Risk Factors for Infections in Myelofibrosis: Role of Disease Status and Treatment. A Study on 507 Patients

Background: Infectious complications represent one of the main causes of morbidity and mortality in patients (pts) with Primary Myelofibrosis (PMF), Post-Essential Thrombocythemia and post-Polycythemia Vera MF (PET/PPV-MF). Up-to-date, very few data are available on incidence and outcome of infectious complications. Also, risk factors of this potentially fatal complication are still to be investigated and defined. This is particularly relevant in the era of targeted therapy in MF, since an increased infectious risk has been reported in pts treated with ruxolitinib (RUX), a JAK1/2 inhibitor.

Aims: To evaluate risk factors for severe infections in a large cohort of MF patients.

Methods: Clinical and laboratory data of pts with MF were retrospectively collected from the database of 5 Italian Hematology Centers. Severe infections were defined  according to the CTCAE. The study was approved by the Ethic Committee of each participating Centers.

Results: Between 1980 and Aug 2014, 507 pts with PMF (362 pts, 71%), or PET-MF (14%) or PPV-MF were diagnosed and followed for a median follow-up of 4.2 yr (0.5-30.1). Baseline characteristics were (median): age, 66 y (range, 26-87); ≥65 y, 54%; male, 59%; hemoglobin (Hb), 11.9 g/dL (4-17.9); Hb <10 g/dL, 25%; PLT, 376×10⁹/L (4-2513); PLT <100×10⁹/L, 8%; spleen enlargement, 71% (spleen length ≥10cm: 20.5%); constitutional symptoms, 20%. International Prognostic Score System (IPSS) was low (21%), intermediate-1 (intm-1, 37%), intermediate-2 (intm-2, 25%), high (17%). Molecular analysis was performed on 321 pts (63%) and was positive in 83% (JAK2V617F), 12% (CALR), 3% (MPLW515K/L); 6 pts (2%) were triple negative. Karyotype was abnormal in 46 (17%) out of 265 evaluable pts (unfavorable in 13 pts; 5%). Three hundred and sixty-five (72%) received cytoreductive therapy (mainly hydroxyurea, 88%) and 71 patients (14%) received corticosteroids (defined as ≥ one cycle of prednisone at the dose of 1mg/Kg/daily for at least a month). Smoking habit was present in 165 patients (37%) and diabetes mellitus in 66 (13%).

Overall, 112 pts (22%) experienced 160 infectious events (grade 3-4, 45%), for an incidence rate of 3.9% pt-y. The cumulative incidence was 20%, 33% and 51% at 5, 10 and 20 y, respectively. Infections were: bacterial (143 events, 89%; pneumonia: 80 cases, 56%); VZV reactivations (11 events, 7%), nodal TBC (3 events, 2%), fungal infections (3 events, 2%). Infectious complications represented the causes of death in 10 (7%) out of 134 deceased pts. Among baseline features, age≥65 y (p=0.001), primary vs secondary MF (p=0.009), spleen length>10 cm below left costal margin (p=0.006), high/intm-2 IPSS (p<0.0001) significantly correlated with higher infectious risk; in multivariate analysis, an high/intm-2 IPSS category and massive splenomegaly confirmed their negative impact (p=0.02 and p=0.04, respectively).

Overall, 128 pts at intm-2/high IPSS risk were treated with RUX for a median time of 23 mos (1-41). Infection-free survival at 5 years was comparable in RUX-treated pts compared to non RUX-treated intm-2/high risk pts (76% vs 67%, p=0.82). In the RUX-treated cohort, age ≥65 y (p=0.008), JAK2 allele burden ≥75% (p=0.03) and steroids exposure before RUX (0.007) correlated with an increased infectious risk. Multivariate analysis confirmed age and corticosteroids utilization as independent negative prognostic risk factors (p=0.003 and p=0.043, respectively). Also, patients who obtained a spleen reduction higher than 50% during RUX therapy were projected to a better infection-free survival compared to non-responders (89% vs 70% at 12 mos, p=0.001)

Summary/Conclusion. This large study confirms severe infections as frequent and potentially fatal events in MF. Also, this study has led to the identification of the main baseline features associated with increased infectious risk, namely baseline IPSS category and massive splenomegaly. Surprisingly, RUX therapy did not seem to significantly increase the risk of infections, despite its immunosuppressant properties. Yet, the successful use of RUX in terms of spleen response was found to correlate with a significant reduction of the probability to develop an infectious complication. Conversely, a combined or sequential use of corticosteroids and RUX may further increase the risk of infectious complications and therefore require a careful evaluation.