Precision and accuracy of clinical and radiological signs in premature infants at risk of patent ductus arteriosus.

OBJECTIVE

To determine the precision (interobserver agreement) and accuracy (agreement with criterion standard) of clinical and radiological signs in premature infants at risk of patent ductus arteriosus (PDA) with left-to-right shunting.

DESIGN

Masked comparison of clinical and radiological examination with Doppler flow echocardiography (criterion standard).

SETTING

Neonatal intensive care unit.

PATIENTS: One hundred infants with birth weights less than 1750 g were studied once between days 3 and 7 of life. A third of the cohort was intubated at the time of study.

INTERVENTION: Five independent observers noted the presence or absence of an increased pulse volume, an active precordium, a heart murmur, a cardiothoracic ratio greater than 60%, increased pulmonary vascular markings on a concurrent chest x-ray film, and a relative increase of the cardiothoracic ratio compared with that from the previous chest x-ray film. Pulsed and color flow Doppler echocardiography was performed within 4 hours. All 100 tapes were reviewed by a second pediatric cardiologist.

RESULTS

Twenty-three infants had a PDA with left-to-right shunting. The precision of clinical signs was modest, with average kappa values of 0.15 for pulse volume, 0.32 for precordium, and 0.41 for murmur. Pulse quality (43%) and murmur (42%) had the highest mean sensitivities. Corresponding specificities were 74% for pulse volume and 87% for murmur. The combination of a cardiac murmur with an abnormal pulse volume had the highest positive predictive value (77%). The radiological examination did not improve the observers' ability to distinguish between patients with and without PDA.

CONCLUSIONS

The precision and accuracy of clinical and radiological signs of a PDA with left-to-right shunting are unsatisfactory. Therefore, Doppler flow echocardiography is required to diagnose PDA confidently in preterm infants between days 3 and 7 of life.