Prevalence of Sleep Disordered Breathing and Sleep Apnea in 50- to 70-Year-Old Individuals
A SurveyZamarrón C.a · Gude F.b · Otero Y.a · Alvarez J.M.a · Golpe A.a · Rodriguez J.R.a
aDivision of Respiratory Medicine and bClinical Epidemiology Unit, Hospital General de Galicia, Santiago de Compostela, Spain
Purpose: To determine the prevalence of sleep disordered breathing (SDB) and sleep apnea syndrome (SAS) in a general population aged from 50 to 70 years. Subjects and Methods: We recruited 76 individuals aged between 50 and 70 years, chosen at random from the electoral census. They were invited to the clinic where a detailed medical history was taken and physical examination, ENT examination, pulmonary function tests and night time recording of respiratory variables performed. Results: The prevalence of SDB (apnea-hypopnea index ≥ 5) was 28.9%, and there were no differences between men (28%) and women (30%). However, the prevalence of SAS was 6.8%, and there were differences between men (5 cases) and women (0 cases) (p = 0.0521). Subjects in the SDB group had higher systolic blood pressure than in the non-SDB group (p < 0.05). Conclusions: SDB and SAS are common among 50- to 70-year olds. The prevalence of SDB was 28.9% and the prevalence of SAS was 6.8%.
Sleep disordered breathing (SDB) refers to periodic impairment of respiration which occurs only during sleep. If these events are associated with symptoms, such intermittent snoring and daytime somnolence, the clinical diagnosis of sleep apnea syndrome (SAS) can be made . SAS is a common disease, affecting more than 4% of the adult population and appears to be associated with a number of forms of morbidity . It is known that SAS, as an illness, is commonly recognized in the 5th to 7th decade . However, relatively few studies exist where the prevalences of SDB and SAS are described in this period of time [2, 4, 5, 6, 7, 8, 9]. The aim of this study was to determine the prevalence of SDB and SAS in a population-based sample of 50- to 70-year-old subjects.
subjects and methods
The study was carried out in Santiago de Compostela, a city in the northwest of Spain, which has a population of 92,364. The subjects were recruited using the electoral census for the district of Santiago. Subjects from 20 to 70 years of age were stratified by age in 10-year groups (20–29, 30–39, etc.). The data were compiled in 1993. A random sample of an equal number of the previously mentioned age groups received a mailed invitation to participate. From the 3,006 invitations sent, 1,556 (51.7%) were returned correctly completed. The distribution pattern is described in table 1.
Table 1. Subjects selection: age distributionof the population, random sample,responders in the first stage (questionnaire),second (hospital consultation) and third(NTRRV) stages
The study consisted of three stages. In the first stage, a structured questionnaire with 26 questions was mailed, asking for predisposing factors, anthropometric details, consumption of alcohol and tobacco, pulmonary and cardiovasuclar illnesses, presence of snoring and daytime sleepiness, other sleep-related characteristics and whether prescribed or nonprescribed medications were being taken. Beforehand, 50 subjects picked at random were given a medical examination with the aim of evaluating problems that the survey could pose. The questionnaires were mailed up to three times in cases where answers were not provided. No differences in age and gender were found between respondents and nonrespondents. Also visits to the homes of 100 nonrespondents were made in order to evaluate the reason for noncompliance. In these cases, nosignificant differences between the primary variables of the study were found.
In the second stage, a sample of 457 subjects came to the hospital for consultation, during which the presence of their partner or spouse was required. We took a complete medical history, performed a general physical examination and completed a sleep structured questionnaire with 112 items. The questionnaire, which was composed mainly of questions (82%) with only two possible answers, was completed. In this second stage, differences were found between those who came for the consultation (mean age ± SD = 53.2 ± 13.5 years; men = 48%, women = 52%) and those who did not (48.6 ± 14.2 years; men = 43%, women = 57%). These differences are significant for the decades between 30 and 50 years of age, but not for those between 50 and 70 years. Alcohol consumption was divided into two categories (<60 and >60 g/day of alcohol). Current smokers were defined as those answering ‘yes’ to the question ‘Do you smoke?’. Furthermore, questions were asked about regular snoring >5 nights/week), whether it was a nuisance to other household members and whether it was linked to respiratory pauses. Questions were also asked about other sleep problems such as interrupted sleep or morning headaches. Daytime sleepiness was defined as mild when present only when little attention was required (e.g. watching television or reading a newspaper), moderate when present in activities requiring more attention and severe when the subject reported inability to work or drive a car because of sleepiness, or when falling asleep while talking. A κ agreement index >0.75 from the test carried out in the postal survey and from the personal interview test in the hospital was obtained for the above-mentioned variables.
After completing the questionnaire, a physical examination was performed checking height, weight, neck circumference (NC), blood pressure and a clinical examination, including the upper airways. Body mass index (BMI) was calculated using the formula (weight/height2). The NC was measured at the level of the cricothyroid membrane, relating it to height. The pharynx was examined and the result recorded as ‘normal’ or ‘abnormal’. Pharyngeal examination was strictly subjective and considered to be abnormal if the pharynx appeared narrow and small; if it was bulky, long and rested on the base of the tongue during phonation or if the tonsils were large enough to compromise the pharyngeal orifice . Forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV1) were performed according standard recommendations  using a Pulmonary Testing System of Collins (USA) spirometer.
In the third stage, a random sample of 80 subjects between 50 and 70 years of age was selected and invited to the hospital sleep laboratory. Seventy-eight subjects showed up, 2 of whom were excluded (1 recent upper airway surgery, 1 decompensated cardiopulmonary disease).
Based on the literature, we anticipated a prevalence of SDB, i.e. an apnea-hypopnea index (AHI ≥5), of 30% in this group [2, 8], with a 95% confidence interval and an error of ±10%. In all these subjects, we performed a totally attended night-time recording of respiratory variables (NTRRV). This record consisted of monitoring respiratory variables throughout one night in a hospital bed using Densa Ltd. Flint (UK) equipment, which registered oronasal flow by means of a thermistor, thoracic and abdominal movements using strain gauges, body position and oximetry. The analogue signals were digitalized, displayed on a computer screen and stored on a computer hard disk. The monitoring apparatus and the validity of the respiratory recordings during sleep taken by this machine have been described elsewhere . Registration commenced a few minutes after turning off the room lights when the patient had fallen asleep (5–20 min). The study finished when the patient awoke. The day after the study, a physician (C.Z. or A.G.) checked the recordings, with each screen showing 3 min of data. The number of respiratory events was determined. Respiratory events were defined as: a cessation of airflow lasting 10 s or more for apneas and 10 s or more of reduced airflow associated with a cyclic SaO2 dip for hypopneas. An AHI was calculated using the whole recording (from turn on to turn off) as the denominator. SDB was defined as AHI ≥5. SAS was considered on the basis of symptoms of disrupted breathing during sleep (snoring, apneas) and daytime sleepiness along with an AHI ≥5.
This study was approved by the Ethics Committee of our hospital.
Data were expressed as mean ± SD. We calculated the prevalence (and 95% confidence intervals) of SDB using cut-off points of ≥5, ≥10, and ≥20 of the AHI. The prevalence of an AHI ≥5, snoring and daytime sleepiness was calculated as the prevalence of SAS meeting the minimal diagnostic criteria [21, 22].
Patients were divided into two groups, SDB (AHI ≥5) and non-SDB (AHI <5). The association between each of the study variables and SDB status was tested using Student’s t test, the Wilcoxon rank sum test for continuous variables, Fisher’s exact test and the =χ2 test of independence for categorical variables. p < 0.05 was considered statistically significant.
Seventy-six subjects were studied. The measured prevalence of various frequencies of abnormal respiratory events is shown in table 2. The prevalence of AHI ≥5 was 28.9%, and there were no differences between men (28%) and women (30%). However, the prevalence of significant SAS was 6.8%, and there were differences between men (5 cases) and women (0 cases) (p = 0.0521). Four of the SAS patients were older than 60 years. Patients with SDB were not significantly older than those in the non-SDB group (63.3 ± 4.6 and 60.8 ± 6.9 years, respectively).
Table 2. Prevalence of SDB and SAS
Physical examination and spirometric data can be seen in table 3. Subjects in the SDB group had a higher systolic blood pressure than the non-SDB group (p < 0.05). The examination of the pharynx revealed differences between both groups. Nine subjects with SDB (41%) had an abnormal pharyngeal examination versus 14 (25.9%) in the non-SDB group. We did not observe any differences in diastolic blod pressure, BMI, NC and spirometric values between the SDB and the non-SDB groups.
Table 3. Physical examination, spirometric values and SDB (AHI ≥5)
The frequency of the different symptoms in subjects with and without SDB is shown in table 4. Regular snoring was more frequent in the SDB group than in the non-SDB group (p < 0.05). Seven subjects (31.8%) with SDB exhibited daytime sleepiness versus 16 (29.6%) in the present non-SDB group. Two cases (9.1%) of the SDB group and 8 (14.8%) of the non-SDB group showed severe daytime sleepiness. In neither category of daytime sleepiness, did we find any significant difference between the centages. BMI = Body mass index; NC = Neck circumference; SBP = systolic blood pressure; DBP = diastolic blood pressure; FVC = forced vital capacity. FEV1 = forced expiratory volume in 1 s; NS = nonsignificant.SDB and the non-SDB groups. In the SDB group, 16 cases (72.7%) demonstrated regular snoring, 7 (31.8%) awakenings, 8 (36.4%) pyrosis and 5 (22.7%) took sleeping pills.
Table 4. Clinical characteristics of SDB (AHI ≥5)
There were no differences between the two groups for symptoms of excessive daytime sleepiness, morning headaches, and tobacco and alcohol use.
The major findings of our study are that undiagnosed SDB, as indicated by 5 or more episodes of apnea or hypopnea per hour of sleep, is prevalent among 50- to 70-year-old individuals (28.9%), and 6.8% present with clinical SAS. We focused our attention on this age group because SAS, as an illness, is commonly recognized in the 5th through 7th decade . This group might be considered the most important referrals to sleep clinics, where the recognition of sleep apnea may be difficult because of age-related increased in comorbidity (e.g. diabetes, heart disease) or in women because of changes in menopausal status. They make up an intermediate group between young people and the eldery. In addition, they collaborate well because they are worried about their general state of health.
Our estimates may be inaccurate because of selection bias. We suspect that there was a tendency for persons with sleep complaints to volunteer to participate, thus the prevalence of SDB in the sample studied is probably higher than that in the source population. It is possible that bias was introduced in the second stage: the sample of 457 subjects who came to the hospital for consultation, because they were significantly older than those who did not come for consultation. However, we stratified age by decade and we did not find any differences in the subjects between 50 and 70 years of age.
The prevalence estimates are imprecise for two reasons: the finite size of the sample and the imperfect sensitivity and specificity of the measurement instrument. With 76 subjects studied, for a 95% confidence interval, the error is ±10%. This is relatively large, but it would be a major undertaking to reduce it since it would have been necessary to increase the sample size fourfold to reduce the confidence interval by one half of its present level. The respiratory monitoring apparatus that we used had a good sensitivity (95%) and specificty (93%) with respect to full polysomnography . We took the AHI as an indicator of SDB and used the results of a single-night polysomnographic study conducted in a hospital. Assessment of SDB from a single-night study may represent an underestimation of SDB, especially in individuals with relatively mild forms of SAS .
The prevalence estimates we reported are similar to those reported by others. The largest community study reported is that of Young et al.  who studied 602 30- to 60-year-old state employees with a polysomnographic study. They observed a prevalence of SAS of 2% for women and 4% for men, while in the 50–70 age group, the prevalence of SDB (≥5 events per hour) was 16% in women and 31% in men. However, since employed people are healthier than those who do not work, their findings may underestimate the prevalence of SDB in the entire population. In a sample of 294 men aged 40–65, Bearpark et al.  found a prevalence of SDB (≥5 events per hour) of 26%, and SAS ranged from 3.1 to 12.2%, depending on the cut-off used for the occurrence of sleepiness. Olson et al.  studied 441 35- to 69-year-olds recruited from different base populations with a portable monitor. They estimated the prevalence of SDB (≥15 events per hour) of 15–20% in 50- to 65-year-olds, however, in subjects aged 65–69 the reported prevalence was high (40%), which may be partly explained by the effect of other chronic illnesses in the elderly. Other studies have examined the prevalence of SDB among the elderly. Ancoli-Israel et al.  found a prevalence of SDB (apnea index ≥5 events per hour) of 24% among a randomly selected cohort of 420 elderly residents (aged ≥65 years) using portable home monitoring. In a separate study, Hoch et al.  evaluated 105 elderly subjects aged ≥60 years with one-night polysomnography, the estimated prevalence of SDB (AHI ≥5) was 26% and the median AHI increased significantly across decades, from age 60 to age 90 years. The differences in prevalence reported by different groups of investigators may be explained by sampling biases, different methodologies and/or the definition used to make the diagnosis of SDB and SAS, but they may also reflect true differences in the prevalence in different communities.
Based on clinic populations, it is well established that male gender is a risk factor for SDB, where the ratio of men:women with diagnosed SDB is in the order of 10:1 . In the population-based study of Young et al. , the magnitude of male excess in SDB prevalence was double in the 50- to 59-year-olds; we found that there is no difference between men and women as concerns the prevalence of SDB, but when we considered clinical SAS, all were men. This finding suggests that men may be more symptomatic than women at the same AHI level or that women are being underevaluated for SDB.
Most of the studies find that the prevalence of SDB and SAS increases with age . The lack of association between age and SDB in our study suggests that age is not a strong risk factor for SDB in adulthood, or it is more likely that we are studying a relatively small age gap, so that the possible effect of age may be lost.
Although some studies have described associations between SAS and obesity, we have not found any differences between the SDB and the non-SDB groups for BMI and NC in this age group. Nor have we found any differences in the parameters that evaluated lung function. However, systolic blood pressure was 10 mm Hg higher in the SDB group than in the non-SDB group, results which are similar to those obtained by Hla et al.  in a community-based study with ambulatory blood pressure monitoring.
Excessive daytime sleepiness is the commonest consequence of SAS, linked to the degree of sleep fragmentation and nocturnal hypoxemia [19, 20]. Associating symptoms of daytime sleepiness with symptoms of breathing disturbances is one strategy that has been employed to improve discrimination between SAS and non-SAS patients . However, in older individuals there is a weaker association between symptoms and measures of apneic activity related to the nonspecificity of daytime sleepiness in this group of subjects . In a recent paper by Olson et al. , it was reported that the prevalence of daytime sleepiness in the general population was very high. Olson et al.  did not find any significant differences in the prevalence of daytime sleepiness between SDB and non-SDB groups. Our work supports this view. Although this fact can be related to the age of our sample: in other studies on middle-aged adults, the proportion of subjects with complaints of daytime sleepiness that had apnea-hypopnea scores <5 is high , although some authors defined SAS as daytime sleepiness with an AHI ≥5 . Given that daytime sleepiness, even when qualified by severity, is poorly discriminating between SDB and non-SDB groups, the presence of snoring improves between-group discrimination [21, 22].
Patients with SAS could have pyrosis. This symptom has been linked to increases in gastric-esophagus pressure during the periods of obstruction of the upper airways . In the cases we have studied, a greater frequency of this symptom has been found.
In this study, no significant differences in both groups with regard to drinking and smoking habits were found. It should be kept in mind that this study was undertaken with healthy members of the population who had not changes their life style for health reasons.
Patients without symptoms would rarely be referred to sleep clinics, since asymptomatic patients would not be motivated to undergo time-consuming diagnostic measurements. Evidently, women have been rarely referred, especially premenopausal women, so the sleep clinic literature has reported an extraordinary predominance of SAS among men, which is inconsistent with the population samples. Thus, preconceptions of SAS may have preselected clinic patients with a clustering of symptoms that we have been unable to confirm as a cluster in the population . Our study is the first to be performed in a representative sample of the general population in which the prevalence of SDB and SAS was studied in subject aged between 50 and 70 years, in whom a complete history was taken and a general exploration, lung function test and NTRRV were carried out. An important fact to mention is that although there are no differences between men and women with regard to the presence of SDB in this age group, SAS was found to be more prevalent in male patients. Other important finding are that SDB is associated with arterial hypertension and pharynx abnormalities. However, no association was found between lung function parameter and SDB.
We conclude that SDB is common among 50- to 70-year-olds. The prevalence of SDB was 28.9% and the prevalence of SAS was 6.8%. This high prevalence of SDB and SAS strengthens the argument that SAS is a widespread population health problem.
This study was made possible thanks to a grant from the Health Research Fund (FIS 93/1310).
Carlos Zamarrón Sanz
Magdalena 4-4oA, Milladoiro
E–15895, La Coruña (Spain)
Tel. +34 981 540036, Fax +34 981 540164
Received: Received: February 16, 1998
Accepted after revision: January 4, 1999
Number of Print Pages : 6
Number of Figures : 0, Number of Tables : 4, Number of References : 24
Respiration (International Review of Thoracic Diseases)
Founded 1944 as ‘Schweizerische Zeitschrift für Tuberkulose und Pneumonologie’ by E. Bachmann, M. Gilbert, F. Häberlin, W. Löffler, P. Steiner and E. Uehlinger, continued 1962–1967 as ‘Medicina Thoracalis’
Vol. 66, No. 4, Year 1999 (Cover Date: July-August 1999)
Journal Editor: C.T. Bolliger, Cape Town
ISSN: 0025–7931 (print), 1423–0356 (Online)
For additional information: http://www.karger.com/journals/res