Journal Logo



Patel, D.1, 2; Knudsen, Alma T.1; Hansen, Lund D.3; Ocias, Frans L.3; Bjerrum, Weis O.4; Brabrand, M.3; Ellervik, C.5; el Fassi, D.6; Frederiksen, M.7; Kjær, L.1; Kristensen, Kielsgaard T.8; Kruse, O.9; Mourits-Andersen, H. T.10; Møller, P.1; Overgaard, Malthe U.4; Severinsen, Tang M.11; Skov, V.1; Sørensen, Lindholm A.1, 2; Stentoft, J.12; Starklint, J.13; de Stricker, K.8; Thomassen, M.9; Larsen, Stauffer T.3; Hasselbalch, Carl H.1

doi: 10.1097/01.HS9.0000567320.93768.1b
Publication Only: Myeloproliferative neoplasms - Clinical

1Department of Hematology, Zealand University Hospital, Roskilde

2Institute for Inflammation Research (IIR), University Hospital Rigshospitalet, Copenhagen

3Department of Hematology, University Hospital Odense, Odense

4Department of Hematology, University Hospital Rigshospitalet, Copenhagen, Denmark

5Harvard Medical School, Boston, United States

6Department of Hematology, University Hospital Herlev, Copenhagen

7Department of Hematology, Hospital of Southern Jutland, Haderslev

8Department of Pathology

9Department of Clinical Genetics, University Hospital Odense, Odense

10Department of Hematology, Hospital of South West Jutland, Esbjerg

11Department of Hematology, University Hospital Aalborg, Aalborg

12Department of Hematology, University Hospital Aarhus, Aarhus

13Department of Hematology, Hospital of West Jutland, Holstebro, Denmark

Back to Top | Article Outline


Smoking is linked to several malignancies, including myeloproliferative neoplasms (MPN). However, the direct impact of smoking on JAK2V617 allele burden (JAK2) has not been examined in MPN.

Back to Top | Article Outline


To compare JAK2 over time among smokers and non-smokers with MPN.

Back to Top | Article Outline


A post-hoc analysis was conducted on the DALIAH-trial (NCT01387763). The study consisted of newly diagnosed patients with either essential thrombocythemia (ET), polycythemia vera (PV), prefibrotic myelofibrosis (pre-MF) or primary myelofibrosis (PMF) according to the WHO criteria. Patients who were naïve to cytoreductive treatment were enrolled in the study and randomized to treatment with recombinant pegylated interferon alpha-2 (r-IFNα) or hydroxyurea (HU). History of smoking prior to inclusion was collected retrospectively. Patients were defined as smokers if they had any daily tobacco consumption at baseline. JAK2 was measured by qPCR on whole blood. Molecular and clinicohematological response assessment was done using the ELN 2009 and EUMNET 2005 criteria. All patients provided written informed consent. A repeated measurements analysis was applied. JAK2 values were logit transformed for testing. Through backward elimination, the model was cleared for factors, which did not affect the model (p > 0.05). The model tested for differences in predicted change of JAK2 by smoking status.

Back to Top | Article Outline


A history of smoking was available for patients from 4/9 study sites (140/203 patients, Table 1). There was a more rapid predicted decrease in JAK2 among non-smokers, which was significant at 12 and 18 months (p < 0.01). Among JAK2-positive patients, 13/14 (93%) smokers received r-IFNα vs. 63/83 (76%) non-smokers, but this did not affect the model. Neither did age, sex, baseline JAK2 or diagnosis. A partial molecular response was achieved in 4/14 (28.6%) smokers vs. 32/83 (38.6%) non-smokers and a complete clinicohematological response was achieved in 14/23 (60.9%) smokers vs. 80/117 (68.4%) non-smokers. A drop-out rate of 15/23 (65.2%) among smokers and 59/117 (50.4%) among non-smokers lead to wider 95% CI with time (Figure 1).



Back to Top | Article Outline


Non-smokers are more likely to have a more rapid decrease in JAK2 particularly in the initial phase of treatment (12 and 18 months, p < 0.01) compared to smokers. Although statistical testing was not performed, partial molecular response and complete clinicohematological response were more incident among non-smokers and drop-out was lower among non-smokers. The study was limited by several factors: (1) The post-hoc design. (2) The few smokers. (3) Smoking status might change in relation to a diagnosis of MPN, thus rendering the initial grouping unprecise. (4) The study design was not powered to detect factors affecting the model. Still, the resulting model was rather conservative. Hence the finding was surprising, and it needs confirmation in additional studies.

Copyright © 2019 The Authors. Published by Wolters Kluwer Health Inc., on behalf of the European Hematology Association.