World Journal of Oncology, ISSN 1920-4531 print, 1920-454X online, Open Access |
Article copyright, the authors; Journal compilation copyright, World J Oncol and Elmer Press Inc |
Journal website https://www.wjon.org |
Original Article
Volume 15, Number 6, December 2024, pages 902-921
The Prevalence of 5-Fluorouracil and Capecitabine Cardiotoxicity: A Systematic Review and Meta-Analysis
Bannawich Sapapsapa , Poomipat Thongnoia , Anchana Pongpuna , Supattra Kitcharoenpanyab , Teerarat Todsarota , Arpa Petchsomrita , Nattawut Leelakanoka, c
aDivision of Clinical Pharmacy, Faculty of Pharmaceutical Sciences, Burapha University, Chonburi, Thailand
bFriend Pharmacy, Chonburi, Thailand
cCorresponding Author: Nattawut Leelakanok, Division of Clinical Pharmacy, Faculty of Pharmaceutical Sciences, Burapha University, Chonburi 20131, Thailand
Manuscript submitted June 26, 2024, accepted August 13, 2024, published online October 30, 2024
Short title: Cardiotoxicity of 5-FU and Capecitabine
doi: https://doi.org/10.14740/wjon1920
Abstract | ▴Top |
Background: The incidence of cardiotoxicity events in patients who use 5-fluorouracil (5-FU) and capecitabine monotherapy remains unclear since previous studies reported the prevalence in patients who used combination regimens. We aimed to systematically review and meta-analyze the incidence of cardiotoxicity in fluorouracil and capecitabine monotherapy users.
Methods: The study protocol was registered with PROSPERO (CRD42023441627). Systematic searches were conducted in five databases (CINAHL, OpenGrey, PubMed, ScienceDirect, and Scopus). The Cochrane Risk-of-Bias tool and the Risk Of Bias In Non-randomized Studies were used to evaluate the risk of bias. Pooled prevalence and 95% confidence interval (CI) were calculated using the DerSimonian-Laird random effect models. The funnel plot was used to assess the publication bias.
Results: Eighty studies were included. There were 24 randomized controlled trials (RCTs) with low to high risk of bias and 56 non-RCTs with critical risk of bias. The pooled prevalence of cardiotoxicity from 5-FU was 3.5% (95% CI: 2.7 - 4.2; P < 0.001; I2 = 73.86%). The pooled prevalence of cardiotoxicity in capecitabine users was 2.8% (95% CI: 1.6 - 4.0; P < 0.001; I2 = 72.62%).
Conclusions: The prevalence of cardiotoxicity from 5-FU and capecitabine was classified as common. Cardiotoxicity may have not been associated with the cumulative dose of 5-FU or capecitabine.
Keywords: 5-Fluorouracil; Capecitabine; Cardiotoxicity
Introduction | ▴Top |
5-Fluorouracil (5-FU) and its oral pro-drug, capecitabine, are anticancer drugs that exert their biological activity by inhibiting DNA synthesis [1]. They are widely used as monotherapy or in combination regimens for the treatment of solid malignancies including colorectal cancer and other types of cancer, e.g., head and neck cancer, esophageal cancer, stomach cancer, and bladder cancer [2]. In addition, they can be combined with radiation therapy or concurrent chemoradiation (CCRT) [3]. 5-FU and capecitabine are significantly crucial for the treatment of cancer that the National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology recommend 5-FU and capecitabine as the first-line treatment for metastatic colorectal cancer [4] and the World Health Organization (WHO) listed 5-FU and its oral pro-drug in the 21st list of the WHO Model List of Essential Medicine [5]. However, the use of 5-FU and its pro-drug can be limited by their toxicities.
The dose-limiting toxicities of 5-FU bolus injection are associated with hematological toxicities but 5-FU continuous infusion causes hand-foot syndrome, diarrhea, and mucositis [6, 7]. In addition, cardiac toxicities, which include chest pain, coronary vasospasm, and myocardial infarction, are common for fluorouracil. Other less common symptoms include congestive heart failure, arrhythmias, pericarditis, and sudden cardiac death [8, 9]. Although cardiotoxicity induced by 5-FU has long been discovered [10], and the prevalence is known to be the second most commonly reported after anthracyclines [11], the exact incidence of fluorouracil-induced cardiotoxicity is difficult to estimate. This is because most of the research reported the prevalence or incidence of cardiotoxicity from fluorouracil-containing regimens whose compositions are largely diverse. In addition, most studies fail to report the total number of patients who are treated with fluorouracil which complicates the estimation of such prevalence [12-15].
The Council for International Organizations of Medical Sciences [16] categorized the prevalence of adverse drug reactions (ADRs) into very common, common (or frequent), uncommon (or infrequent), rare, and very rare. The categorization of the risk of cardiotoxicities from 5-FU and capecitabine assists the risk communication to the patients, which affects how patients perceive the risk and adhere to the anticancer treatment [17]. Since the incidence of adverse events is crucial for patient care, we aimed to systematically review and meta-analyze the incidence of cardiotoxicity during the use of fluorouracil.
Materials and Methods | ▴Top |
Search strategy
This study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 [18]. The study protocol was registered with PROSPERO (CRD42023441627). The Institutional Review Board (IRB) approval does not apply to this study. Systematic searches were conducted in five databases (CINAHL, OpenGrey, PubMed, ScienceDirect, and Scopus) without language and study design restrictions. The search was conducted from their inception to June 1, 2024. The following search concept was used: “5-fluorouracil” OR “capecitabine” AND “cardiotoxicity” (Supplementary Material 1, wjon.elmerpub.com).
Study eligibility criteria
The systematic review was conducted using the following eligibility criteria: 1) were clinical trials or observational studies; 2) explicitly reported details of 5-FU or capecitabine treatment regiments; 3) used 5-FU or capecitabine as monotherapy for any cancer treatment indication; and 4) explicitly reported the number of cardiotoxicity events. We excluded studies that did not have sufficient data on cardiotoxicity incidence.
Data extraction
Three authors independently screened retrieved articles’ titles, abstracts, and full texts. Disagreements were resolved by consulting the senior author in the team. Also independently, three authors extracted the author’s name, year of publication, region, study design, sample size, duration of the study, mean age, sex, characteristics of participants, comorbidity of participants, cancer type, treatment regimen, cumulative dose, and cardiotoxicity events. Data deemed important but unavailable in the publications were retrieved by contacting the corresponding authors. The articles were excluded if the corresponding authors did not respond in a reasonable time. The median age (with interquartile range) was converted to mean and standard deviation using a formula by Wan et al [19]. The cumulative dose was calculated from the total 5-FU or capecitabine dose received before cardiotoxicity occurred. ADRs were classified according to the Council for International Organizations of Medical Sciences [16] into five groups as follows: very common (frequency higher than 1/10), common (less than 1/10 to higher than 1/100), uncommon (less than 1/100 to higher than 1/1,000), rare (less than 1/1,000 to higher than 1/10,000), and very rare (less than 1/10,000).
Quality assessment
Three authors independently assessed the risk of bias in the included studies. The risk of bias in randomized controlled trials and non-randomized trials was evaluated using the Cochrane Risk-of-Bias tool 2.0 (RoB 2.0) [20], and the Risk Of Bias In Non-randomized Studies (ROBINs) [21], respectively.
Statistical analysis
The pooled prevalence of cardiotoxicity induced by 5-FU or capecitabine and 95% confidence interval (CI) were calculated using the DerSimonian-Laird random effect models [22] (OpenMetaAnalyst for Windows 8). Heterogeneity was assessed using Cochrane’s Q statistic and I2 values. The P-value of Cochrane’s Q of less than 0.10 was considered significant. I2 of greater than 75% indicated high heterogeneity while I2 of less than 25% indicated high homogeneity [23]. The funnel plot was used to observe publication bias.
Subgroup analysis and meta-regression
The influence of baseline characteristics, which may have caused heterogeneity, was determined by meta-regression (OpenMetaAnalyst for Windows 8 [24]). The effect of the following variables was planned for the analysis a priori: age, sex, study site, total cumulative dose, and year of study.
Results | ▴Top |
Study characteristics
From 34,705 articles from the systematic search, 80 studies were selected for meta-analysis [8, 15, 25-102]. Twenty-four were randomized studies and 56 were non-randomized studies. Details of the systematic search and screening are shown in the PRISMA diagram (Fig. 1). This systematic review included 18,524 participants, 5,836 of which were males. The average age of the participants was approximately 53.65 ± 7.32 years old. Most had gastrointestinal cancer such as colorectal cancer or gastric cancer. Almost all studies did not report patients’ medications for other comorbidities. Baseline characteristics are shown in Table 1 and additional participant characteristics are in Supplementary Material 2 (wjon.elmerpub.com). Information on dosage regimens and cardiotoxicity outcomes in patients who received 5-FU or capecitabine monotherapy is shown in Supplementary Materials 3 and 4 (wjon.elmerpub.com).
Click for large image | Figure 1. The PRISMA flow chart of the study selection. |
Click to view | Table 1. Characteristics of the Included Studies |
Risk of bias assessment
We found two randomized studies with a high risk of bias [27, 57] since more than 10% of data were missed and reasons for missing patients were not reported. The remaining had some concerns (N = 17) and a low risk of bias (N = 5) (Fig. 2). All non-randomized studies (N = 56) reported a crude prevalence of cardiotoxicity outcomes, so they were evaluated as having a critical risk of bias (Fig. 3).
Click for large image | Figure 2. The evaluation of the risk of bias in randomized control trials using the Cochrane Risk-of-Bias tool 2.0 (RoB 2.0). |
Click for large image | Figure 3. The evaluation of the risk of bias in non-randomized studies using The Risk Of Bias In Non-randomized Studies (ROBINs). |
Prevalence of 5-FU and capecitabine cardiotoxicity
Of 80 included studies, 70 reported the prevalence of cardiotoxicity in 5-FU users. The pooled prevalence of cardiotoxicity was 3.5% (95% CI: 2.7 - 4.2; P < 0.001; I2 = 73.86%; Fig. 4). In addition, 14 studies reported the prevalence of cardiotoxicity in capecitabine users. The pooled prevalence of cardiotoxicity was 2.8% (95% CI: 1.6 - 4.0; P < 0.001; I2 = 72.62%; Fig. 5). Cardiotoxicity from 5-FU and capecitabine was classified as common by the Council for International Organizations of Medical Sciences criteria. The funnel plot showed that most included studies were small. The symmetry was not evaluable since the pooled prevalence was close to zero causing the plot to distribute at the positive side of the funnel (Fig. 6).
Click for large image | Figure 4. Forest plot of cardiotoxicity in patients receiving 5-FU monotherapy. 5-FU: 5-fluorouracil. |
Click for large image | Figure 5. Forest plot of cardiotoxicity in patients receiving capecitabine monotherapy. |
Click for large image | Figure 6. A funnel plot for the publication bias. |
Meta-regression
Meta-regression revealed that the heterogeneity in the analysis of cardiotoxicity in 5-FU users was not caused by age, cumulative dose, place of study, sex, and year of study (P-value ≥ 0.099 in all analyses). The heterogeneity in the analysis of cardiotoxicity in capecitabine users was caused by place of study (P-value = 0.001) but not age, cumulative dose, sex, and year of study (P-value > 0.1 in all analyses). Details of the meta-regression analysis are shown in Supplementary Materials 5 and 6 (wjon.elmerpub.com).
Discussion | ▴Top |
The prevalence of any cardiotoxic events in 5-FU users was 3.5% and 2.8% in capecitabine users. Although cardiotoxicity was classified as common, the prevalence was still lower than other common ADRs. For example, the most common ADRs to 5-FU were diarrhea (64%), stomatitis (60%), and nausea/vomiting (51%), while the most common ADRs to capecitabine were hand-foot syndrome (62%), diarrhea (46%), and nausea/vomiting (36%) [103]. The prevalence of cardiotoxicity from 5-FU and capecitabine was lower than drugs well-known for cardiotoxicity such as anthracyclines, cyclophosphamide, and docetaxel which have a prevalence of, approximately 9% [104], 7-28%, and 2.3-8% [105], respectively. While our meta-analysis reported the estimated prevalence of cardiotoxicity from fluoropyrimidine monotherapy regimen as approximately 3%, a systematic review of cardiotoxicity from 5-FU and capecitabine as either monotherapy or combination regimens reports the prevalence as 0-20% and 3-35%, respectively [106].
Cardiotoxicity from fluoropyrimidines may be considered type A since evidence supporting pharmacological mechanisms is available. First, the most recognized mechanism is that fluoropyrimidines induce the release of vasoconstrictive mediators. For example, kinase C causes endothelium-independent vasoconstriction [107], and endothelin-1 is a potent vasoconstrictor that can induce coronary artery disease [108, 109]. Second, fluoropyrimidines induce vascular endothelial dysfunction and impaired oxygen delivery. Animal studies have shown that 5-FU can have direct toxic effects on vascular endothelial cells. This results in direct endothelial damage, fibrin, and platelet accumulation. Additionally, studies have shown that 5-FU can alter erythrocyte membranes, resulting in decreased oxygen transport in the blood and myocardial ischemia [110, 111]. Third, fluoropyrimidines degrade to alpha-fluoro-beta-alanine (FBAL) which causes a direct toxic effect on cardiomyocytes [112, 113]. From these mechanisms, clinical symptoms from the most common cardiotoxicity are angina and the less common are arrhythmias, myocardial infarction, heart failure, acute pulmonary edema, and cardiac arrest [114].
The cardiotoxicity of 5-FU and capecitabine can occur in the first cycle of use, 12 - 48 h after receiving the first dose [115]. The risk factors are not well understood. For example, the effect of pre-existing cardiovascular diseases [8, 25, 26, 116] on cardiotoxicity is inconclusive. Anyhow, there are some limitations in our study. First, our study analyzed the data from fluoropyrimidine monotherapy only. This may not reflect the cardiotoxicity in patients who used combination regimens with fluoropyrimidine and other drugs. Second, the risk of bias in the included studies was high. This is expected since prevalent studies are highly affected by biases in nature. Third, most studies did not report the severity of cardiotoxicity so the authors could not incorporate the severity data in the meta-analysis. Next, we cannot distinguish the cardiotoxicity from bolus versus continuous 5-FU since most studies used both types of administration. Last, we included seven studies [15, 26, 41, 47, 52, 74, 83] whose minor fraction of participants had pre-existing cardiac conditions. These studies did not specify whether the cardiotoxicity occurred in patients with pre-existing cardiac conditions. In addition, the small number of studies discouraged the meta-regression to determine the effect of pre-existing cardiac conditions on the pooled estimate.
This is the first systematic review and meta-analysis that identifies the prevalence of cardiotoxicity in 5-FU and capecitabine monotherapy users. The number of included participants is large so the prevalence can be reported more accurately. There are several applications for this study. First, although cardiotoxicity from fluoropyrimidines may be dose-dependent, this study did not support the association between the cumulative dose of 5-FU or capecitabine and the cardiotoxicity prevalence. The management of cardiotoxicity should be based on how type B reactions are managed including discontinuation. Non-dihydropyridine calcium channel blockers and nitrates should also be provided [114]. Second, cardiotoxicity is common and can include serious events. Therefore, patients should be followed up from the first cycle until the end of treatment. The follow-up should include an electrocardiography (EKG) which allows the detection of subclinical cardiotoxicity. Patients with underlying heart diseases should be closely monitored. Some studies suggest that colorectal cancer patients with dihydropyrimidine dehydrogenase (DPD) deficiency have an increased risk of cardiotoxicity, and therefore pretreatment screening of DPD activity may be considered [117]. Future studies should include studies that evaluate risk factors for cardiotoxicity from 5-FU or capecitabine more accurately.
Conclusion
The prevalence of cardiotoxicity from 5-FU and capecitabine was 3.5% and 2.8%, respectively. We did not find evidence that cardiotoxicity was associated with the cumulative dose of 5-FU or capecitabine.
Supplementary Material | ▴Top |
Suppl 1. Search term.
Suppl 2. Additional baseline characteristics of the included studies.
Suppl 3. Details of the dosage regimen and cardiotoxicity outcomes in patients who received 5-FU monotherapy.
Suppl 4. Details of the dosage regimen and cardiotoxicity outcomes in patients who received capecitabine monotherapy.
Suppl 5. Meta-regression for an association between the cardiotoxicity prevalence in 5-FU users and demographic data (age, cumulative dose, place of study, sex, and year of study).
Suppl 6. Meta-regression for an association between the cardiotoxicity prevalence in capecitabine users and demographic data (age, cumulative dose, place of study, sex, and year of study).
Acknowledgments
None to declare.
Financial Disclosure
This work was financially supported by the Research Grant of the Faculty of Pharmaceutical Science, Burapha University (grant no. 2/2566 (extra)).
Conflict of Interest
There is nothing to declare.
Informed Consent
Not applicable.
Author Contributions
BS and NL contributed to the research idea and design. BS, PT, AnP, SK, and TT contributed to data collection. BS and NL contributed to the statistical analysis and interpretation of data. BS wrote the first draft of the manuscript. NL and ArP edited the draft of the manuscript. All authors contributed to the critical revision of the manuscript for important intellectual content and approved and reviewed the final manuscript.
Data Availability
The data supporting the findings of this study are available from the corresponding author upon reasonable request.
Abbreviations
ADRs: adverse drug reactions; 95% CI: 95% confidence interval; CCRT: concurrent chemoradiation; EKG: electrocardiography; FBAL: alpha-fluoro-beta-alanine; 5-FU: 5-fluorouracil; NCCN: National Comprehensive Cancer Network; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses; RoB 2.0: the Cochrane Risk-of-Bias tool 2.0; ROBINs: the Risk Of Bias In Non-randomized Studies; WHO: World Health Organization
References | ▴Top |
This article is distributed under the terms of the Creative Commons Attribution Non-Commercial 4.0 International License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
World Journal of Oncology is published by Elmer Press Inc.