Treatment for lymphoma and late cardiovascular disease risk: A systematic review and meta‐analysis

Abstract Background and aims Lymphoma patients are frequently treated with cancer therapies that may increase the risk of adverse health outcomes later in life, including cardiovascular disease (CVD) mortality. We sought to investigate the long‐term risk of CVD incidence in this survivor population relative to the general population to quantify this health burden. Methods A systematic review and meta‐analysis was conducted using EMBASE, MEDLINE, and CINAHL databases, from date of inception to November 2016, with additional searches completed through June 2018. Included reports were observational studies assessing CVD incidence in patients of either Hodgkin or non‐Hodgkin lymphoma (HL, NHL) who survived for at least 5 years from the time of diagnosis or if the study had a median follow‐up of 10 years. Meta‐analyses were performed using random effects models, and subgroup analyses were conducted to determine the incidence of specific CVD subtypes (coronary heart disease, pericardial disease, valvular heart disease, myocardial disease, cardiac dysrhythmia, and cerebrovascular disease). Heterogeneity was assessed using I 2 statistics and prediction intervals. Results Of the 7734 studies identified, 22 studies were included in this review, representing 32 438 HL and NHL survivors. Relative to the general population, lymphoma survivors had statistically significant two to threefold increases in the risk for nearly all subtypes of CVD examined. Lymphoma survivors appeared to be particularly susceptible to pericardial diseases (HL: 10.67, 95% confidence interval (CI), 7.75‐14.69; NHL: 4.70, 95% CI, 2.08‐10.61) and valvular diseases (HL: 13.10, 95% CI, 7.41‐23.16; NHL: 3.76, 95% CI, 2.12‐6.66). Although the 95% CIs were suggestive of increased risks, the 95% prediction intervals often included the null, reflecting the high heterogeneity of the estimates. Conclusion Given the suggested increased risks of cardiovascular outcomes in lymphoma survivor populations relative to the general population, tailored screening and prevention programmes may be warranted to offset the future burden of disease.


| INTRODUCTION
Hodgkin Lymphoma (HL) and non-Hodgkin Lymphoma (NHL) are solid tumours of the immune system common in both adults and children, 1 accounting for an estimated 79 990 and 509 590 cases of cancer worldwide in 2018, respectively. 2 Improvements in treatment and control strategies have resulted in an increased number of survivors, with 5-year survival estimates of 86% and 70% for HL and NHL, respectively. 3 Though many therapies have proven to be curative, there is increasing epidemiological evidence to suggest that individuals treated for cancer have an increased risk of adverse health outcomes, including fertility issues, cardiovascular diseases (CVD), and secondary cancers relative to the general population. [4][5][6][7][8][9][10][11][12] In general, treatment for lymphoma involves chemotherapy alone or in combination with radiation, stem cell transplantation, or biologic therapies. 3 The long-term cardio-toxic effects of these treatments, especially chemotherapy regimens utilizing anthracyclines and radiation therapy, have become more apparent in cancer survivors over the past decade. [13][14][15][16][17][18][19] We previously conducted a meta-analysis and found that the number of deaths due to CVDs within HL and NHL survivors were 7.31 (95% CI, 5.29-10.10) and 5.35 (95% CI, 2.55-11.24) times greater than the general population, respectively. 20 In acknowledging that there is a substantially increased risk of mortality because of cardiovascular-related events, we sought to further investigate if there is also an increased risk of CVD incidence within this population. It is possible that both the CVD incidence and mortality rates experienced by this survivor group relative to the general population are different because of cardio-toxic effects of treatment and damage to the cardiovascular system. Additionally, given that HL and NHL account for 3.2% of all cancers globally, 2,21 there is a need to quantify the long-term risk of CVD development among these survivors. Currently, international guidelines recommend lifelong follow-up and surveillance of paediatric survivors treated with either high-dose anthracyclines or highdose radiotherapy to the chest to decrease the burden of CVDs attributed to these treatments. 22 To our knowledge, no meta-analyses have previously examined the long-term risk of CVD incidence among HL and NHL survivors compared with the general population. As such, in the current systematic review and meta-analysis, we sought to examine the association of CVD development after treatment for HL and NHL, with particular emphasis on the type of CVD. We hypothesized that long-term HL and NHL survivors will have an elevated risk of incident CVD events relative to the general population, and that the incidence would differ by type of CVD.

| Protocol and registration
This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 23 The protocol was registered in PROSPERO (registration number: CRD42016052342).

| Data sources and search strategy
We The initial screening was completed by two reviewers (D.J.B. and A.T.M.), who independently assessed articles in a two-stage process.
In the first stage, the title and abstract of each study were screened, and studies were then considered for full-text assessment if they met the following criteria: (a) the study was published in a peer-reviewed journal, (b) original data were presented, (c) human participants were under investigation, and (d) the article was relevant to the objectives of this review. In the second stage, studies were assessed in their entirety to determine whether or not they were eligible for inclusion into the systematic review. To be included in this review, all of the following criteria had to be met: (a) the population studied were patients with a diagnosis of and prior treatment for lymphoma; (b) the patients survived a minimum of 5 years after diagnosis, the study had a median follow-up of at least 10 years from the time of diagnosis, or the study presented risk estimates specific to individuals who survived for 5 years or more after their diagnosis; (c) there was a comparator group that was representative of the general population; (d) the outcomes reported included risk, hazards, or odds ratios, or sufficient data were provided for their calculation; (e) the study was of a cohort, case-control, nested case-control, case-cohort, or cross-sectional design. A third reviewer updated the search utilizing the same two-stage process (AM), consulting with C.R.S. and D.J.B. to ensure consistency.
At each stage of review, percent agreement and kappa (κ) statistics were used to quantify agreement between the two reviewers.
Any disagreements were resolved by consensus between the reviewers. In cases where there were multiple studies using the same study population and assessing the same outcome, the study with the largest sample size was retained in the review and the study with the smaller sample size was excluded.

| Data extraction and study quality assessment
A data extraction form to collect study information was created spe-  (CBVD). If authors reported several incident outcomes that would be categorized into the same CVD subtype (eg, reporting estimates for two types of myocardial diseases: heart failure and cardiomyopathy), all relevant clinical study outcomes were extracted and included in the analysis, to maximize validity. It was assumed that reported outcomes would be largely independent from one another within a given publication (ie, few people would have developed multiple clinical CVD subtypes within a single study). The CVD groups are detailed in Table S1.
A single reviewer (A.M.) assessed study quality using the Newcastle-Ottawa Scale for case-control and cohort studies. 26 This scale assessed the quality of included studies with scores ranging from 0 (indicating low quality studies) to 9 (indicating high quality studies).
These scores came from three domains: selection (maximum of four points), comparability (maximum of two points), and outcome (maximum of three points).

| Statistical analysis
Individual study results were pooled overall to derive a standardized incidence ratio (SIR) to estimate the risk of cardiovascular incidence among lymphoma survivors relative to the general population. All analyses were performed using Stata version 14.3. Meta-analyses were conducted using a DerSimonian and Laird random-effects model to acknowledge the clinical heterogeneity present in this body of literature. 27 Meta-analyses were stratified and conducted across CVD subtypes. Cumulative meta-analyses were conducted within CVD subtypes to understand how the associations between lymphoma and types of CVD incidence changed over time.
Heterogeneity in the literature was assessed visually using forest plots, and statistically using I 2 statistics and prediction intervals using the Stata "rfdist" command. 28,29 To assess publication bias, we visually appraised funnel plots for asymmetry and also quantified asymmetry of funnel plots using Begg 30

| Study characteristics
The study characteristics for the 22 included reports are summarized in

| Study quality assessment
Attributes reflecting study quality are provided for all 22 studies in Table S2. Overall, the included studies were of high quality: six studies received 8 out of a possible 9 points on the Newcastle-Ottawa Scale; 12 studies were scored at 7, three at 6, and only one study received a score of 5. All studies had a representative cohort of lymphoma survivors and a nonexposed comparator group drawn from the same community. Thirteen studies did not explicitly demonstrate that individuals with a history of CVD were excluded at baseline. 33

| DISCUSSION
This systematic review and meta-analysis suggests that, compared with the general population, lymphoma survivors are at an elevated risk of developing cardiovascular events. Together with the metaanalysis that we previously completed, in which we investigated cardiovascular mortality in lymphoma survivors, 20 it is apparent that both HL and NHL survivors have both a higher incidence and severity of cardiovascular events compared with the general population. Though there were high levels of unexplained heterogeneity present, a novel finding in our meta-analysis is the differences in observed magnitude of increased risk between the various CVD subtypes, notably, the 10-fold and 13-fold increases in risk for PD and VHD, respectively, in HL. PD in lymphoma survivors may be more severe in magnitude compared with the general population because of the use of chemotherapeutic drugs including anthracyclines, as well as mediastinal radiation.
It is possible that our inclusion criterion of results from studies of survivors who were at least 5 years post-treatment could explain, in part, the higher incidence of pericardial disease that we observed. Delayed pericardial diseases can develop from 6 months post-radiation treatment to 15-years post treatment. 55 Cardiac valves are not directly damaged by chemotherapeutic agents, however, radiation-induced VHD is a relatively common side effect reported for lymphoma survivors. 35 Interestingly, coronary heart disease in NHL was found to be the only cardiovascular subtype that did not have a statistically significant increased risk compared with the general population. In a consensus statement by Lancellotti et al, the authors state that coronary artery disease (which is captured within our CHD subtype), is latent until at least 10-years after radiation exposure. 56 This latency period may account for the nonapparent increased risk found in this subtype, since the patients included in our review may not have survived long enough to experience this outcome.
It is unlikely that the associations found in this meta-analysis are spurious, for several reasons. First, temporality is evident, since all survivors must have been treated for HL or NHL to be subsequently assessed for CVD incidence within each included study. Second, the to the increased incidence of CVD in lymphoma survivors relative to the general population. Several chronic inflammatory conditions might also be associated with increased CVD risk. 58 We did not, however, T A B L E 2 Preliminary analyses for pooled SIR of incident CVD in HL and NHL survivors (using random-effects models) introduce some bias to our pooled estimates. To address this concern, the most adjusted measure of risk/incidence was used in the meta-F I G U R E 2 Forest plot of the risk of pericardial disease among lymphoma survivors, sorted in ascending order by median treatment era of each study F I G U R E 3 Forest plot of the risk of valvular heart disease among lymphoma survivors, sorted in ascending order by median treatment era of each study analysis. Another limitation of this study was that we did not restrict to studies only looking at contemporary treatments, and there have been changes in treatments over time. Therefore, it is possible that the large effects found in our analyses may be overestimating the effects that truly occur in current practice with improved treatment modalities. 60 In conclusion, this systematic review and meta-analysis is the first to investigate the long-term risks of CVD subtype incidence among HL and NHL survivors compared with the general population. Even if these risk estimates are overestimated because of uncontrolled confounding or heterogeneous studies, the overall magnitude of associations is strong enough to support the importance of utilizing cardiovascular screening, prevention, and surveillance programmes within this population of lymphoma survivors to potentially mitigate the future burden of CVD.

FUNDING
There was no specific funding source for this study.

Doreen M Rabi received travel reimbursement from Hypertension
Canada. This funding did not influence the study design, collection, analysis and interpretation of data; writing of the manuscript; or the decision to submit for publication. The corresponding author had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.

TRANSPARENCY STATEMENT
The corresponding author affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.