Recently developed prognostic models for follicular lymphoma have proven useful in predicting overall survival (OS), but most have used data from multiple centers. Our goal was to look at prognostic factors within a single institution.We conducted a retrospective study on 77 newly diagnosed patients with follicular lymphoma, focussing on clinical characteristics, symptom duration before diagnosis, pathologic findings, including grade, laboratory data, imaging studies at initial presentation and management. The study population was 53% male. Ages ranged from 25 to 87 years (median 51). By Ann Arbor classification 4% were stage I, 8% stage II, 18% stage III and 69% stage IV. Initial therapy was deferred in 39%. The remaining patients received stage-appropriate therapy. Survival was measured from time of diagnosis to death. Prognostic factors at initial diagnosis that were statistically significant in univariate log-rank comparisons of Kaplan-Meier survival curves were used to build a multi-variate proportional hazard regression model of OS. Median OS for these patients was 10.3 years. OS differed only with high (>12 g/dl) versus low (<12 g/dl) hemoglobin (p=0.001) and in younger (<60 years) versus older (>60 years) patients (p=0.05), as indicated by univariate log-rank tests. Both hemoglobin and age were also significant in a multivariate proportional hazards analysis. Low hemoglobin and increased age were independent predictors of lower OS with hazard ratios of 6.6 (95% CI, 2.2-20.1) and 3.7 (95% CI, 1.2-11.7), respectively. Median survival for older patients who also had anemia was only 3.1 years. A test for interaction between age and hemoglobin was negative (p=0.35). The estimated hazard ratio for an older individual with low hemoglobin was 24.7 (95% CI, 4.0-153.3). To assess the proportional hazards assumption we tested for an interaction between time and both age (p=0.92) and hemoglobin (p=0.66) and found no evidence against proportionality. A hemoglobin <12 g/dl and age >60 years at diagnosis are significant predictors of worse OS.

Related Articles