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


Precordial percussion for detecting cardiomegaly

ACP J Club. 1992 Jan-Feb;116:20. doi:10.7326/ACPJC-1992-116-1-020

Source Citation

Heckerling PS, Wiener SL, Moses VK, et al. Accuracy of precordial percussion in detecting cardiomegaly. Am J Med. 1991 Oct;91:328-34.



To assess the accuracy of precordial percussion and apical impulse palpation in detecting cardiomegaly radiographically.


Physical measurements of cardiomegaly were compared with roentgenographic evidence.


Medical wards and outpatient clinics of a university hospital.


Patients had posteroanterior chest radiographs as part of routine admission or screening procedures. Patients were excluded (n = 15) if they had signs of chronic obstructive pulmonary disease, thoracic deformities, or previous thoracic surgery; or if the roentgenogram showed inadequate inspiration. The 72 inpatients and 28 outpatients included had a mean age of 52 years; 27% were obese. Of 97 patients providing historical information, 49% had exertional dyspnea, 29% had edema, 25% had orthopnea, 14% had resting dyspnea, and 16% had angina pectoris.

Description of test and diagnostic standard

Supine patients were percussed and palpated. Examiners used light, indirect percussion to define the heart borders in the second through fifth (right) and sixth (left) intercostal spaces. Dullness was distinguished by changes in sound and vibration. Distances were measured from the midsternal line.

The diagnostic standard was the radiographic cardiothoracic ratio (CTR), the ratio of the maximum diameter of the cardiac silhouette to the maximum inner diameter of the thorax. Radiographs were read independently of the physical examinations, which were blinded to radiographic results. The cut point for cardiomegaly was a CTR > 0.5.

Main results

36 patients had a CTR > 0.5. Percussion distance (PD) in the left fifth intercostal space discriminated cardiomegaly most accurately (area under the ROC curve, 0.95). With PD > 10.5 cm or > 11 cm along the fifth intercostal space, sensitivities for cardiomegaly were 94% (95% CI, 80% to 99%) and 89% (CI, 73% to 96%), respectively, and specificities were 67% (CI, 54% to 78%) and 91% (CI, 80% to 96%). After adjustment for blood pressure and clinical symptoms, PD remained significantly correlated with CTR (r = 0.68).

2 independent examiners had 57% agreement beyond chance (kappa; CI, 18% to 96%) in their classification, by percussion in the fifth interspace, of a 14-patient subset. Distance of the apical impulse from the midsternum discriminated cardiomegaly with an ROC area of 0.95, but an impulse was palpated in only 40 patients.


Percussion in the left fifth intercostal space accurately identified patients with cardiomegaly detected radiographically. Apical impulse palpation appeared to be accurate but could be measured in only 40% of patients.

Source of funding: Not stated.

Address for article reprint: Dr. P.S. Heckerling, Department of Medicine, University of Illinois, Box 6998 M/C 787, Chicago, IL 60680.


A sure sign of advancing age is when an intern, poring over an ancient medical record, asks you, "What does the abbreviation 'LBCD' mean?," and you know the answer. (LBCD: Archaic term for "left border of cardiac dullness," as derived by the percussion technique. Cardiac percussion was a discredited technique until recently.)

Heckerling and colleagues suggest that cardiac percussion is another "old" test with "new" value. Through the use of signal detection theory methodology, they have determined that percussion of the heart accurately discriminates between patients with and without cardiac enlargement. Because the left ventricular apical impulse often is not palpable, cardiac percussion is a useful adjunct in making this determination.

The next issue is one of practicality. Does the value of the information acquired through the use of cardiac percussion justify the time expended, or should this maneuver be reserved for cases where the yield from the test should be high? In a healthy population, the inference from Bayes theorem is that the test would provide relatively little information, because the prevalence of cardiomegaly is low. In an elderly cardiac population, the test would yield more information (1). Those who endeavor to define a "minimalist physical examination" (2), where each maneuver can be justified on the basis of cost (as expressed in units of time) and benefit (as expressed in units of information obtained), need to address this question very exactly.

Heckerling and colleagues reinforce the impression that, like all diagnostic tests, cardiac percussion is useful if one understands its characteristics and the populations in which it is applied. Knowing these facts can facilitate the personal decision about whether and when to include the maneuver in one's own repertoire.

Thomas A. Parrino, MD
Brown University Providence, Rhode Island