|Home | About | Journals | Submit | Contact Us | Français|
Patients with metabolic syndrome have a greater risk of cardiovascular disease, although their susceptibility to chemotherapy-induced cardiac disease is not well documented. The aim of this meta-analysis was to assess associations between obesity or being overweight and cardiotoxicity from anthracyclines and sequential anthracyclines and trastuzumab in patients with breast cancer.
We performed a random-effects analysis and a network meta-analysis and assessed publication bias. We included 15 studies and 8,745 patients with breast cancers who were treated with anthracyclines and sequential anthracyclines and trastuzumab.
Combination of obesity and being overweight was significantly associated with a greater risk of developing cardiotoxicity after anthracyclines and a sequential anthracyclines and trastuzumab regimen in patients with breast cancer. Pooled odds ratio for cardiotoxicity was 1.38 (95% CI, 1.06 to 1.80; I2 = 43%; N = 8,745) for overweight or obesity (body mass index > 25 kg/m2), 1.47 (95% CI, 0.95 to 2.28; I2 = 47%; n = 2,615) for obesity, and 1.15 (95% CI, 0.83 to 1.58; I2 = 27%; n = 2,708) for overweight. Associations were independent of study design, year of publication, drug regimen (anthracyclines alone v sequential anthracyclines and trastuzumab), or definitions of cardiotoxicity and of overweight or obesity. There was no evidence of publication bias; however, we could not separate the contributions of obesity-related cardiovascular risk factors, such as diabetes and hypertension, from that of obesity itself in this largely unadjusted analysis.
Our findings in a largely unadjusted analysis suggest that overweight and obesity are risk factors for cardiotoxicity from anthracyclines and sequential anthracyclines and trastuzumab.
Anthracyclines are anticancer agents with a broad spectrum of activity in oncological practice1; however, use of anthracyclines is limited by the acute and chronic dose-dependent cardiotoxicity they may induce.2-4 After administration of anthracyclines in an adjuvant setting of breast cancer, trastuzumab (Herceptin; Genentech, South San Francisco, CA), a humanized monoclonal antibody against the human epidermal growth factor receptor 2 (HER2 or ErbB2) protein,5 is effective in patients who overexpress this receptor. Combination of trastuzumab and chemotherapy provides a clinical benefit in terms of disease-free survival and overall survival compared with chemotherapy alone.6-9 Nonetheless, whereas preclinical studies did not reveal cardiotoxicity,10 later clinical studies showed that treatment with trastuzumab led to an unexpected incidence of cardiac adverse effects,11,12 the most frequent of which was reduced left systolic ventricular function, which was asymptomatic or associated with heart failure.13
Being overweight and obesity are complex health problems that affect a growing number of populations and which increase the risk of several chronic diseases, including cancer.14 Although it is has been known for years that patients with metabolic syndrome have a higher risk of cardiovascular disease,15 their susceptibility to chemotherapy-induced cardiac disease is not well documented. Moreover, obesity has recently been associated with poor outcomes—overall survival and disease-free survival—in patients with breast cancer who were treated with doxorubicin-based chemotherapy.16 Conversely, other studies have indicated that body mass index (BMI) had no effect on disease-free survival, but only a significant negative impact on overall survival,17 which suggested that obese women with breast cancer have poorer overall survival, possibly as a result of a higher risk of noncancer-related causes of death, as well as an unfavorable socioeconomic status and genetic background.18 Therefore, it is important to explore in greater depth the influence of overweight and obesity as aggravating factors in the development of cardiotoxicity, in particular, in the context of anthracyclines and sequential anthracyclines and trastuzumab-based therapy for breast cancer.
We conducted comprehensive literature searches in PubMed, Embase, Web of Science, Scopus, and ClinicalTrials.gov using predefined keywords [(anthracyclines or doxorubicin or epirubicin or trastuzumab) and (cardiac toxicity or cardiotoxicity or heart failure) and (BMI or body mass index or overweight or underweight or obesity or body weight) and (breast or breast cancer)]. We used both medical subject headings and text words in PubMed, and we used Emtree terms and text words in Embase. We used text words in Web of Science, Scopus, and ClinicalTrials.gov. We included all languages, even though we did not identify any eligible non-English papers. We excluded case reports, reviews, guidelines, editorials, and letters. We checked reference lists of included articles for additional studies. Moreover, we looked at the full text of studies on the associations of risk factors and cardiotoxicity of either anthracyclines or trastuzumab to identify additional studies.
The first author (C.G.) assessed the titles, abstracts, and full texts of the studies found and determined whether the studies examined the associations of weight-related factors with cardiotoxicity of anthracyclines and/or trastuzumab. We included cross-sectional and cohort studies as well as both population-based and hospital-based case-control studies in the systematic review. To be eligible for meta-analysis, studies had to report quantitative data on the association between overweight and/or obesity and cardiotoxicity of anthracyclines and/or trastuzumab. We also contacted authors of studies that were potentially eligible for inclusion in the meta-analysis to obtain additional information or new results.19-21
Two reviewers (C.G. and A.L.) independently assessed the quality of the study by using the Newcastle-Ottawa Scale adapted to the specific design of the study.22 Disagreements between the two reviewers were resolved by consensus. The definition used for cardiotoxicity was also evaluated as adequate or not according to the classic definition and to the recent expert consensus on Cancer Therapeutics–Related Cardiac Dysfunction.23
First, the proportion of cardiotoxicity was estimated in a meta-analysis by pooling the proportion of patients with cardiotoxicity in individual studies, which was calculated as the proportion of patients with cardiotoxicity divided by the number of patients who were treated with anthracyclines and/or trastuzumab. An arcsine transformation was used to stabilize the variation of proportions.24 A random-effect model was used according to DerSimonian-Laird’s method.25,26
Second, paired and network meta-analyses were conducted to pool the effect sizes of the weight status (normal weight, overweight, or obese) on the cardiotoxicity of anthracyclines and/or trastuzumab in patients who were treated for breast cancer. Use of meta-analyses as an adjustment method avoids a bias called Simpson’s paradox, which can be encountered in simple additions of patients and populations of each individual study.27 Network meta-analyses combine the information from direct evidence (pairwise comparisons) and indirect evidence (comparing B to C from comparisons of A to B and A to C). Odds ratios (OR), expressed with 95% CI, were calculated. A random-effect model with the DerSimonian-Laird method was used. The estimate from a random-effects meta-analysis is more conservative than that from a fixed-effect meta-analysis.28 Paired meta-analyses were performed by using STATA 10 (metan add-on; STATA, College Station, TX; Computing Resource Center, Santa Monica, CA).29 Network meta-analyses were performed by using R 3.1.0 software (netmeta package; The R Foundation, Vienna, Austria) according to the method described by Rücker.30,31 This is a frequentist method that is based on electrical networks and graph theory. It performs as well as the classic Bayesian network analysis, without the need to specify priors, and is easier to implement.32 Inconsistency was evaluated with Cochran’s Q.33
We assessed the presence of heterogeneity across the studies by means of the I2 statistic.34 The I2 statistic shows the total variation across studies that is not a result of chance. An I2 statistic < 25% indicates a small amount of inconsistency, and > 50% indicates a large amount of inconsistency.35 We used meta-regression to determine whether study-level covariates accounted for the observed heterogeneity. Main analyses were performed to compare the impact of overweight and obesity with that of normal weight. Subgroup analyses were conducted according to the treatment (anthracyclines only v trastuzumab with or without anthracyclines). The specific cardiotoxicity of trastuzumab alone was not assessable in the absence of individual data as all of the studies that included trastuzumab had a proportion of patients who were also treated with anthracyclines. We also performed subgroup analyses according to WHO recommended cutoff points for BMI [weight (kg)/height (m2)] and defined overweight as BMI 25 to 29.9 and obesity as BMI ≥ 30.36
To assess publication bias, we used a funnel plot, which compared the sizes of the overweight or obesity effects with their standard errors. We used the Egger regression test to examine funnel plot asymmetry. Statistical significance for publication bias was based on a P value < .1037
Preferred Reporting Items for Systematic reviews and Meta-Analyses was used as guidance for reporting this meta-analysis.38
The review flow is depicted in Figure 1. Our searches initially identified 581 abstracts. There were 123 relevant studies on the associations of cardiac risk factors with anthracyclines and sequential anthracyclines and trastuzumab cardiotoxicity in patients with breast cancer. Obesity was evaluated as a potential risk factor of cardiotoxicity in 15 studies; three of these were not included in the meta-analysis because the study population also included patients with nonbreast malignancies. We contacted the authors of 14 other studies that were eligible for the assessment of cardiotoxicity but did not mention obesity; three of them were able to provide additional data about the proportion of obese patients in the cardiotoxicity and noncardiotoxicity groups of their studies. Overall, we included 15 studies and 8,745 patients with localized and metastatic breast cancers who were treated with anthracyclines and sequential anthracyclines and trastuzumab (Table 1): seven were prospective cohort studies and eight were retrospective cohort studies. Most of these studies only included patients with localized breast cancer (11 of 15 studies) with adjuvant and neo-adjuvant chemotherapies. Patients with metastatic breast cancer were treated with anthracyclines and trastuzumab in the metastatic setting and accounted for a relatively small proportion of the patients included. Of the 15 studies included in this meta-analysis, four focused on anthracycline cardiotoxicity and 11 focused on trastuzumab with or without anthracycline cardiotoxicity—rates of anthracycline use in the trastuzumab studies are described in Table 1. Of the 15 studies, 11 reported results for overweight, 15 reported results for overweight or obesity, and 11 reported results for obesity. Similar results were reported from cross-sectional, case-control, and cohort studies; therefore, we combined all designs in a single analysis.
Overall study quality was satisfactory, although most studies stemmed from institutional databases and were retrospective (Table 2).
Mean rate of cardiotoxicity of the 8,745 patients analyzed was 17% (95% CI, 11 to 25). Patients treated with anthracyclines only (n = 3,898) had a rate of cardiotoxicity of 20% (95% CI, 5 to 43), whereas patients treated with trastuzumab with or without anthracyclines (n = 4,847) had a 16% (95% CI, 10 to 24) rate of cardiotoxicity.
By using a meta-analysis of the 11 studies that included patients treated with trastuzumab, we found that 91% (95% CI, 83 to 97) of the included patients were also treated with anthracyclines.
As shown in Figure 2, the combination of overweight and obesity was significantly associated with a greater risk of developing cardiotoxicity after an anthracyclines and sequential anthracyclines and trastuzumab regimen in patients with breast cancer.
Pooled odds ratio for cardiotoxicity was 1.38 (95% CI, 1.06 to 1.80; I2 = 43%; N = 8,745) for overweight or obesity (BMI > 25), 1.47 (95% CI, 0.95 to 2.28; I2 = 47%; n = 2,615) for obesity, and 1.15 (95% CI, 0.83 to 1.58; I2 = 27%; n = 2,708) for overweight.
Moreover, there was no significant effect of the drug regimen on the impact of overweight and obesity on the risk of cardiotoxicity (Fig 3).
By using network meta-analysis, subgroup analysis showed a gradual increase in the risk with higher BMI (P = .05): normal-weight patients had the lowest risk of developing cardiotoxicity, overweight patients had an intermediate risk [OR, 1.13 (95% CI, 0.80; 1.61) v normal-weight patients], and obese patients had the highest risk [OR, 1.27 (95% CI, 0.88; 1.86) v overweight and OR, 1.44 (95% CI, 0.99; 2.10) v normal-weight patients].
In this meta-analysis, estimates of studies on the association between obesity or overweight and cardiotoxicity were moderately heterogeneous (I2 = 43%; Fig 2).
In the metaregression of the 15 studies on the association between overweight or obesity and cardiotoxicity, heterogeneity across studies was not explained by study design, year of publication, geographic origin, drug regimen (anthracyclines alone v trastuzumab with or without anthracyclines), or the definition of cardiotoxicity and overweight or obesity. The proportion of anthracycline use in the studies was not associated with changes in the effect of overweight and obesity on cardiotoxicity by metaregression analysis (β = 2.09; P = .22).
The funnel plot of data from the 15 studies included in the meta-analysis was symmetrical (Appendix Fig A1, online only). The P value for the Egger test was .43.
This network meta-analysis on the basis of 15 studies and 8,745 patients shows that overweight and obesity are significantly associated with the risk of cardiotoxicity from anthracyclines and sequential anthracyclines and trastuzumab. In addition, drug regimen, that is, anthracyclines alone or sequential anthracyclines and trastuzumab, is not associated with a different effect of weight parameters on cardiotoxicity risk.
Most of the included studies used cardiotoxicity definitions (Table 1) that were considered adequate according to the consensual definition49 used before the new expert consensus.23 Moreover, in the metaregression of the 15 studies on the association between overweight or obesity and cardiotoxicity, heterogeneity across studies was not explained by the definition of cardiotoxicity. Thus, we can consider that the results were not significantly influenced by the slightly different definitions used.
In a recent meta-analysis on anthracycline cardiotoxicity, clinically overt cardiotoxicity occurred in 6% (95% CI, 3 to 9) of patients, whereas subclinical cardiotoxicity developed in 18% (95% CI, 12 to 24) of patients.50 These rates are consistent with ours in patients who received anthracycline [20% (95% CI, 5 to 43)] and can be explained by the high rates of asymptomatic left ventricular ejection fraction alterations. Concerning trastuzumab cardiotoxicity rates, the largest adjuvant clinical trials (Breast Cancer International Research Group,51 NSABP B-318) reported rates of left ventricular ejection fraction decline from 14% to 20%.52 These are likely an underestimate given that they are estimates in a clinical trial population that is likely healthier and with fewer cardiovascular comorbidities compared with the real world patient; however, these rates were comparable to those found in our study [trastuzumab plus anthracyclines, 16% (95% CI, 10 to 24)]. In a retrospective cohort analysis, Bowles et al53 reported a cumulative incidence of heart failure and cardiomyopathy among recipients of anthracycline and trastuzumab of 6.2% (95% CI, 4.1 to 8.2) after 1 year of follow-up, which increased to 20.1% (95% CI, 14 to 25.6) by 5 years. Thus, the duration of follow-up could also explain the different cardiotoxicity rates observed in the literature data.
Regarding the cardiotoxicity rates according to the drug regimen, we did not find a potentiation of anthracyclines cardiotoxicity when trastuzumab was also used. These results are discordant with those recently published,53,54 which demonstrated in large population-based retrospective cohort studies that risk of clinical cardiotoxicity, as assessed by congestive heart failure or cardiac death, was higher in patients with breast cancer who received a sequential therapy—anthracyclines followed by trastuzumab—than in those who were treated with other drug regimens. However, these studies used administrative billing codes to identify a congestive heart failure end point, usually associated with a much higher rate of cardiotoxicity. Moreover, neither our study nor our search strategy was designed to specifically evaluate the rates of cardiotoxicity according to the different drug regimens, as suggested by the relatively large CIs reported in the subgroup analysis. Thus, a potential lack of power as a result of population size and selection bias in the studies selected could have reduced the accuracy of our cardiotoxicity rates.
The link between obesity and heart failure (HF) in the general population is well established. Kenchaiah et al55 conducted the first large, epidemiologic study by analyzing 5,881 individuals from the Framingham Heart Study and showed that obesity was an independent risk factor for the development of HF. It concluded that for each increment of 1 kg/m2 in BMI, there was an increase in the risk of HF of 7% for women and 5% for men.55 The risk of developing HF was independent of age, alcohol, cigarette use, and comorbidities that included diabetes, hypertension, and a history of myocardial infarction.56 The present unadjusted analysis is the first to assess the potential role of overweight and obesity in cardiac dysfunction induced by anthracyclines and trastuzumab. The large population size allowed an accurate assessment of the excess risk associated with weight-related abnormalities. Obesity has been associated with a poor outcome in patients with breast cancer who were treated with anthracycline-based chemotherapy,17 affecting both overall survival and disease-free survival, and remained an important prognostic factor, independent of treatment dose.16,57 This, in part, could be explained by the higher rates of cardiac adverse events encountered by obese patients who are treated with anthracyclines or trastuzumab.
In fact, several studies have shown that obese patients58 or rats fed with a high-lipid diet59 are more sensitive to the cardiotoxic effects of anthracyclines. In an in vivo model of overweight induced by postnatal programming, we recently demonstrated that moderately overweight adult mice were more sensitive to cardiac systolic impairment and confirmed the potentiating action of trastuzumab on doxorubicin-induced cardiotoxicity in lean mice.60 Mechanisms by which overweight or obesity is able to promote anthracycline-induced cardiotoxicity may involve adiponectin downregulation.61,62 The dysregulation of adipocyte-derived hormones, adipocytokines, promotes the development of diverse obesity-linked diseases.63 Plasma adiponectin levels are decreased in obese patients.64 Adiponectin-KO mice showed exacerbated left ventricle contractile dysfunction after doxorubicin injection, whereas exogenous adiponectin improved doxorubicin-induced left ventricular dysfunction in wild-type and adiponectin-KO mice.62 However, no study has yet been performed to elucidate the role of adiponectin on trastuzumab-induced cardiac dysfunction.
The current study has several limitations. First, it was not possible to take into account the associated cardiac risk factors or the other classic risk factors of anthracycline and trastuzumab cardiotoxicity. Metaregressions to control for other risk factors would have required individual data for each study, which we did not obtain. Adjusted ORs were available in few studies and could not be used either; however, adjusted ORs were similar to crude ORs when they were available.20,40,44 Thus, the higher risk of cardiotoxicity in obese patients who were treated with these drugs could also be related to a higher prevalence of other risk factors in these patients. This could explain the moderate heterogeneity observed among the studies included in the analysis both regarding the rates of cardiotoxicity and the effect of overweight and obesity on cardiotoxicity. However, in unpublished data that combined breast cancer populations from two studies by Cardinale et al,19,20 we found that among patients with breast cancer who suffered from anthracycline and sequential anthracyclines and trastuzumab cardiotoxicity (n = 108), obese patients (n = 13) were not statistically different from normal-weight patients regarding the rates of hypertension, dyslipidemia, family history of coronary artery disease, and smoking, which thus suggested that the cardiovascular risk profile of obese patients with breast cancer was not significantly different from that of normal-weight patients.
Most of the studies included were not designed to specifically address the matter of obesity as a risk factor of drug-induced cardiotoxicity, thus leading to a possible bias in data collection. Moreover, definitions of both obesity and overweight, as well as cardiotoxicity, differed slightly among the studies selected. Finally, subgroup analysis was limited by the relative lack of power related to the small number of studies that included all three of the groups being compared. In addition, many studies included here reported limited details, and for several others we combined the occurrence of clinical and subclinical cardiotoxicity as a single end point to increase statistical power.
Our findings in a largely unadjusted analysis suggest that overweight and obesity are risk factors for cardiotoxicity from treatment with anthracyclines and sequential anthracyclines and trastuzumab. A careful assessment of cardiovascular risk factors is crucial to improve risk stratification for cardiotoxicity in patients with breast cancer who are treated with anthracyclines and trastuzumab. Further studies are needed to establish the independent predictive value of obesity on cardiotoxicity, but also to identify potential therapeutic targets to mitigate the increased risk of cardiotoxicity associated with overweight and obesity.
We thank Philip Bastable for English assistance.
Authors’ disclosures of potential conflicts of interest are found in the article online at www.jco.org. Author contributions are found at the end of this article.
Conception and design: Charles Guenancia, Annick Lefebvre, Sylvain Ladoire, Luc Rochette, Yves Cottin, Catherine Vergely
Provision of study materials or patients: Daniela Cardinale, Anthony F. Yu
Collection and assembly of data: Charles Guenancia, Anthony F. Yu, Marianne Zeller
Data analysis and interpretation: Charles Guenancia, Annick Lefebvre, Daniela Cardinale, Anthony F. Yu, Sylvain Ladoire, François Ghiringhelli, Marianne Zeller, Yves Cottin
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or jco.ascopubs.org/site/ifc.
Travel, Accommodations, Expenses: Biotronik International, Medtronic, Boston Scientific, Liva Nova, St. Jude Medical
No relationship to disclose
No relationship to disclose
Consulting or Advisory Role: Bristol-Myers Squibb, Clovis Oncology (I), AstraZeneca (I)
Research Funding: Clovis Oncology (I), Incyte (I), Astellas Pharma (I), AstraZeneca (I)
No relationship to disclose
No relationship to disclose
No relationship to disclose
No relationship to disclose
No relationship to disclose
No relationship to disclose