RÉSUMÉ
Objective: To develop a clinical strategy for detection of Congenital heart disease (CHD) in the newborn through a combination of clinical signs and pulse oximetry. Design: Prospective longitudinal study. Setting: Community level hospital in the city of Kochi, Kerala. Participants and interventions: All consecutive newborns between June 2006 and February 2009 were prospectively screened for CHD, 48 hours after birth. The on-site pediatrician performed clinical screening. A study nurse recorded pulse oximetry in a lower extremity; value of <94% was defined as abnormal. Echocardiography was performed on site by a trained research officer. A 6- week clinical follow-up evaluation was done for all. Main outcome measure: Detection of CHD by echocardiography. Results: Of 5487 babies screened, 425 (7.75%) had CHD. 17 (0.31%) had major CHD, two of whom (one ALCAPA and one large VSD) were missed during the initial evaluation. The rest were minor CHD (408 patients, 7.44%), most of which normalized by 6 weeks. On multivariate analysis, murmur, central cyanosis, abnormal precordial pulsations and abnormal pulse oximetry emerged as significant predictors of CHD. The sensitivity of clinical evaluation and pulse oximetry combined was 19% for all CHDs and 20% for major CHD; specificity was 88%. Conclusions: In the community setting of a developing country, clinical evaluation and pulse oximetry after birth had a very low sensitivity for detection of CHD. Though an abnormal screening warrants prompt echocardiography, a 6-week clinical evaluation is recommended to ensure that major CHD is not missed.