Objectives: To describe the characteristics of salmonella infections in sickle cell disease and to compare the features of osteomyelitis and those with bacteremia/septicemia without obvious bone changes. To search for risk factors for osteomyelitis, and to draw attention to the frequency and significance of salmonella bacteremia/septicemia. Study design: A retrospective review of all salmonella isolations from the blood, pus, or aspirates during a 22-year period. Setting: The sickle cell clinic at the University Hospital of the West Indies, Kingston, Jamaica. Subjects: Patients with all genotypes of sickle cell disease. Incidence data and the frequency of associated cholelithiasis were derived from the cohort study based on follow-up of all children detected by neonatal screening. Main outcome measures: Osteomyelitis and bacteremia/septicemia. Results: Of 55 patients with salmonella infections, 28 initially had osteomyelitis and 27 had bacteremia/septicemia. Three of the first group later had bacteremias for a total of 30 episodes of bacteremia/septicemia, and 4 of the second group later had osteomyelitis for a total of 32 episodes of osteomyelitis. The incidence of salmonella infection was 8.6% by 15 years, and 96% of infections occurred before the age of 10 years. Preceding episodes of avascular necrosis of bone were more frequent ( p <0.006) in patients with osteomyelitis. Patients with osteomyelitis were not more prone to gallstones. High fever (temperature ≥40° C or 104° F) occurred in 41%, and occasionally marked bone marrow suppression mimicked the aplastic crisis. Twenty Salmonella serotypes were isolated; Salmonella enteritidis accounted for 36% of infections, but no serotype difference occurred between those with osteomyelitis and those with bacteremia/septicemia. There were no deaths in the 32 patients with osteomyelitis, but 7 (23%) of 30 patients with septicemia died. Conclusions: Anti-salmonella prophylaxis requires assessment in the management of bone necrosis. Anti-salmonella agents may be indicated in undiagnosed septic conditions in sickle cell disease pending culture results. (J Pediatr 1997;130:394-9)