A clinical study was conducted in three Chinese community hospitals to investigate the reliability of respiratory rate and various clinical signs in the diagnosis of pneumonia among 54 children less than 5 years of age. Anteroposterior chest film was used as the diagnostic standard. The cutoff criterion for rapid breathing was 50 breaths/minute for infants ages 2 to 11 months and 40/minute in children 1 to 5 years old. Rapid breathing was a better predictor of pneumonia than rales (positive predictive values of 74.5 and 66.9%). Nasal flaring, chest indrawing, stridor and cyanosis of the tongue had predictive values of > 86%, but these clinical signs were observed in only a small proportion of patients. We recommend that village health workers use rapid breathing for diagnosis of pneumonia, rather than auscultation which is difficult and has proved unreliable. Sensitivity, specificity and positive and negative predictive values are presented for seven signs and symptoms of pneumonia.
PIP: During October 1989 to March 1990 pediatricians and radiologists conducted a clinical study in Shunyi, Nanbu, and Shifang County Hospitals, China, of 160 healthy children and 541 children with fever and cough to examine the reliability of respiratory rate and various clinical signs in the diagnosis of radiologically confirmed pneumonia among 54 children under 5 years of age. The mean respiratory rate among children with cough and fever was 50 breaths/minute for infants aged 0-11 months and 40 breaths/minute for children aged 1-5 compared with about 40 breaths/minute and 30 breaths/minute, respectively, for healthy children . The researchers deemed these rates to be the cutoff criterion for rapid breathing. Rapid breathing could better predict pneumonia than rales could (positive predictive values, 74.5% vs. 66.9%). Nasal flaring, chest indrawing, and cyanosis of the tongue had high specificities (86.5%, 92.5%, and 93.5%, respectively). Yet these signs occurred in only a small percentage of the population (25.2% for nasal flaring and 10% for chest indrawing and cyanosis of the tongue). Based on these findings, the researchers call for village health workers to use rapid breathing to diagnose pneumonia rather than ausculatory signs, especially rales. Ausculatory signs are unreliable predictors of pneumonia.