Background: Determination of the etiology of pneumonia in young children is difficult because blood culture, the usual method of diagnosis, is positive in only a small proportion of cases. For this reason vaccine trials that include bacterial pneumonia as an endpoint must be large.
Objectives: To determine whether a diagnostic test based on a polymerase chain reaction could be used as an alternative to conventional blood culture for diagnosis of invasive Haemophilus influenzae type b (Hib) infections in young children investigated during the course of a large vaccine trial.
Methods: DNA was extracted from blood culture supernatants and probed for the presence of Hib DNA with a PCR assay with primers derived from the cap gene locus of Hib. Results of the PCR assay were compared with those obtained by conventional culture techniques.
Results: Blood cultures were obtained from 1544 children with suspected pneumonia, meningitis or septicemia and from 31 healthy control children who were contacts of cases. Blood culture supernatants were tested for Hib DNA in the PCR test. The sensitivity and specificity of a positive PCR test in blood culture supernatant as against culture of Hib from any normally sterile site were 100 and 99%, respectively. Eleven children had positive Hib PCR tests on blood culture supernatants but were negative by culture. In one of these cases Hib was isolated from a lung aspirate and in two other patients H. influenzae strains other than Hib were obtained from the cerebrospinal fluid. Eight of these 11 children were in the control group. When the results of the PCR assay were used to determine vaccine efficacy, a value of 86% was obtained compared with a figure of 95% obtained when conventional culture techniques were used.
Conclusions: An Hib PCR assay on blood culture supernatants proved to be sensitive and specific for the diagnosis of Hib disease in children. The distribution of PCR-positive, culture-negative cases between Hib-vaccinated and control groups paralleled that of culture-positive cases, suggesting that most of these children had been infected with Hib. A trial of a highly efficacious vaccine provides a novel way for evaluating new diagnostic tests for which there is no standard diagnostic test of 100% reliability.