Risk of Congestive Heart Failure in Early Breast Cancer Patients Undergoing Adjuvant Treatment With Trastuzumab: A Meta-Analysis

This meta-analysis showed that adjuvant trastuzumab therapy was strongly associated with an increased risk of significant congestive heart failure in patients with early breast cancer, particularly after 2 years of use.

Abstract

Background.

The use of trastuzumab has proven to be a successful strategy in patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer; however, it is associated with an increased risk of cardiac dysfunction. We performed an up-to-date, comprehensive meta-analysis to clarify the risk of congestive heart failure (CHF) in patients with early breast cancer receiving different durations of adjuvant trastuzumab with the longest-term follow-up.

Methods.

Eligible studies included randomized control trials of HER2-positive early breast cancer patients with or without trastuzumab in adjuvant chemotherapy. Adequate reporting of CHF data were required for inclusion. Statistical analyses were conducted to calculate the overall incidence, relative risk (RR), and 95% confidence interval (CI) by use of a fixed-effects model.

Results.

Six randomized control trials including 18,111 patients were identified. The overall incidence of high-grade CHF in patients treated with trastuzumab versus placebo was 1.44% (95% CI, 0.79%–2.64%) and the RR was 3.19 (95% CI, 2.03–5.02; p < .00001). In subgroup analysis, the difference in CHF incidence failed to achieve significance. The RR for 8 mg/kg trastuzumab (high dose) was greater than that for 4 mg/kg (low dose) (RR, 6.79, 95% CI, 2.03–22.72, p = .0001; versus RR, 2.64; 95% CI, 1.61–4.32; p = .002). Additionally, higher RRs were observed for patients receiving trastuzumab for 1 year (RR, 3.29; 95% CI, 2.07–5.25) and 2 years (RR, 9.54; 95%CI, 2.19–41.43), but not 9 weeks (RR, 0.50; 95% CI, 0.05–5.49) compared with control groups. No evidence of publication bias was observed.

Conclusion.

Adjuvant trastuzumab therapy was strongly associated with an increased risk of significant CHF in patients with early breast cancer, particularly in 2-year use.

Implications for Practice:

This comprehensive meta-analysis evaluated the risk of congestive heart failure with a usage profile of adjuvant trastuzumab in patients with early breast cancer. Before initiating treatment with trastuzumab, a risk-benefit analysis for individual patients should be critically evaluated, considering that the prognosis is closely related to drug dose and duration of use. Cardiac function should be monitored throughout the treatment period and also during follow-up. Thus, early identification of trastuzumab-related cardiac dysfunction can allow effective medical intervention, elimination of symptoms, recovery of function, and continuation of trastuzumab therapy.

Introduction

The use of trastuzumab has proven to be a successful strategy in patients with human epidermal growth factor receptor 2 (HER2)-positive early-stage and metastatic breast cancer [1, 2]. It has also been to shown to improve clinical outcomes in several malignancies, including gastric cancer [3, 4], urothelial carcinoma [5], advanced bladder cancer [6], and non-small-cell lung cancer [7], and it is currently approved by the U.S. Food and Drug Administration for the treatment of these malignancies.

However, the treatment is associated with an increased risk of cardiac dysfunction, which may be attributed to the blockage of HER2 signaling in cardiac myocytes [8]. In particular, congestive heart failure (CHF) is the most serious adverse effect of trastuzumab.

Three previous meta-analyses [9–11] have reported the cardiac risk associated with trastuzumab in several trials of trastuzumab in breast cancer. However, those studies were limited to shorter-duration trastuzumab use, and the CHF profile of 2-year trastuzumab was not reported. Patients with early-stage breast cancer may be treated with adjuvant trastuzumab over long periods; hence this topic is of special importance in these patients. We sought to investigate the incidence and risk of CHF in the most up-to-date meta-analysis of randomized controlled trials (RCTs) of adjuvant trastuzumab in early breast cancer (EBC) patients.

Methods

Data Sources and Selection of the Studies

We reviewed PubMed citations from January 1966 to July 2015. The search criteria included only randomized controlled trials published in the English language and the keywords trastuzumab (as well as its commercial name, Herceptin) and breast cancer or adenocarcinoma of the breast. We also searched abstracts and virtual meeting presentations from American Society of Clinical Oncology (ASCO) conferences from January 2004 to July 2015 and abstracts from the San Antonio Breast Cancer Symposium (SABCS) from 2007 to 2015. When more than one publication was identified from the same clinical trial, we used the most recent or complete report of that trial. An independent search using the citation database Web of Science (developed by the Institute for Scientific Information) was conducted to ensure that no clinical trials were missed. The most recent trastuzumab package insert was accessed to identify relevant information [12].

Data Extraction and Clinical Endpoints

Data abstraction was conducted independently by three investigators (H.-D. Long, Y.-E. Lin, and J.-J. Zhang) according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [13], and any discrepancies among reviewers were resolved by consensus. For each study, we extracted the following information: study characteristics, year of publication, patient characteristics, trastuzumab treatment information, median follow-up duration, and reported CHF events. The primary outcomes were high-grade CHF, including symptomatic CHF, definite cardiac death (due to CHF, myocardial infarction, or primary arrhythmia), or probable cardiac death (sudden death without documented etiology). High-grade CHF (namely symptomatic CHF) was defined as New York Heart Association class III or IV.

Statistical Analysis

For the calculation of incidence, the number of patients with CHF events and patients treated with trastuzumab were extracted from the safety profiles of selected RCTs. The incidence of patients with CHF and 95% confidence intervals (CIs) were derived from each trial. We also calculated relative risk (RR) and 95% CI of CHF events in patients assigned to trastuzumab versus controls in the same trial.

Statistical heterogeneity among trials was estimated by using the Cochran Q statistic [14]. Heterogeneity was considered statistically significant at p_{heterogeneity} <.1. If heterogeneity existed, data were analyzed using a random-effects model. In the absence of heterogeneity, a fixed-effects model was used. A two-tailed p value of less than .05 was considered statistically significant. We also conducted subgroup analysis for trastuzumab loading dose and duration of treatment. Finally, publication bias was evaluated by using Begg’s and Egger’s tests [15, 16]. Analyses were conducted using Review Manager version 5.2 (Cochrane Collaboration, Oxford, U.K., http://www.cochrane.org), Stata/SE version 12.0 (Stata Corp., College Station, TX, http://www.stata.com), and R statistical software, version 2.15.0 (https://www.r-project.org).

Results

Search Result

Our search yielded a total of 601 trastuzumab studies: 536 abstracts from PubMed and 65 from ASCO or SABCS meetings. After evaluating each study, 543 were initially excluded. We carefully screened each of the remaining 27 trials and excluded an additional 20 trials for being duplicates, subgroup analysis only, or combined analysis. In the FinXX trial, CHF events were not reported [17]. Finally, six trials were selected for inclusion in the meta-analysis [18–24]. HERA [23] and BIG1-01 [24] were the same trial, and reports of CHF events were the same. A detailed selection process is represented in Figure 1.

Figure 1.

Flow diagram of studies identified, included, and excluded.

Abbreviation: HER2, human epidermal growth factor receptor 2.

Trial Characteristics

The baseline characteristics of each trial are listed in Table 1. Six RCTs (five large [18, 19, 21, 22, 24] and one small [20]) were open, phase III, prospective, multicenter trials. Participants were randomly assigned to receive treatment with or without trastuzumab.

Table 1.

Characteristics of randomized controlled clinical trials included in the meta-analysis

Data from 18,111 patients were available for the meta-analysis. All selected trials included early-stage patients with breast cancer, with adequate organ function, coagulation, and hematologic function. All the trials typically excluded patients with uncontrolled hypertension, clinically significant congestive heart failure, cerebrovascular disease, peripheral vascular disease, unstable angina, or recent history of myocardial infarction. All trials allowed inclusion of patients with prior anthracycline exposure; moreover, all trials treated patients with concomitant anthracycline and patients with HER2-positive breast disease. CHF events were reported in each trial (Table 2).

Table 2.

Summary of CHF events with trastuzumab

The longest median follow-up was 8 years in the HERA trial, which was the only one with a 2-year duration of trastuzumab, followed by the NASBP B-31, with a median follow-up of 7 years. FinHer and BCIRG 006 both had 5-year median follow-up times. In short, two-thirds of the trials had a median follow-up of ≥5 years.

Incidence of CHF

The overall incidence of CHF among 8,615 patients receiving trastuzumab from six trials was 1.44% (95% CI 0.79%–2.64%) according to the random-effects model (Q = 45.19; p < .0001; I² =89%) (Table 3).

Table 3.

Incidence and RR of CHF for patients with breast cancer who were treated with trastuzumab, stratified by trastuzumab loading dose and duration of treatment

RR of CHF

To define the particular contribution of trastuzumab to the development of CHF and exclude the influence of confounding factors, we determined the overall RR of CHF from the six RCTs. The overall RR of high-grade CHF with trastuzumab was 3.04-fold higher (95% CI 1.12- to 7.85-fold; p < .00001) than for patients who did not receive trastuzumab. No significant heterogeneity among the included trials was observed (Q = 6.92; p = .23; I² =28%) (Fig. 2).

Figure 2.

Relative risk of high-grade congestive heart failure events associated with trastuzumab versus control among patients with breast cancer.

Abbreviations: BCIRG, Breast Cancer International Group; CI, confidence interval; FinHer, Finland Herceptin; HERA, Herceptin Adjuvant trial; M-H, Mantel-Haenszel; PACS, Program Adjuvant Cancer Sein.

Influence of Trastuzumab Dose

There are two approved doses for trastuzumab, 4.0 mg/kg (low loading dose) [18–20, 22] and 8.0 mg/kg (high loading dose) [21, 24]. Therefore, we attempted to look at incidence and RR stratified by trastuzumab dose. The incidence rates of CHF were 0.95% (95% CI, 0.68%–1.33%) and 1.64% (95% CI, 0.74%–3.57%) (Table 2) for the high and low doses, respectively, with no statistically significant difference (p = .23). Nevertheless, the overall RRs of CHF for the high and low doses were 6.79 (95% CI, 2.03–22.73) and 2.64 (95% CI, 1.62–4.33), respectively, for which a significant difference was found (p < .00001) (Table 3).

Influence of Duration of Trastuzumab

The duration of trastuzumab treatment was 1 year [18, 19, 22], 2 years [23, 24], or 9 weeks [21], with CHF incidence of 1.45% (95% CI 0.78%–2.66%), 1.09% (95% CI, 0.67%–1.78%), and 0.94% (95% CI, 0.19%–4.51%), respectively. The respective RRs were 3.29 (95% CI, 2.07–5.25), 9.54 (95% CI, 2.19–41.43), and 0.50 (95% CI, 0.05–5.49). The difference in incidence was not statistically significant (p = .75), but the RR was (p < .00001) (Table 3).

Sensitivity Analysis

We deleted each study one at a time from the overall pooled analysis to check the influence of the removed data set to the overall RR. Two studies (BCIRG 006 [18] and HERA [24]) materially changed the between-study heterogeneities in trastuzumab dose and duration of trastuzumab, respectively. After the deletion of either of the two studies, the heterogeneity vanished, whereas the conclusion stayed significant (Fig. 3).

Figure 3.

Sensitivity analysis for the influence of individual studies on the overall RR for the association between trastuzumab and congestive heart failure events. The vertical midline indicates the overall RR; the lateral vertical lines indicate its 95% CI; and circles indicate the pooled RR when the left study is omitted in this meta-analysis. The two ends of every dotted line represent the respective 95% CI.

Abbreviations: BCIRG, Breast Cancer International Group; CI, confidence interval; FinHer, Finland Herceptin; HERA, Herceptin Adjuvant trial; PACS, Program Adjuvant Cancer Sein; RR, relative risk.

Publication Bias

No obvious publication bias was detected for the RR of CHF in trastuzumab-treated patients with breast cancer by either Egger’s (p = .711) or Begg’s (p = .707) test (Fig. 4).

Figure 4.

Funnel plot studies evaluating the relative risk of high-grade congestive heart failure events associated with trastuzumab versus control patients with breast cancer. Egger’s test, p = .711; Begg’s test, p = .707.

Discussion

To our knowledge, this is the first long-follow-up report to show a significant increase in the risk of CHF in EBC patients treated with adjuvant trastuzumab. The overall incidence of clinically significant (high-grade) CHF was 1.44%. Despite the reasonably low absolute incidence, our results demonstrate an overall increase of threefold in the CHF risk of trastuzumab-treated patients compared with controls. Patients with EBC may be at an especially increased risk of CHF because of their trastuzumab use profile or other exposure factors in the adjuvant setting.

Stratified analyses were performed to evaluate subgroups with potentially higher incidence of CHF. No different incidence rate was found between high and low doses of trastuzumab, but sevenfold and threefold higher RRs were observed with high and low doses versus no trastuzumab, indicating that dose is closely related to risk of CHF. This also suggests that the low dose may already be reaching the saturation level to impair cardiac function. Other crucial findings include a 3.29-fold higher risk when using trastuzumab for 1 year and a 9.54-fold higher risk for 2 years compared with patients who did not receive trastuzumab. However, caution should be taken when analyzing these subgroups; because of the RCT setting, we are not able to directly determine whether the analysis is more significant or simply equivalent.

Consonant with previous meta-analysis [11], the incidence of CHF was still low among patients receiving trastuzumab, but the risk of CHF was increased by approximately three to four times compared with controls. Unlike previous studies limited to shorter-term follow-up that did not simultaneously quantify heterogeneity according to trastuzumab dose and duration of use, our study has updated currently available data of trastuzumab-associated CHF in EBC patients with longer-term follow-up, for example 5 years [18, 20], 7 years [22], and 8 years [24]. In addition, the incidence of CHF in the 2-year use duration of the HERA trial was also low; intriguingly, CHF events tended to become more uncommon at the later 5-year follow-up.

Unlike that of anthracycline, trastuzumab-related cardiac dysfunction is not dose dependent and is generally reversible [25]. As recommended in the package insert [12], trastuzumab is administered in an initial loading dose of 4 mg/kg and a once-weekly maintenance dose of 2 mg/kg until disease progression or metastasis. Nonetheless, in the adjuvant setting, trastuzumab can be given with an initial loading dose of 8 mg/kg and 6 mg/kg every 3 weeks in addition to this weekly dosing. From our available data for analysis (only HERA and PACS-04 trials used 8 mg loading dose and 6 mg maintenance; the other trials [NCCTG N9831, FinHer, BCIRG 006, and NSABP B-31] used the 4-mg dose), the higher dose carried a 6.79-fold higher risk for CHF, compared with 2.64-fold for the lower dose, versus the observation group. This difference between the two dose regimens is significant (p < .00001), suggesting that dose may be closely associated with the risk of CHF. Because of the RCT setting, CHF data could have been recorded more strictly than in general practice. As a result, our study provides evidence that high-dose trastuzumab indirectly causes a greater risk of CHF than low-dose therapy.

To determine whether duration of adjuvant trastuzumab treatment can affect the risk of CHF, we carried out a subgroup analysis. Similar improvements in progression-free and overall survival were found for for 2 years versus 1 year of trastuzumab use [18, 22–24]. The 2-year arm of HERA showed the highest RR for CHF during adjuvant trastuzumab; 1-year treatment demonstrated that the cumulative RR of CHF tends to plateau after treatment completion; and the 9-week duration in FinHer failed achieve statistically different RRs for CHF, with no benefit to overall survival. These data suggest that 1 year might be the ideal use of trastuzumab for early-stage breast cancer patients (supplemental online Fig. 1).

Because the patients enrolled in clinical trials had already been screened for cardiac function, and age has been regarded as a risk factor of cardiac toxicity [26], our meta-analysis might underreport the cardiac toxicity, because some older patients with cancer might be excluded owing to suboptimal cardiac function or other complications, leading to a lower observable incidence of CHF. Moreover, all of these studies excluded patients with any cardiac disease, and this is partly why the incidence seems lower than in clinical experience. Recently proposed guidelines for patients receiving trastuzumab recommend a monitoring schedule that assesses baseline and on-treatment cardiac function and intervention strategies with cardiovascular medication to improve cardiac status before, during, and after treatment. If severe CHF occurs, it is strongly recommended to stop trastuzumab therapy [27, 28].

Recently proposed guidelines for patients receiving trastuzumab recommend a monitoring schedule that assesses baseline and on-treatment cardiac function and intervention strategies with cardiovascular medication to improve cardiac status before, during, and after treatment. If severe CHF occurs, it is strongly recommended to stop trastuzumab therapy.

The strengths of our pooled analysis are the large sample, the assessment of CHF risk according to trastuzumab use, and the consistency of the primary findings and the internal validation study. The principal limitation of our analysis is the unavailability of individual patient information. Therefore, establishment of risk factors for the development of CHF (including prior exposure to cardiotoxic agents) or of potentially contributing cocomorbidities (including prior cardiovascular disease) is not possible in this analysis. In contrast, in a retrospective cohort study of trastuzumab- and anthracycline-based breast cancer therapy in 12,500 patients, an increased rate of CHF (12.1%) was observed in a community-based setting, in which a higher prevalence of cardiac risk factors might exist [29]. It is important to point out that expanded access programs are usually designed to include confounding factors, so the incidence of CHF might not truly reflect the actual trastuzumab-related CHF risk among EBC patients, whether consistent with the clinical trials or not. In a recent report, trastuzumab treatment was found to be associated with a significant increase in the risk of severe hypertension [30]. Therefore, an increase in the risk of CHF may have been secondary to increased incidence of severe hypertension. However, we could not relate the incidence of CHF to secondary hypertension, as neither causality nor association was reported in any trial. It is also important to note that death, significant disability from high-grade CHF, and timing of CHF were not captured in the studies analyzed; therefore, the impact of trastuzumab on these aspects cannot be determined. These are typical disadvantages inherent to meta-analysis of literature-based studies, and only individual patient data will be able to answer these questions and establish clinical risk factors for trastuzumab-induced CHF. Nevertheless, meta-analysis from individual patient data can also carry significant bias, as data are available to limited numbers of research groups and for only a few types of studies that have high public health priority; consequently, when an opportunity arises, pooled analysis may be possible in the future.

Conclusion

Trastuzumab significantly increases risk of high-grade CHF in the longest-term follow-up of HER2-positive patients with EBC compared with control groups. A higher risk of trastuzumab-associated CHF compared with no trastuzumab was seen with a loading dose of 8 mg/kg and treatment duration of 2 years, irrespective of other benefits. The findings have important implications for physicians and investigators to optimize the balance between oncologic clinical benefit and life-threatening adverse events; this adverse effect should be closely monitored. Importantly, more optimal therapeutic strategies need to be investigated.

See http://www.TheOncologist.com for supplemental material available online.

Author Contributions

Conception/Design: Hui-Dong Long

Provision of study material or patients: Hui-Dong Long

Collection and/or assembly of data: Hui-Dong Long, Yun-En Lin, Rui-Nian Zheng

Data analysis and interpretation: Juan-Juan Zhang, Rui-Nian Zheng

Manuscript writing: Hui-Dong Long, Wen-Zhao Zhong

Final approval of manuscript: Hui-Dong Long

Disclosures

The authors indicated no financial relationships.