Objective: To examine the presentation, course, and outcome of pneumococcal bacteremia in children infected with human immunodeficiency virus (HIV).
Methods: A retrospective series of HIV-infected children less than 18 years of age with Streptococcus pneumoniae bacteremia from four urban, tertiary care hospitals was evaluated. The main outcome measures included persistent bacteremia, the development of a focal infection, and death.
Results: Seventy-two episodes of pneumococcal bacteremia were identified in 59 patients. Fifty-four first episodes were included; 26/54 were occult. Mean temperature was 39.8 degrees C. In patients with bacteremia, white blood cells (WBCs) > or = 15,000 and > or = 10,000 had sensitivities of 40% and 75%, respectively. At the time of bacteremia, age >3 years old was associated with a lower mean WBC count compared with episodes occurring in patients <3 years old (11.2 vs 16.1, P < 0.05). Patients with occult bacteremia who were discharged with antibiotics (12 i.m., 7 p.o.) were less likely than patients without antibiotic treatment to have persistent bacteremia at a return visit within 72 hours (0/19 vs 2/5, P < 0.05). No patient with occult bacteremia died, progressed to clinical meningitis, or had other sequelae. Two of fifty-four patients died as a result of their first episode of invasive pneumococcal disease. Both patients who died had meningitis and appeared ill on initial presentation.
Conclusions: Neither a WBC count > or = 15,000 nor > or = 10,000 is a sensitive indicator of pneumococcal bacteremia in HIV-infected children. Empiric antibiotics are useful to decrease the risk of persistent bacteremia. Children infected with HIV who have occult pneumococcal bacteremia appear to do well with appropriate antibiotics. Patients who are afebrile and well appearing on reevaluation may be safely treated as outpatients.