Acute febrile illnesses developed in 2 children with rheumatoid arthritis and death occurred within 72 hr. Both children were receiving indo-methacin as part of their therapeutic regimen. Physical examination shortly before death did not suggest the impending danger. In 1 patient, postmortem examination showed no definite cause of death. Group A /3-hemolytic streptococcal septicemia appeared to be the immediate cause of death in the other. No comparable experiences were reported in adults. It would seem reasonable to use indomethacin only in children not responsive to other agents, to limit the dose, and to be alert to the possibility of overwhelming infection. Two additional cases of unexpected death in arthritic children have been reported. One was an 8-year-old girl who received indomethacin, 50 mg daily for 1 month. She was discharged from her community hospital 48 hr. prior to her unexpected death. At home she suddenly became acidotic, comatose, and icteric. She was readmitted to the hospital but died in 8 hr. Autopsy revealed large zones of hepatic necrosis with heavy periportal inflammatory cell infiltrates, some membranous glomerulitis and tubular necrosis in the kidneys, and findings of juvenile rheumatoid arthritis. Cultures of the blood showed no growth. The administration of predni-sone had been discontinued 1 month previously. The other patient was an 11-year old boy who developed jaundice and severe gastrointestinal bleeding after receiving indomethacin, 100 mg daily for 2 months and an unknown dose of prednisone. Cultures of the blood showed no growth. Despite therapeutic efforts, he died within 2 weeks. Autopsy showed hepatic necrosis, ascites, and juvenile rheumatoid arthritis.