Long-term maintenance steroid therapy (MST) is frequently required for repeated relapses of cryptogenic organizing pneumonia (COP); however, the optimal minimal dose has not been clarified. Therefore, this study evaluated the minimal MST dose required to prevent repeated relapses and identify relapse predictors.

We retrospectively reviewed the medical records of patients with steroid-treated COP and compared background factors between the non-relapse and relapse groups. We also reviewed the treatment course in the relapse group and determined the minimal effective steroid dose based on the MST dose at relapse events and the current relapse prevention dose.

In total, 48 patients were identified, including 27 (56%) in the non-relapse group and 21 (44%) in the relapse group. Receiver operating characteristic curve analysis identified prednisolone at 5 mg/day as the optimal cut-off value in the relapse group. Relapse-free time in patients with relapsed COP was significantly longer in the MST dose ≥5 mg/day group than in the <5 mg/day group (log-rank P = 0.003; hazard ratio, 0.19; 95% confidence interval [CI], 0.04–0.60). Multivariate logistic regression analysis demonstrated that a high eosinophil percentage and CD4/CD8 ratio in bronchoalveolar lavage fluid (BALF) were predictors of relapse (odds ratio [OR], 1.12; 95% CI, 1.02–1.23; P = 0.008 and OR, 3.87; 95% CI, 1.29–11.6; P = 0.008, respectively).

Our results indicate that 5 mg/day of prednisolone may be the minimal effective dose for preventing repeated relapses, and a high BALF eosinophil percentage and CD4/CD8 ratio are independent predictors of relapse.