Despite previously carrying out a first open study of sacroiliac injection of long-acting corticosteroid, it was not possible to evaluate the role of a placebo effect. We therefore performed a double-blind study in 10 patients/13 articulations, suffering painful sacroiliitis. At 1 month, 5/6 sacroiliac joints injected with corticosteroid described a relief of > 70%, in comparison to 0/7 of the placebo group (P < 0.05). Dolometry showed a marked decrease in the corticosteroid group from (mean +/- S.E.M.) 6.8 +/- 0.6 to 1.3 +/- 0.3, and decreases were mild in the placebo group: 7.0 +/- 0.6 to 5.2 +/- 0.5 (P < 0.005). Six of the seven sacroiliac joints of the placebo group and two patients with failure and relapse of the corticosteroid group were reinjected with corticosteroid. At 1 month, 12/14 (85.7%) were assessed as having a good result. Results were still significant at 3 months (62%) and 6 months (58%). Tolerance was good or very good in 86% of the cases, and we did not report any notable complication. This technique is safe and very efficient, and it has to be considered more widely in patients with contraindications or complications with NSAID, or if the medical treatment is unable to control sufficiently the active sacroiliitis.