at menopause that range from 0.31– 0.53 in an
Australian sample (5) and 0.63 in a United Kingdom sample                 Study sample. The Framingham Heart Study was established in 1948,
                                                                          when 5209 residents of Framingham, MA, including 2873 women, aged
                                                                          28 – 62 yr, were enrolled in a prospective cohort study. Original cohort
  First Published Online March 15, 2005                                   members have undergone follow-up examinations every 2 yr. In 1971,
  Abbreviation: CI, Confidence interval.                                  5124 offspring of the original cohort members and offspring spouses
JCEM is published monthly by The Endocrine Society (http://www.           were enrolled in the Framingham Offspring Study. Of the offspring, the foremost professional society serving the en-      cohort, 2641 were female ranging in age from 5–70 yr at study enroll-
docrine community.                                                        ment. These participants have undergone follow-up examinations ap-

3428   J Clin Endocrinol Metab, June 2005, 90(6):3427–3430                                              Murabito et al. • Heritability of Menopausal Age

proximately every 4 yr. Study design and entry criteria have been re-        cients for first-degree relative pairs using the equation h2     2r (h2 in-
ported elsewhere (14, 15). The Institutional Review Board at Boston          dicates heritability and r correlation among first-degree relative pairs).
Medical Center approved the examination content for all examinations.        To estimate heritability from second-degree relatives, the correlation
   Of the 2873 original cohort and 2641 offspring women, 618 original        coefficient is quadrupled using the equation h2      4r (h2 indicates her-
cohort and 833 offspring women were excluded because they were not           itability and r correlation among aunt-niece pairs). Next, heritability of
part of a biological family. An additional 755 original cohort and 876       age at natural menopause was calculated using the variance-compo-
offspring women were excluded because none of the women underwent            nents methods implemented in the Sequential Oligogenic Linkage Anal-
a natural menopause. The final study sample is composed of 1296              ysis Routines (SOLAR) computer package (19) because it takes into
families that include 1500 original cohort and 932 offspring women           account all familial relations together. The variance-components model
reporting natural menopause. The study sample includes 622 mother-           assumes that variation in the trait can be partitioned into genetic and
daughter pairs, 474 sister-sister pairs, 29 grandmother-granddaughter        random environmental components. Heritability was estimated as the
pairs, 258 avuncular, and 165 first-cousin pairs. For multivariable anal-    ratio of genetic variance to total phenotypic variance via a maximum
yses, the sample was reduced to 1022 families that include 984 original      likelihood method.
cohort and 680 offspring women after excluding women with missing               Heritability was calculated separately for offspring women and orig-
covariate data.                                                              inal cohort women first, and then calculated for a pooled sample of
                                                                             offspring and original cohort women.
Definition of age at natural menopause
   At each examination, women were queried as to their menopausal
status with the following questions: whether her periods had stopped         Clinical characteristics
for 1 yr or more, age periods stopped, cause periods stopped (natural,          The distribution of menopausal ages in our study sample
surgical, or other), hysterectomy (yes or no), and number of ovaries
removed (0, 1, 2, or unknown). For the purposes of this study, natural       is shown in Fig. 1. The mean age at natural menopause was
menopause occurred after a woman had ceased menstruating naturally           similar in original cohort and offspring women, 49.1 yr
for 1 yr, and the age at natural menopause was the self-reported age at      (range, 29 – 60 yr) and 49.4 yr (range, 25– 61 yr), respectively.
last menstruation. Offspring women who attended more than one ex-            No important differences were noted with respect to body
amination after the onset of natural menopause may have reported
different ages at which periods stopped at each postmenopausal exam-
                                                                             mass index, cigarette smoking, alcohol intake, or parity.
ination attended. We used the age periods stopped at the first report to
minimize recall error if the reported stop ages varied by five or fewer      Correlation coefficients
years (16). If the reported stop ages varied by more than 5 yr, one
investigator (J.M.M.), blinded to family relationship, reviewed each            Correlation coefficients for mother-daughter pairs, sister-
woman’s research chart to ascertain the best report of age at natural        sister pairs, and aunt-niece pairs were calculated using Sta-
menopause.                                                                   tistical Applications for Genetic Epidemiology family corre-
                                                                             lations. For crude and multivariable-adjusted analyses, the
Potential environmental factors related to age at                            correlation coefficients for age at natural menopause were
natural menopause                                                            0.21 and 0.21, respectively, for mother-daughter pairs, 0.32
   At each examination, height and weight were obtained and body             and 0.22, respectively, for sister-sister pairs, and 0.03 and
mass index was defined as the weight in kilograms divided by the height      0.12, respectively, for aunt-niece pairs. Using the multivari-
in meters squared. Women were asked whether they smoked cigarettes           able-adjusted mother-daughter, sister-sister, and aunt-niece
regularly in the year preceding each examination, and if yes, the number     correlation coefficients, the estimate of heritability is similar
of cigarettes smoked per day was recorded. Women were asked to report
the average number of alcoholic drinks consumed per week. Parity was         0.42, 0.44, and 0.48, respectively.
dichotomized as zero (nulliparous) vs. one or more live births (parous).
Offspring women were queried about age at menarche at exam 2 and             Heritability estimates
about oral contraceptive use at every exam.
                                                                                Heritability estimates using variance components for age
Statistical analysis                                                         at natural menopause are shown in Table 1. The crude model
                                                                             resulted in heritability estimates of 0.65 [95% confidence
   Analyses were performed on crude age at natural menopause and             interval (CI), 0.42, 0.89] in original cohort, 0.59 (95% CI, 0.37,
standardized residuals from multiple linear regressions in SAS (SAS/
STAT Software, Version 8.2, 1999, SAS Institute, Cary, NC) that adjusted
age at natural menopause for covariates of interest. Thus, crude and
multivariable-adjusted age at natural menopause were used as pheno-
types in subsequent analyses. Because the crude trait was skewed and
the variance component method assumes normality, we also computed
heritability for the Winsorized crude trait. Winsorization is a widely
accepted methodology for reducing skewness and kurtosis (17, 18). In
this study, we replaced values that were four times the sd above or below
the mean by the mean plus/minus four times the sd. Covariates in the
multivariable models included mean number of cigarettes smoked,
mean body mass index, and mean alcohol intake measured across at-
tended examinations, parity (0 vs. 1 or more live births), oral contra-
ceptive use ever, and age at menarche.
   Heritability was calculated using two methods. First, we examined
familial aggregation of age at natural menopause by calculating corre-
lations for relatives of different types. We used the family correlations
procedure in Statistical Applications for Genetic Epidemiology (SAGE,
release 3.1; Case Western Reserve University, Cleveland, OH) to calcu-
late the intraclass correlations for mother-daughter pairs, sibling pairs,
and aunt-niece pairs using equal weights for each pedigree. A simple
estimate of heritability is obtained by doubling the correlation coeffi-     FIG. 1. Distribution of age at natural menopause in the study sample.
Murabito et al. • Heritability of Menopausal Age                                         J Clin Endocrinol Metab, June 2005, 90(6):3427–3430   3429

TABLE 1. Estimated heritability of age at natural menopause using variance-components methodology in SOLAR: 1296 Framingham
Heart Study families

                                                            n                 H2, mean                     95% CI                   P value
        Original cohort
          Crude                                           1500                  0.65                     0.42, 0.89                 0.0001
          Multivariable-adjusteda                          984                  0.74                     0.31, 1.00                 0.002
        Offspring cohort
          Crude                                            932                  0.59                     0.37, 0.81                 0.0002
          Multivariable-adjustedb                          680                  0.48                     0.15, 0.81                 0.003
        Pooled original and offspring cohorts
          Crude, generation adjusted                      2432                  0.49                     0.37, 0.61                 0.0001
          Multivariable-adjustedc                         1672                  0.52                     0.35, 0.69                 0.0001
      Adjusted for mean body mass index, cigarette smoking, and parity.
      Adjusted for mean body mass index, cigarette smoking, parity, mean alcohol intake, oral contraception use ever, and age at menarche.
      Adjusted for generation (original cohort/offspring), mean body mass index, cigarette smoking, and parity.
0.81) in offspring, and 0.49 (95% CI, 0.37, 0.61) in the pooled         of our findings. Most national survey data have not reported
offspring and cohort sample. The heritability estimate for the          an association between race or ethnicity and menopause (21,
multivariable-adjusted age at natural menopause in the orig-            22), but other studies have found menopause to occur later
inal cohort was 0.74 (95% CI, 0.31, 1.00), for the offspring was        in some racial/ethnic groups (10, 23). Second, age at natural
0.48 (95% CI, 0.15, 0.81), and for the pooled sample of off-            menopause was ascertained by self-report, raising the con-
spring and cohort women was 0.52 (95% CI, 0.35, 0.69).                  cern of recall bias. This is most challenging for women who
Therefore, at least 50% of the interindividual variability in           are menopausal on study entry. We did not remove these
menopausal age appears to be attributable to genetic effects.           women because this would have resulted in removing
                                                                        women with younger ages at menopausal onset. Third, par-
Components of variance analysis                                         ity is higher in the original cohort women in our study
   In offspring, the contribution of genetic factors to overall         sample (95%) than in U.S. population estimates for women
variation in age at natural menopause was 45% (heritability),           born at the same general time period. Because nulliparity is
and the contribution of the covariates accounted for 8% of the          associated with earlier menopause, it is possible that the
total variation, leaving a residual of 47%. In the original             distribution of menopausal ages in our sample is shifted
cohort, the genetic factors accounted for 74% (heritability) of         toward later onset. Fourth, some offspring women remain
the variation in age at natural menopause, whereas the co-              premenopausal. Therefore, our sample may not include as
variates accounted for 4% of the total variation, leaving a             many women at more extreme ages of menopausal onset
residual of 22%.                                                        with a greater likelihood of an underlying genetic mecha-
                                                                        nism. This may result in an underestimation of heritability.
                              Discussion                                Last, family studies are not able to distinguish between ge-
                                                                        netic factors and shared environmental factors within fam-
   We have found that about 50% of the variation in age at
                                                                        ilies. It is possible that lifestyle factors such as diet and
natural menopause can be explained by genetic factors in
                                                                        exercise or other household exposures shared among family
women in Framingham families. Family history studies have
                                                                        members might have influenced menopausal age.
focused on women with early-onset or premature meno-
pause (3, 4), and almost all previous reports of heritability
focused on twin registries, which may overestimate herita-              Conclusions
bility because heritability estimates cannot effectively dis-
                                                                           A substantial proportion of the variability in age at natural
tinguish between genetic factors and the early shared envi-
                                                                        menopause is explained by genetic factors. Additional stud-
ronmental influences that are more similar among twins (20).
                                                                        ies of genetic linkage and candidate gene association are
Our data demonstrate that heritability estimates using cor-
                                                                        warranted to identify the specific genetic variants associated
relation coefficients and variance-components methodology
                                                                        with menopausal onset. Confirmation of the first report of
are comparable, and the magnitude of the heritability for
                                                                        linkage analysis in sibling pairs with extreme concordance or
natural menopausal age (48 –74%) is similar to reports from
                                                                        discordance for age at menopause is needed (24). Determin-
twin registries in Australia, the United Kingdom, and The
                                                                        ing the genetic factors associated with age at menopause may
Netherlands. The similarities between our estimates of her-
                                                                        have important clinical implications because of the potential
itability and those derived from twin registries supports our
                                                                        ability to intervene early in many diseases in postmeno-
finding that the environmental contribution to age at natural
                                                                        pausal women.
menopause is relatively small. Our estimate of heritability of
age at natural menopause for mother-daughter pairs (0.42) is
almost identical to the only other report of mother-daughter                                     Acknowledgments
pairs derived from a Dutch sample (0.44) (13).
                                                                           Received January 27, 2005. Accepted March 4, 2005.
Limitations                                                                Address all correspondence and requests for reprints to: Joanne Mu-
                                                                        rabito, M.D., Sc.M., National Heart, Lung, and Blood Institute’s Fra-
  Our study has several important limitations. First, our               mingham Heart Study, 73 Mount Wayte Avenue, Suite 2, Framingham,
sample is primarily Caucasian, limiting the generalizability            Massachusetts 01702-5827. E-mail:
3430    J Clin Endocrinol Metab, June 2005, 90(6):3427–3430                                                       Murabito et al. • Heritability of Menopausal Age

   National Heart, Lung, and Blood Institute’s Framingham Heart Study                13. van Asselt KM, Kok HS, Pearson PL, Dubas JS, Peeters PH, te Velde ER, van
is supported by contract number N01-HC-25195.                                            Noord PA 2004 Heritability of menopausal age in mothers and daughters.
                                                                                         Fertil Steril 82:1348 –1351
                                                                                     14. Dawber TR, Meadors GF, Moore Jr FE 1951 Epidemiological approaches to
                                References                                               heart disease: the Framingham Study. Am J Public Health 41:279 –281
 1. van der Schouw YT, van der GY, Steyerberg EW, Eijkemans JC, Banga JD             15. Feinleib M, Kannel WB, Garrison RJ, McNamara PM, Castelli WP 1975 The
    1996 Age at menopause as a risk factor for cardiovascular mortality. Lancet          Framingham Offspring Study. Design and preliminary data. Prev Med 4:518 –
    347:714 –718                                                                         525
 2. van Noord PA, Dubas JS, Dorland M, Boersma H, te Velde ER 1997 Age at            16. Colditz GA, Stampfer MJ, Willett WC, Stason WB, Rosner B, Hennekens CH,
    natural menopause in a population-based screening cohort: the role of men-           Speizer FE 1987 Reproducibility and validity of self-reported menopausal
    arche, fecundity, and lifestyle factors. Fertil Steril 68:95–102                     status in a prospective cohort study. Am J Epidemiol 126:319 –325
 3. Cramer DW, Xu H, Harlow BL 1995 Family history as a predictor of early           17. Fernandez JR, Etzel C, Beasley TM, Shete S, Amos CI, Allison DB 2002
    menopause. Fertil Steril 64:740 –745                                                 Improving the power of sib pair quantitative trait loci detection by phenotype
 4. Togerson DJ, Thomas RE, Reed DM 1997 Mothers and daughters menopausal                winsorization. Hum Hered 53:59 – 67
    ages: is there a link? Eur J Obstet Gynecol Reprod Biol 74:63– 67                18. Shete S, Beasley TM, Etzel CJ, Fernandez JR, Chen J, Allison DB, Amos CI
 5. Treloar SA, Do KA, Martin NG 1998 Genetic influences on the age at meno-             2004 Effect of winsorization on power and type 1 error of variance components
    pause. Lancet 352:1084 –1085                                                         and related methods of QTL detection. Behav Genet 34:153–159
 6. Snieder H, MacGregor AJ, Spector TD 1998 Genes control the cessation of a        19. Almasy L, Blangero J 1998 Multipoint quantitative-trait linkage analysis in
    woman’s reproductive life: a twin study of hysterectomy and age at meno-             general pedigrees. Am J Hum Genet 62:1198 –1211
    pause. J Clin Endocrinol Metab 83:1875–1880                                      20. Austin MA, King MC, Bawol RD, Hulley SB, Friedman GD 1987 Risk factors
 7. de Bruin JP, Bovenhuis H, van Noord PA, Pearson PL, van Arendonk JA, te              for coronary heart disease in adult female twins. Genetic heritability and
    Velde ER, Kuurman WW, Dorland M 2001 The role of genetic factors in age              shared environmental influences. Am J Epidemiol 125:308 –318
    at natural menopause. Hum Reprod 16:2014 –2018                                   21. Cooper GS, Baird DD, Darden FR 2001 Measures of menopausal status in
 8. Rodstrom K, Bengtsson C, Milsom I, Lissner L, Sundh V, Bjourkelund C 2003            relation to demographic, reproductive, and behavioral characteristics in a
    Evidence for a secular trend in menopausal age: a population study of women          population-based study of women aged 35– 49 years. Am J Epidemiol 153:
    in Gothenburg. Menopause 10:538 –543                                                 1159 –1165
 9. Midgette AS, Baron JA 1990 Cigarette smoking and the risk of natural meno-       22. Brett KM, Cooper GS 2003 Associations with menopause and menopausal
    pause. Epidemiology 1:474 – 480                                                      transition in a nationally representative US sample. Maturitas 45:89 –97
10. Kato I, Toniolo P, Akhmedkhanov A, Koenig KL, Shore R, Zeleniuch-                23. Gold EB, Bromberger J, Crawford S, Samuels S, Greendale GA, Harlow SD,
    Jacquotte A 1998 Prospective study of factors influencing the onset of natural       Shurnick J 2001 Factors associated with age at natural menopause in a mul-
    menopause. J Clin Epidemiol 51:1271–1276                                             tiethnic sample of midlife women. Am J Epidemiol 153:865– 874
11. Hardy R, Kuh D, Wadsworth M 2000 Smoking, body mass index, socioeco-             24. van Asselt KM, Kok HS, Putter H, Wijmenga C, Peeters PH, van der Schouw
    nomic status and the menopausal transition in a British national cohort. Int J       YT, Grobbee DE, te Velde ER, Mosselman S 2004 Linkage analysis of ex-
    Epidemiol 29:845– 851                                                                tremely discordant and concordant sibling pairs identifies quantitative trait
12. van Noord PA, Peeters PH, Grobbee DE, Dubas JS, te Velde ER 1999 Onset               loci influencing variation in human menopausal age. Am J Hum Genet
    of natural menopause. J Clin Epidemiol 52:1290 –1292                                 74:444 – 453

       JCEM is published monthly by The Endocrine Society (, the foremost professional society serving the
                                                         endocrine community.

Shared By:
hkksew3563rd hkksew3563rd http://