Accuracy of precordial percussion in detecting cardiomegaly

To assess the value of precordial percussion in detecting cardiomegaly, and to compare it with palpation of the apical impulse.

Descriptive study.

Hospitals and clinics of a university medical center.

Light indirect percussion of the precordium was performed on 72 inpatients and 28 outpatients. All patients had a posteroanterior radiograph of the chest. Percussors were unaware of the clinical history and of chest roentgenogram results.

Thirty-six patients (36%) had cardiomegaly, defined as a cardiothoracic ratio of greater than 0.5 on chest roentgenogram. The cardiothoracic ratio was significantly correlated with percussion distance from the midsternal line in the left fourth (r = 0.35, p less than 0.0006), fifth (r = 0.65, p less than 0.00001), and sixth (r = 0.40, p less than 0.0001) intercostal spaces. After adjustment for clinical symptoms and systolic and diastolic blood pressures, percussion distance in the left fifth intercostal space remained a significant independent predictor of the cardiothoracic ratio. Percussion distance in the left fifth interspace discriminated cardiomegaly with a receiver-operating characteristic (ROC) area of 0.95. Percussion dullness more than 10.5 cm from the midsternal line in the left fifth interspace had a sensitivity of 94.4% (95% confidence interval [CI], 79.9% to 99.0%) and a specificity of 67.2% (CI, 54.2% to 78.1%). Distance of the apical impulse from the midsternal line discriminated with an ROC area of 0.95, but an impulse was palpated in only 40% of cases.

Percussion in the left fifth intercostal space accurately discriminates patients with and without cardiomegaly, and adds information beyond that obtainable from the history and other parts of the physical examination.