Immunocompromised pediatric patients with HIV infection and other immunosuppressive disorders are susceptible to Pneumocystis carinii pneumonia. Bronchoscopy with bronchoalveolar lavage (BAL) may be technically more difficult in pediatric patients. Therefore, an alternative, yet safe and reliable, diagnostic method had to be determined. We have continued our prospective assessment of the utility of sputum induction to diagnose P. carinii pneumonia in children. An ultrasonic nebulization technique using hypertonic sodium chloride with expectoration, oropharyngeal, or nasopharyngeal suctioning has been employed. The ongoing project's major goal was to further improve the sampling technique in order to increase sensitivity. The medical records of 121 patients aged 4 months to 18 years who underwent sputum induction from January 1990 to December 1994 were reviewed. Of 196 sputum inductions in 121 patients; 28 were positive for P. carinii, 167 were negative, and one collection was insufficient. Of the 167 negative sputum inductions, 57 were followed by BAL, of which 10 were P. carinii positive. The remaining 157 patient encounters that were negative for P. carinii by sputum induction or BAL were followed for 1 month: 150 remained negative for P. carinii, 2 became P. carinii positive, and 5 had no information available. The overall sensitivity of sputum induction was 70 percent; the negative predictive value was 93 percent. We concluded that sputum induction is effective and fairly sensitive in this group of pediatric patients and should be the initial diagnostic procedure in immunosuppressed pediatric patients suspected of having P. carinii pneumonia. This project has extended the clinical applicability of a diagnostic procedure, developed largely by the National Institutes of Health, which now offers many centers the option of a test that is less expensive, less invasive, and less associated with complications than bronchoscopy.
