Trends in overall mortality among US veterans with intermediate (Int)/high-risk primary myelofibrosis (PMF).

Abstract

e19534 Background: PMF is a heterogeneous disorder with variable outcomes and differential effects on overall survival (OS) based on risk. Data on the OS of veterans with MF are limited and represent a unique opportunity given the stability of longitudinal retrospective follow up in patients (pts) with multiple comorbid conditions. In phase 3 trials, ruxolitinib (RUX; Janus kinase [JAK]1/JAK2 inhibitor) improved OS in pts with MF [int2/high DIPSS] leading to FDA-approval in 2011. In this study, we investigated mortality trends among US veterans with PMF pre- and post-RUX approval. Methods: De-identified PMF pt-level data obtained from Veterans Integrated Service Networks using PMF diagnosis codes (ICD-9: 238.76, ICD-10: D47.1) were collected from 2 Veterans Health Administration (VHA) data periods (01JAN2006–31DEC2010 and 01JAN2014–30SEP2018) reflecting years pre- and post-RUX approval. The first PMF diagnosis in the identification periods (pre-RUX, 01JAN2007–31DEC2010; post-RUX, 01JAN2015–30SEP2018) was the index date. Pts were required to be continuously enrolled in the VHA plan from 1 calendar year prior to 6 mo after the index date. Pts also must have had int/high-risk PMF defined as ≥1 risk factor at diagnosis (age > 65 y, hemoglobin [Hb] < 10 g/dL, white blood cells [WBCs] > 25 × 109/L). Demographic and clinical characteristics were recorded. All-cause 1-y, 2-y, and overall mortality rates and time to death from index until end of data availability were estimated using Kaplan-Meier (KM) methodology; univariate Cox regression was used to examine the hazard ratio (HR) for all-cause mortality for the pre- and post-RUX cohorts. Results: 244 and 1005 pts with PMF were included in the pre-and post-RUX cohorts, respectively. Most pts were age ≥65 y (84% vs 95%), male (99% vs 99%), and white (64% vs 77%). 64% vs 32% of pts had Hb < 10 g/dL at PMF diagnosis and 20% vs 14% had WBCs > 25 × 109/L for pre- and post-RUX cohorts, respectively. Based on KM analysis, the 1-y, 2-y, and overall mortality rates for pre- vs post-RUX periods were 45%, 63%, and 89% vs 27%, 37%, and 57%, respectively. Time from index PMF diagnosis to death was significantly longer for the post- vs pre-RUX cohort (median [95% CI], 1091 [989‒NR] vs 430 [347‒568] d; P< 0.001), and risk of death was 47% lower in the post- vs pre-RUX period (HR, 0.53; P< 0.001). Conclusions: In this retrospective analysis, patients with PMF had a high mortality rate; a 47% lower risk of mortality was observed post-RUX approval. Multiple factors may have influenced the lower mortality observed, which will be further explored.