Introduction Consistent evidence suggests that obstructive sleep apnoea (OSA) is associated with increased cardiovascular risk. However, it is unclear whether OSA is underdiagnosed in the cardiology outpatient setting. In the present study, we prospectively evaluated the potential underdiagnosis of OSA in several subspecialties from a tertiary cardiology university hospital.
Methods Consecutive outpatients from five subspecialties (hypertension, coronary, arrhythmia, heart failure (HF), valvular heart disease) were studied. We performed anthropometric measurements, assessed the risk of OSA using the Berlin Questionnaire and evaluated the prior diagnosis and treatment for OSA. In a subset of patients randomly selected, we performed portable sleep monitoring to objectively evaluate the presence of OSA (defined by an apnoea–hypopnoea index ≥15 events/h of sleep).
Results We evaluated 500 patients (100 from each subspecialty). The mean age and body mass index (BMI) were 59±13 years and 28.2±5.3 kg/m2, respectively. We found that 51.6% (258 patients) had a high risk for OSA (Berlin Questionnaire). However, only 13 (3.1%) of these patients had a previous diagnosis of OSA. Of those, only six patients were receiving specific OSA treatment. Fifty patients (10 from each specialty) participated in sleep studies. No differences were found in patients who underwent sleep monitoring and those who did not. We found a high frequency of OSA (66%), varying from 50% (hypertension group) to 80% (HF group).
Conclusions Despite significant scientific evidence pointing to OSA as an emerging cardiovascular risk factor, OSA is still underdiagnosed in several cardiology subspecialties.
Statistics from Altmetric.com
Obstructive sleep apnoea (OSA) is characterised by partial or total upper airway obstruction during sleep, leading to increased negative intrathoracic pressure, intermittent hypoxia and sleep fragmentation.1 OSA is a common condition worldwide, mainly explained by the ongoing obesity epidemic. The most recent analysis of the Wisconsin Cohort indicates that one-third of sleep studies report some degree of OSA (apnoea–hypopnoea index ≥5 events/h of sleep).2 This scenario is similar in Europe3 and Latin America.4
Obesity is a major risk factor for OSA.5 Excessive weight is frequently observed in patients with cardiovascular disease.6 The pandemic of obesity may partially explain the OSA burden in patients with cardiovascular diseases. Therefore, it is reasonable to expect a high frequency of OSA in the cardiology setting. Consistent evidence from the literature suggests that OSA is more than a common condition and is associated with increased cardiovascular risk independently of obesity7–10 and that the main treatment for OSA, namely continuous positive airway pressure (CPAP), reduces cardiovascular events.11–14 Despite all these evidences, it is possible that OSA is frequently underdiagnosed by cardiologists and consequently undertreated. For instance, the vast majority of reports addressing the impact of obesity on cardiovascular diseases completely ignored OSA as a potential confounding factor.5 Previous studies that evaluated the underdiagnosis of OSA were frequently limited to a specific condition, such as hypertension.15–18 In order to evaluate the potential underdiagnosis of OSA in the cardiology setting, we evaluated the presence of OSA in a large sample of consecutive outpatients from a tertiary cardiology university hospital.
This study was approved by our local Ethics Committee (# 3795/12/051), and all participants gave a written informed consent. During 6 months of recruitment, we targeted to evaluate 500 consecutive outpatients undergoing regular treatment at the Heart Institute (InCor), University of São Paulo Medical School. These patients came from five subspecialties of the Heart Institute (100 patients each): (1) hypertension, (2) coronary, (3) arrhythmia, (4) heart failure (HF) and (5) valvular heart disease. We excluded patients who refused to participate as well as patients with significant cognitive impairment. All participants underwent clinical evaluation by the same investigator (LEC), which included demographic data, anthropometric measurements, questionnaires for assessing the risk of OSA (Berlin Questionnaire)18 and excessive daytime sleepiness (Epworth Sleepiness Scale).19 Briefly, the Berlin Questionnaire evaluates the risk of OSA on the basis of the responses in three symptom categories. In category 1, high risk is defined as persistent symptoms (>3–4 times/week) for ≥2 questions about snoring. In category 2, high risk is defined as persistent (>3–4 times/week) daytime tiredness or fatigue. In category 3, high risk is defined as a history of high blood pressure or body mass index (BMI) >30 kg/m2. High risk of OSA is defined when ≥2 symptom categories are positive. In the Epworth Sleepiness Scale, we determined the general level of daytime sleepiness by having patients rate the likelihood of dozing during eight different daytime situations. The scale ranges from 0 to 24, and scores >10 were considered associated with excessive daytime sleepiness.
In addition, we evaluated the awareness of OSA in high-risk individuals as well as specific treatment in those with a high risk for OSA. In addition, we randomly selected 50 patients (10 patients from each specialty) to undergo portable sleep monitoring (ApneaLink Plus, ResMed, ResMed, Sydney, Australia). This is a simple, automated two-channel (oximetry and nasal pressure) screening device. ApneaLink Plus is a validated tool for diagnosing OSA that is used in patients with cardiovascular disease in the ongoing SAVE trial.20 In our study, hypopnoea was defined as 50% airflow lasting ≥10 s, associated with an oxygen desaturation of >3%. Apnoea was defined as a drop in the peak signal excursion of airflow by ≥90% of pre-event baseline lasting ≥10 s.21 OSA was defined by an apnoea–hypopnoea index ≥15 events/h of sleep.
The data were analysed using the Statistical Package for Social Sciences, V.10.0, statistical software (SPSS, Chicago, Illinois, USA). The sample size calculation was estimated based on the fact that these five subspecialties together are following around 30 000 outpatients in our Institute. Our primary goal was to evaluate the overall underdiagnosis of OSA in the cardiology setting (and not by specific diseases). We estimated that 50% of them presented a high risk for OSA. Our sample size of 500 patients allowed us to have a 95% CI of 45.7% to 54.3%.
Quantitative variables are expressed as the mean±SD. A comparison of continuous variables between patients with and without OSA was performed using the Student's t test or Wilcoxon test, as appropriate. Categorical variables were expressed as frequency distribution and were compared using the χ2 test or Fisher’s exact test. Frequency comparisons between the five different cardiology subspecialties were analysed by logistic regression analysis. For all these procedures, a significance level of 5% was adopted.
We initially selected 510 patients from the aforementioned cardiology subspecialties. We excluded five who declined to participate and five who had provided incomplete responses. Therefore, the total study sample included 500 patients. Table 1 lists the demographic, anthropometric and clinical characteristics of the participants. Overall, the sample included middle-aged and overweight patients with a significant percentage of comorbidities, including hypertension, diabetes and obesity. Half of them were men. We found that 51.6% had a high risk for OSA as evaluated by the Berlin Questionnaire. Excessive daytime sleepiness (Epworth Scale>10) was present in 38%. Patients at high risk for OSA were more obese and had a greater percentage of hypertension and diabetes than patients with a low risk for OSA. The high risk for OSA group also had higher scores on the Epworth Sleepiness Scale and a higher percentage of excessive daytime sleepiness (table 1).
Despite the high frequency of high risk for OSA (258 of 500 patients), only 3.1% (13 patients) had a previous diagnosis of OSA. Of those, only 46% (six patients) were receiving specific OSA treatment (CPAP: n=4 and mandibular advancement splints: n=2). Figure 1 shows the frequency of high risk of OSA, excessive daytime sleepiness and previous OSA diagnosis in the total population as well as in the five cardiology subspecialties.
Fifty randomly selected patients underwent sleep studies. Only two patients presented examination failure and both of them agreed to repeat the sleep study. Table 2 shows the characteristics of participants who underwent or did not undergo sleep monitoring. There was no significant difference in several variables, suggesting that this subset of patients was representative of the total sample studied. Considering the whole group, the mean apnoea–hypopnoea index was 19.4±10.6 events/h (no OSA group: 7.7±3.3; OSA group: 25.4±7.6 events/h; p<0.01). Data from sleep studies confirmed a high frequency of OSA in the total sample and in the five cardiology subspecialties (figure 2). The frequency of excessive daytime sleepiness was not as high as that of OSA diagnosis. Interestingly enough, no single patient from this subsample had a previous diagnosis of OSA.
To the best of our knowledge, our study is the first to systematically explore the potential underdiagnosis of OSA in a large sample of patients undergoing treatment in several cardiology subspecialties from a tertiary cardiology hospital. We found that more than 50% of patients with several cardiovascular diseases receiving regular treatment were at high risk for OSA, but the vast majority of them were unaware of this possibility. These findings were further confirmed in a representative subset of patients who underwent objective sleep studies. As a consequence, patients had not been treated. Even in those with a previous OSA diagnosis, approximately half of them were receiving specific OSA treatment. Taken together, our results support the notion that OSA is still underdiagnosed and consequently undertreated in the cardiology outpatient setting.
OSA underdiagnosis has been previously explored in the general population. Kapur et al22 evaluated data on medical disorders, symptoms of sleep disorders and cardiovascular risk factors in several US communities. Participants were classified into three groups: those with a self-reported physician diagnosis of OSA, those with self-reported physician-diagnosed and physician-treated OSA and those reporting both frequent snoring and daytime sleepiness (two-symptom group). In an adjusted logistic model, only male sex and BMI were independently associated with physician-diagnosed and physician-treated OSA. They concluded that disparities (especially in women and in those with lower BMI) exist between recognition rates for OSA and the estimated prevalence by symptom reports across the USA.22 In the cardiology scenario, the evidence specifically addressing OSA underdiagnosis is scanty. As previously mentioned, previous investigations were limited by studying one specific condition. One interesting study conducted by Konecny et al23 reviewed the medical records of 798 consecutive patients hospitalised with a diagnosis of acute myocardial infarction in a tertiary cardiology centre.23 Electronic records, including admission and discharge notes, were searched for diagnosed or suspected OSA during myocardial infarction hospitalisation. Only 12% of patients had a previous or suspected OSA. They further invited 74 patients who were hospitalised with acute myocardial infarction to undergo a formal sleep study. Fifty-one patients (69%) had at least mild OSA and 41% had moderate to severe OSA. More recently, McCormack et al24 used another screening questionnaire for OSA (STOP-BANG) in 135 consecutive patients admitted with acute myocardial infarction in a tertiary referral centre. A STOP-BANG score suggestive of OSA was present in 100 (74%) patients. In patients with HF, an analysis from Medicare beneficiaries showed that comorbid sleep apnoea was clinically suspected in only 4%.25 Of these, 553 (2% of the total cohort) received sleep apnoea testing and 545 received treatment. After adjustment for age, sex and comorbidities, subjects with HF who were tested, diagnosed and treated for sleep apnoea had a better 2-year survival rate compared with subjects with HF who were not tested.25 Similarly, among subjects who were tested and diagnosed, those who were treated had a better 2-year survival rate than those who were not treated.25 Even considering that OSA and central sleep apnoea are common in patients with HF, these results clearly suggest that active OSA screening and treatment might have significant cardiovascular impact. Our study extends previous findings by exploring the potential underdiagnosis of OSA in consecutive outpatients from five cardiology subspecialties. All subspecialties presented a significant proportion of high risk for OSA. Consequently, most of these patients remained untreated.
Based on the growing recent evidence that OSA is an emerging cardiovascular risk factor,5 we strongly believe that it is crucial to have screening tools for OSA in the cardiology setting. The Berlin Questionnaire and the Epworth Sleepiness Scale are easy to apply and therefore may not interrupt clinic flow significantly. In order to confirm clinical suspicions and determine the severity of OSA, we will need formal sleep studies. Patient access to appropriate investigation and treatment for OSA is clearly restricted. Full polysomnography is costly and the growing demand in clinical practice has led to unacceptable waiting periods. In this sense, portable sleep monitoring has gained increasing interest, especially because it is simple and can be performed at home. However, it is important to mention that interpretation of portable sleep monitors requires sleep experts that also indicate appropriate treatments for OSA. Therefore, cardiologists should be trained to screen for OSA and refer to sleep specialists for appropriate management of these patients.
The precise reasons for this underdiagnosis in the cardiology setting are probably multiple and may include lack of knowledge about sleep medicine, low frequency of daytime sleepiness (as observed in our study), obstacles for diagnosis and access to treatment and reluctance to consider OSA as a cardiovascular risk factor by the cardiologists. The lack of daytime symptoms in a significant proportion of patients with cardiovascular diseases is probably a key point in the underestimation of the diagnosis and possibly to the cardiovascular impact of OSA. Another important issue is that the demonstrated impact of OSA treatment with CPAP on cardiovascular events is still based on observational studies.11–14 Lack of randomised clinical trials may help explain the relatively low awareness of OSA. This important issue should be explored in future investigations.
There are strengths to the present study: We included outpatients receiving regular treatment for several cardiovascular diseases. Very few exclusion criteria were applied, which may mimic a ‘real-life’ condition in the cardiology scenario. The following limitations should be addressed: (1) Because of logistic and costs issues, we performed screening questionnaires for all participants and a portable sleep study for a subset of them. Although it was a representative sample (table 2), we did not perform full polysomnography, considered the gold standard for OSA diagnosis. Because portable monitors may underestimate OSA severity, we may argue that the frequency and severity of OSA may be even higher than what we reported. (2) Because portable monitoring was performed in only 10% of the studied population, the current study was not designed to check the sensitivity and specificity of the Berlin Questionnaire for screening OSA in all cardiovascular conditions studied. The Berlin Questionnaire seems to be useful in patients with hypertension,15 but this may not be true for other cardiovascular diseases. Despite this limitation, our sleep study data from a representative subsample of patients showed a huge frequency of moderate to severe OSA. (3) We may argue that the present results may be valid only for this tertiary cardiology university hospital. The organisation of subspecialties reflects the structure of our tertiary hospital and may not be universal. In addition, several patients who attended one subspecialty may also have the typical clinical conditions of others. However, this fact reflects ‘real-life’ clinical practice. If we selected ‘pure’ patients with valvular diseases, HF and so on, we may limit the external validation of our findings. Although we agree that the present results may not be applied to secondary and primary care, the low awareness of OSA observed in three different studies from the USA and UK, including patients with myocardial infarction23 ,24 and HF25 suggests that the present results may be applicable to other tertiary cardiology hospitals.
We conclude that despite significant scientific evidence suggesting that OSA is independently associated with increased cardiovascular risk, OSA is still underdiagnosed in several cardiology subspecialties. Because CPAP treatment seems to reduce cardiovascular events,11–14 exploring potential causes and improving recognition and treatment of OSA in cardiology might have benefits in the reduction of cardiovascular morbidity and mortality observed in patients with OSA.
What is already known on this subject?
Consistent evidence suggests that obstructive sleep apnoea (OSA) is independently associated with increased cardiovascular risk.
What might this study add?
Despite significant scientific evidence pointing to OSA as an emerging cardiovascular risk factor, OSA is still underdiagnosed in several cardiology subspecialties.
How might this impact on clinical practice?
Exploring potential causes and improving recognition as well as OSA treatment in cardiology might have benefits in the reduction of cardiovascular morbidity and mortality.
Contributors LEC, CHGU, RRH: data acquisition and analysis; LAB and GL-F: data analysis and critical review of the manuscript; LFD: study concept and design; data analysis and drafting of the manuscript. All authors: final approval of the version published.
Funding LEC is supported by FAPESP (#2012/10324-5); LFD is supported by a Young Investigator Award, FAPESP (# 2012/02953-2).
Competing interests None declared.
Patient consent Obtained.
Ethics approval Heart Institute Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.