Menopausal Characteristics and Physical Functioning in Older Adulthood in the NHANES III

Abstract

Objective

We hypothesized that natural menopause would be related to better physical functioning compared to surgical menopause and that later age at menopause would be related to better physical functioning.

Methods

Our sample comprised 1765 women aged ≥ 60 years who participated in the National Health and Nutrition Examination Survey III, a cross-sectional study representative of the United States population. Women recalled age at final menstrual period and age at removal of the uterus and ovaries and reported age, race and ethnicity, height, weight, educational attainment, smoking status, number of children, and use of estrogen therapy. Respondents completed a walk trial and chair rises and reported functional limitations.

Results

Women with a surgical menopause had chair rise times that were an average of 4.4% slower than those of women with natural menopause (95% CI 0.56, 8.27). Women with natural menopause at age ≥ 55 years had an average walking speed 0.05 meters/second (95% CI 0.01, 0.10) faster than women with natural menopause at age < 45 years. Later ages at natural and surgical menopause were also related to lower self-reported functional limitation. Women with surgical menopause at age ≥ 55 years had odds of functional limitation 0.52 times (95% CI 0.29, 0.95) those of women with surgical menopause at age < 40 years, with similar patterns for natural menopause.

Conclusions

Women with surgical menopause and earlier age at menopause had worse physical function in older adulthood. These groups of women may benefit from interventions to prevent functional decline.

INTRODUCTION

Maintaining good physical functioning with age is a vital component of independence in later life, as poor physical functioning is associated with institutionalization, hospitalization, and mortality.^1–3 Identifying characteristics associated with poor physical functioning could contribute to prevention and management strategies that help older people to maintain their independence and also therefore their quality of life. Women consistently have lower levels of physical functioning than men in adulthood,^4–6 with evidence that physical functioning begins to decline at a faster rate among women than men from midlife onwards.^6–8 The timing of the onset of more rapid decline in functioning among women coincides with the transition to menopause, during which time endogenous hormone production decreases. Changes in levels of hormones such as estrogen and progesterone may influence the decline in physical functioning, as these hormones are beneficial to muscle performance.^8–10 Women who have menopause later may therefore have better subsequent physical functioning than women who have earlier menopause because of their longer period of exposure to endogenous hormones. Additional factors associated with surgical menopause may also influence physical functioning. For instance, the events and conditions leading to surgical menopause, physical recovery from the surgery itself, or the abrupt¹¹ or premature alterations to hormone levels could result in lower physical functioning levels among women with surgical menopause compared to women who have undergone natural menopause.

A limited number of studies have examined the relationship of menopausal status with physical functioning in populations of women at different stages of the menopausal transition and have produced inconsistent findings. For instance some studies have found similar levels of physical functioning, assessed using physical performance tests or by self-report, among women with natural and surgical menopause,^{12, 13} whereas another study of self-perceived physical functioning found that women with surgical menopause experienced faster rates of decline in functioning over five years than women with natural menopause.¹¹ When comparing women at different stages of the menopausal transition, women who are naturally postmenopausal or surgically postmenopausal have often been found to have lower levels of physical functioning^13–15 and faster rates of decline over five years¹¹ than perimenopausal and premenopausal women. Perimenopausal and postmenopausal women have also been reported to experience faster rates of decline in self-reported physical function than premenopausal women over 2–4 years of follow-up.¹⁶ These studies include populations of women covering a wide range of ages. Given that premenopausal women are likely to be younger than postmenopausal women, differences in chronological age may therefore at least partially explain the associations found between menopausal status and physical functioning. In a birth cohort study of women born in the same week who completed physical performance tests at the same age, naturally menopausal women had similar physical functioning levels as premenopausal and perimenopausal women,¹² supporting this idea that variations in study member’s ages may have explained findings in other studies.

To our knowledge no studies have investigated whether the differences in physical functioning by menopausal status found in previous studies persist once all women are postmenopausal. Once all women within a population are postmenopausal, the potential confounding effects of age are not as great as they are in studies of women during the menopausal transition, when menopausal stage reflects age. Further, by examining postmenopausal women only it is possible to clarify whether type and timing of menopause are related to physical functioning later in the aging process and whether changes to physical functioning during menopause are transient or persistent. We examined the relationships of type and timing of menopause with physical functioning in a nationally representative sample of American women who were postmenopausal and tested two hypotheses: 1) women who undergo natural menopause would perform better in physical performance tests and report lower levels of functional limitation than women with surgical menopause and 2) women who experience menopause at later ages would perform better in physical performance tests and report lower levels of functional limitation than women who experience menopause at earlier ages.

METHODS

Participants

The National Health and Nutrition Examination Survey III (NHANES III), is a cross-sectional, cluster sample of the civilian, non-institutionalized U.S. population aged ≥ 1 year that took place between 1988 and 1994.^{17, 18} The in-person evaluation included a home interview and a physical examination. A total of 31,311 individuals participated in the NHANES III. This analysis utilizes the Public Use Data File.¹⁹

Physical function

Measures of physical function included two timed physical performance tests and self-reported functional limitations for respondents aged 60 years and over. The physical performance examination took place in a mobile examination center or in the home if respondents were too disabled or otherwise unable to attend the examination center.²⁰ Each respondent completed two trials of an 8-foot walk at her usual walking pace and five chair rises from an armless chair. Technicians recorded the time in seconds to complete each task. Respondents could use assistive devices for the 8-foot walk trials but not another person’s assistance.²⁰ These analyses utilized the faster of the two 8-foot walk times to calculate speed in meters/second (m/s). Because the distributions of chair rise times were skewed, we used a natural log transformation in analyses. Respondents reported level of difficulty in walking for a quarter of a mile; walking up 10 steps; stooping, crouching or kneeling; lifting or carrying something as heavy as 10 pounds; and standing up from an armless straight chair.^{21, 22} As in previous analyses, we considered women who reported at least some difficulty in at least three of the five tasks to have self-reported functional limitation, coded as a binary variable.²³

Menopause characteristics

We based characteristics of menopause on recalled age at final menstrual period and, when applicable, age at removal of the uterus and ovaries. Women with natural menopause did not have a hysterectomy or bilateral oophorectomy prior to their final menstrual period. This group also included women who responded that they did not have hysterectomy prior to their final menstrual period but had missing information on oophorectomy or did not have bilateral oophorectomy prior to their final menstrual period but had missing information on hysterectomy. Exclusion of this subset of women with missing information in sensitivity analyses did not alter results. Women whose periods stopped because of hysterectomy and/or bilateral oophorectomy comprised the group with surgical menopause. Sensitivity analyses that separately analyzed women with hysterectomy who retained both ovaries or had a unilateral oophorectomy and women with bilateral oophorectomy with or without hysterectomy produced similar results. Therefore, we combined these two groups into one. We used age at final menstrual period for age at menopause. A total of 175 women reported only a 5 -year age category for final menstrual period. In this case, we used the mid-point of the age category, except for the category of age ≥ 55 years (n = 10), for which we used age 55 years.

Covariates

We also considered the following potential confounding variables: age at interview, race/ethnicity, height, weight, educational attainment, smoking status, number of children, and use of estrogen therapy. The race/ethnicity variable consisted of non-Hispanic white, non-Hispanic Black, Hispanic, and other. Technicians measured the respondent’s height in centimeters and weight in kilograms. Educational attainment categories were ≤ 8, 9 – 11 years, 12 years, and ≥ 13 years. Each respondent reported her smoking status as never, former, and current. The categories for number of children were no births, 1, 2, 3, and 4 or more. Women reported never, past, or current use of estrogen therapy.

A total of 2968 women aged 60 – 90 years participated in the NHANES III. Of these women, 2397 women provided relevant information about self-reported physical functioning and completed the chair rise task and/or 8 foot walk task, and of these, 1765 women provided information on type of and timing of menopause and covariates.

Statistical Analysis

First we examined the associations between each of the potential confounding factors and type of menopause, using t-tests and chi-squared tests as appropriate. We then evaluated the multivariate associations of type of menopause and age at menopause with the physical performance tests using linear regression. We report the regression coefficients for chair rise time multiplied by 100, as the interpretation of this product (i.e. 100*log_ex) is the percentage difference in time (“sympercent”).²⁴ We then tested the multivariate associations of type and timing of menopause with self-reported functional limitation using logistic regression. When analyzing timing of menopause, we examined age at natural menopause and age at surgical menopause in separate models, as the distributions for age at surgical and natural menopause differ. We conducted separate analyses by type of menopause to clarify the relationships between age and type of menopause with physical functioning. For each menopausal characteristic and outcome pair, we present a set of models that first adjust for age and then also race/ethnicity, height, weight, educational attainment, smoking status, number of children, and use of estrogen therapy. All analyses were weighted and accounted for the complex survey design of the NHANES and survey non-response. We performed analyses in Stata Version 10.1 SE (StataCorp LP, College Station, TX).

RESULTS

Around one-third of the sample had undergone surgical menopause (Table 1). The mean age at natural menopause was 49 years, and at surgical menopause was 42 years. Women with natural menopause were more likely to be older, of Hispanic ethnicity, shorter and lighter, nonsmokers, and to be less educated than women with surgical menopause. Women with natural menopause also had more children (≥ 4) and were less likely to currently use estrogen therapy than women with surgical menopause.

Table 1.

Description of Analysis Sample by Menopause Type

Variable		Natural Menopause %b (SE)	Surgical Menopause%b (SE)	p valuec
na		1157	608
Covariates
Mean age (y)		70.6 (0.4)	68.7 (0.4)	< 0.001
Mean age (y) at menopause		48.8 (0.2)	42.0 (0.4)	< 0.001
Race/ethnicity
Non-Hispanic white		83.5 (0.02)	85.7 (0.02)	0.28
Non-Hispanic black		8.6 (0.01)	9.1 (0.01)
Hispanic		2.4 (0.003)	2.0 (0.003)
Other		5.5 (0.01)	3.2 (0.01)
Mean height in cm		158.4 (0.3)	160.3 (0.3)	< 0.001
Mean weight in kg		67.3 (0.6)	71.3 (0.8)	< 0.001
Education (years)
≤ 8		23.0 (0.02)	19.3 (0.02)	0.19
9 – 11		15.9 (0.02)	20.2 (0.02)
12		36.2 (0.02)	37.7 (0.02)
≥ 13		24.9 (0.02)	22.9 (0.03)
Smoking				0.46
Never		58.0 (0.02)	54.3 (0.03)
Past		27.3 (0.02)	29.8 (0.03)
Present		14.8 (0.02)	15.9 (0.02)
Number of Children				< 0.001
0		1.3 (0.002)	6.0 (0.01)
1		16.0 (0.02)	12.3 (0.01)
2		28.5 (0.02)	28.0 (0.02)
3		18.3 (0.02)	20.4 (0.02)
≥ 4		36.0 (0.02)	33.4 (0.02)
Use of Estrogen Therapy				< 0.001
Never		78.4 (0.02)	43.9 (0.03)
Past		18.7 (0.02)	31.7 (0.02)
Present		2.9 (0.007)	24.4 (0.02)
Outcomes
Mean walking speed in meters/second (SD)d		0.8 (0.2)	0.8 (0.2)	0.71
Mean chair rise time in seconds (SD)e		13.1 (1.4)	13.4 (1.3)	0.17
Self-reported functional limitation		27.1 (0.02)	28.6 (0.02)	0.63

^aFull sample of those with information on at least one physical performance measure and self-reported functional limitation. Outcome-specific samples are slightly smaller as denoted.

^bunless otherwise indicated

^cfrom a chi-squared test for categorical variables; t test for age, age at menopause, height, and weight, and rank sum test for chair rise and walk time

^dn natural = 1157; n surgical = 579

^en natural = 1157; n surgical = 606

In age-adjusted analyses, women with surgical menopause had an average chair rise time that was 4.3% slower than that of women with natural menopause. Type of menopause was not related to walking speed or self-reported functional limitations (Table 2). Adjustment for further confounding variables did not alter the results.

Table 2.

Age-Adjusted and Fully-Adjusted Associations between Menopause type and Physical Functioning Outcomes

		n	Age-adjusted	Fully-adjusteda
			Mean Differenceb (95% CI)	Mean Differenceb (95% CI)
Walking speed (m/s)
Natural menopause		1148	0	0
Surgical menopause		606	−0.01 (−0.04, 0.01)	−0.02 (−0.04, 0.01)
			% differencec (95% CI)	% differencec (95% CI)
Chair rise time
Natural menopause		1106	0	0
Surgical menopause		579	4.32 (0.89, 7.76)	4.42 (0.56, 8.27)
			Odds Ratio of Limitation (95% CI)	Odds Ratio of Limitation (95% CI)
Self- reported functional limitation
Natural menopause		1157	1	1
Surgical menopause		608	1.21 (0.86, 1.70)	1.23 (0.86, 1.77)

^aAdjusted for age, race/ethnicity, height, weight, smoking, education, number of children, and use of estrogen therapy.

^bA positive coefficient indicates poorer performance, compared to the reference group.

^cCalculated by multiplying the regression coefficients by 100 (i.e. 100*log_ex), which is the percentage difference in time (“sympercent”).²⁴ A negative coefficient indicates poorer performance, compared to the reference group.

Later age at natural menopause was related to faster walking speed and lower odds of self-reported functional limitation than earlier age at natural menopause in age- adjusted models (Table 3). For example, women with natural menopause at age 50 – 54 years had a walking speed that was 0.08 m/s faster and odds of functional limitation that was 0.55 times lower than women with natural menopause at age < 45 years, adjusting for age. Results for walking speed and functional limitation were similar in the fully-adjusted models.

Table 3.

Associations of Age at Natural Menopause with Physical Functioning Outcomes

		n	Age-adjusted	Fully-adjusteda	p –valueb
Age at menopause (y)			Mean Differencec (95% CI)	Mean Differencec (95% CI)
Walking speed (m/s)					<0.01
< 45		262	0	0
45 – 49		307	0.05 (−0.01, 0.1)	0.04 (−0.01, 0.09)
50 – 54		413	0.08 (0.03, 0.1)	0.06 (0.02, 0.09)
≥ 55		166	0.08 (0.03, 0.1)	0.05 (0.01, 0.10)
			% differenced (95% CI)	% differenced (95% CI)
Chair rise time
< 45		247	0	0	0.79
45 – 49		293	−3.85 (−12.31, 4.61)	−2.98 (−10.11, 4.14)
50 – 54		403	−4.81 (−10.80, 1.18)	−1.90 (−7.73, 3.93)
≥ 55		163	−4.98 (−12.59, 2.63)	−1.46 (−9.14, 6.20)
			Odds Ratio of Limitation (95% CI)	Odds Ratio of Limitation (95% CI)
Self- reported functional limitation
					0.06
< 45		264	1	1
45 – 49		310	0.62 (0.35, 1.13)	0.66 (0.35, 1.24)
50 – 54		415	0.55 (0.38, 0.81)	0.61 (0.40, 0.95)
≥ 55		168	0.51 (0.32, 0.79)	0.60 (0.35, 1.03)

^aAdjusted for age, race/ethnicity, height, weight, education, number of children, and use of estrogen therapy.

^bFrom a test of trend in the fully-adjusted model.

^cA positive coefficient indicates poorer performance, compared to the reference group.

^dCalculated by multiplying the regression coefficients by 100 (i.e. 100*log_ex), which is the percentage difference in time (“sympercent”).²⁴ A negative coefficient indicates poorer performance, compared to the reference group.

Later age at surgical menopause was related to moderately decreased odds of self-reported functional limitation but not to differences in walk speed or chair rise time (Table 4). Women with an age at surgical menopause of ≥ 45 years had odds of self-reported functional limitation nearly half of that of women with age at surgical menopause < 40 years in age-adjusted models. Adjustment for other potential confounding factors did not attenuate this relationship. While the direction of the coefficients for walk speed suggested that women with later surgical menopause had faster speeds, the differences were not statistically significant.

Table 4.

Associations of Age at Surgical Menopause with Physical Functioning Outcomes

		n	Age-adjusted	Fully-adjusteda	p –valueb
Age at menopause (y)			Mean Differencec (95% CI)	Mean Differencec (95% CI)
Walking speed (m/s)
< 40		100	0	0	0.26
40 – 44		115	0.008 (−0.07, 0.08)	0.01 (−0.05, 0.08)
45 – 49		134	0.04 (−0.03, 0.11)	0.03 (−0.04, 0.09)
≥ 50		257	0.03 (−0.03, 0.10)	0.03 (−0.02, 0.08)
			% differenced (95% CI)	% differenced (95% CI)
Chair rise time
< 40		94	0	0	0.54
40 – 44		109	1.54 (−8.28, 11.36)	3.57 (−5.30, 12.45)
45 – 49		130	1.82 (−8.46, 12.10)	4.96 (−4.45, 14.36)
≥ 50		246	−1.41 (−7.28, 4.46)	0.49 (−5.98, 4.99)
			Odds Ratio of Limitation (95% CI)	Odds Ratio of Limitation (95% CI)
Self- reported functional limitation
< 40		101	1	1	0.01
40 – 44		115	1.06 (0.51, 2.17)	1.11 (0.56, 2.19)
45 – 49		135	0.51 (0.24, 1.09)	0.58 (0.27, 1.28)
≥ 50		257	0.57 (0.30, 1.10)	0.52 (0.29, 0.95)

^aAdjusted for age, race/ethnicity, height, weight, education, number of children, and use of estrogen therapy.

^bFrom a test of trend in the fully-adjusted model.

^cA positive coefficient indicates poorer performance, compared to the reference group.

^dCalculated by multiplying the regression coefficients by 100 (i.e. 100*log_ex), which is the percentage difference in time (“sympercent”).²⁴ A negative coefficient indicates poorer performance, compared to the reference group.

DISCUSSION

In a nationally representative survey of postmenopausal American women, age at menopause was related to physical functioning. Women with surgical menopause had slower chair rise times than women with natural menopause. Women who transitioned to natural menopause at older ages had faster walking speeds and less self-reported functional limitation in later life than women who transitioned to natural menopause earlier. Early age at surgical menopause was also related to increased levels of self-reported functional limitation. These differences remained after adjusting for potential confounding factors. We believe the differences in walk speed and chair rise time found may be clinically significant in relation to future health outcomes.²

Comparison with other studies

Previous studies have produced mixed results concerning the relationship between type of menopause and physical functioning. Our results contrast with those of a previous study during the menopausal transition that found no difference in chair rise time¹² between women who were naturally postmenopausal and those who were surgically postmenopausal. Another study during the menopausal transition showed that women with surgical menopause experienced a faster rate decline of walk speed than women with natural menopause,¹¹ while we found no relationship between type of menopause and walk speed. The lack of association between menopause type and walking speed and self-reported limitation in the NHANES supports the previous finding of no difference in SF-36 physical component score¹³ between women who were naturally postmenopausal and those who were surgically postmenopausal. The finding of a slower chair rise time among women with surgical menopause supports a faster of decline over 5 years in the SF-36 physical component found in a study during the menopausal transition.¹⁰

Our results suggest that timing of menopause is related to physical functioning even once all women in a population have undergone the menopausal transition. Our findings for walking speed are in line with those of one paper showing that women with natural menopause who did not use HT had greater decreases over 5 years in walk velocity than women who remained premenopausal or perimenopausal.¹¹ The same paper showed a similar relationship for surgical menopause, but we found only weak evidence for a relationship between age at surgical menopause and walk speed. Our results for chair rise time are also consistent with a study that showed no association between menopausal status and chair rise time for 10 chair rises.¹² However, our findings contrast those from a study using a single chair rise test that showed that compared to women who were perimenopausal or premenopausal, women with natural menopause using HT had a slower decline and women with surgical menopause had a faster decline over 5 years.¹¹ Our findings for self-reported limitation are consistent with results from studies that found greater levels^{13, 16} and faster rates of development¹¹ of self-reported limitation associated with later stages in the menopausal transition, for both natural and surgical menopause. Early age at surgical menopause was related to higher risk of self-reported functional limitations but not walking speed or chair rise time. This discordance in findings between objectively assessed and self-reported outcome measures could be explained by the fact that women with early hysterectomy may have worse psychological symptoms over the life course,²⁵ which may negatively influence self-reported functional limitation.

Our results may differ from previous results that focused on comparing women at different stages of the menopausal transition for several methodological reasons. We may not have detected more consistent differences in physical functioning by type of menopause if some effects were short term. For example, physical recovery from surgery itself¹¹ may temporarily compromise physical function. As time passes women with surgical menopause may experience improvement in their physical functioning such that they have levels of physical functioning in line with women with natural menopause in later life. Similarly, once the body adapts to the abrupt¹¹ or premature decline in hormonal levels, it is possible that the impact on walk speed dissipates. Discrepancies for chair rise time may relate to the difference in multiple chair rises versus a single chair rise in another study.¹¹ The repeated chair rise task may capture types of capacity such as endurance and cardio-respiratory function,^{4, 12} which may differ from those required to perform a single chair stand.

Mechanisms

Type and timing of menopause may have different impacts on the body systems required to perform different tasks. For example, walking involves coordination¹¹ as well as muscular strength. Rising from a chair requires considerable muscular strength, particularly of the knee extensor muscles,^{11, 26, 27} neuromuscular speed and control, and integration of the central nervous system and cardiovascular and respiratory function.^{4, 12}

Early age at menopause means an earlier decline in exposure to estrogen and progesterone, which act in complex manners directly^8–10 or in combination with other hormones²⁸ on muscle performance. Therefore, exposure to hormones could play an important role in explaining the association between early age at menopause and poor physical functioning. The evidence supporting the role of hormones was weaker among women with surgical menopause, as a relationship between early surgical menopause and poor physical function existed only for self-reported functional limitation. The estimates from models suggested that women with early age at surgical menopause had slower walk speed, although the differences were not statistically significant. It is possible that insufficient statistical power prevented the detection of the differences in age at surgical menopause due to the smaller size of the sample of women who had experienced this type of menopause.

Early age at menopause and type of menopause may also indicate a poor overall health profile that exists prior to or following menopause. Women with early age at menopause, particularly surgical menopause, have been shown to have increased levels of risk factors across the life course for poor health in older adult years, including adverse early life development^{29, 30} and lower childhood^{12, 31} and adult¹² socioeconomic position. Early age at menopause may also reflect premature ovarian aging, which may be associated with general aging³² that might relate to functional decline.

Methodological considerations

This analysis has several limitations. Women recalled age at menopause and reproductive surgical procedures. However, previous studies have found that most women accurately recall age at surgical and natural menopause³³ and whether they had a hysterectomy.³⁴ In addition, recall of age at natural or surgical menopause is unlikely to be related to the physical functioning outcomes. Thus, any existing measurement error in age at menopause would likely lead to an underestimation between age at menopause and physical functioning outcomes.

As the survey is cross sectional, we were unable to examine declines of physical functioning within individuals, which would have allowed us to more fully explore the relationship between menopausal characteristics and physical functioning in later life. Another limitation of the cross-sectional analysis is that we cannot account for survival differences by timing and type of menopause between the menopausal transition and the timing of the survey. For example, women with early age at menopause may have greater mortality risk than women with later age at menopause.^35–37 The women in our sample who survived to at least age 60 years and had early age at menopause may be more robust than women with early age at menopause who died. The analysis included only approximately half of women aged 60 years who participated in NHANES III because of missing information on variables included in the analysis. However, 82% of the women who provided information on self-reported functional limitation and walk speed or chair rise time also provided information on type of menopause or age at menopause.

Results concerning surgical menopause may not be generalizable to more recent generations of women. The use of hysterectomy has become more limited over the 20^th century in the United States. At the time that women of the NHANES III underwent surgical menopause, hysterectomy was used for some gynecological conditions which now would be treated with less invasive techniques, including hysteroscopic surgery and uterine artery fibroid embolization. Screening programs for conditions resulting in hysterectomy have also become more frequent.^38–42 Therefore, women with surgical menopause in the NHANES are likely to be a more heterogeneous group than younger women with surgical menopause.

Conclusions

To our knowledge, this study is the first to consider the relationship between characteristics of menopause and physical functioning beyond the menopausal transition. We examined both physical performance measures and self-reported questions, both of which contribute to one’s realized health experience.⁴³ The analysis included women with surgical menopause, the timing of which had differing relationships to physical functioning than natural menopause. Our results add to the evidence that health at older ages is related to life course risk factors. Women who had surgical menopause and who underwent menopause earlier had worse physical functioning in older adulthood than women with natural menopause and later age at menopause. Women who have surgical menopause and undergo early menopause may therefore benefit from interventions to limit declines in physical functioning with age.