Low-dose maintenance steroid treatment could reduce the relapse rate in patients with type 1 autoimmune pancreatitis: a long-term Japanese multicenter analysis of 510 patients

Background: The effect of maintenance steroid treatment (MST) in reducing the risk of relapse in patients with autoimmune pancreatitis (AIP) remains under debate. The aim of this study was to validate the effect of MST on AIP administered in accordance with the 2010 Japanese consensus guidelines. Methods: The clinical data of patients with (n = 510) from 22 high-volume centers in Japan were studied. The primary endpoints were the relapse rates (RRs) in patients administered MST versus those not administered MST. The secondary endpoints were the optimal dose and duration of MST in terms of steroid toxicity and the predictors of relapse. Results: The RRs were 10.0% within 1 year, 25.8% within 3 years and 35.1% within 5 years. The RR in the steroid therapy group reached a plateau at 42.7% at 7 years. In terms of the optimal dosage, the overall RR in the MST 5 mg/day group was 26.1%, which was significantly lower than that in the group which had discontinued steroid therapy (45.2%; p = 0.023) or was receiving MST at 2.5 mg/day (43.4%, p = 0.001). The RRs in the group receiving MST at ≥5 mg/day versus the patient group receiving MST at <5 mg/day were 10.6 vs. 10.3% within 1 year, 23.5 vs. 32.9% within 3 years and 32.2 vs. 41.3% within 5 years, respectively (log-rank, p = 0.028). The best cutoff value of the total steroid dose for serious steroid toxicity was 6405 mg, with a moderate accuracy of 0.717 determined using the area under the curve. Presence of diffuse pancreatic swelling [odds ratio OR) 1.745; p = 0.008) and MST at >5 mg/day were identified as predictors of relapse (OR 0.483; p = 0.001). Conclusions: The RR could continue to increase for 7 years even under MST. Based on our analysis of the side effects of steroid therapy, MST at 5 mg/day for 2 (total 4625 mg) to 3 (total 6425 mg) years might be a rational and safe therapeutic strategy in terms of keeping the RR to <30% while avoiding potential steroid toxicity.