FREQUENCY OF PERFORATION & IMPACT OF BOWEL REST IN AGGRESSIVE NON‐HODGKIN LYMPHOMA WITH GASTROINTESTINAL INVOLVEMENT: AN INTERNATIONAL, MULTI‐CENTER RETROSPECTIVE STUDY

Abstract

Introduction: The gastrointestinal (GI) tract is involved in 10-15% of patients with newly diagnosed non-Hodgkin lymphoma (NHL) and can lead to perforation, peritonitis and death. Some physicians advocate admission for the first cycle of chemotherapy to facilitate early recognition and rapid surgical intervention. Others also employ bowel rest and prescribe total parenteral nutrition (TPN) to reduce peritoneal contamination in the event of perforation. However, it is unclear if these measures are effective. Methods: We performed a multi-centre, retrospective analysis of patients with newly diagnosed aggressive NHL with GI involvement defined by either tissue biopsy or imaging between 1st January 2006 and 1st January 2016. Two centers employed bowel rest as a routine measure, while two did not. The Kaplan-Meier method was used to measure time from diagnosis to perforation and death or last followup. Univariate and multivariate analysis of factors associated with perforation and survival was performed using Cox regression. Results: We identified 419 patients, 204 (49%) treated as outpatients and 215 (51%) as inpatients. of inpatients, 106 (49%) received bowel rest; 109 (51%) did not. After a median follow-up of 3.6 years (range 0.1-11.9), 41 (9.8%) perforated; 28 (68%) at presentation or prior to chemotherapy. Excluding these, the median time to perforation was 28 days (2-877). Diffuse large B-cell lymphoma accounted for 85% of patients (357), high-grade B-cell lymphoma 3% (13), Burkitt lymphoma 5% (21), peripheral T-cell lymphoma 2% (9) and enteropathy associated T-cell lymphoma (EATL) 4% (16). There were one case each of plasmablastic lymphoma and anaplastic large cell lymphoma (0.4%). (Figure Presented) The perforation rate varied according to site (Figure 1A) and histology (Figure 1B). By multivariate analysis, small bowel involvement (HR 3.3; 95% CI 1.4-7.5, P = 0.005), large bowel involvement (HR 3.25; 95% CI 1.2-9.2, P = 0.026), EATL (HR 3.2; 95% CI 1.3-8.0, P = 0.012) and ECOG > 1 (HR 2.0; 95% CI 1.1-3.9; P = 0.035) were associated with increased perforation risk. Bowel rest was not associated with differences in rates of perforation (8.5% v 10.2%, p = 0.635; Figure 1C), peritonitis (75% v 58%, p = 0.448), surgery (75% v 97%, P = 0.092) or overall survival (HR = 1.1, 95% CI 0.7-1.6, P = 0.721; Figure 1D). Conclusions: Perforation occurs in 9.8% of patients with aggressive NHL and GI involvement, mostly at initial presentation, prior to chemotherapy. Small and large bowel involvement, EATL and poor ECOG are associated with increased risk of perforation. These data do not support a benefit for bowel rest and TPN in the management of unselected patients with aggressive NHL and GI involvement.