Abstract
The prevalence of sleep-disordered breathing has not been well studied in women, especially in terms of the effects of age, body mass index (BMI), and menopause. We evaluated this question using a two-phase random sample from the general population. In Phase I, 12,219 women and 4,364 men ranging in age from 20 to 100 yr were interviewed; and in Phase II, 1,000 women and 741 men of the Phase I subjects were selected for one night of sleep laboratory evaluation. The results of our study indicated that, for clinically defined sleep apnea (apnea/hypopnea index ⩾ 10 and daytime symptoms), men had a prevalence of 3.9% and women 1.2%, resulting in an overall ratio of sleep apnea for men to women of 3.3:1 (p = 0.0006). The prevalence of sleep apnea was quite low in premenopausal women (0.6%) as well as postmenopausal women with hormone replacement therapy (HRT) (0.5%). Further, in these women the presence of sleep apnea appeared to be associated exclusively with obesity (BMI ⩾ 32.3 kg/m2). Postmenopausal women without HRT had a prevalence of sleep apnea that was significantly higher than the prevalence in premenopausal women with HRT (2.7 versus 0.6%, p = 0.02) and was more similar to the prevalence in men (3.9%), although it remained significantly less when controlling for age and BMI (p = 0.001). These data combined indicate that menopause is a significant risk factor for sleep apnea in women and that hormone replacement appears to be associated with reduced risk.
Sleep-disordered breathing (SDB) has been assumed to be a condition associated primarily with men. In clinical samples, the ratio between men and women of the prevalence of SDB has been considered to be about 10:1 (1-3). In general population samples, it has been shown that sleep apnea occurs more frequently than assumed in women (4-8). Most estimates of the male/female ratio in the general public range between 2:1 and 4:1.
It has been suggested that sleep apnea is uncommon in premenopausal women but increases in prevalence after menopause (9). However, data supporting this conclusion are mixed. In studies that have assessed the relationship between SDB and menopause, using random samples of the community, some have reported an association (5, 10) whereas others have not (7). No differences have been observed between pre- and postmenopausal women in terms of responsiveness to challenges to breathing by nasal occlusion or alcohol (11, 12).
It is commonly assumed that normal hormonal function in premenopausal women is associated with a lower prevalence of SDB (9). In some studies the use of hormonal replacement therapy (HRT) has been reported to be effective in the treatment of SDB (13-15), whereas in others it has had no effect (16, 17). To date, no random sample of the general population has been used to assess whether HRT in postmenopausal women is associated with a lower prevalence of SDB.
The prevalence of sleep apnea in general samples of women has only been evaluated using restricted age ranges, i.e., either middle age or elderly. Further, the relationship between HRT and SDB has not been evaluated in a general sample. Thus, the purpose of this study was to contrast the prevalence of SDB in a large sample of women compared with men from the general population across a wide age range (20 to 100 yr) while controlling for age, obesity, and, in women, menopause and HRT.
In the first phase of this study, a sample of adult women (age ⩾ 20 yr) was randomly selected from telephone households in two counties of southern Pennsylvania (Dauphin and Lebanon). Telephone households were selected using the Mitofsky–Waksberg two-stage random digit dialing procedure, excluding households previously evaluated in the study of men (18). A within-household selection procedure described by Kish was used to select the woman to be interviewed (19). Telephone interviews were conducted with 12,219 age-eligible women residing in the sample households; the response rate among age-eligible women was 74.1% which was similar to the response rate (73.5%) in the age-eligible men previously reported (20). The questionnaire used in this interview included basic demographic information as well as questions assessing the five risk factors for sleep apnea (snoring, daytime sleepiness, obesity, hypertension, and menopause).
In the second phase of this study, a random sample from the 12,219 women previously interviewed by telephone was selected for study in our sleep laboratory to assess the presence of sleep apnea. This sample was chosen by counting the number of the five risk factors reported by each interviewed subject. Those with higher counts of risk factors were oversampled (2, 4, 10, 27, 51, and 56% of those with zero, one, two, three, four, and five symptoms reported, respectively). A total of 1,000 subjects have been evaluated in the sleep laboratory to date. The response rate, based on those we attempted to contact, of this phase of the study in women was 65.8% compared with 67.8% in the previous sample of men. We contrasted those subjects who were recorded in the laboratory with those who were selected but who were not recorded in terms of age and the risk factors used to select them. There were no significant differences between these two groups on any of these variables.
Each female subject selected for laboratory evaluation completed a comprehensive sleep history, a physical examination, and a psychometric battery that assessed psychologic and cognitive function, exactly as was completed by the male sample (20). The psychometric data will be reported in a future publication. This study was reviewed by the institutional review board of the Penn State University College of Medicine, and signed informed consents were obtained from all subjects. All subjects were evaluated for one night in the sleep laboratory in sound-attenuated light- and temperature-controlled rooms. During this evaluation, they were continuously monitored for 8 h using 16-channel polygraphs (model 78d; Grass Instrument, Quincy, MA). The electroencephalogram, a three-channel electrooculogram, and an electromyogram were recorded. The sleep records were subsequently scored independently according to standardized criteria (21).
Respiration was monitored throughout the night by use of thermocouples at the nose and mouth (model TCT 1R; Grass Instrument) and thoracic strain gauges. All-night recordings of hemoglobin oxygen saturation (SaO2 ) were obtained with an oximeter (model 8800; Noonin Medical, Plymouth, MN) attached to the finger. An apnea was considered present if a breath cessation exceeded 10 s. Each apnea was categorized in terms of obstructive (chest wall movement present) or central (chest wall movement absent). In addition, hypopneas were considered present when an airflow reduction of approximately 50% was indicated at the nose or mouth and was associated with a reduction of 4% SaO2 .
Because the sampling procedures departed from equal probability selection, to improve the efficiency of subject identification and screening, compensatory weights were developed for the analysis to obtain estimates for the original target population of men and women in the two-county study area. Specifically, three weights were created for the men. First, in the telephone sample, 32 of the 963 clusters of phone numbers in the first stage were “exhausted” before the target sample size was obtained. A compensatory weight was computed which corrected for this problem. A second weight was computed because the within-household screening deliberately introduced unequal probabilities of selection across the three age groups in order to oversample the middle age group. The final weight for the men was computed to account for the oversampling of subjects for the sleep laboratory study (Phase II); those with larger counts of the four possible risk factors had substantially higher probability of being selected. For the women, the only weight required was to account for the oversampling of subjects for the sleep laboratory study. To eliminate any suggestion of possible sampling bias, we calculated 32 unique weights for the women and 16 unique weights for the men corresponding to all possible combinations of the five risk factors for the women and four for the men. Any individual weight that had too small of a cell size was combined with adjacent cells so that less than 10% of the cells had a sample size < 25 and no cell had a size less than 10 (19).
As we began evaluating our sample of women it appeared that the mean body mass index (BMI) was markedly too high compared with the national population. Because of the strong association between BMI and sleep apnea, it was felt that a poststratification-population control weight needed to be established to present a more accurate estimate of prevalence. We used the BMI and race distributions by age decade from the NHANES III laboratory data as the standard (22) to adjust both the men and women in terms of BMI and race to be more representative of the national population. However, those strata that had too small of a cell size (less than 30 yr age group) were collapsed into their adjacent strata so that less than 10% of the cells had a sample size < 25 and no cell had a size less than 10 (19).
For analysis purposes, we calculated various apnea indices on the basis of the specific types of apneas per hour of sleep. Three indices of sleep apnea were calculated for each subject: the apnea/hypopnea index (A/HI), an obstructive apnea/hypopnea index (OA/HI), and a central apnea index (CI). For the purposes of this article we focused on obstructive sleep apnea, primarily because women did not appear to have central sleep apnea. The first of four mutually exclusive groups of SDB that we defined was the more severe form of SDB (OA/HI ⩾ 15). A second group was considered to have mild SDB (SNORE+). This group was defined as those who reported snoring of a moderate to severe degree during the phone interview (Phase I) and were also observed to snore in the laboratory (Phase II) and had mild indices of SDB (i.e., 0 < OA/HI < 15 or breathing-related arousals). The third group of subjects had “simple snoring” (SNORE); i.e., they snored without the presence of any SDB (OA/HI = 0 and no breathing-related arousals). The final group of subjects was those with no indication of SDB.
Obstructive sleep apnea (OSA), as well as central sleep apnea, was also diagnosed using Sleep Disorders Clinic (SDC) criteria, which employed sleep laboratory plus clinical findings. This diagnosis was made by a Sleep Disorders Medicine specialist (one author [A.N.V.]) on the basis of whether immediate treatment was considered appropriate. This diagnosis required an A/HI ⩾ 10 plus the presence of clinical symptomatology, e.g., daytime sleepiness, hypertension, or other cardiovascular complication. Central apnea was diagnosed when central apnea was the predominant type of apnea.
The women for this study were also divided into three groups: (1) premenopausal, defined as those having regular menstrual periods and not on any HRT; (2) postmenopausal without HRT, defined as those having no menstrual period in the past 12 mo, age > 40 yr, and not on any HRT; and (3) postmenopausal with HRT, defined as those who met the criteria in (2) but who were on some kind of HRT. HRT was defined as regular use of estrogen alone or estrogen combined with progesterone. Finally, those women who were considered possible perimenopausal were included in the premenopausal group because that sample size was too small (∼ 4%)
Differences in risk of apnea among the age, sex, or menopausal groups were analyzed with age-, sex-, or menopause-specific prevalence estimates, respectively. All estimates of prevalence and relative risk are presented with corresponding 95% confidence intervals in parentheses. When mean values are reported, the variance estimate is in terms of the standard deviation. Finally, various statistical models using logistic regression were built in an attempt to assess the risk factors while simultaneously controlling for potential confounders. Given the low prevalence of sleep apnea in the sample, the odds ratios from the multivariate analysis can also be interpreted as relative risks. The weights were incorporated into the analyses using SAS (version 6.12; SAS Institute, Carey, NC). SUDAAN was not needed because SAS Proc Logistic, with SAS weight statement, will yield the same estimates and standard errors as SUDAAN when the weights are standardized in SAS to sum to the sample size and no cluster sampling is involved.
The laboratory subsample of the original phone sample was established on the basis of oversampling those subjects at higher risk for sleep apnea. It was assumed that those with a higher number of risk factors would be at a higher risk for sleep apnea. To assess the validity of this assumption we plotted the cumulative distribution of sleep apnea (OA/HI ⩾ 15) for each of the six sampling strata (zero to five risk factors) in Figure 1. The cumulative distributions of the six sampling strata strongly support this assumption. To assess the representativeness of the laboratory subsample compared with the original phone sample, we calculated the prevalence of each risk factor for each of the five sampling strata (one to five risk factors) for both the phone and laboratory samples (Table 1). The two samples had similar distributions, including the joint distributions (e.g., snoring and menopause).
Fig. 1. Cumulative distribution of OA/HI for both men and women for each sampling stratum.
[More] [Minimize]| Risk Factor | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Obesity | Snore | Daytime Sleepiness | Hypertension | Postmenopause | ||||||
| Phone sample | ||||||||||
| 1 | 29.7 | 8.8 | 23.3 | 7.1 | 31.1 | |||||
| 2 | 58.8 | 25.2 | 30.7 | 33.5 | 51.8 | |||||
| 3 | 85.9 | 45.6 | 41.2 | 60.8 | 66.5 | |||||
| 4 | 96.4 | 72.8 | 62.4 | 86.3 | 82.1 | |||||
| 5 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | |||||
| Laboratory sample | ||||||||||
| 1 | 32.9 | 12.9 | 20.1 | 3.0 | 31.1 | |||||
| 2 | 63.6 | 32.0 | 35.6 | 27.6 | 41.2 | |||||
| 3 | 85.9 | 53.1 | 46.2 | 55.7 | 59.1 | |||||
| 4 | 96.1 | 76.7 | 62.6 | 85.2 | 79.4 | |||||
| 5 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | |||||
The sample of subjects recorded in the sleep laboratory included 1,000 women and 741 men. The mean age for the women was 48.8 ± 14.1 yr and for the men 48.2 ± 12.9 yr. The mean BMI, unadjusted for race and BMI, was 29.6 ± 7.5 for the women and 26.9 ± 4.2 for the men. After adjustment for BMI and race on the basis of the NHANES III sample (22), the BMI was 27.9 ± 6.6 for women and 27.2 ± 4.4 for men.
The unadjusted prevalence for sleep apnea on the basis of SDC criteria for women was 1.8% (1.0, 2.6) and for men 3.3% (2.2, 4.8). The overall adjusted prevalence of OSA diagnosed using SDC criteria was found to be 1.2% of the women in this sample (Table 2). This prevalence is contrasted with an adjusted prevalence of 3.9% in our previous sample of men. Thus, a ratio of 3.3:1 was observed between men and women in terms of OSA diagnosed using SDC criteria (risk ratio [RR] = 0.3 [0.2, 0.6], p = 0.0006). The overall prevalence of sleep apnea using only sleep laboratory criteria (OA/HI ⩾ 15) was 2.2% compared with 7.2% in men (RR = 0.3 [0.2, 0.5], p < 0.0001). Thus, a ratio of 3.3:1 was again observed between men and women in terms of obstructive sleep apnea diagnosed using only sleep laboratory criteria (OA/HI ⩾ 15).
| Sleep Apnea | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| n† | SDC‡ | OA/HI ⩾ 15 | SNORE+§ | SNORE‖ | ||||||
| Women | 1,000 | 1.2 (0.7, 2.2) | 2.2 (1.5, 3.3) | 5.4 (4.1, 6.9) | 10.4 (8.7, 12.5) | |||||
| Men | 741 | 3.9 (2.8, 5.6) | 7.2 (5.6, 9.3) | 17.3 (14.7, 20.2) | 17.4 (14.8, 20.3) | |||||
| Women | ||||||||||
| Age, yr | ||||||||||
| 20–44 | 455 | 0.7 (0.2, 2.0) | 0.6 (0.2, 2.0) | 2.3 (1.3, 4.2) | 8.2 (6.0, 11.2) | |||||
| 45–64 | 376 | 1.1 (0.4, 2.9) | 2.0 (1.0, 4.0) | 7.2 (5.0, 10.3) | 14.5 (11.3, 18.4) | |||||
| 65–100 | 169 | 3.1 (1.3, 7.1) | 7.0 (4.0, 11.9) | 9.4 (5.8, 14.8) | 7.1 (4.1, 12.1) | |||||
| BMI, kg/cm2 | ||||||||||
| < 32.3 | 814 | 0.4 (0.1, 1.2) | 1.1 (0.5, 2.1) | 4.0 (2.8, 5.5) | 8.8 (7.0, 11.0) | |||||
| ⩾ 32.3 | 186 | 4.8 (2.5, 9.0) | 7.2 (4.3, 11.9) | 11.5 (7.7, 17.0) | 17.5 (12.7, 23.6) | |||||
| Menopause | ||||||||||
| Pre | 503 | 0.6 (0.2, 1.8) | 0.6 (0.2, 1.8) | 3.2 (2.0, 5.2) | 7.9 (5.8, 10.6) | |||||
| Post | 497 | 1.9 (1.0, 3.6) | 3.9 (2.5, 6.0) | 7.5 (5.5, 10.2) | 13.0 (10.3, 16.2) | |||||
| Hormone replacement | ||||||||||
| With | 183 | 0.5 (0.1, 3.8) | 1.1 (0.3, 4.3) | 3.8 (1.8, 7.7) | 9.8 (6.3, 15.1) | |||||
| Without | 314 | 2.7 (1.4, 5.3) | 5.5 (3.4, 8.6) | 9.7 (6.9, 13.5) | 14.8 (11.3, 19.2) | |||||
The age-specific prevalence of OSA in women using SDC criteria was 0.7% in those aged 20 to 44 yr compared with 1.1% in those aged 45 to 64 yr (RR = 1.7 [0.4, 6.3], p = 0.50) and 3.1% in those aged ⩾ 65 yr (RR = 4.6 [1.2, 15.9], p = 0.03). However, when the age-specific prevalence was evaluated by decade, a peak at age 60 to 69 yr was suggested. The prevalence was 1.3% (0.2, 9.6) for those 20 to 29 yr compared with 3.3% (1.3, 8.3) for those 60 to 69 yr (RR = 2.7. [0.3, 12.4], p = 0.50). The age-specific prevalence was slightly lower in the ⩾ 70-yr groups (1.9% [0.5, 7.1]) compared with the 60-to-69-yr age group (RR = 0.6 [0.2, 2.8], p = 0.56). This age-specific distribution of OSA was somewhat similar to that observed in men (20). The prevalence was higher for the 50-to-59-yr-old men compared with to those 20 to 29 yr old, and higher compared with men 60 to 69 and ⩾ 70 yr old.
The age-specific prevalence of OA/HI ⩾ 15 was 0.6% for those aged 20 to 44 yr compared with 7.0% for those aged ⩾ 65 yr (RR = 11.1 [3.0, 32.6], p = 0.0002). When the age-specific prevalence was evaluated by decade, a peak was again suggested at age 60 to 69 yr. The age-specific prevalence of OA/HI ⩾ 15 was 1.3% (0.2, 9.6) compared with 8.4% (4.7, 14.7) for the 20-to-29- and 60-to-69-yr age groups, respectively (RR = 6.8 [0.8, 25.9], p = 0.09). For those women older than 70 yr, the prevalence was only 3.7% (1.0, 6.4) compared with the 60-to-69-yr-old group (RR = 0.4 [0.2, 1.4], p = 0.16).
The BMI-specific prevalence of the most severe form of SDB, sleep apnea diagnosed using SDC criteria (see Table 2), was 0.4% for the nonobese and 4.8% for the obese women (BMI ⩾ 32.3) (RR = 11.5 [3.3, 33.8], p = 0.0001). In men, the BMI-specific prevalence for sleep apnea was 2.0% (1.1, 3.4) for the nonobese compared with 13.8% (8.7, 20.9) for the obese (BMI ⩾ 31.1) (RR = 7.0 [3.4, 13.9], p < 0.0001). The BMI-specific prevalence in women with the mildest form of SDB, i.e., simple snoring, was 8.8% for the nonobese compared with 17.5% for the obese women (RR = 2.0 [1.4, 2.9], p = 0.0004). The BMI-specific prevalence of simple snoring in men was 16.4% (10.9, 23.9) for the nonobese and 17.6% (14.8, 20.8) for the obese (RR = 0.9 [0.6, 1.5], p = 0.81). Those with SDB of intermediate severity (SNORE+) had intermediate ratios.
An unexpected finding observed in this sample of women was the general absence of the central type of sleep apnea. There was no central apnea diagnosed using SDC criteria, whereas for men there was a prevalence of central apnea of 0.4% (0.1, 1.2). The overall prevalence of central apnea (CI > 0) in women was only 0.3% (0.0, 0.6) compared with 7.8% (5.9, 9.7) in men (RR = 0.04 [0.0, 0.05], p = 0.004).
The prevalence of OSA using SDC criteria for premenopausal women was 0.6% compared with 1.9% for postmenopausal women (RR = 0.3 [0.1, 1.2], p = 0.04). Further, those postmenopausal women with HRT had a prevalence of sleep apnea similar to those who were premenopausal (0.5 versus 0.6%) (RR = 0.8 [0.2, 7.9], p = 0.91). In contrast, the prevalence of sleep apnea for those without HRT was higher than the prevalence in premenopausal women (2.7 versus 0.6%) (RR = 4.7 [1.2, 14.5], p = 0.02) and was more similar to the prevalence in men (2.7 versus 3.9%) (RR = 0.7 [0.4, 1.5], p = 0.39). Further, those women who were postmenopausal with HRT had a prevalence of sleep apnea that was less than that in postmenopausal women without HRT (0.5 versus 2.7%) (RR = 0.2 [0.2, 1.7], p = 0.14).
The prevalence of sleep apnea using only sleep laboratory criteria (OA/HI ⩾ 15) for premenopausal women was 0.6% compared with 3.9% for postmenopausal women (RR = 0.2 [0.1, 0.5], p = 0.003). Those women who were postmenopausal with HRT had a prevalence that was similar to the prevalence in premenopausal women (1.1 versus 0.6%) (RR = 1.9 [0.4, 10.1], p = 0.40). In contrast, those women who were postmenopausal without HRT had a prevalence that was higher than the prevalence in premenopausal women (5.5 versus 0.6%) (RR = 9.3 [2.6, 25.8], p = 0.003) and was similar to the prevalence in men (5.5 versus 7.2%) (RR = 0.8 [0.5, 1.3], p = 0.32). Further, those women who were postmenopausal with HRT had a prevalence of sleep apnea that was less than that in postmenopausal women without HRT (1.1 versus 5.5%) (RR = 0.2 [0.1, 0.9], p = 0.03).
In women who were premenopausal or postmenopausal with HRT, the prevalence of obesity was different than in postmenopausal women without HRT and in men. In premenopausal and postmenopausal women with HRT who had sleep apnea (OA/HI ⩾ 15), 100% were obese (BMI ⩾ 32.3 kg/ m2). In contrast, in those women who were postmenopausal without HRT and had sleep apnea (OA/HI ⩾ 15), only 49.4% (27.4, 71.5) were obese. This latter distribution of sleep apnea in relation to BMI was similar to that in men (42.1% [29.7, 55.6]) (RR = 1.2 [0.7, 2.1], p = 0.54).
Postmenopausal women who were taking estrogen alone were, on average, the same age as those postmenopausal women who were taking estrogen plus progesterone (57.1 ± 0.9 versus 57.8 ± 1.0, p = 0.60). They were also approximately of similar BMI (27.6 ± 0.5 versus 27.9 ± 0.8, p = 0.75). The percentage of postmenopausal women who were estrogen-alone users with OA/HI ⩾ 15 was 1.5%, compared with 0.3% of postmenopausal women taking estrogen plus progesterone with OA/HI < 15 (p = 0.65). Similarly, the percentage of postmenopausal women who were obese (BMI ⩾ 32.3), had sleep apnea (OA/ HI ⩾ 15), and were estrogen-alone users was 7.5%, compared with 1.4% of those who were estrogen-plus-progesterone users, were obese, and had sleep apnea (p = 0.95).
Another way to understand the relationship between gender and prevalence of SDB was to evaluate these data using logistic regression to adjust for possible confounding. Various different models were evaluated on the basis of the entire cohort of men and women across all age ranges (⩾ 20 yr), using both backward and stepwise procedures at α = 0.05. All of the models that were tested resulted in the same final model. Gender, menopause status (pre-, post- with HRT, and post- without HRT), age, log(BMI), and all possible interactions of these variables were included, as well as alcohol use, smoking, and race. Age and BMI were treated as continuous variables. Log(BMI) was used to avoid the need to model a quadratic form of BMI. None of the interactions of gender with age and BMI entered into the final model. Alcohol, smoking, and race did not change the strength of the relationship between gender and prevalence, nor did they interact with gender. Similarly, they did not alter the relationship between log(BMI) and prevalence. Thus, the final model included just the three levels of menopause, age, and log(BMI).
The first question posed of these data was whether the prevalence of sleep apnea in women was similar to that in men. For this comparison the reference group employed was men (Table 3). These data suggest that the prevalence of sleep apnea in women in all three categories of menopause status is significantly less than that in men but less so in postmenopausal women without HRT.
| Menopause Status | SDC† | OA/HI > 15 | SNORE+‡ | SNORE§ | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR (95% CI)‖ | p Value | OR (95% CI) | p Value | OR (95% CI) | p Value | OR (95% CI) | p Value | |||||||||
| Baseline, men | ||||||||||||||||
| Pre | 0.1 (0.03, 0.4) | 0.001 | 0.1 (0.01, 0.2) | 0.0001 | 0.1 (0.1, 0.2) | < 0.0001 | 0.3 (0.2, 0.4) | < 0.0001 | ||||||||
| Post, with HRT | 0.1 (0.01, 0.5) | 0.008 | 0.04 (0.01, 0.2) | 0.0001 | 0.1 (0.04, 0.2) | < 0.0001 | 0.4 (0.2, 0.6) | 0.0003 | ||||||||
| Post, without HRT | 0.2 (0.1, 0.5) | 0.001 | 0.2 (0.1, 0.4) | 0.0001 | 0.3 (0.2, 0.4) | < 0.0001 | 0.7 (0.4, 0.98) | 0.04 | ||||||||
| Baseline, premenopause | ||||||||||||||||
| Post, with HRT | 0.5 (0.04, 5.6) | 0.56 | 0.9 (0.1, 5.8) | 0.89 | 0.8 (0.3, 2.1) | 0.65 | 1.4 (0.7, 2.6) | 0.32 | ||||||||
| Post, without HRT | 1.9 (0.4, 8.7) | 0.39 | 4.3 (1.1, 17.3) | 0.04 | 2.2 (1.1, 4.5) | 0.03 | 2.4 (1.4, 4.2) | 0.001 | ||||||||
The second question posed of these data was the influence of menopause status and HRT in women. To address this question the premenopausal women were used as the reference group (Table 3). The odds ratio of sleep apnea (OA/HI ⩾ 15) in postmenopausal women with HRT compared with premenopausal women was not significant, whereas for postmenopausal women without HRT there was a significant difference (p = 0.04). It should be noted that if only women were used in the analysis, the relationships in terms of odds ratios among the three groups of women were nearly identical to those using the entire sample including men; however, the p values were somewhat higher.
This is the first sleep laboratory study of a large representative random sample of women and men from the general population to investigate the difference in prevalence of SDB between women and men with a particular emphasis on the impact of menopause and HRT. The results of this study indicate that women have a prevalence of sleep apnea that is less than the prevalence in men. The prevalence of sleep apnea in women using SDC criteria was 1.2%, compared with 3.9% in men (p = 0.0006). These values are similar to the prevalence values based on estimating “minimal diagnostic criteria for sleep apnea” reported by Young and colleagues (4). The overall ratio of men to women for sleep apnea, using our SDC criteria, was 3.3:1. Similarly, the male/female ratio for sleep apnea using only sleep laboratory criteria (OA/HI ⩾ 15) was 3.3:1 (p < 0.0001). This is consistent with previous estimates, which have reported ratios of prevalence of SDB ranging between 2:1 and 4:1 (4-8).
HRT appears to be associated with a lower prevalence of sleep apnea. Using the SDC criteria, postmenopausal women with HRT had a prevalence of sleep apnea that was similar to the prevalence in premenopausal women (0.5 versus 0.6%, p = 0.91). This same relationship was also observed when using only the sleep laboratory criteria (1.1 versus 0.6%, p = 0.40).
The absence of HRT in postmenopausal women appears to be associated with a higher prevalence of sleep apnea than that in premenopausal women. When using the SDC criteria, the prevalence of sleep apnea was higher in postmenopausal women without HRT than in premenopausal women (2.7 versus 0.6%, p = 0.02), as well as when using only sleep laboratory criteria (5.5 versus 0.6%, p = 0.003). Further, the prevalence of sleep apnea in postmenopausal women without HRT was higher than in postmenopausal women with HRT using SDC criteria (2.7 versus 0.5%, p = 0.14) as well as when using only sleep laboratory criteria (5.5 versus 1.1%, p = 0.03). It should be noted that the relative risk in these latter two analyses is similar, suggesting that the lack of significance with the criteria may be due to sample size.
The prevalence of sleep apnea in postmenopausal women without HRT remained significantly less than that in men, when controlling for age and BMI. Using the SDC criteria, the prevalence of sleep apnea in postmenopausal women without HRT remained less than that in men (2.7 versus 3.9%, p = 0.001). Using the sleep laboratory criteria alone resulted in a similar relationship (5.5 versus 7.2%, p = 0.0001).
Women who had sleep apnea (OA/HI ⩾ 15) and were premenopausal or postmenopausal with HRT were all obese (BMI ⩾ 32.3 kg/m2). This finding is consistent with previous reports (5, 23). In contrast, postmenopausal women without HRT who had sleep apnea (OA/HI ⩾ 15) had a prevalence of obesity that was similar to men (49.4 versus 42.1%). This finding that obesity and menopause are significant risk factors for SDB lends further support to the hypothesis that hormonal/ metabolic factors play a role in the etiology of SDB (24).
Women had an age distribution for the prevalence of sleep apnea that is similar to that in men. Specifically, the prevalence of sleep apnea using our SDC criteria in women appears to increase with age to approximately 65 yr, after which it appears to decline. This age distribution of the prevalence of sleep apnea is similar to that in men, except that the peak is later (i.e., ∼ 65 yr for women versus ∼ 55 yr for men). Sample size is probably the reason why the decrease in prevalence in women did not reach significance. The fact that central apnea was prominent in elderly men (20) and almost absent in women may partially account for the gender difference in the age distribution of the prevalence of A/HI ⩾ 15. The significance of the lack of central apnea in women is not well understood.
Both BMI and age are significant risk factors of sleep apnea. In addition, the relationship between age and BMI closely resembles the relationship between age and sleep apnea, indicating that BMI confounds the age/sleep apnea relationship. In other words, the decline in prevalence of sleep apnea in the elderly might be at least partially accounted for by the associated decrease in BMI rather than the associated increase of age. When controlling for BMI, the association between age and sleep apnea was diminished. This finding supports previous reports by Gislason and colleagues (personal communication). However, it should be pointed out that the decrease in severity of sleep apnea with age that we have previously reported (20) may also play a role in the underrepresentation of older ages observed in the SDC population.
In women it is thought that progesterone levels may play a role in protecting them from sleep apnea before menopause (25). Women experience an increase in ventilatory drive during the luteal phase of the menstrual cycle when progesterone levels are the highest (26). Oral progesterone has been associated with slight but definite improvement in ventilatory indices during sleep in both male and female sleep apnea patients (16, 27). One study found that combined estrogen and progesterone supplementation improved indices of SDB in a group of healthy, nonobese, postmenopausal women (14). It has been suggested that estrogen may increase the sensitivity of the ventilatory centers to the stimulant effect of progesterone (23). Also, administration of estrogen in postmenopausal women is associated with decreased plasma interleukin (IL)-6, which is elevated in patients with sleep apnea (24). Thus, it is possible that the lower prevalence of sleep apnea in the HRT group is also associated with a suppression of IL-6. Within the postmenopausal women with HRT of this study, a nonsignificant difference was suggested in terms of prevalence of sleep apnea associated with estrogen-plus-progesterone use. This difference must, however, be considered with caution due to the low prevalence of sleep apnea in these two groups. In contrast to the role of female hormones in women, in men, exogenous administration of testosterone levels have been shown to induce sleep apnea (28, 29). Testosterone levels have also been shown to be associated with upper airway collapsability in patients with sleep apnea (29, 30). Finally, in women, sleep apnea has been induced by exogenous administration of testosterone (31) and resolved after removal of testosterone-producing tumors (32). Much more study is required to understand the complex relationship between sleep apnea and hormonal function.
Estrogen deficiency after menopause results in many physiologic changes in a woman's body which can have severe consequences on her health and quality of life. One serious consequence of the loss of estrogen after menopause is the increase in cardiovascular disease seen in postmenopausal women, making cardiovascular disease, particularly coronary heart disease, the number-one cause of death in this age group (33). Several studies have shown that estrogen use by postmenopausal women decreases coronary heart disease by approximately 30 to 50% (33). Also, the majority of postmenopausal women experience a reduction in blood pressure with HRT (34). Given that SDB is a risk factor for hypertension, one could speculate that the reduction of cardiovascular morbidity/mortality in postmenopausal women with HRT may be related, at least in part, to the reduction of SDB in this population as demonstrated by our study.
The authors gratefully acknowledge the strong support in terms of the design of this project by Professor J. Richard Landis, Ph.D., Center for Epidemiology and Biostatistics, University of Pennsylvania. In addition, James M. Lepkowski, Ph.D., Senior Research Scientist, Survey Research Center, Institute for Social Research, University of Michigan, contributed expert guidance for the sampling strategy, as well as the establishment of the sample weights.
Supported by the National Institutes of Health Grants: HL40916 and HL51931.
| 1. | Block AJ, Boysen PG, Wynne JW, Hunt LASleep apnea, hypopnea and oxygen desaturation in normal subjects. N Engl J Med3001979513517 |
| 2. | Guilleminault C, Quera-Salva MA, Partinen M, Jamieson AWomen and the obstructive sleep apnea syndrome. Chest931988104109 |
| 3. | Vgontzas AN, Tan TL, Bixler EO, Martin LF, Shubert D, Kales ASleep apnea and sleep disruption in obese patients. Arch Intern Med154199417051711 |
| 4. | Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr SThe occurrence of sleep-disordered breathing among middle aged adults. N Engl J Med328199312031205 |
| 5. | Redline S, Kump K, Tishler PV, Browner I, Ferrett VGender differences in sleep disordered breathing in a community-based sample. Am J Respir Crit Care Med1491994722726 |
| 6. | Marin JM, Gascon JM, Carrizo S, Gispert JPrevalence of sleep apnoea syndrome in the Spanish adult population. Int J Epidemiol261997381386 |
| 7. | Kripke DF, Ancoli-Israel S, Klauber MR, Wingard DL, Mason WJ, Mullaney DJPrevalence of sleep-disordered breathing in ages 40–64 yr: a population-based survey. Sleep2019976576 |
| 8. | Olson LG, King MT, Hensley MJ, Saunders NAA community study of snoring and sleep-disordered breathing: prevalence. Am J Respir Crit Care Med1521995711716 |
| 9. | Block AJ, Wynne JW, Boysen PGSleep-disordered breathing and nocturnal oxygen desaturation in postmenopausal women. Am J Med6919807579 |
| 10. | Gislason T, Benediktsdóttir B, Björnsson JK, Kjartansson G, Kjeld M, Kristbjarnarson HSnoring, hypertension, and the sleep apnea syndrome: an epidemiologic survey of middle-aged women. Chest103199311471151 |
| 11. | Carskadon MA, Bearpark HM, Sharkey KM, Millman RP, Rosenberg C, Cavallo A, Carlisle C, Acebo CEffects of menopause and nasal occlusion on breathing during sleep. Am J Respir Crit Care Med1551997205210 |
| 12. | Block AJ, Hellard DW, Slayton PMinimal effect of alcohol ingestion on breathing during the sleep of postmenopausal women. Chest881985181184 |
| 13. | Kimura H, Tatsumi K, Kunitomo F, Okita S, Tojima H, Kouchiyama S, Masuyama S, Shinozaki T, Honda Y, Kuriyama TProgesterone therapy for sleep apnea syndrome evaluated by occlusion pressure responses to exogenous loading. Am Rev Respir Dis139198911981206 |
| 14. | Pickett CK, Regensteiner JG, Woodard WD, Hagerman DD, Weil JV, Grindlay Moore L. Progestin and estrogen reduce sleep-disordered breathing in postmenopausal women. J Appl Physiol 1989;66:1656–1661. |
| 15. | Franklin K, Lundgren R, Rabben TSleep apnoea syndrome treated with oestradiol and cyclic medroxyprogesterone. Lancet3381991251252 |
| 16. | Block AJ, Wynne JW, Boysen PG, Lindsey S, Martin C, Cantor BMenopause, medroxyprogesterone and breathing during sleep. Am J Med701981506510 |
| 17. | Cistrulli PA, Barnes DJ, Grunstein RR, Sullivan CEEffect of short term hormone replacement in the treatment of obstructive sleep apnoea in postmenopausal women. Thorax491994699702 |
| 18. | Waksberg JSampling methods for random digit dialing. J Am Stat Assoc7319784046 |
| 19. | Kish L. Survey sampling. New York: John Wiley and Sons; 1965. |
| 20. | Bixler EO, Vgontzas AN, Ten Have T, Tyson K, Kales A. Effects of age on sleep apnea in men: I. Prevalence and severity. Am J Respir Crit Care Med 1998;157:144–148. |
| 21. | Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Washington, DC: U.S. Government Printing Office; 1968. |
| 22. | U.S. Department of Health and Human Services (DHHS). National Center for Health Statistics. Third National Health and Nutrition Examination Survey, 1988–1994, NHANES III laboratory data file. Hyattsville, MD: Centers for Disease Control and Prevention; 1996. |
| 23. | Wilhoit SC, Surratt PMObstructive sleep apnea in premonopausal women: a comparison with men and with postmenopausal women. Chest911987654658 |
| 24. | Vgontzas AN, Papanicolaou DA, Bixler EO, Hopper K, Lotsikas A, Lin H-M, Kales A, Chrousos GPSleep apnea and daytime sleepiness and fatigue: relation to visceral obesity, insulin resistance, and hypercytokinemia. J Clin Endocrinol Metabol85200011511158 |
| 25. | Skatrud J, Dempsey J, Kaiser DVentilatory response to medroxyprogesterone acetate in normal subjects: time course and mechanism. J Appl Physiol441978939944 |
| 26. | White D, Douglas N, Pickett C, Weil J, Zwillich CHypoxic ventilatory response during sleep in normal premenopausal women. Am Rev Respir Dis1261982530553 |
| 27. | Hensley M, Saunders N, Strohl KMedroxyprogesterone treatment of obstructive sleep apnea. Sleep31980441446 |
| 28. | Schneider BK, Pickett CK, Zwillich CW, Weis JV, McDermott MT, Santen RJ, Verano LA, White DPInfluence of testosterone on breathing during sleep. J Appl Physiol611986618623 |
| 29. | Cistulli PA, Grunstein RR, Sullivan CEEffect of testosterone administration on upper airway collapsibility during sleep. Am J Respir Crit Care Med1491994530532 |
| 30. | Grunstein RRMetabolic aspects of sleep apnea. Sleep191996S218S220 |
| 31. | Johnson MW, Anch AM, Remmers JEInduction of the obstructive sleep apnea syndrome in a woman by exogenous androgen administration. Am Rev Respir Dis129198410231025 |
| 32. | Dexter DD, Dovre EJObstructive sleep apnea due to endogenous testosterone production in a woman. Mayo Clin Proc731988246248 |
| 33. | Mosca L, Manson JE, Sutherland SE, Langer RD, Manalio T, Barret-Connor ECardiovascular disease in women: a statement for healthcare proffesionals from the American Heart Association. Circulation96199724682482 |
| 34. | Laragh JH, Brenner BM, editors. Hypertension, pathophysiology diagnosis, and management. New York: Raven Press; 1995. |
