Smoking Is a Prognostic Factor for Survival in Chronic Myeloid Leukemia

Introduction

Smoking is suspected to be a risk factor for the development of a chronic myeloid leukemia (CML) (Musselmann, Cancer Epidemiol. 2013). However, it is currently not clear whether smoking is also an unfavorable prognostic factor, since the last analyses on this topic have been made in the pre-tyrosine kinase inhibitor era. Thus, our aim was to investigate the impact of smoking in chronic phase CML patients on survival and progression to advanced phases.

Methods

Out of 1,536 evaluable patients from the German CML study IV, the categorization as “smokers” or “non-smokers” was available for 1,326 patients.  261 patients (20%) were smokers. The median observation time was 7.0 years (range: 0.1 – 11.9 years).
Survival times were calculated starting with the date of diagnosis. Patients that were still alive were censored at the date of last observation. For the analysis of progression to accelerated phase or blast crisis, death without prior progression was considered as a competing event. Cox models were estimated to assess the impact of smoking adjusted for the following covariates: sex, prognostic category (EUTOS score), age at diagnosis, comorbidity (using the Charlson Comorbidity index without consideration of age) and type of treatment setting. Additionally, the models were stratified according to randomised treatment.  As a sensitivity analysis, the interaction effect between smoking and age was included.

Results

Smokers and Non-Smokers had significantly different baseline values concerning hemoglobin (median: 12.1 for non-smokers vs. 13.0 for smokers; p<0.001), leukocytes (median: 81.8 vs. 63.8; p<0.001), blasts (median 3 vs. 2; p=0.012), myelocytes (median: 16 vs. 14, p=0.026) and age at diagnosis (median: 54 vs. 46, p<0.001). Smoking was more frequent in males (28%) than in females (16%). Additionally smokers had by tendency a lower education, with only 13% of the patients having a higher degree compared to 25% in non-smokers (p<0.001). Distributions of EUTOS score, comorbidities, spleen size, platelets, basophils and eosinophils were similar for smokers and non-smokers.

The 8-year survival probability of a non-smoking patient was 91% (95%-confidence interval (CI): 89-93%), the 8-year probability of a smoker was 86% (95%-CI: 82-91%). In a multivariate model for survival, we found a strong negative effect of smoking, with a hazard ratio (HR) of 1.94 (95%-CI: 1.3-2.8, p<0.001). Besides smoking, increasing age, male sex, comorbidities and EUTOS high risk were confirmed as poor prognostic factors.

As a sensitivity analysis, we included an interaction term between smoking and age. With increasing age, the disadvantage of smoking got lost. While we found a HR of 5.27 (95%-CI: 1.7-15.9, p<0.001) for e.g. a 16-year old smoker compared to a 16 year-old non-smoker, a 80-year old smoker did not have a higher risk than a 80-year old non-smoker.

When analyzing the risk for progression to advanced phase, smoking was found to be the only relevant prognostic factor besides EUTOS risk. Smokers had a 2.02-times higher cause-specific hazard of progression (95%-CI: 1.3-3.1, p=0.002) than non-smokers. Molecular response rates were similar between smokers and non-smokers. We did not find significant differences with regard to adverse events.

Discussion

Smoking is a prognostically unfavorable factor even in the era of TKIs. This result also held in a multivariate setting. We did not find a direct effect of smoking on the TKI treatment, e.g. on molecular remission rates.
Higher cumulative incidences of progression to accelerated phase and blast crisis suggest that the lower survival probabilities of smokers are not just the result of smoking-related mortality, e.g. lung and cardiovascular diseases. Smoking is often connected to a lower socioeconomic status and an unhealthy general lifestyle. These variables may contribute to the higher mortality of smokers. Our results suggest that CML patients should be encouraged not to smoke and receive adequate support to quit smoking.