Incidence of cardiovascular events with fluoropyrimidines in gastrointestinal cancers

INTRODUCTION

Fluoropyrimidines (FP) such as 5-fluorouracil (5FU), capecitabine, S1, and Uracil-tegafur (UFT) form the backbone of the management for a majority of gastrointestinal (GI) cancers. One of the less common, but more dreaded, of the FP-related adverse events is cardiac abnormalities (cardiotoxicity).^[1] Some mechanisms are considered to be responsible for FP-related cardiotoxicity, including coronary vasospasm with related myocardial ischemia, FP-induced thrombotic occlusive disease, and direct cardiotoxic activity of a FP-related metabolite, fluoroacetate. Additionally, there are no definite proven risk factors for the development of FP-related cardiotoxicity, though a number of factors, such as pre-existing cardiac dysfunction, use of infusional 5-FU (IFL) as opposed to bolus 5-FU, and comorbidities such as hypertension (HTN) and smoking, may increase the risk of cardiotoxicity.^[2]

Besides IFL, oral FPs such as capecitabine, S1 and UFT are also commonly used in the management of GI cancers. There is contrasting data with regard to the potential for cardiac toxicity with the oral analogues of FP in contrast to IFL.^[3,4]

With this background, the investigators conducted a retrospective study evaluating the incidence of cardiotoxicity with IFL, capecitabine, and S1 and whether there were factors predicting the development of cardiotoxicity with these agents.

MATERIAL AND METHODS

RESULTS

Cardiac events and outcomes

Cardiac events, as pre-defined, were seen in 40 patients (1.6%). The most common cardiac abnormalities noted were self-limiting bradycardia or tachycardia during IFL infusions, seen in 15 patients (38%; n = 40), while chest pain without a diagnosis of MI (angina/angina equivalents) was seen in 14 patients (35%; n = 40). Major cardiac events were seen in 25 patients (1%). Nine of the cardiac events (23%) were finally labelled as coronary vasospasm. Seven patients (18%) (2 patients with Non ST-segment elevation Myocardial Infarction, two patients with high ejection fraction heart failure, 1 patient with acute onset pericarditis and shock, 1 patient with cardiogenic shock, and 1 patient with supraventricular tachycardia) required an intensive care unit (ICU) admission. Of the 25 patients with major cardiac events, 19 patients were continued on the same therapy post recovery, 3 patients shifted from IFL to S1, 2 patients were shifted from capecitabine to S1, while one patient was not continued with further chemotherapy. Five of these patients were started on nitrates during the course of their further FP-based chemotherapy after Cardio-oncology consultation. Further details of cardiac events are illustrated in Figure 1.

Close

Figure 1: Types of cardiac events in the entire cohort. MI: Myocardial infarction.

Export to PPT

DISCUSSION

The current study of nearly 2500 patients with GI cancers suggests that the incidence of FP-related cardiotoxicity was 1.6%, while the incidence of major cardiac events (excluding transient bradycardia/tachycardia) was 1%.

FPs are the second most common cause of chemotherapy-related cardiac events, with anthracyclines being the most common cause. The major clinical presentation is chest pain due to underlying vasospasm, but other cardiac abnormalities like MI and CHF are also known to occur with lesser frequency. Though the overall frequency of FP-related cardiotoxicity is not high, the fact that these drugs are ubiquitously used in GI cancers makes this a clinically relevant problem. Raltitrexed, an anti-folate thymidylate synthase inhibitor, can be used as a substitute in patients at high-risk for cardiac toxicities with FP, but availability is scarce.^[5]

This entails the use of FPs in a majority of GI malignancies, even with potential risk factors for the development of cardiotoxicity.

The baseline characteristics of the study cohort in this study have a few noteworthy points to expand upon. Firstly, the median age of patients was 51 years, and only 26% of patients were classified as ‘elderly’ in the Indian context (>60 years). This is at variance with the median age of presentation in a majority of GI cancers, as per previously published data from India.^[6] Secondly, while the presence of HTN and DM was seen in more than 20% of patients, the prevalence of baseline proven cardiac dysfunction in the cohort was only 4%. Pre-existing cardiac dysfunction has been a suggested predictor for 5-FU-induced cardiac dysfunction, and the current study cohort provides a population wherein the baseline risk of FP-induced dysfunction is already low. Finally, a significant proportion of patients received IFL and capecitabine, thereby allowing for a comparison of their propensity to cause cardiac events.

A significantly high proportion of patients (38%) in the study had transient bradycardia or tachycardia as their cardiac abnormality. All these events were self-limiting without any need for cessation of 5-FU infusion. Once these self-limiting events were excluded, the overall incidence of cardiac events in the study was 1%. This is one of the lower ends of the spectrum with regard to the occurrence of cardiac events, as evinced in published literature.^[4,7] A recent high quality retrospective study by Zafar et al., evaluating a large cohort of more than 4000 patients found the incidence of 5-FU induced vasospasm to be 2.16%; the current study has a baseline cohort which is similar to the quoted study in terms of incidence of baseline comorbidities in patients experiencing a major cardiac event (hypertension- 36% vs. 40% and cardiac dysfunction- 12% vs. 16%).^[8] However, another larger study (n = 103,110) using a target trial emulation framework to link national cancer, cardiac, and hospitalisation registry data in patients with GI cancers found that the 1-year risk of acute composite cardiovascular events was about 15.8% in patients receiving FPs. However, even in patients not receiving FPs, the event rate was 14.9%, with the differences being of marginal statistical significance.^[9] Such contrasting data likely reflect differences in baseline characteristics across the populations under study.

There appeared to be no major differences between infusional IFL and capecitabine in terms of incidence of major cardiac events, and this is one of the major takeaways from the current study. While IFL was associated with transient rate abnormalities, there were no statistical or clinically relevant differences in terms of the incidence of major cardiac abnormalities. While S1 statistically seemed to be associated with an increased number of cardiac events, this was based on an analysis of only 55 patients and needs further evaluation. Additionally, a greater proportion of patients receiving capecitabine or S1 had baseline cardiac dysfunction, and this may have contributed to the increased incidence of cardiac events with S1. It also needs to be mentioned that a recent study has actually shown that patients with 5-FU/capecitabine-induced cardiotoxicity had very low incidences of repeated cardiotoxicity (4%) when switched to S1.^[3] This is in concurrence with previously published low incidences of cardiotoxicity with S1; the results of our study are probably limited by the limited usage of S1.

Despite the large number of patients evaluated, the study failed to find any statistically relevant predictors for major cardiac toxicity. Though traditionally recognised risk factors such as HTN, DM, and pre-existing cardiac dysfunction correlated with the overall incidence of cardiac events, levels of statistical significance were not achieved in terms of predicting major cardiac abnormalities. Previous studies have also been inconclusive in showing the role of these risk factors in predicting cardiac abnormalities. This highlights two noteworthy points- primarily, the current study included a cohort that was probably carefully selected and monitored while receiving FP-based therapy and might have some degree of optimisation prior to FP-based therapy and secondarily, it suggests that alternate mechanisms or factors beyond traditional cardiovascular risk factors need to be identified for the prediction of FP-induced cardiotoxicity.^[10,11]

Though our study has the advantage of a large study cohort, some caveats need to be recognised. We cannot offer any information on the correlation of bolus 5-FU or UFT-based regimens with cardiotoxicity. There is a very limited representation of patients who have received S1. We also have not provided any cancer-related outcomes in these patients who had 5-FU-related cardiac events.

TAKE HOME MESSAGE

In conclusion, the current study of almost 2500 patients with GI cancers receiving predominantly IFL or capecitabine suggests that 1% of all patients develop major cardiac abnormalities. The incidence of cardiac abnormalities appears to be similar between IFL and capecitabine. Cardiovascular comorbidities appeared to have a weak correlation with the incidence of FP-induced cardiac events, highlighting the need for further research in this area of unmet need, as well as constant vigilance with regard to cardiotoxicity in patients receiving FP-based therapy.