High prevalence and incidence of cardiovascular disease in chronic lymphocytic leukaemia: a nationwide population‐based study

Abstract

Chronic lymphocytic leukaemia (CLL) is a disease primarily affecting the elderly, with a median age of 72 years at diagnosis.1 Thus, the incidence and prevalence of cardiovascular disease (CVD) is expected to be high in this population. The use of Bruton tyrosine kinase (BTK) inhibitor (BTKi), ibrutinib, in patients with CLL has increased rapidly due to high efficacy and tolerability in patients with poor prognostic features.2, 3 It is used also in elderly and frail patients.4, 5 Patients treated with ibrutinib have an increased risk of adverse events (AEs) affecting the cardiovascular system, mainly atrial fibrillation (AF) and hypertension,6 the highest risk in patients with previous CVD.7-9 Population-based data on CVD in patients with CLL are scarce. Institutional data from the Mayo Clinic10 reported a prevalence of AF of 6·1% at diagnosis, and an annual incidence of 1% (n = 2444, median follow-up of 7·3 years). As the Mayo Clinic is a tertiary referral centre, this cohort might diverge from the general CLL population. We therefore decided to acquire data on the prevalence of CVD in an unselected cohort, using nationwide population-based registries. The aim of the present study was to investigate: (i) the prevalence of CVD, especially AF and hypertension, at the time of CLL diagnosis, (ii) the prevalence of CVD at the time of treatment initiation in patients with CLL requiring therapy, (iii) the incidence of CVD within the first 5 years after start of first-line therapy with chemotherapy or chemoimmunotherapy (CIT). Patients diagnosed with CLL [International Classification of Diseases, Tenth Revision (ICD-10): C91.1] between 2007 and 2010 were identified in the nationwide Swedish Cancer Register11 and the Swedish CLL Register. To acquire data in patients not exposed to BTKi, patients treated with ibrutinib 2015–2016 were excluded (n = 56). Data on CVDs were collected from the National Patient Register,12 using the ICD system. This contains data on inpatient hospitalisations from 1987 onwards, as well as specialist outpatient care consultations from 2001. A search for ICD codes (Table S1) was done for: (i) all CVDs diagnosed within 10 years prior to and until the date of CLL diagnosis in all patients, (ii) all CVDs diagnosed within 10 years prior to and until the date of treatment initiation in those in need of treatment either at diagnosis or during follow-up, (iii) all CVDs diagnosed within 5 years after start of first-line treatment for CLL in patients with and without previous CVD, i.e. cumulative incidence. The present study was approved by the Regional Ethics Committee in Uppsala (Dnr 2016/178) and conducted according to the Helsinki Declaration. Total CVDs (ICD I00-I99), as well as the specific CVDs, within 10 years prior to the time of CLL diagnosis and the time of initiation of first-line therapy for CLL were calculated in numbers and percentages. The cumulative incidence of new CVDs at 3 and 5 years after starting first-line therapy was calculated for all patients with and without previous CVD, each diagnosis counted once only. The different time points of analysis and the number of patients are outlined in Figure S1. In total, 2078 patients were diagnosed with CLL in Sweden during 2007–2010, 828 of these initiating therapy between 2007 and 2016. In all, 32% (n = 675/2078) had a diagnosis of CVD within 10 years prior to CLL diagnosis and 37% (n = 307/828) prior to treatment initiation (Table I). The most prevalent CVD was AF (9% at diagnosis and 8% at start of treatment) and hypertension (22% and 21%). The majority, 81% (n = 547/675), of patients with previous CVD had three or more different cardiovascular diagnoses (Table S2). Each diagnosis only counted once. All patients (N = 2078) n (%) Yes, n (%) n = 675 (32) No, n (%) n = 1403 (68) Yes, n (%) n = 307 (37) No, n (%) n = 521 (63) History of CVD at time of CLL diagnosis (N = 2078) n (%)† History of CVD at time of start of first-line therapy (N = 828) n (%)† At start of first-line therapy, 521 patients of 828 had no previous history of CVD. Within 5 years after start of treatment, 145 of these 521 (28%) were diagnosed with a new CVD (Fig 1, Table S3). The cumulative incidences for patients with previous CVD at the time of treatment initiation are presented in Table S4. Data regarding type of first-line therapy and incidence of CVD separated on type of therapy are shown in Tables S5 and S6 and Figure S2. CVD was registered as the main cause of death in 131 patients of 678 deceased (19%) (Table S7). A Danish study13 has investigated changes in causes of death in CLL over time. They described a declining proportion of CVDs as the main cause of death; however, differences in diagnoses included time-span of inclusion and follow-up, makes direct comparisons difficult. Compared with data from the Mayo Clinic, we found a higher percentage of AF (9% vs. 6%) at time of CLL diagnosis. This might be explained by a lower median age at CLL diagnosis in the Mayo data, 65 years, compared with 71 years in our unselected CLL population. A strength of our present study is the population-based design capturing virtually all patients with CLL in Sweden in the different registries. A limitation is that the data reported to the National Patient Register does not include visits to general practitioners, which might lead to under-reporting, especially regarding data on mild hypertension. In summary, using population-based data from the pre-BTKi era, we found that one-third of all patients with CLL had a high burden of CVD at the time of diagnosis and treatment initiation, and a high incidence of CVD after treatment with chemotherapy and CIT. This is a cause for concern, due to the increased risk of cardiovascular events using BTKi, especially in patients with pre-existing CVD. Another concern is that the use of anti-coagulants and platelet inhibitors can be troublesome due to the bleeding risks associated with BTKi. We suggest awareness and a thorough cardiovascular evaluation to be performed before initiating treatment with BTKi. Our present data may also be useful as a basis for future studies on CVD associated with the CLL disease itself, and the different treatments used. The authors would like to thank the Swedish CLL group and the Uppsala U-CAN biobank for supporting this project. None of the authors have any relevant conflicts of interest. Mattias Mattsson and Karin Larsson equally contributed in interpretation of data, outline of figures and writing of the paper, with support from Ingrid Glimelius and Martin Höglund. Mattias Mattsson with the help of Ingrid Glimelius and Martin Höglund collected data from the registries. Fereshte Ebrahim linked the registries and performed the statistical analyses. Table S1. ICD diagnoses according to ICD-10. Table S2. History of CVD at time of CLL diagnosis and start of first-line therapy Table S3. Number of patients without previous history of CVD, who were diagnosed with a new CVD within 5 years after start of first-line therapy for CLL. Each year and in total after 5 years. Table S4. Number of patients with previous history of CVD who were diagnosed with a new CVD within 5 years after start of first-line therapy for CLL. Each year and in total after 5 years. Table S5. Type of first-line therapy and baseline characteristics of patients each group. Table S6. Patients without CVD at treatment initiation developing a CVD. Separated on type of therapy. Table S7. CVD as cause of death within 5 years after CLL diagnosis. Table S8. CVD as cause of death within 5 years after initiating CLL therapy. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts