Defining Criteria for Steroid Response to Distinguish Primary Sclerosing Cholangitis (PSC) From IgG4-Sclerosing Cholangitis (IgG4-SC)

Abstract

Introduction: Elevated serum IgG4 levels are seen in 10-15% of PSC, raising suspicion of IgG4-SC. However, differentiating PSC from IgG4-SC is a challenge. While IgG4-SC responds to steroids, it is unknown if liver test response to steroid therapy can distinguish the two entities. We evaluated the performance characteristics of reduction in serum alkaline phosphatase (ALP) following steroid treatment to distinguish PSC from IgG4-SC. Methods: We retrospectively identified 40 patients with SC (PSC: n=38, 27 M, mean age 46 yrand IgG4-SC: n=18, 15 M, mean age 66 yr) who received steroid treatment (≥30 mg prednisone with slow taper) and had ALP measured before and after initiation of treatment. PSC patients were treated with steroids because of elevated serum IgG4 (n=18), or, in those with normal IgG4 (n=20), for concomitant inflammatory bowel disease (n=15) or other indications (n=5). Pre- and post-steroid (median 5 weeks) ALP values were analyzed as fold upper limit of normal (ULN) and percentage decrease from baseline. Results: The pre- and post-treatment ALP distribution is reported in Table 1. Post-treatment ALP decreased in 91% of PSC and 96% of IgG4-SC with abnormal baseline ALP regardless of serum IgG4 status. If baseline ALP was < 2X ULN, its decrease after treatment did not distinguish IgG4-SC from PSC. If baseline ALP was ≥2X ULN, 0/28 PSC and 8/12 (67%) IgG4-SC had post-treatment normalization of ALP (p < 0.0001), 11/28 (39%) PSC vs 4/12 (33%) IgG4-SC had post treatment ALP 1-2X ULN (p=0.33), and post-treatment ALP persisted at >2X ULN in 17/28 (61%) in PSC and 0/12 in IgG4-SC (p=0.0002). Among patients with baseline ALP ≥2X ULN, the sensitivity and specificity for IgG4-SC was 67% and 100%, respectively, for ALP normalization, 83% and 86%, respectively, for reduction in ALP to < 1.5X ULN, and 100% and 61%, respectively, for reduction in ALP to < 2X ULN after steroid treatment. ALP reduction of >50% had a sensitivity of 67% and specificity of 89% for IgG4-SC.Table 1: The pre- and post-treatment ALP distribution in PSC and IgG4-SC.Conclusion: Following steroid treatment ALP decreases in both PSC and IgG4-SC; if baseline ALP is < 2X ULN, this response is not useful in distinguishing between the two entities. In SC patients with baseline ALP ≥2 fold ULN, a 6 week steroid trial may be diagnostic in 60%; normalization of ALP is diagnostic of IgG4-SC and its persistence >2 fold ULN is diagnostic of PSC; an intermediate ALP response is non-discriminatory. Percentage decrease in ALP did not reliably distinguish the two entities.