Selected features of bedside clinical evaluation were predictive of obstructive airways disease
ACP J Club. 1993 July-Aug;119:24. doi:10.7326/ACPJC-1993-119-1-024
Holleman DR Jr, Simel DL, Goldberg JS. Diagnosis of obstructive airways disease from the clinical examination. J Gen Intern Med. 1993 Feb;8:63-8.
To evaluate the operating characteristics of selected items in the pulmonary history and physical examination and of bedside measures of airflow obstruction and to develop and validate a predictive model for assessing airflow obstruction.
A blinded comparison and validation of clinical assessment with spirometry.
Preoperative assessment clinic in a Veterans Affairs hospital.
Data from 164 consecutive men (mean age 66 y) were used to develop a model for diagnosis with validation carried out in 84 other men.
Description of tests and diagnostic standard
An internist and anesthesiologist examined patients using a standardized history (smoking history; symptoms of chronic bronchitis; current cough, wheezing, and dyspnea) and a standardized physical examination (wheezing, rales, cough, subxyphoid cardiac impulse, and breath sounds). Physicians documented an overall clinical impression and made a bedside assessment of pulmonary airflow (forced vital capacity [FVC] time and peak expiratory flow rate). A trained respiratory technician, blinded to patient results, did standard spirometry (forced expiratory volume in 1 second [FEV1], FVC, and FEV1 to FVC ratio [FEV1/FVC]). Diagnosis of obstructive airways disease used both low FEV1 and FEV1/FVC.
44% of the study sample and 42% of the validation sample had airflow obstruction. Sensitivity and specificity for history items were smoking (94%, 24%); dyspnea (82%, 33%); cough (51%, 71%); wheezing (51%, 84%); chronic bronchitis (42%, 89%); and orthopnea (19%, 88%). Sensitivity and specificity for physical examination items were decreased breath sounds (29%, 85%); wheezing (14%, 99%); cough (14%, 93%); subxyphoid impulse (4%, 99%); and rales (1%, 99%). Overall clinical impression correlated with disease (likelihood ratio [LR] of 0.42 for no disease and 4.2 [95% CI 2.2 to 8.0] for moderate or severe disease). Independent predictors of airflow obstruction were number of years of smoking and patient-reported wheezing (positive LR 3.1, CI 1.9 to 5.2; negative LR 0.58, CI 0.45 to 0.75) and auscultated wheezing (positive LR 12, CI 1.7 to 98; negative LR 0.87, CI 0.79 to 0.96). For bedside measures of airflow obstruction, peak expiratory flow rate was the most accurate (area under the receiver operating characteristic curve [ROC] 0.81), whereas forced expiratory time added little to the predictive value of history and physical examination (increment in ROC area 0.03).
Selected features of bedside clinical evaluation were predictive of obstructive airways disease in patients examined in a preoperative assessment clinic.
Source of funding: In part, A. W. Mellon Foundation.
For article reprint: Dr. D.R. Holleman, Medical Service (111K), Veterans Affairs Medical Center, 2250 Leestown Road, Lexington, KY 40511, USA. FAX 606-258-1197.
For a more recent review of the evidence for this question, see Holleman DR Jr, Simel DL. Does the clinical examination predict airflow limitation? JAMA. 1995;273:313-19.
Holleman and colleagues have attempted, using an elegant mathematical analysis (including ROC curves), to define the characteristics that can enable clinicians to best diagnose obstructive airways disease without resorting to pulmonary function tests. The authors developed a nomogram to predict the probability of airways obstruction from years of cigarette smoking, peak flow rates, and objective and subjective evidence of wheezing. Peak flow, however, is dependent on age, sex, and height; therefore, broad use of the nomogram appears questionable. Additionally, wheezing occurs in many diseases other than obstructive airways diseases.
Disturbingly, there was only marginal interobserver agreement on many items of history and physical examination. It was not surprising, therefore, that smoking and the objective peak-flow rate were the best determinants of obstructive airways disease.
Should a patient with shortness of breath, wheezing, or a 50-year smoking history be cleared for surgery or begin treatment without pulmonary function studies? Do we need a nomogram to tell us what we already know from our clinical judgment? Clausen (1) agrees that smoking is a major risk factor but suggests that even simple spirometry will not clearly define the "type" and severity of obstructive pulmonary disease. Although severe disease is easy to diagnose, mild disease is not.
The mathematical approach used in this study may be useful in clinical decision making in complicated situations. I feel, however, that this report adds little to the clinician's skills to diagnose and treat obstructive airways disease.
George J. Heymach, PhD, MD
University of PittsburghPittsburgh, Pennsylvania, USA
We are pleased that science confirmed the value of clinical judgment. Clinicians who question peak flows can substitute wheezing in our nomogram. Students, residents, and even experienced clinicians should compare their current qualitative assessment to our more quantitative approach. We feel spirometry should supplement clinician assessment, not replace it.
Donald R. Holleman, Jr., MD