Predicting Factors of Long-term Outcome of Gastrointestinal Behcet's Disease: A Chinese Retrospective Study.

PURPOSE: Behcet's disease (BD) is a complex disorder affecting multiple systems and organs, and gastrointestinal BD is poorly understood. We aimed to identify factors influencing the long-term outcomes of patients with gastrointestinal BD. METHODS: Consecutive patients with gastrointestinal BD were analyzed retrospectively. Data on the following clinical characteristics were collected: sex, age at diagnosis, symptoms, endoscopic findings, medical treatments, and surgery. Mucosal healing and surgical rates at 1, 2, and 5 years were evaluated. Log-rank test and Cox proportional hazards regression models were used to evaluate the factors affecting long-term outcomes. FINDINGS: Baseline data of 175 patients with gastrointestinal BD were included. The mean (SD) age at diagnosis was 38.3 (12.9) years. The typical clinical symptoms were oral ulcer (72.6%), abdominal pain (71.4%), and weight loss (41.1%). The most commonly involved location was the ileocecum; isolated oval ulcer was the most common ulcer type. Seventeen patients (9.7%) underwent 18 surgeries after inclusion. The cumulative surgical rates were 8.6% (n/N = 15/175), 8.6% (n/N = 15/175), and 9.1% (n/N = 16/175) in 1, 2, and 5 years, respectively. Data from 101 patients who underwent at least 2 endoscopies were included in the analysis for mucosal healing. Kaplan-Meier curve showed that the cumulative mucosal healing rates at 1, 2, and 5 years were 34.7% (n/N = 35/101), 41.6% (n/N = 42/101), and 61.4% (n/N = 62/101), respectively. We compared cumulative mucosal healing rates between 4 treatment groups, including 5-aminosalicylic acid (3% [n/N = 3/101]), mono-immunosuppressant (31.7% [n/N = 32/101]), combined therapy (36.6% [n/N = 37/101]), and escalation therapy (28.7% [n/N = 29/101]), and found that mono-immunosuppressant achieved earlier mucosal healing than combined therapy (P = 0.0008) and escalation therapy (P = 0.0008). The univariate analysis showed that moderate to severe disease activity (P = 0.013, P = 0.004), diameter of the maximal ulcer >4 cm (P = 0.002), and nonsimple esophageal involvement (P < 0.001) were risk factors, and number of ulcers between 2 and 5 was the protective factor of mucosal healing (P = 0.001). Multivariate regression analysis indicated that nonsimple esophageal involvement (P < 0.001) and the maximal ulcer >4 cm (P = 0.041) were independent risk factors of mucosal healing. IMPLICATIONS: Most patients with gastrointestinal BD need long-term treatment to achieve mucosal healing. The location and size of ulcers have a significant impact on the mucosal healing of gastrointestinal BD.