Evaluation of Cardiovascular Disease Risk in Chronic Myelogenous Leukemia Patients Using Electronic Medical Records from Community-Based Oncology Practices in the United States

PURPOSE: Current guidelines (Version I.2015) from the National Cancer Comprehensive Network recommend that patient’s comorbidities, including cardiovascular disease (CVD), be taken into consideration when evaluating tyrosine kinase inhibitor (TKI) therapy options for chronic myelogenous leukemia (CML). The guidelines also recommend using certain TKIs “with caution in patients with cardiovascular risk factors” and evaluating patients “for vascular risk factors prior to initiating and during treatment.” Recent research has evaluated the prevalence of CVD in the CML patient population, but there is a lack of published data on the risk of developing CVD in this population. The study purpose was to assess the risk for CVD among CML patients using an oncology-based electronic medical record (EMR) database.

METHODS: A retrospective cohort analysis was conducted using data from the International Oncology Network (ION) EMR Database (01/2005 to 01/2015) of patients treated by community-based oncologists in the U.S. The study population included patients with a diagnosis of CML (ICD-9 CM: 205.1) aged 18 years and older, having no other primary cancer or acute myeloid leukemia, and not enrolled in a clinical trial anytime during the study period. Study index was the date of the first observed CML diagnosis during 01/2005 to 10/2014. Prevalence of CVD and its risk factors were estimated for baseline and during a follow-up of up to 5 years. Baseline included all available medical information prior to and on the index date. CVD and its risk factors were defined as in the Framingham Heart Study, and were identified using diagnoses and/or treatment as applicable from pre-defined fields in standard EMR tables or through electronic text search of the physician progress notes available in the database. The Framingham risk score for CVD was computed at baseline using age, gender, smoking status, hypertension, dyslipidemia, and diabetes. To compare with annual estimates from the U.S. general population, all prevalence estimates at 1 year for the CML population were age and gender standardized to the U.S. 2010 census population aged 20 and older. Standardized rate ratios and z-tests were computed to assess the magnitude and significance, respectively, of the differences in annual rates between CML patients and the general U.S. population.

RESULTS: A total of 1,639 CML patients were included with mean±SD and median follow-up time (in years) of 2.4±1.7 and 2.1, respectively. Average age was 59±16 and 52.2% were male. CVD was present in 18.7% of patients at baseline. On average, CML patients without CVD at baseline had a Framingham CVD risk score of 12.8% compared to 8.7% in the general US adult population without CVD and nearly 44% of CML patients without CVD were at intermediate to high risk (>10%). Among patients without CVD, in the 5-year follow-up period 77.7% had at least 1 CVD risk factor up from 52.7% at baseline. Relative to the general U.S. adult population, annual standardized prevalence rates in CML patients were 20%, 30%, and 40% higher for hypertension, diabetes and obesity, respectively, while being 20% and 60% lower for dyslipidemia and smoking. In the 5-year follow-up period, the prevalence of CVD in the study sample rose to 33.0% of the sample of patients available at 5 years. Annual standardized prevalence rates of myocardial infarction, atherosclerosis, and heart failure was 2.7 to 3.5 times higher (p<0.001) relative to the general U.S. adult population.  

CONCLUSION: This study suggests that CML patients may have both a higher risk and prevalence of CVD compared to the general U.S. adult population, likely due to several contributing factors which should be evaluated in further research. Given the chronic nature of CML and the need for long term TKI treatment, it is important to regularly evaluate and monitor patients for CVD and its risk factors.