5‐Fluorouracil‐induced acute coronary syndrome

Clinical record

A 70‐year‐old man was admitted to Gosford District Hospital with exertional chest pain on his way back home after starting receiving 5‐fluorouracil (5‐FU) infusion for treatment of recently diagnosed stage III nasopharyngeal cancer. He did not have a significant cardiovascular history. His risk factors included a 17 pack‐year smoking history, hypertension and hyperlipidaemia. His medications on presentation were amlodipine, rosuvastatin and fenofibrate. His cardiovascular examination was unremarkable. His initial electrocardiograms (ECGs) in the emergency department (ED) revealed a sinus rhythm without any ST segment or T wave abnormalities (Box 1). Serial high sensitivity troponins were 39 ng/L and 69 ng/L, respectively. His chest x‐ray showed normal cardiac silhouette and lung vasculature. Based on his risk factors and mild troponin rise, he was diagnosed with non‐ST‐elevation myocardial infarction. He received loading doses of aspirin and ticagrelor, started taking therapeutic low molecular weight heparin, and was admitted to a monitored cardiology bed. Shortly after admission to the wards, he developed severe chest pain radiating to the left arm, with associated diaphoresis. His ECG revealed anterolateral ST segment elevation consistent with transmural ischaemia (Box 2). He was given sublingual glyceryl trinitrate (GTN), with immediate relief. Two hours later, he developed further chest pain after which he was commenced on GTN infusion, with optimal pain relief. He then underwent a coronary angiogram that revealed only minor coronary artery disease (Box 3). A transthoracic echocardiogram on the same day did not reveal any major wall motion abnormality, with an overall mildly impaired systolic function with an ejection fraction of 45–50%. Subsequently, a presumptive diagnosis of coronary vasospasm secondary to 5‐FU was made. The patient was reviewed by his regular oncologist and 5‐FU was ceased from his chemotherapy regimen.

Discussion

Fluoropyrimidines, which include 5‐FU and capecitabine, form the backbone of chemotherapy regimens for a variety of cancers. In fact, 5‐FU is the third most commonly used chemotherapeutic agent for the treatment of solid malignancies.1 Unfortunately, 5‐FU is also the second most common chemotherapeutic agent causing cardiotoxicity, with a reported incidence between 2% and 18%. Some of the reported difference in incidence can be accounted for by the differences in treatment duration, dose, additional agents, and patient characteristics.2 Capecitabine, which is an oral pro‐drug of 5‐FU, has also been associated with cardiotoxicity, with a reported incidence between 3% and 9%.2 While the most common manifestation of 5‐FU‐induced cardiotoxicity is chest pain, presentations can vary from arrhythmias, myocarditis and pericarditis, heart failure and even death.2 Of note is that 5‐FU cardiotoxicity tends to occur most commonly during the first cycle of administration. The median time to symptoms is 12 hours following an infusion, but symptoms could occur any time after commencing 5‐FU, even up to 1–2 days after infusion, thus complicating causal analysis. Clinical studies have shown that infusional regimens are associated with greater risk compared with bolus therapies.3

The potential mechanisms via which 5‐FU causes cardiotoxicity remain a matter of debate. While multiple pathways have been implicated, the two most likely are coronary vasospasm leading to myocardial ischaemia and direct drug‐related myocardial toxicity. Indeed, both experimental and physiological studies have hinted at these being the primary pathways. However, further vigorous prospective investigations in appropriate clinical settings as well as experimental models are required to provide evidence.2

Further complicating the use of a potentially life‐saving drug is the lack of predictive, clinically relevant risk factors that would enable clinicians to stratify risk and prescribe appropriately. It is interesting that previous history of cardiac diseases does not necessarily predict 5‐FU‐related toxicity, as a review demonstrated that, out of the 377 cases, only 14% had a previous history of cardiac disease.4

Despite 5‐FU‐associated cardiotoxicity being well documented, there is no widely established treatment strategy. While it is universally agreed that the first step should be to cease the infusion immediately and then administer medical anti‐anginal therapy aimed at symptomatic relief with calcium channel blockers (CCBs) and/or nitrates, the challenge of establishing a causal relationship between exposure and symptoms remains daunting. Fortunately, CCBs and nitrates have been shown to abort symptoms in up to 69% of affected patients.2 However, there remains a distinct lack of a definitive test to establish 5‐FU as the cause. Temporal relationship, while a necessary condition, does not exclude other possible contributory factors.

If the patient is at high risk of coronary artery disease, as in our example, it is advised to perform a prompt coronary angiogram to identify an alternative pathological process driving the patient's symptoms. Patients without conventional cardiac risk factors may be screened using a non‐invasive test; for example, a computed tomography coronary angiography. 5‐FU‐induced cardiotoxicity can be presumed in patients with clinically insignificant coronary disease or normal arteries.

However, presence of significant coronary stenosis does not necessarily exclude the possibility of superimposed 5‐FU‐related toxicity. To further investigate causality, invasive pharmacological provocation using acetylcholine or nitroglycerin during coronary angiography may be useful by demonstrating excessive vasospasm, thereby suggesting either coronary endothelial or primary smooth muscle dysfunction. However, this is not widely available and its role in this setting has not been validated. Moreover, there is no evidence to establish that positive provocation tests predict cardiotoxicity.2

Furthermore, rechallenging these patients with 5‐FU drugs remains controversial, as randomised clinical trials comparing different strategies are lacking.2 Rates of recurrence of cardiotoxicity as high as 90% have been reported, with death up to 13%.2 Considering the potential implications of rechallenge, it is advised to have a detailed discussion with the patient, the cardiologist and the oncologist regarding the risk–benefit analysis, taking into account the original probability of 5‐FU‐related cardiotoxicity, availability of alternative non‐5‐FU drugs, intent of therapy (ie, curative versus palliative) and the patient's own expectations. Since rechallenge may be fatal, it should only be pursued in a subset of patients for whom the benefits are thought to significantly outweigh the risk, and even then, only in a carefully monitored setting with a low threshold for discontinuation. One example would be in patients with chest pain after exposure to 5‐FU who are found to have significant coronary artery disease and undergo revascularisation and for whom no other reasonable treatment options are available.

Box 1 – Electrocardiogram on arrival at the emergency department — the patient was asymptomatic

---

 

Box 2 – Electrocardiogram (ECG) during severe chest pain on arrival at the ward — the ECG demonstrates anterolateral ST segment elevation

---

 

Box 3 – Coronary angiogram showing only minor coronary artery disease

---

Caud = caudal; Cran = cranial; Cx = circumflex; LAD = left anterior descending; OMA = obtuse marginal artery; RAO = right anterior oblique; RCA = right coronary artery.