Objective
We sought to describe the pattern of age at menopause and factors associated with type of menopause.
Study Design
This was a prospective cohort study of 5113 postmenopausal health survey respondents in the Royal College of General Practitioners’ Oral Contraception Study. Logistic regression was used to evaluate associations between sociodemographics, lifestyle, and medical history and menopause type.
Results
Median age at natural menopause (n = 3650) was 49.0 years (interquartile range, 45.0–51.0), and at surgical menopause (n = 1463) was 42.4 years (38.0–46.4). Early natural menopause was associated with smoking, ever-use of oral contraception, sterilization, and history of endometriosis (all increased odds ratios) and ever-use of hormone replacement therapy (decreased). Surgical menopause was associated with manual social class, sterilization, and having a history of endometriosis, menorrhagia, or painful menstruation (all increased), and ever-use of hormone replacement therapy (decreased).
Conclusion
Age at natural menopause was younger in this cohort than in other studies. More associations were found for surgical menopause than early natural menopause.
Menopause is an important point in a woman’s life because it signals the end of her reproductive life. The menopause transition is often accompanied by a number of symptoms, such as vasomotor symptoms and vaginal atrophy. A woman’s age at menopause may affect the type and severity of her menopausal symptoms. A prospective cohort study of 6917 Swedish women found that those aged >53 years at natural menopause had reduced frequency and intensity of vaginal dryness and atrophy, but increased frequency of hot flashes than women experiencing menopause at a younger age. Age at menopause may also influence subsequent health. A prospective study of 5287 Californian women showed that young age at natural menopause (<40 years) was associated with an increased risk of subsequent all-cause mortality compared to natural menopause occurring at 50-54 years. Other studies have shown that earlier age at both natural and surgical menopause is a risk factor for future osteoporosis and cardiovascular disease.
Age at menopause is likely to be affected by numerous genetic, environmental, and sociodemographic factors that differ between populations. Cigarette smoking has been consistently shown to affect menopause onset, with smokers frequently reaching menopause as much as 1.5 years sooner than nonsmokers, and smokers at greater risk of surgical menopause than nonsmokers. Evidence regarding the effects of other sociodemographic and lifestyle factors (eg, social class, alcohol consumption, body mass index [BMI], parity, prior oral contraceptive [OC] use) on age and type of menopause are conflicting.
Age at menopause has been investigated in US, Dutch, and Australian women, but only to a small extent in United Kingdom women. Due to its possible association with future ill health, it is important to know the mean age at, and factors associated with, menopause in United Kingdom women. We report here our findings from the ongoing Royal College of General Practitioners’ (GPs) Oral Contraception Study (RCGP OCS).
Materials and Methods
The RCGP OCS
Beginning in May 1968, 1400 GPs across the United Kingdom recruited approximately 23,000 women who were using OCs and a comparable number of women who had never done so. Participants were mostly white (98%) and all were either married or living as married. Both pill users and never-users were of a similar age (mean 29 years, SD 6.6 years) at recruitment. Information collected at baseline included social class (as determined by the woman’s husband’s occupation), significant medical history, previous OC use, and parity. Following recruitment, information was received from the participating GPs every 6 months about: hormonal preparations prescribed, pregnancies, surgery, new episodes of illness, and, if applicable, date and cause of death. Information from the GPs was coded using the International Classification of Diseases (8th revision) for morbidity and deaths, and the Code of Surgical Operations and Procedures (1969) for any surgery. Women were lost from GP follow-up if they left the recruiting practice, they received OC from a different source (eg, a family planning clinic), they died, or their GP withdrew from the study.
From November 1994 through July 1995, most of the 12,303 women who were still under GP observation were sent a health survey questionnaire by their GP on our behalf. The health survey collected information about: parity, lifetime cigarette smoking, weekly alcohol consumption, current height and weight, weight at age 30 years, weekly hours of physical activity, whether the respondent had ever had a hysterectomy or been through menopause, and the month and year of her final menstrual period (FMP).
The study sample
The study sample consisted of all postmenopausal women who returned the health survey questionnaire and who reported valid information about the timing of their FMP ( Figure 1 ). To check that the health survey was sent to the correct woman, health survey data were compared with RCGP OCS data and any women with discrepant information excluded (n = 48). Type of menopause (natural or surgical) and age at menopause was determined using the process detailed in Figure 2 .
The RCGP OCS records were checked for a recording of hysterectomy and/or oophorectomy (Surgical Operations and Procedures R codes R671-672, R681, R690-694, R696) by October 1994. When a woman had no operations recorded, her responses to the health survey were examined. If a woman had reported having gone through menopause and had provided a date for the event, the midpoint of the corresponding year (June) was taken when calculating her age at menopause. We took the year midpoint because many women reporting their FMP failed to give the month as well as the year of the event. These women were allocated to the natural menopause group. We excluded women without a recording of an operation who had not finished their periods (n = 3574), those who did not provide the year of their FMP (n = 1210), those whose calculated age at menopause seemed particularly young (<25 years) or old (>60 years; n = 24) as we suspect these data might be incorrect, women who had a hysterectomy prior to recruitment to the RCGP OCS (n = 25), and women who had a pregnancy after her reported date of menopause (n = 13).
If a woman had a record of hysterectomy and/or oophorectomy, her health survey responses were examined to see whether she also reported a date of menopause. If the respondent did not report the date of FMP on the health survey, the date of her operation was used to calculate the age at menopause and her menopause was considered to be surgical (n = 434). Women aged ≥55 years at time of operation (n = 58) were excluded from this group because they were likely to have been through natural menopause before this age although the age at menopause was unknown. If a date of FMP was reported in the health survey, both age at operation and age at reported date of menopause were calculated. If the age at operation was less than or equal to the reported age of menopause, the woman was allocated to the surgical menopause group (n = 1029) and the age at operation taken as the age at menopause. Women were excluded (n = 8) from this group if their age at operation was particularly young (<25 years) or old (>60 years). If a woman’s age at operation was older than her age at menopause reported in the health survey, her menopause was considered to be natural and the age at menopause based on the survey response used (n = 541).
For each postmenopausal woman, information recorded on the RCGP OCS database between recruitment to the main study and age at menopause were extracted for: social class, OC and hormone replacement therapy (HRT) use, any record of tubal sterilization, recorded menstrual difficulties (ie, endometriosis, menorrhagia, and painful menstruation), any cancer, and autoimmune disease (ie, diabetes, rheumatoid arthritis, inflammatory bowel disease, and thyroid disease). This information was supplemented by health and lifestyle data reported by respondents in the health survey: pack-years smoked, weekly alcohol consumption, weekly physical activity, BMI at age 30 years, and parity.
Statistical analysis
Data were analyzed using software (SPSS for Windows, version 17.0.1; SPSS, Chicago, IL). Age at menopause was examined among all postmenopausal respondents in the health survey, those with a natural menopause and those with a surgical menopause. We then completed separate analyses to examine the relationship between age at menopause and lifestyle, personal characteristics, reproductive, gynecological, and other medical history among women having a natural menopause, and similar relationships between natural or surgical menopause. When investigating naturally postmenopausal women, 2 groups were created, stratified by the median age at menopause in that group. Thus, women who were younger than (referred to as “early natural menopause”) were compared with those older than the median age (“later natural menopause”). The characteristics of women in these 2 natural menopause groups were compared using χ 2 tests for categorical variables. The χ 2 tests were also used to compare the characteristics of women with natural menopause to women with surgical menopause. Statistical significance was set at P ≤ .01 in all analyses because of the large number of comparisons being made.
Unadjusted odds ratios (ORs) and their 99% confidence intervals (CIs) were calculated for each variable using binary logistic regression. The 2 outcomes were early natural menopause and surgical menopause. Confounders were decided upon a priori, based on frequent associations in the literature between age at menopause and OC use, social class, and smoking. Age in 1995 was also included in the model examining factors associated with surgical menopause, to account for the influence of aging. Adjusted ORs (and their 99% CIs) were calculated once the confounders were entered into the model. For consistency a missing value category was included for each variable.
Details of ethics approval
The RCGP OCS began before multicenter research ethics committees were introduced in the United Kingdom. The procedure we used, however, maintained participant confidentiality. All correspondence between participating GPs and the RCGP OCS used a unique study number for each woman, the key to which only the GPs held.
Results
Median age at menopause
Median age at menopause among all 5113 postmenopausal respondents to the health survey was 47.2 years (interquartile range [IQR], 43.0–50.2) ( Figure 3 ). There were 3650 women with a natural menopause, occurring at a median age of 49.0 years (IQR, 45.0–51.0) ( Figure 4 ). The median age at menopause among the 1463 women with a surgical menopause was 42.4 years (IQR, 39.0–46.4) ( Figure 5 ).
Median age at natural menopause, stratified by birth decade, is shown in Table 1 . Although birth years of respondents with natural menopause ranged from 1919 through 1950, there was an insufficient number of women born in 1919 and 1950 for inclusion in the analysis. Table 1 shows that the median age at natural menopause was similar among each of the birth decades examined. Trends in surgical menopause, however, have changed throughout these 3 decades. While 8.3% of women born from 1920 through 1929 experienced a surgical menopause, 18.0% and 52.9% of women born from 1930 through 1939 and 1940 through 1949 experienced a surgical menopause, respectively.
| Birth year decade | Age at health survey in 1995, y | n | Median age at natural menopause (interquartile range) |
|---|---|---|---|
| 1920 through 1929 | 66–75 | 538 | 49.0 (46.0–52.0) |
| 1930 through 1939 | 56–65 | 2321 | 49.0 (46.0–52.0) |
| 1940 through 1949 | 46–55 | 786 | 47.0 (44.0–49.0) |
Characteristics associated with an early rather than late menopause
Women with an early natural menopause were significantly more likely to be of manual social class than women with a later natural menopause, when assessed univariately but not after adjustment ( Table 2 ). Fifteen or more pack-years of smoking was significantly more common on both univariate and multivariate analysis among women in the early natural menopause group than those in the later group. Overall, smokers in our study had a mean age at menopause of 45.6 years (SD 6.04 years) while nonsmokers’ mean age at menopause was 46.9 years (SD 5.7 years). Age at natural menopause was not associated with weekly alcohol consumption, physical activity, BMI at age 30 years, or parity.
| Characteristic | All | Age at menopause | |||||||
|---|---|---|---|---|---|---|---|---|---|
| <49 y | ≥49 y | ||||||||
| (n = 1691) | (n = 1959) | Unadjusted OR | Adjusted OR a | ||||||
| n | (%) | n | (%) | (99% CI) | P value | (99% CI) | P value | ||
| SOCIAL CLASS AT RECRUITMENT TO RCGP OCS | |||||||||
| Nonmanual | 1375 | 582 | (34.4) | 793 | (40.5) | 1.00 b | 1.00 b | ||
| Manual | 2268 | 1104 | (65.3) | 1164 | (59.4) | 1.29 (1.08–1.54) | < .001 | 1.17 (0.97–1.40) | .029 |
| Missing | 7 | 5 | (0.3) | 2 | (0.1) | — | — | ||
| PACK-YEARS SMOKED | |||||||||
| 0 | 1949 | 810 | (47.9) | 1139 | (58.1) | 1.00 b | 1.00 b | ||
| <15 | 528 | 231 | (13.7) | 297 | (15.2) | 1.09 (0.85–1.41) | .366 | 1.07 (0.83–1.39) | .477 |
| 15–30 | 552 | 313 | (18.5) | 239 | (12.2) | 1.84 (1.43–2.37) | < .001 | 1.75 (1.36–2.25) | < .001 |
| >30 | 466 | 263 | (15.6) | 203 | (10.4) | 1.82 (1.39–2.38) | < .001 | 1.71 (1.30–2.24) | < .001 |
| Missing | 155 | 74 | (4.4) | 81 | (4.1) | — | — | ||
| NO. OF U OF ALCOHOL DRUNK/WK | |||||||||
| 0 | 862 | 413 | (24.4) | 449 | (22.9) | 1.00 b | 1.00 b | ||
| <1 | 535 | 258 | (15.3) | 277 | (14.1) | 1.01 (0.76–1.35) | .910 | 1.11 (0.83–1.48) | .370 |
| 1–4 | 969 | 442 | (26.1) | 527 | (26.9) | 0.91 (0.72–1.16) | .325 | 0.96 (0.75–1.23) | .671 |
| >4 | 918 | 406 | (24.0) | 512 | (26.1) | 0.86 (0.68–1.10) | .119 | 0.82 (0.64–1.06) | .044 |
| Missing | 366 | 172 | (10.2) | 194 | (9.9) | — | — | ||
| PHYSICAL ACTIVITY, H/WK | |||||||||
| Light, <16 | 1139 | 549 | (32.5) | 590 | (30.1) | 1.17 (0.94–1.45) | .065 | 1.13 (0.91–1.41) | .146 |
| Moderate, 16–29 | 1286 | 596 | (35.2) | 690 | (35.2) | 1.08 (0.88–1.33) | .327 | 1.08 (0.87–1.33) | .361 |
| Heavy, >29 | 1183 | 525 | (31.0) | 658 | (33.6) | 1.00 b | 1.00 b | ||
| Missing | 42 | 21 | (1.2) | 21 | (1.1) | — | — | ||
| BODY MASS INDEX (KG/M 2 ) AT 30 Y OF AGE | |||||||||
| <18.50 | 94 | 48 | (2.8) | 46 | (2.3) | 1.22 (0.71–2.10) | .338 | 1.09 (0.63–1.89) | .687 |
| 18.50–24.99 | 2786 | 1283 | (75.9) | 1503 | (76.7) | 1.00 b | 1.00 b | ||
| 25.00–29.99 | 572 | 265 | (15.7) | 307 | (15.7) | 1.01 (0.80–1.28) | .904 | 1.02 (0.80–1.30) | .807 |
| ≥30 | 147 | 68 | (4.0) | 79 | (4.0) | 1.01 (0.65–1.56) | .961 | 0.97 (0.62–1.52) | .880 |
| Missing | 51 | 27 | (1.6) | 24 | (1.2) | — | — | ||
| PARITY | |||||||||
| 0 | 159 | 73 | (4.3) | 86 | (4.4) | 1.00 b | 1.00 b | ||
| 1 | 401 | 179 | (10.6) | 222 | (11.3) | 0.95 (0.59–1.54) | .785 | 0.85 (0.52–1.39) | .389 |
| 2 | 1430 | 654 | (38.7) | 776 | (39.6) | 0.99 (0.64–1.53) | .966 | 0.89 (0.57–1.39) | .504 |
| 3+ | 1558 | 726 | (42.9) | 832 | (42.5) | 1.03 (0.67–1.58) | .869 | 0.87 (0.56–1.35) | .407 |
| Missing | 102 | 59 | (3.5) | 43 | (2.2) | — | — | ||
| ORAL CONTRACEPTIVE USE UP TO AGE AT MENOPAUSE | |||||||||
| Never | 1544 | 632 | (37.4) | 912 | (46.6) | 1.00 b | 1.00 b | ||
| Ever | 2004 | 1000 | (59.1) | 1004 | (51.3) | 1.44 (1.21–1.71) | < .001 | 1.37 (1.14–1.63) | < .001 |
| Missing | 102 | 59 | (3.5) | 43 | (2.2) | — | — | ||
| HORMONE REPLACEMENT THERAPY USE UP TO AGE AT MENOPAUSE | |||||||||
| Never | 3249 | 1564 | (92.5) | 1685 | (86.0) | 1.00 b | 1.00 b | ||
| Ever | 299 | 68 | (4.0) | 231 | (11.8) | 0.32 (0.22–0.46) | < .001 | 0.28 (0.20–0.41) | < .001 |
| Missing | 102 | 59 | (3.5) | 43 | (2.2) | — | — | ||
| TUBAL STERILIZATION | |||||||||
| No | 3308 | 1499 | (88.6) | 1809 | (92.3) | 1.00 b | 1.00 b | ||
| Yes | 342 | 192 | (11.4) | 150 | (7.7) | 1.55 (1.15–2.07) | < .001 | 1.38 (1.02–1.87) | .006 |
| ENDOMETRIOSIS | |||||||||
| No | 3352 | 1626 | (96.2) | 1926 | (98.3) | 1.00 b | 1.00 b | ||
| Yes | 98 | 65 | (3.8) | 33 | (1.7) | 2.33 (1.34–4.07) | < .001 | 2.49 (1.42–4.37) | < .001 |
| MENORRHAGIA | |||||||||
| No | 2533 | 1157 | (68.4) | 1376 | (70.2) | 1.00 b | 1.00 b | ||
| Yes | 1117 | 534 | (31.6) | 583 | (29.8) | 1.09 (0.91–1.31) | .234 | 1.09 (0.90–1.31) | .255 |
| PAINFUL MENSTRUATION | |||||||||
| No | 3403 | 1567 | (92.7) | 1836 | (93.7) | 1.00 b | 1.00 b | ||
| Yes | 247 | 124 | (7.3) | 123 | (6.3) | 1.18 (0.84–1.66) | .207 | 1.13 (0.80–1.59) | .369 |
| RECORD OF CANCER | |||||||||
| No | 3575 | 1657 | (98.0) | 1918 | (97.9) | 1.00 b | 1.00 b | ||
| Yes | 75 | 34 | (2.0) | 41 | (2.1) | 0.96 (0.53–1.76) | .861 | 1.03 (0.56–1.91) | .888 |
| DIABETES | |||||||||
| No | 3555 | 1647 | (97.4) | 1908 | (97.4) | 1.00 b | 1.00 b | ||
| Yes | 95 | 44 | (2.6) | 51 | (2.6) | 1.00 (0.58–1.71) | .998 | 1.00 (0.58–1.73) | .985 |
| RHEUMATOID ARTHRITIS | |||||||||
| No | 3565 | 1646 | (97.3) | 1919 | (98.0) | 1.00 b | 1.00 b | ||
| Yes | 85 | 45 | (2.7) | 40 | (2.0) | 1.31 (0.74–2.31) | .217 | 1.23 (0.69–2.19) | .354 |
| INFLAMMATORY BOWEL DISEASE | |||||||||
| No | 3627 | 1680 | (99.3) | 1947 | (99.4) | 1.00 b | 1.00 b | ||
| Yes | 23 | 11 | (0.7) | 12 | (0.6) | 1.06 (0.36–3.12) | .885 | 1.00 (0.33–2.99) | .996 |
| THYROID DISEASE | |||||||||
| No | 3404 | 1575 | (93.1) | 1829 | (93.4) | 1.00 b | 1.00 b | ||
| Yes | 246 | 116 | (6.9) | 130 | (6.6) | 1.04 (0.74–1.46) | .788 | 1.03 (0.73–1.45) | .833 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
