AUTHORS
Manal Ahmidouch, MD, Laneisha Maldonado, MD2, Jhonsley Frederic, MD3, Marina Kremyanskaya, MD, PhD4, Hannah Levavi, MD5, Daniel Nathan, MD6, Ronald Hoffman, MD7, Anthony Michael Hunter, MD8, Douglas Tremblay, MD9
AFFILIATIONS
1 Icahn School of Medicine at Mount Sinai, New York, New York, United States
2 Emory School of Medicine, Atlanta, United States
3 Emory School of Medicine, Atlanta, United States
4 Icahn School of Medicine at Mount Sinai, New York, New York, United States
5 Icahn School of Medicine at Mount Sinai, New York, New York, United States
6 Icahn School of Medicine at Mount Sinai, New York, New York, United States
7 Icahn School of Medicine at Mount Sinai, New York, New York, United States
8 Emory School of Medicine, Atlanta, United States
9 Icahn School of Medicine at Mount Sinai, New York, New York, United States
ABSTRACT
Background
Patients with essential thrombocythemia (ET) are at risk of thrombosis, bleeding, progression to myelofibrosis (MF) or accelerated/blast phase disease, and death. Interferon (IFN) has long been used in ET, but real-world studies comparing IFN with other cytoreductive strategies often rely on fixed treatment classifications. These approaches do not account for treatment switching, potentially misclassifying exposure and obscuring associations with outcomes. We evaluated cytoreductive therapy as a time-dependent covariate for event-free survival (EFS) in a multicenter retrospective ET cohort, with particular attention to IFN.
Methods
We performed a multicenter retrospective cohort study of patients with ET. Clinical, treatment, and outcome data were manually abstracted from the electronic medical record. Confirmed treatment dates were used to construct start-stop intervals that accounted for starting, stopping, and switching specific cytoreductive agents. EFS was defined as time from ET diagnosis to disease transformation/progression, thrombosis, bleeding, or death. Cytoreductive therapy was modeled as a time-dependent covariate, with intervals classified as no cytoreduction, hydroxyurea (HU), IFN, or other cytoreductive therapy. Multivariable Cox proportional hazards models clustered by patient compared EFS risk across cytoreductive therapy groups, using HU as the reference group. A sensitivity analysis included only intervals on active cytoreductive therapy.
Results
A total of 500 patients with ET were included. Cytoreductive therapy was documented in 386 patients (77.2%), and 114 (22.8%) received no cytoreductive therapy. Among initial cytoreductive therapies, HU was most common (n=299, 77.9%), followed by IFN (n=36, 9.4%), ruxolitinib (n=26, 6.8%), and anagrelide (n=23, 6.0%). Cytoreductive therapy was initiated at a median of 0.2 years from ET diagnosis (IQR 0-3.6) and 119 patients (31%) received multiple cytoreductive agents during follow-up.
Median age at diagnosis was 52.1 years (IQR 36.0–63.0), 327 patients (65.4%) were male, and 52 (10.4%) had prior thrombosis. Driver mutations included JAK2 in 294 patients (58.8%), CALR in 114 (22.8%), MPL in 19 (3.8%), and triple-negative disease in 72 (14.4%). Revised IPSET-thrombosis risk at diagnosis was lower-risk in 343 patients (68.6%) and high-risk in 157 (31.4%). Patients initially treated with IFN were younger than those treated with HU (median age 41.9 vs. 57.4 years), had lower-risk disease (27.8% vs. 40.5%), and had similar median follow-up (9.9 vs. 8.9 years).
During a median follow-up of 11.5 years, 63 patients experienced thrombosis (17 arterial, 46 venous), 130 had disease transformation/progression (85 MF, 34 PV, and 11 AP/BP), 18 had bleeding, and 42 died. Five-year EFS was 92.8% with no cytoreduction (95% CI, 87.3 - 98.7), 82.4% among patients initially treated with IFN (95% CI, 70.5 - 96.3), and 83.2% among those initially treated with HU (95% CI, 78.8 - 87.7); overall log-rank p = 0.015).
The median cumulative time on therapy was 3.1 years for IFN and 5.5 years for HU. Compared with HU, IFN exposure was not significantly associated with EFS in unadjusted analyses (HR 0.86, 95% CI 0.45 - 1.65; p=0.654) or after adjustment for age, sex, JAK2 status, prior thrombosis, and platelet count (HR 0.34, 95% CI 0.08 - 1.49; p=0.152). Findings were similar in a sensitivity analysis restricted to intervals on active cytoreductive therapy (adjusted HR 0.35, 95% CI 0.08 - 1.54; p=0.165). In exploratory subgroup analyses, IFN exposure was not significantly associated with EFS among patients with JAK2-mutated disease or within high or lower-risk revised IPSET-thrombosis groups.
For secondary outcomes, IFN exposure was not significantly associated with progression-free survival (adjusted HR 0.36, 95% CI 0.05 - 2.79), MF-free survival (adjusted HR 0.37, 95% CI 0.05 - 2.97), or thrombosis-free survival (adjusted HR 0.56, 95% CI 0.07 - 4.53), compared with HU.
Conclusions
In this multicenter ET cohort, IFN exposure modeled as a time-dependent covariate was not significantly associated with EFS compared with HU. Modeling cytoreductive therapy as a time-dependent covariate more accurately reflects real-world practice by accounting for treatment switching and cumulative exposure. Findings were consistent across sensitivity, subgroup, and secondary outcome analyses.
DISCLOSURES
Manal Ahmidouch, MD: No relevant conflicts of interest to declare.