Sleep-disordered breathing in a middle-aged and older Polish urban population

Authors


Robert Pływaczewski, 2nd Department of Respiratory Medicine, National TB and Lung Diseases Research Institute, Warsaw, Poland. Tel.: +48-22-4312307; fax: +48-22-4312454; e-mail: r.plywaczewski@igichp.edu.pl

Summary

Obstructive sleep apnoea syndrome (OSAS) is a frequent disorder; however, the prevalence of sleep-disordered breathing is not well known in many countries. The aim of our investigation was to assess the prevalence of sleep-disordered breathing (SDB) in a representative sample of the population of Warsaw. We studied 676 subjects (57.1% of the randomised cohort from the Monica II study). The sample comprised 356 males (52.7%) and 320 females (47.3%), whose mean age was 56.6 ± 8.2 years (range 41–72 years). Mean number of apnoeas and hypopnoeas per hour of time in bed (AH) in males was 7 ± 9.5 and in females 3.9 ± 6.6 (P < 0.001). SDB (cut-off point AH > 5 or >10) was identified in 188 subjects (27.8%) and 97 subjects (14.3%), respectively. SDB was established twice as frequently in males as in females (respectively 36.5 versus 18.5%; P < 0.001 for AH > 5 and 19.8 versus 8.5%; P < 0.001 for AH > 10). A diagnosis of OSAS AH > 10 and Epworth Sleepiness Score ≥11 points was established in 51 subjects (7.5%). The prevalence of OSAS was nearly four times higher in males (40 subjects, 11.2%) than in females (11 subjects, 3.4%; P < 0.001). The severity of OSAS was similar in both sexes (AH: males 32.3 ± 14.9 and females 31.4.1 ± 15.4). Older age and male sex were predictors of SDB. SDB and OSAS were independent predictors of coronary artery disease after adjusting for age, sex, body mass index, neck circumference and smoking habit.

Introduction

Obstructive sleep apnoea syndrome (OSAS) is characterised by recurrent episodes of pharyngeal collapse during sleep and significant reduction (hypopnoea) or absence (apnoea) of airflow despite the increased work of respiratory muscles. Periods of airflow limitation are associated with repetitive hypoxaemia and loud snoring (American Academy of Sleep Medicine, 1999). Arousals terminate apnoeas and hypopnoeas leading to sleep fragmentation and a reduction in slow wave and REM sleep (American Academy of Sleep Medicine, 1999). Changes in sleep architecture result in daytime sleepiness, impaired work performance and cognitive deficits. Subjects suffering from OSAS have an increased risk of arterial hypertension (AHP) (Fletcher et al., 1985; Millman et al., 1991; Nieto et al., 2000; Young et al., 1997), coronary artery disease (CAD) (Hung et al., 1990; Koskenvuo et al., 1987, Mooe et al., 1996a,b), stroke (Bassetti et al., 1997; Dyken et al., 1996; Koskenvuo et al., 1987; Palomaki et al., 1989) and road accidents (Findley et al., 1988, 1991; Masa et al., 2000; Stoohs et al., 1994). Unfortunately, snoring and daytime sleepiness are frequently considered benign conditions and OSAS is recognised very late, when clinical symptoms and complications of the disease are very severe.

The prevalence of OSAS is not well established in many countries and only a few studies have been performed in large, non-selected populations, chiefly in the USA (Young et al., 1993, 2002).

In Poland (over 38 million inhabitants) there are around 30 sleep centres and only half of them had full polysomnography (PSG) facilities. Prevalence of obese and overweight people in the Polish population is still growing (in 1996 in adults over 15 years obesity was found in 10.3% of males and 12.4% of females; 18.7% of males and 14.2% of females were overweight) (Demographic Yearbook of Poland 2001). Study results may encourage Polish Health Authorities to provide more funds for the development of a network of sleep labs/clinics.

The aim of our investigation was to assess the prevalence of sleep-disordered breathing (SDB) in a representative sample of the population of Warsaw, and to assess the relationship between SDB and sex and age.

Materials

The study was performed in a randomly selected sample of adult inhabitants of Warsaw. In 1993, 2090 males and females over 30 years of age were selected from electoral registers. The recruited population (1503, 49.4% males), formed the Polish sample of the MONICA II study (Multinational Monitoring of Trends and Determinants of Cardiovascular Disorders) (Tunstall-Pedoe et al., 1994).

All participants in the Monica II Poland study received a questionnaire on sleep problems, daytime sleepiness, snoring and observed apnoeas in 1997. Completed questionnaires were returned by 1186 participants (79%).

The local ethics committee approved the protocols and written consent was obtained from the subjects participating in the study.

Between 2000 and 2002, all 1186 subjects were invited to attend overnight, in-laboratory PSG. An invitation letter was sent out twice at 6-month intervals. We studied 676 subjects (57.1% of the total); the remaining 510 (42.9%) were not investigated because of the following reasons: 46 (3.9%) had died, 119 (10%) refused to participate, 201 (16.9%) did not respond to the three invitations, 113 (9.5%) gave their consent but did not attend and 31 (2.6%) were lost to follow-up.

The sample consisted of 356 males (52.7%) and 320 females (47.3%), and was representative of the initial group studied in 1993 (Table 1).

Table 1.   Distribution of subjects according to sex and age groups in 1993 (MONICA II) and in the present study (2000–2002)
Age categoryYearsMales (n)Females (n)χ2/P
  1. Age categories from MONICA II study.

Group I199334–441771910.006/0.94
2000–200241–52103108
Group II199345–542052211.52/0.23
2000–200251–62138121
Group III199355–641961961.61/0.20
2000–200261–7211591

Methods

A standardised questionnaire was used to evaluate sleep problems (Zieliński et al., 1999). Daytime sleepiness was assessed using the Epworth Sleepiness Scale (ESS) (Johns, 1992, 1993). Scores equal to or higher than 11 out of 24 were classified as excessive daytime sleepiness (EDS). A physical examination was performed and the following parameters were recorded: neck circumference, waist/hip ratio (WHR), body mass index (BMI, kg m−2). Central (visceral) obesity was recorded as a WHR > 1 and peripheral obesity as a WHR < 0.8.

Sleep studies were performed between 22:00 and 06:00 hours in our Sleep Laboratory. Full PSG was performed in 288 subjects (42.6%) and limited PSG in 388 participants (57.4%) (subjects were randomly selected for full or limited PSG). Full PSG was conducted using 16-channel Somnostar Alpha equipment (SensorMedics, Yorba Linda, CA, USA). All polysomnographies were analysed manually by one of the three investigators (RP, MB, LJ), who were blinded to clinical status.

Limited sleep studies (without EEG analysis) were performed using an 8-channel Poly-Mesam recorder (MAP, Martinsried, Germany). Technical data and high sensitivity and specificity of Poly-Mesam in the diagnosis of SDB have been described elsewhere (Marrone et al., 2001; Verse et al., 2000). The final score and report was corrected by one of us. We lost data for five limited PSG (0.7%) because of recording failure or difficulties in interpretation (those five subjects refused to participate in a second sleep study in our laboratory).

Sleep was analysed according to Rechtschaffen and Kales’ (1968) rules. Breathing disorders were classified as apnoeas (complete cessation of airflow ≥10 s) or hypopnoeas (event of decreased breathing amplitude ≥50% combined with desaturation of ≥4%). SDB was diagnosed as the average number of apneoas and hypopnoeas per hour of time in bed (AH) being higher than five or higher than ten. Obstructive sleep apnoea syndrome (OSAS) was diagnosed when AH > 10 and ESS ≥ 11 points.

Arterial hypertension was diagnosed when:

  • • the mean blood pressure recording of four separate measurements was greater than 140/90 mmHg or
  • • subjects gave a positive history of a diagnosis of AHP diagnosis and/or undertook appropriate treatment.

Diabetes was defined as casual plasma glucose concentration ≥200 mg dL−1 or fasting plasma glucose ≥126 mg dL−1 or 2-h postload glucose ≥200 mg dL−1 (during OGTT) or current history of antidiabetic medication.

Coronary artery disease was diagnosed when subjects gave a positive history of disease symptoms and/or had abnormal ECG (ischaemia, scar, left bundle block).

Statistical analysis

Data were analysed using Statistica 5.0 software (Statsoft, Inc., Tulsa, OK, USA). The Kolmogorov–Smirnov test with Lilliefors correction was performed to evaluate data distribution. Pearson’s chi-square test (with corrections for N) was used to compare categorical variables, and the Student’s t-test was used to compare continuous data. Results are reported as mean ± SD. Analysis of variance was used to compare three or more groups. Differences were considered significant at P < 0.05. Levene’s test was used to check homogeneity of variance. In the case of normal data distribution and homogeneous variance test F-anova was used. Post-hoc comparisons were performed using Scheffe’s test and Tukey’s verification. When data distribution was not normal or not homogenous variance non-parametric the Kruskal–Wallis H test was applied. Pearson’s correlation described relationships between variables.

Logistic regression analyses were applied to identify independent predictors of SDB. These analyses were done using spss (SPSS Inc., Chicago, IL, USA) version 10.0 for Windows.

Results

Anthropometry

The mean age of the population studied was 56.6 ± 8.2 years and ranged from 41 to 72 years. There were no differences in age between males and females (57 ± 8.2 and 56.2 ± 8.2 years, respectively, P = 0.22). To evaluate the relationship between frequency and severity of SDB and age, we divided the studied cohort into four subgroups according to age (group 1: 41–49 years, group 2: 50–59 years, group 3: 60–69 years and group 4: ≥70 years) (Table 2).

Table 2.   Percentage of females and males in the four age categories
Age groups (years)FemalesMales
n%n%
1. 41–498727.28323.3
2. 50–5911335.313136.8
3. 60–6910733.412635.4
4. ≥70134.1164.5

The mean BMI was abnormal for the whole group – 28.7 ± 4.7 kg m−2. The mean BMI in males and females was similar averaging 28.5 ± 4.3 kg m−2 and 29 ± 5.3 kg m−2, respectively. Two hundred and thirty-five subjects (34.8%) were overweight (26 ≤ BMI ≤ 30 kg m−2) and 237 subjects (35%) were obese (BMI > 30 kg m−2).

Relationships between body weight and age and sex is shown in Table 3.

Table 3.   Body weight in four age categories of males and females
GenderBody mass index (BMI) (kg m−2)P
Group 1Group 2Group 3Group 4
  1. NS, not significant.

Males28.5 ± 4.028.8 ± 4.628.2 ± 4.228.6 ± 4.3NS
Females27.3 ± 5.229.3 ± 5.430.1 ± 4.928.4 ± 4.8<0.01 (groups 1 and 3)

Snoring

Subjects were divided into three classes according to their answers on the questionnaire: non-snorers, moderate snorers and habitual snorers. There were 171 (25.3%) non-snorers (answer: never or rarely), 221 (32.7%) moderate snorers (answer: sometimes) and 284 (42%) habitual snorers (answers: often and always). The prevalence of habitual snoring was significantly higher in males than females (48 and 35.3%, respectively, P < 0.01).

Sleepiness

In the group as a whole the ESS score was within the normal range – 6.4 ± 3.9 points. Scores for the ESS were higher in males when compared with females (7.1 ± 4 and 5.5 ± 3.5, respectively, P < 0.001). In neither sex was there any difference in the ESS score across the four age groups. EDS was found in 109 subjects (16.1%) (ESS ≥ 11 points). EDS was more frequent in males (78 subjects: 21.9%) than in females (31 subjects: 9.7%), P < 0.001.

Sleep-disordered breathing

The mean AH was 5.5 ± 8.4 in the population as a whole. There was a wide range of recorded AH; the lowest AH was 0 and the highest 74. Division into quartiles revealed that the majority of subjects had a normal AH (25% had AH < 1, 50% had AH < 2.7, 75% had AH < 6). In the upper quartile were people with AH between 6 and 74.

Overnight oxygenation was satisfactory in the investigated population as a whole. Mean arterial blood saturation (SaO2) was 93.2 ± 2.6%. Time spent with saturations below 90% (T 90) averaged 9.0 ± 18.1%. Severe desaturation time, SaO2 below 80% (T80), was only 0.4 ± 2.7%.

Analysis of breathing disorders conducted separately for males and females revealed significant differences in SDB between both sexes (Table 4).

Table 4.   Sleep disordered breathing in females and males (F-anova test)
VariableFemalesMalesP
AH (n/h)3.9 ± 6.67 ± 9.5<0.001
Mean SaO2 (%)93.5 ± 2.692.8 ± 2.6<0.001
Lowest SaO2 (%)81.5 ± 9.981.3 ± 10.30.92
T 90 (%)8.9 ± 19.19.1 ± 17.20.09

There was an increase in SDB related to ageing. In females, the lowest AH was found in group 1 and the highest in group 4. In males, the lowest AH was recorded in group 1 and the highest in group 4, but differences between groups were not significant (Fig. 1).

Figure 1.

 Characteristics of AH in age subgroups of females and males. Mean value of apnoea and hypopnoea per hour of registration (AH)±standard deviation, separately presented for each age group of males and females (group 1: 41–49 years, group 2; 50–59 years, group 3: 60–69 years and group 4 ≥ 70 years).

The mean overnight SaO2 and time spent in desaturation below 90% (T90) are described in Table 5.

Table 5.   Mean overnight SaO2 and time spent in desaturation below 90% (T90) in age categories
VariableGenderGroup 1Group 2Group 3Group 4P
  1. NS, not significant.

Mean SaO2 (%)Males93.3 ± 2.192.5 ± 2.992.8 ± 2.593.3 ± 1.9NS
Females94.5 ± 2.293.2 ± 2.793.1 ± 2.693.1 ± 2.40.0016 (groups 1 and 2)
0.005 (groups 1 and 3)
T90 (%)Males6.5 ± 13.78.3 ± 15.711.8 ± 20.78.6 ± 12.4NS
Females4.2 ± 129.1 ± 18.311.6 ± 23.317.2 ± 19.8NS

Comparison of full and limited PSG revealed that subjects studied by Poly-Mesam were older (age: 58.3 ± 8.2 versus 54.4 ± 7.8 years; P < 0.001), more obese (BMI: 29.5 ± 4.9 versus 27.6 ± 4.3 kg m−2; P < 0.001) and had more apnoeas and hypopnoeas (AH: 7.3 ± 9.5 versus 3.2 ± 5.8; P < 0.001).

Diagnosis of SDB

AH > 5 was identified in 188 subjects (27.8%). The diagnosis was established twice as frequently in males (129 subjects, 36.5%) as in females (59 subjects, 18.5%) (P < 0.001). An AH > 10 was found in 97 subjects (14.3% of the total). There were 70 males (19.8%) and 27 females (8.5%) with an AH > 10 (P < 0.001). Risk of SDB in males (after adjusting for age and BMI) was three times higher when compared with females (OR 3.1, 95%CI 2.09–4.31; P < 0.0001 and OR 2.98, 95% CI 1.87–4.75; P < 0.0001, respectively for AH > 5 and >10). Prevalence of SDB in males and females increased with age (Table 6).

Table 6.   Sleep-disordered breathing (SDB) (AH > 5 or >10) in separate age categories of males and females (number of subjects and percent of category)
VariableGenderGroup 1 (n/%)Group 2 (n/%)Group 3 (n/%)Group 4 (n/%)P
  1. NS, not significant.

  2. *Differences were significant between all groups.

  3. **Reference group.

  4. Odds ratio (OR) and 95% confidence intervals (95% CI) for AH > 10 in age groups.

AH > 5 (n/h)Males21 (25.3)41 (31.5)56 (44.8)11 (73.3)<0.004*
Females6 (6.9)15 (13.5)32 (29.9)6 (46.1)0.00001*
AH > 10 (n/h)Males10 (12)28 (21.5)28 (22.4)4 (26.7)NS
Females1 (1.1)7 (6.3)15 (14)4 (30.8)0.0003*
   OR (95% CI)
AH > 10 (n/h)Both sexes1.0**1.93 (1.02–3.64)2.76 (1.49–5.12)4.13 (1.55–10.9)0.04 (group 2)
0.001 (group 3)
0.004 (group 4)

Obstructive sleep apnoea syndrome

A diagnosis of OSAS (AH > 10 and ESS ≥ 11 points) was established in 51 subjects (7.5%). The prevalence of OSAS was four times higher in males (40 subjects, 11.2%) than in females (11 subjects, 3.4%) (P < 0.001). Risk of developing OSAS in males (after adjusting for age and BMI) was 4.2 times higher than in females (95%CI 2.07–8.63; P = 0.0001). The mean age of subjects with OSAS was 59.5 ± 7.7 years. Females with a diagnosis of OSAS were older than males (64.3 ± 5.1 and 58.2 ± 7.9 years, P = 0.02). The mean AH in the OSAS group was 32.1 ± 14.9, and was similar in both sexes (males: 32.3 ± 14.9 and females: 31.4.1 ± 15.4). Subjects with OSAS were obese. Mean BMI equals 30.9 ± 5.3 kg m−2 and was lower in males (respectively, 30.4 ± 5.2 and 32.6 ± 5.9 kg m−2 in both sexes; NS). Mean SaO2 in OSAS subjects was 93.1 ± 2.6% (males: 91.8 ± 3.8% and females: 92.1 ± 3.5%; NS). Females spent more time in desaturation (T90 = 27.3 ± 35.7%) than males (T90 = 19.6 ± 22.6%; NS). The severity of daytime sleepiness in OSAS subjects assessed by ESS score (13.3 ± 2.1 points in the whole subgroup) was similar in males (13.5 ± 2.2 points) and females (12.7 ± 1.4 points).

The severity of OSAS in age groups is shown in Table 7.

Table 7.   Severity of OSAS (AH) in separate age categories
GenderAH (n/h)P
Group 1Group 2Group 3Group 4
  1. NS, not significant.

Males (number of subjects)29.9 ± 8.7 (7)35.7 ± 17.9 (17)31.6 ± 15.3 (14)26.7 ± 2.1 (2)NS
Females (number of subjects)– (0)22.5 ± 3.5 (2)34.3 ± 17.4 (8)26 ± 0 (1)0.04 (groups 2 and 3)

Comorbidities

Arterial hypertension was found in 308 subjects (45.6%). The majority of subjects with AHP presented with a history of the disease or took appropriate treatment (199, 64.6%). A similar percentage of males and females was affected by AHP (173 subjects – 48.7% and 135 subjects – 42.2%, respectively). The distribution of AHP in relation to age is shown in Fig. 2.

Figure 2.

 Arterial hypertension (AH+) in males and females in relation to age. Prevalence of arterial hypertension (AH+) in all age groups (on left side in females; on right side in males) – presented as a percent of group.

Sixty-three percent (32 subjects) with OSAS and 44.2% (276 subjects) without OSAS had AHP (P < 0.01). The risk for AHP in OSAS subjects was 2.1 times higher (95% CI, 1.17–3.82) when compared with non-OSAS subjects (P = 0.01). In females with confirmed OSAS, the risk of AHP was higher than in males [respectively, OR 14.6 (95% CI, 1.85–114.8; P = 0.01) and OR 1.3 (95% CI, 0.68–2.57; P = 0.4)]. OSAS after adjustment for age, sex, BMI, neck circumference and smoking habit was not associated with hypertension (OR 1.15, 95% CI 0.6–2.22; P = 0.65). Logistic regression analysis revealed a significant relationship between EDS and hypertension on unadjusted test (OR 1.56, 95% CI 1.04–2.34; P = 0.03). Adjusted test (confounders – age, sex, BMI, neck circumference and smoking habit) did not confirm this association (OR 1.22, 95% CI 0.72–2.07; P = 0.45).

Coronary artery disease was present in 200 subjects (29.6%). There was no difference in the frequency of CAD between males and females [101 subjects (28.4%) and 99 subjects (30.9%), respectively]. The frequency of CAD increased with age (Fig. 3).

Figure 3.

 Coronary artery disease (CAD+) in males and females in age groups. Occurrence of coronary artery disease (CAD+).

Coronary artery disease was present in 32 subjects with OSAS (62.7%). In the non-OSAS group, only 168 subjects (26.9%) had an established diagnosis of CAD (P < 0.001). The probability of CAD in the OSAS group was 4.6 times higher (95% CI 2.52–8.28; P < 0.0001) than in the non-OSAS group. The risk of developing CAD in males with OSAS compared to non-OSAS males was very high (OR 5.9, 95% CI 2.95–11.96; P < 0.0001). In females, the OR for CAD in OSAS subjects was 2.8, compared with non-OSAS females (95% CI 0.82–9.36; P = 0.1).

Obstructive sleep apnoea syndrome after adjustment for age, sex, BMI, neck circumference and smoking habit was associated with CAD (OR 3.98, 95% CI 2.05–7.73; P < 0.0001).

Only two subjects (3%) with OSAS and 20 subjects (3.4%) without OSAS had a history of stroke.

Diabetes (D) was reported by 10 OSAS subjects (19.6%) and 39 non-OSAS subjects (6.3%). The OR for diabetes in the OSAS group was 3.6 times higher than in non-OSAS subjects (95% CI 1.7–7.85; P = 0.0009). The risk of diabetes in OSAS males was 4.1 times higher when compared to non-OSAS males (95% CI 1.66–10.23; P = 0.002). In OSAS females, the OR for diabetes was higher (but without statistical significance) than in non-OSAS females (OR 3, 95% CI 0.61–15; P = 0.17).

Obstructive sleep apnoea syndrome after adjustment for age, sex, BMI, neck circumference and smoking habit was not associated with diabetes (OR 2.28, 95% CI 0.97–5.38; P = 0.06).

AH (similar to OSAS) was associated only with CAD after adjustment for age, sex, BMI, neck circumference and smoking habit. The relationship between increasing scores of AH and AHP, CAD, diabetes and excessive daytime sleepiness adjusted for age, sex, BMI, neck circumference and smoking habit is shown in Table 8.

Table 8.    Odds ratio (OR) and 95% confidence intervals (95% CI) for arterial hypertension, coronary artery disease, diabetes and excessive daytime sleepiness by AH category based on sleep studies
Odds ratioAH category
0 (n = 86)0.1–4.9 (n = 340)5–14.9 (n = 166)≥15 (n = 84)
  1. Adjusted for age, sex, BMI, neck circumference and smoking habit.

  2. *Not significant.

 Arterial hypertension [OR (95% CI)]
Crude1.01.49 (0.89–2.5)2.44 (1.4–4.3) (= 0.002)3.1 (1.6–5.9) (< 0.001)
Adjusted†,*1.01.1 (0.6–1.9)1.3 (0.7–2.45)1.2 (0.6–2.54)
 Coronary artery disease [OR (95% CI)]
Crude1.01.2 (0.7–2.1)1.9 (1.02–3.6) (P = 0.04)4.3 (2.1–8.6) (P < 0.0001)
Adjusted1.00.9 (0.5–1.6)1.06 (0.5–2.1)2.6 (1.2–5.5) (P = 0.015)
 Diabetes [OR (95% CI)]
Crude1.01.8 (0.5–6.2)1.7 (0.4–6.2)5.2 (1.4–18.8) (P = 0.01)
Adjusted†,*1.01.2 (0.3–4.5)0.8 (0.2–3.1)2.1 (0.5–8.5)
 Excessive daytime sleepiness [OR (95% CI)]
Crude1.00.6 (0.27–1.36)1.28 (0.56–2.92)16.9 (7.36–38.7) (P < 0.0001)
Adjusted1.00.54 (0.23–1.24)1.11 (0.46–2.67)15.2 (6.17–37.6) (P < 0.0001)

Of females, 40% (127 subjects) were either premenopausal or on hormone replacement therapy. Hormonal replacement therapy was continued up to the age of 62 years, so we divided females into two groups: those ≤62 years (oestrogen positive) and >62 years (oestrogen negative). Postmenopausal females had a significantly higher AH than premenopausal females or those on hormone replacement therapy (5.3 ± 8 and 1.9 ± 2.4, respectively; P = 0.005).

Discussion

Our study confirms that SDB is prevalent in the Warsaw population. The frequency of SDB increased with age and was twice as high in males as in females. The incidence of OSAS was near four times higher in males when compared with females, however, severity of disease was similar in both sexes. SDB and OSAS were independent predictors of coronary artery disease after adjusting for age, sex, BMI, neck circumference and smoking habit.

The results of our investigations are comparable with the Wisconsin Sleep Cohort Study in the corresponding age groups (Young et al., 1993) (PSG performed in 602 subjects). In the WSCS study, SDB (AHI ≥ 5) was found in 24% males and 9% females, OSAS (AHI ≥ 5 and excessive daytime sleepiness) in 4% males and 2% females. Incidence of SDB and OSAS was higher in our group. This difference may be explained by the health status of the population (healthy employees in WSCS study), age range (WSCS subjects were younger, 30–60 years versus 41–72 years in our study) and genetic/ethnic factors (only white people in current study).

The design of our investigation is similar to that of the Sleep Heart Health Study (SHHS) (Young et al. 2002), which assessed the prevalence of SDB and cardiovascular complications in a large group of subjects. Polysomnography (performed at home) was recorded in 5615 subjects aged 40–98 years (mean age 63.5 ± 10.7 years). SDB (AHI ≥ 15) was diagnosed in 18% of subjects (25% of males and 11% of females). Independent risk factors of SDB in the SHHS group were habitual snoring, obesity, WHR and a greater neck circumference.

An AHI ≥ 5 was found in 47% of subjects versus 27.8% in our group (however, AH is different to an AHI and describes the average number of apnoeas and hypopnoeas per hour of time in bed). In the SHHS study, a higher frequency of SDB in older subjects (>60 years) was not found. This may be explained by differences in methods (SHHS group was more representative – 5615 versus 676 subjects in our group, PSG in all subjects in Young’s study), age (older age in SHHS study, only 29 subjects ≥70 years in our group) and genetic make-up (different populations).

A high prevalence of SDB in the elderly has been found by other authors. Ancoli-Israel et al. (1989) found a respiratory disturbance index (RDI) ≥ 5 in 70% of subjects aged around 80 years. In another investigation (Ancoli-Israel et al., 1991) an RDI ≥ 10 was found in 62% subjects aged more than 65 years. Bixler et al.(1998) reported the highest frequency of SDB (AHI ≥ 5) in males older than 65 years (30.5%) compared with the group aged 45–64 years (19.7%) and the group aged 20–44 years (7.9%).

While our study was in progress, the prevalence of SDB in the general population of Spain was published by Duran et al. (2001) who used different technology to the WSCS and SHHS studies. They performed limited PSG (Mesam IV) [this system was validated by Stoohs and Guilleminault (1992) and Esnaola et al. (1996)] in a random selection of 2148 subjects aged from 30 to 70 years. In the second phase of the study authors invited for PSG 305 subjects with negative results from the Mesam IV study and all 442 subjects with suspicion of OSAS (positive results). Finally, PSG was performed in 555 subjects. A diagnosis of OSAS (AHI ≥ 10) was established in 19% of males and 15% of females (these findings are very similar to our data, especially for males). In the Spanish study, prevalence of OSAS increased with age in both sexes similar to in the Warsaw population. Duran et al. revealed a link between OSAS and hypertension but not to excessive daytime sleepiness. In contrast, our study underlines the association between SDB and EDS and coronary artery disease.

Ip et al. (2001) assessed SDB in middle-aged Chinese men (sleep questionnaires sent to 1542 subjects in the age range 30–60 years/784 respondents/full PSG performed in 150 respondents). AHI ≥ 5, ≥ 10 and ≥ 15 was found in 8.8, 6.3 and 5.3%, respectively. OSAS (AHI ≥ 5, ≥ 10 or ≥ 15 and excessive daytime sleepiness) was confirmed in 4.1, 3.2 and 3.1%, respectively. Independent predictors of SDB were BMI, habitual snoring, time taken to fall asleep and age. Chinese males were younger (mean age 41 years), had normal body weight (BMI 23.9 kg m−2) and the percentage of habitual snores was smaller (23%) when compared with our data (respectively, 57 ± 8.2 years, 28.5 ± 4.3 kg m−2, 48% of habitual snorers).

Three years later, Ip et al. (2004) published the results of SDB prevalence in Chinese women (1532 females aged 30–60 years/854 questionnaires returned/full PSG performed in 106 respondents). SDB (AHI ≥ 5) and OSAS (AHI ≥ 5 and EDS) were found in 3.7 and 2.1%, respectively. The authors described a 12-fold increase of SDB prevalence from the fourth and sixth decades of life. BMI and age were identified as significant independent predictors of SDB. Chinese females were younger, thin and only 15% of them were classified as habitual snorers (these factors explain the lower incidence of SDB in the aforementioned group).

Kim et al. (2004) performed polysomnography in 457 Korean males and females randomly allocated from an initial cohort of 5020 participants between 40 and 69 years. The prevalence of SDB (AHI ≥ 5) was 27% in males and 16% in females (results similar to our data). The incidence of OSAS (AHI ≥ 5 and EDS) was 4.5% in males and 3.2% in females (lower than in our subjects). BMI, sex and hypertension were associated with the risk of SDB in a middle-aged Korean population. Smaller BMI, younger age, lower occurrence of habitual snoring (21.9% in males and 12.5% in females) and ethnic factors may explain the differences between the Korean and Polish studies.

Udwadia et al. (2004) conducted an epidemiological study in healthy, Indian urban males aged 35–65 years (658 subjects completed sleep questionnaires; 250 of them had PSG at home). SDB (AHI ≥ 5) was found in 19.5% and OSAS (AHI ≥ 5 and EDS) in 7.5%. BMI, neck circumference and diabetes were independent predictors of SDB. The prevalence of SDB and OSAS was lower than in Warsaw males. Normal BMI in the whole Indian group, younger age (47.8 years) and less prevalent habitual snoring (26%) may be responsible for the differences between the two studies.

Study limitations

Our study has several limitations. We investigated only 57% of the randomly selected group. Despite modest participation, the age distribution in respondents was similar to the whole randomised cohort (Table 1). We have no data concerning actual BMI and concomitant diseases in non-respondents.

Slightly more than half of our subjects underwent limited polysomnography and the remainder were investigated with full PSG. Several studies have confirmed very high sensitivity and specificity (>90%) of the equipment we used (Poly-Mesam) in the diagnosis of SDB (Marrone et al., 2001; Verse et al., 2000).

Subjects studied by Poly-Mesam were significantly older and more obese and had more apnoeas and hypopnoeas that the subgroup studied by PSG (these groups were non-homogeneous and this may decrease the value of our results).

Our population comprised only an urban population (analysis of the same number, gender and age proportions of rural inhabitants would significantly improve our results and conclusions).

The quality of our recordings was high; all investigations were performed in the sleep laboratory and were supervised by a technician.

Conclusions

In this representative sample of adult inhabitants of Warsaw, SDB (AH > 10) was found in 14.3% of subjects. OSAS affects 7.5% of the middle-aged and older Polish urban population. SDB was established twice as frequently in males as in females (respectively, 36.5% versus 18.5%; P < 0.001 for AH > 5 and 19.8% versus 8.5%; P < 0.001 for AH > 10). Risk of SDB increased with age (highest in the oldest group) and was three times higher in males than in females. A diagnosis of OSAS AH > 10 and Epworth Sleepiness Score (ESS) ≥ 11 points was established in 51 subjects (7.5%). The prevalence of OSAS was nearly four times higher in males (40 subjects, 11.2%) than in females (11 subjects, 3.4%) (P < 0.001). Risk of OSAS development in males was 4.2 times higher (after adjusting for age and BMI) than in females, however the severity of OSAS was similar in both sexes (AH = 32.3 ± 14.9 in males and 31.4.1 ± 15.4 in females).

Sleep-disordered breathing and OSAS were independent predictors of coronary artery disease after adjusting for age, sex, BMI, neck circumference and smoking habit.

Acknowledgements

The authors thank the technicians Mrs Barbara Cieplechowicz and Mrs Krystyna Łukawska for their assistance during sleep studies. We give special thanks to Dr Renata Riha for essential and editorial advice during manuscript preparation. This study was supported by the State Committee for Scientific Research, Grant 4 PO5D 082 16.

Ancillary