Original Reports
Age at Natural Menopause and Mortality

https://doi.org/10.1016/S1047-2797(97)00207-XGet rights and content

Abstract

PURPOSE: The purpose of this study was to examine the association between age at menopause and mortality in a population-based sample of women in the United States.

METHODS: This study was based on data from the National Health and Examination Survey (NHANES) Epidemiologic Follow-up Study; 3191 women aged 50–86 years were included. There were 345 deaths over a mean follow-up time of 4.0 years. We used age-stratified and Poisson regression procedures to assess mortality risk by age at natural menopause, with adjustment for age, duration of follow-up, race, education, smoking, and use of hormone replacement therapy. We conducted a separate analysis for surgical menopause with bilateral oophorectomy.

RESULTS: Compared with women who were menstruating to age 50 or later, the adjusted mortality rate ratios (RR) were 1.50 (95% confidence interval (CI), 0.97–2.34) for women with a natural menopause at age < 40, 1.04 (95% CI, 0.72–1.51) for those with menopause at age 40–44, and 0.96 (95% CI, 0.72–1.26) for those with menopause at age 45–49. Women with a natural menopause at age 40–44 years experienced an increased risk of cancer-related mortality (adjusted RR 2.34, 95% CI, 1.20–4.58). No age-related increased mortality risk was seen among women who had surgical menopause with bilateral oophorectomy.

CONCLUSIONS: This study provides some support for the concept that age at natural menopause serves as a biological marker of health and aging, with potential implications extending beyond cardiovascular diseases.

Introduction

Age at natural menopause has been proposed as a marker for factors that may directly or indirectly damage the follicular pool [1]and as a marker of aging and health [2]. The frequency of ovarian failure before age 40 is ⩽ 1% 3, 4, 5, but by age 45, about 5–10% of women have experienced natural menopause 4, 5, 6. Thus an increased risk of mortality associated with early menopause, if confirmed, could represent an important factor affecting risk of disease and mortality among older women. We used data from the National Health and Examination Survey Epidemiologic Follow-up Study to examine the association between age at natural menopause and mortality. Our analysis included 3191 women aged 50–86 years at baseline.

The first National Health and Examination Survey (NHANES I) was a national population-based study of 23,808 persons aged 1–74 years; the survey was conducted between 1971 and 1975 by the National Center for Health Statistics. This survey evaluated nutritional and health status using an extensive questionnaire and clinical and laboratory assessments 7, 8. NHANES I included 8596 noninstitutionalized civilian women 25–74 years of age. Additional assessments of this cohort were carried out in 1982–1984 (subjects who were 25–74 years old during NHANES I), 1986 (subjects who were initially 55 to 74 years old), and 1987 (subjects who were initially 25–74 years old). Thus there were three follow-up assessments of subjects aged 55–74 years at baseline and two follow-up assessments for subjects aged 25–54 years at baseline. These additional follow-up studies are referred to as the NHANES Epidemiologic Follow-up Study (NHEFS). Of the age-eligible NHANES I subjects, 93% were traced and included in the 1982–1984 NHEFS [9]. During the 1986 NHEFS, 95% of the eligible participants were traced, 96% of those traced were interviewed, and death certificates were obtained for 97% of those who had died [10]. Corresponding figures for the 1987 NHEFS were 94% traced, 91% interviewed, and 94% of the death certificates obtained [11].

The National Death Index and the Health Care Financing Administration records were used to trace subjects who were not initially located at the time of each NHEFS survey. Death certificates were sought from state vital records offices for all subjects identified as deceased from these sources or from a proxy interview. The mortality data for this analysis are based on information from death certificates. Death certificate coding was based on the International Classification of Diseases, Ninth Revision (ICD-9), and was completed by NCHS. We used the underlying cause of death code for this analysis. We included ICD-9 codes 410–414 in the coronary heart disease category, 430–438 as cerebrovascular deaths, 140–239 as cancer deaths, and the remaining codes as all other causes of death.

The NHANES I interview asked if menstrual periods had entirely stopped, and if so, at what age; information on type of menopause was not obtained. The 1982–1984 NHEFS assessment included more detailed menopause-related questions in a self-administered booklet. Women were asked if they were still having periods, if periods were regular or irregular, and if irregular, whether this was because she was going through the change of life or for some other reason. Women who indicated they were not still having periods were asked the age at last period and whether periods stopped naturally, because of surgery, or for some other reason, and whether both ovaries were still present. Women who had stopped having periods or who indicated periods were irregular because of change of life were asked about use of female hormones for reasons related to menopause. We used the woman's age at her last menstrual period as the age at menopause.

The questions on menopause were not included in the proxy questionnaire; therefore, information on type of menopause was not available for anyone who had died between NHANES I and the 1982–1984 NHEFS assessment. Since we thought that this limitation could bias the results of an analysis of mortality in relation to age at natural menopause between NHANES I (1971–1975) and the 1982–1984 NHEFS, our analysis began with the 1982–1984 NHEFS assessment as our baseline period. We limited the analysis to female self-respondents who were at least 50 years old at this assessment (n = 3926). Other exclusions were 606 women who had a surgical menopause before age 50 and who retained at least one ovary, 39 who indicated “don't know” to type of menopause, 61 who indicated the type of menopause was other than natural or surgical, 20 who had a natural menopause but were missing information on age at menopause in both NHANES I and NHEFS, five women with inconsistent answers about menopausal status (i.e., those who reported being postmenopausal in 1971–1975 NHANES I and premenopausal in 1982–1984 NHEFS), three women with inconsistent answers about current age and age at menopause, and one woman who never menstruated. These exclusions left 629 women with a surgical menopause involving bilateral oophorectomy before age 50, 1087 women with a natural menopause before age 50, and 1475 women who reported the presence of menstrual cycles at least to age 50 and who had a natural menopause at age ⩾ 50 (n = 1303) or who were still menstruating (n = 172). This group of 1475 constitutes the referent group for our analysis.

We conducted separate analyses for the mortality risk associated with “early” (at age < 50) natural menopause or bilateral oophorectomy. We calculated age-specific and age-adjusted mortality risk ratios by age at natural menopause group using Mantel-Haenszel procedures for stratified analyses. We also modeled the observed number of deaths in relation to person-time at risk and other covariates using a Poisson regression procedure [12]. The exponentiated coefficients calculated in the regression analysis can be interpreted as estimates of the rate ratios (RR). Person-time at risk was counted from age at baseline (1982–1984 interview) to age at death, last interview, or loss to follow-up. The mean length of follow-up was 4.0 years. We also used this Poisson regression analysis to examine mortality in relation to age at surgical menopause. Information about race, education level, current weight, use of hormone replacement therapy (HRT), pregnancy history, and smoking history was collected at the 1982–1984 interview. Height was recorded during the 1971–1975 NHANES I assessment. We examined these factors as possible confounding variables in our regression analyses, with body mass index calculated as kg/m2. We used the NHANES I data on age at menopause when it was available (n = 2172), but for 714 women with natural menopause and 133 women with bilateral oophorectomy, data on age at menopause was only available from the 1982–1984 NHEFS questionnaire. For the 1548 women with age at natural menopause information from both NHANES I and NHEFS, agreement between the two data sources was 43.3% within 1 year, 57.4% within 2 years, 66.9% within 3 years, and 81.1% within 5 years.

Section snippets

Results

The total number of deaths in this cohort of women during the follow-up period was 345 (10.8%), for an overall death rate of 27 per 1000 person-years. The mean age at natural menopause was 48.8 (sd, 5.08) years; 115, 268, and 704 women reported a natural menopause before ages 40, 40–44, and 45–49, respectively. There was no clear secular trend in age at menopause, with mean ages at natural menopause of 48.6, 48.9, and 48.7 years for women who were 50–59, 60–69, and ⩾ 70 years old at baseline.

Discussion

We observed a small increased mortality rate among women who experienced natural menopause before age 40, with adjusted rate ratios of 1.50 (95% CI, 0.97–2.34) compared with women who were menstruating at age 50. However, there was little indication of an association between total mortality and natural menopause at age 40–44 or 45–49 years, and there was no association between mortality and bilateral oophorectomy before age 40. In a 6-year longitudinal study of 5287 Seventh Day Adventist women

Acknowledgements

We thank Rebecca Darden of Westat for her assistance in data analysis, Joan Cornoni-Huntley for advice concerning the NHEFS study design, and Allen Wilcox and Mathew Longnecker for their reviews of this manuscript.

References (21)

There are more references available in the full text version of this article.

Cited by (0)

View full text