Patients with sickle cell disease have been documented to be particularly susceptible to osteoarticular infections. Controversy exists concerning the bacteriology, etiology, and clinical presentation in differentiating osteoarticular infections from bone infarct. We retrospectively reviewed all cases from our institution over the past 22 years of osteoarticular infections in children who carry the diagnosis of sickle cell disease. Two thousand consecutive patient charts of children enrolled in the Pediatric Sickle Cell Clinic of our institution between 1973 and 1995 were evaluated. There were 14 cases of bone or joint infections (10 osteomyelitis, four septic arthritis). There was one case of multicentric osteomyelitis and one case of meningitis complicating the septic arthritis. There were nine male and five female patients with ages ranging from 6 months to 17 years (mean, 8.0). All patients were noted to have hemoglobin SS. The predominant presenting symptoms were pain (79% of cases) and swelling (71% of cases). The most frequent physical findings were fever >38.2 degrees C (71% of cases) and tenderness (86% of cases). Ninety-three percent of the children had a white blood count exceeding 15,000/mm3 (range, 7,900-32,300). Westergren sedimentation rates ranged from 14 to 89 mm/h with 93% of the children exceeding the normal value in our hospital. Cultures were positive in 75% of tissue biopsies, 58% of the blood cultures, and 70% of the bone or joint aspirates. The most common offending organism found in osteomyelitis was Salmonella (eight of 10 cases); however, no predominant organism found was identified in cases of septic arthritis. Radiographs and bone scans were of limited value in the differential diagnosis between osteoarticular infections and bone infarction. Early diagnosis and treatment of osteoarticular infections is key to satisfactory outcome. This study suggests that an ill-appearing patient with a fever >38.2 degrees C, pain, and swelling should prompt the physician to aspirate or biopsy the area and not rely on diagnostic studies that we found to be unreliable.