OBJECTIVETo explore the correlation between the dosage of corticosteroid, time of onset and incidence of osteonecrosis (ON) in patients with SARS.

METHODSFrom July 2003 to January 2004, general survey carried out for ON in 551 patients with SARS. Five hundred and fifty-one patients except 12 were administrated by corticosteroid from 80 mg to 30 000 mg. The age of patients was (33 +/- 9) years old ranging from 19 to 59 years old. One hundred and thirty-one were male, and four hundred and twenty were female. MRI and X-ray film were taken in all patients including both hips, knees, shoulders, ankles and wrists. CT scan was taken in partial patients. Common classification system were used for staging of hip (ARCO), knee (Lotka) and shoulder (Cruess). Independent test, rank-sum test and multiple factor logistic regression analysis were used for statistical analysis.

RESULTSNo osteonecrosis was detected in 12 patients without corticosteroid. Osteonecrosis was detected in 176 patients (32.7 percent) among 539 patients. There were ON of femoral head in 130 cases (210 hips), ON of knee in 98 cases (130 knees), ON of humeral head in 21 cases (36 shoulders), ON of talus and calcaneus in 16 cases (26 ankles), ON of scaphoid and lunate in 11 cases (17 wrists), ON of patella in 3 cases (4 patella), ON of ilium in 1 case and bone infarction (femur, tibia) in 18 cases. One hundred and nineteen cases (195 hips) with ONFH were in stage I (IA 45 hips, IB 77 hips, IC 73 hips). Eleven cases (15 hips) were in stage II. All osteonecrosis of the knee and humoral head was stage I. Thirty-four patients with ON had one joint affected, 45 patients had 2 joints, 93 patients had more than 3 joints. The dosage of corticosteroid was (5842 +/- 4988) mg in ON group and (2719 +/- 2571) mg in non-ON group (P < 0.0001). The duration of steroid was (38 +/- 17) d in ON group and (27 +/- 15) d in non-ON group (P < 0.01). The dosage of pulse treatment was (340 +/- 207) mg/d in ON group and (211 +/- 160) mg/d in non-ON group (P < 0.01). The duration of pulse treatment was (28 +/- 13) d in ON group and (18 +/- 11) d in non-ON group (P < 0.01). All patients with ON were detected within 6 months from administration.

CONCLUSIONAbout one-third patients with SARS who were treated with a high dose of corticosteroid occurred osteonecrosis. ON is frequently multiple focuses. The actual time of onset of ON is early of steroid used. MRI is golden standard for early diagnosis of ON. The patients who were treated with a high dose of corticosteroid should be inspected initially by MRI.