Current issues of ACP Journal Club are published in Annals of Internal Medicine


Diagnosis

Sleep apnea screening using history and physical examination

ACP J Club. 1992 Jan-Feb;116:22. doi:10.7326/ACPJC-1992-116-1-022


Source Citation

Viner S, Szalai JP, Hoffstein V. Are history and physical examination a good screening test for sleep apnea? Ann Intern Med. Sep


Abstract

Objective

To evaluate clinical history, pharyngeal examination, and overall clinical impression as screening tests for sleep apnea.

Design

Blinded comparison of history, physical examination, and clinical impression with the results of the diagnostic standard, nocturnal polysomnography.

Setting

Sleep clinic at a tertiary care medical center.

Patients

410 consecutive patients referred because of suspected sleep apnea syndrome. The mean age was 47 years; 338 were men.

Description of tests and diagnostic standard

Before nocturnal polysomnography, 1 physician interviewed all patients for snoring, restless sleep, nocturnal choking, daytime sleepiness and drowsiness, morning symptoms, and impotence. The pharynx was examined and considered abnormal if it was narrow and small or if the uvula or tonsils were large or obstructive. With this information the clinician predicted the presence or absence of the sleep apnea syndrome. Nocturnal polysomnography (chest wall and abdominal movements, oronasal flow and O2 saturation, electrocardiogram, electro-oculogram, electroencephalogram, and electromyogram) was done, and the results were assessed blind to clinical data. Patients with > 10 respiratory events (apneas and hypopneas) per sleeping hour were considered to have the sleep apnea syndrome.

Main outcome measures

Stepwise logistic linear regression was used to develop 2 predictive models for diagnosing sleep apnea, based on clinical impression or on the previously listed clinical symptoms as well as age, sex, and body mass.

Main results

190 patients (46%) were apneic. Apneic patients, compared with nonapneic patients, were more likely to snore (P = 0.04) or choke at night (P = 0.02), have an abnormal pharyngeal examination (P = 0.002), be male (P = 0.02), be older (P < 0.001), and have a higher body mass index (P < 0.001). However, when logistic regression was used, only age, body mass index, male sex, and snoring were found to predict sleep apnea and, with these variables as covariates, sensitivity was 28% and specificity was 95% using a cut point of 70% calculated probability of apnea. Using a calculated probability of 20% as the cut point, sensitivity was 94% and specificity was 28%. Subjective clinical assessment had a sensitivity of 52% and specificity of 70%.

Conclusion

Clinical features, alone or in combination, do not reliably predict sleep apnea among patients referred to a sleep clinic with suspected sleep apnea.

Source of funding: Not stated.

Address for article reprint: Dr. V. Hoffstein, St. Michael's Hospital, 30 Bond Street, Toronto, Ontario M5B 1W8.


Commentary

The authors used sophisticated statistical methods to develop a model for screening patients suspected of having sleep apnea. However, regardless of whether the logistic regression model incorporated the clinician's subjective impression or the identified clinical determinants of sleep apnea (age, male sex, body mass index, and snoring), there was insufficient sensitivity or specificity to allow confident prediction of disease. Thus, the practical usefulness of the model is limited.

The study was done among patients who had already been referred to a sleep disorder clinic. It is important to note that the results of this study provide no new guidance for the practicing internist who must make a clinical decision whether or not to refer a patient with suspected sleep apnea for diagnostic polysomnography.

Sleep apnea can be a highly morbid disease, which may negatively affect survival, for which there is effective therapy. It is, therefore, imperative that the disease be diagnosed or excluded in suspected patients. Unfortunately, the results of this study support anecdotal experience that clinical impression is often inaccurate; highly suspect patients may not have sleep apnea whereas the reverse is often true for patients with a less-than-striking presentation. Thus, once sleep apnea is suspected based on the history and physical examination, the diagnosis must be confirmed by nocturnal polysomnography, the present diagnostic standard.

Melvin Lopata, MD
University of Illinois Chicago, Illinois