A systematic search across PubMed, EMBASE and Cochrane databases were performed from inception of the databases until November 2021 for relevant studies. Supplementary file provides the search strategy and MeSH terms used in each database (Supplementary Table S1). In addition to database search, we also reviewed the reference lists of included articles to identify additional studies.

The present analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and American Heart Association guidelines (10, 11). Since, only studies with prior ethical committee clearance were included in the present analysis, no separate ethical clearance was required.

After checking for duplicates, the searched articles were screened for relevant studies. The inclusion criteria were as follows: (a) observational studies, reporting right ventricular echocardiography outcomes at baseline and post Anthracycline and/or Trastuzumab chemotherapy among breast cancer patients; (b) no exclusion based on sample size; (c) reporting one of the following right ventricular echocardiography outcomes; right ventricle ejection fraction (RVEF), right ventricular fractional area change (RVFAC), right ventricular global longitudinal strain (RVGLS), right ventricular free wall longitudinal strain (RVFWLS), tricuspid annular plane systolic excursion (TAPSE); (d) exclusion of studies conducted exclusively in hematologic malignancies and pediatric cancer survivors and (e) finally, studies published in English language. The screening for inclusion was performed independently by two reviewers (A.J., M.S.) at two separate levels. At level one, title and abstracts of searched citations were screened for relevance. At the second level, articles identified from level one screening were subjected to complete manuscript review and considered when they met the inclusion criteria.

Data extraction from included studies was performed independently by two reviewers (A.J., M.S.), and later examined for similarity. The following information was extracted from original primary publications; author, year of study, sample size, diagnosis, cancer chemotherapy agent used, percentage of patients treated with Anthracyclines, follow-up period, baseline and follow-up values of RVEF, RVFAC, RVGLS, RVFWLS, TAPSE, and left ventricular ejection fraction (LVEF). The data extraction was performed as per a pre-designed data extraction form. Any disparity was resolved by mutual consensus and after consultation with other authors. Studies included observational studies without any interventions, therefore bias tools such as the ROBINS-I tool were not employed. The National Institute of Health (NIH) Quality Assessment Tool for Before-After (Pre-Post) Studies with No Control Group was utilized for risk of bias assessment in the current analysis.

We used the National Institutes of Health (NIH) Quality Assessment Tool for Before-After (Pre-Post) Studies with No Control Group. It included the following assessments; pre-specified entry criteria, sample size, blinded assessors of data and clear descriptors of outcome measurements and others. Two authors independently judged each domain, with a third author resolving conflicts.

We used the inverse variance method with a random effect model and DerSimonian and Laird method of Tau2 generation to calculate mean difference [MD] with 95% confidence interval [CI]. We used the method provided by Hozo et al., to calculate mean and standard deviation from studies reporting median and inter quartile range (12). To obtain a conservative result with a wide confidence interval, we used a correction (r = 0) between the pre-treatment and post treatment mean while calculating the standard deviation of the pooled estimate. Statistical heterogeneity was assessed using Higgins's I² or chi-square P-value (13). The pooled estimate was deemed statistically heterogeneous if Higgins's I² was >50% or chi-square P-value was < 0.05. Meta-analysis results were reported graphically using forest plots: the measure of effect (MD) was represented by a square, with the area being proportional to study weight. A p value < 0.05 was considered significant. The analysis was carried out using RevMan Version 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014).

The database search identified a total of 1,449 articles after checking for duplicates. Figure 1 represents the PRISMA flow chart for the inclusion of studies. 1,426 studies were excluded at the first level of screening and 15 studies were excluded at the second level of screening. Seven studies were included after manual search of reference list of the included studies. Finally, fifteen studies met the inclusion criteria and were included in the final analysis (14, 15, 24–26, 16–23). The fifteen studies constituted a total of 644 patients. All studies included breast cancer patients who were treated with Anthracycline and/or Trastuzumab chemotherapy, other than study by Cottin et al., 1996 which included women with breast cancer and lymphoma. Follow-up period of all included studies was less than 12 months. Baseline characteristics of patients from the included studies are presented in Table 1. We included studies which assessed right ventricular parameters by 2D echo parameters including RVEF, RVFAC, RVGLS, RVFWLS, TAPSE, as defined by the American Society of Echocardiography definitions and reference ranges for each parameter in Table 2. Summaries of individual studies included in the present systematic review and meta-analysis are presented in Table 3.

Per the NIH Quality Assessment Tool for Before-After (Pre-Post) Studies with No Control Group, all the included studies had sample size not sufficiently large to provide confidence in the findings, and people assessing the outcomes were not blinded to the participants’ exposures/interventions, hence potentially leading to bias in the reported results.

Our study reported significant findings. Anthracycline and/or Trastuzumab chemotherapy led to a statistically significant decrease in Right Ventricular Ejection Fraction (RVEF), with six studies involving 168 patients providing data on RVEF [MD: 2.70, 95% CI: 0.27 to 5.13, P = 0.03, I² = 71%, χ2 P < 0.05] (Figure 2A). Additionally, Right Ventricular Fractional Area of Change (RVFAC) decreased significantly following the chemotherapy treatment with Anthracycline and/or Trastuzumab, as depicted in five studies [MD: 3.74, 95% CI: 1.33 to 6.15, P < 0.01, I² = 68%, χ2 P < 0.05] (Figure 2B). Chemotherapy with Anthracycline and/or Trastuzumab had no effect on Right Ventricular global longitudinal strain (RVGLS) [MD: -1.97, 95% CI: -4.16 to 0.23, P = 0.08, I² = 51%, χ2 P = 0.15] (Figure 2C). Right Ventricular free wall longitudinal strain (RVFWLS) declined at follow-up, [MD: -1.00, 95% CI: -1.86 to -0.15, P < 0.05, I² = 0%, χ2 P = 0.40] (Figure 3D). Furthermore, Left Ventricular Ejection Fraction (LVEF) was also reduced at follow-up [MD: 4.04, 95% CI: 2.08 to 6.01, P < 0.01, I² = 91%, χ2 P < 0.05] (Figure 3F). Tricuspic annular plane systolic excursion (TAPSE) was not affected by the treatment [MD: 0.53, 95% CI: -0.11 to 1.17, P = 0.11, I² = 98%, χ2 P < 0.05] (Figure 3E). However, most of the pooled estimates were associated with significant statistical heterogeneity, mostly because of the heterogeneity in the magnitude of effect measures rather than variation in effect of chemotherapy with Anthracycline and/or Trastuzumab on echocardiography parameters.

Results from one of the earlier trials investigating the effect of Anthracyclines on RV systolic and diastolic function using radionuclide angiography showed no significant change in RV EF during the one-year follow-up (42, 43). This is non-congruent to our own findings, which demonstrate a statistically significant drop in RV EF at follow-up. A study by Grover et al. using CMR to assess RV and LV function, not only concurred with our findings, demonstrating a decline in RV EF at 12-month follow-up but also, demonstrated a steeper decline in RV EF as compared to reduction in LVEF (20). Our meta-analysis established a statistically significant reduction in RVFAC, the parameter for RV radial systolic function among patients treated with Anthracyclines and/or Trastuzumab. This is on par with prior reports (19, 23, 26). Evaluation of RV strain is an important component of quantitating RV dysfunction, however it is a measurement that is subject to several methodological complexities. Consistent with our findings, several studies have demonstrated a significant decline in RVFWLS. Chang et al. (27) reported that RVFWLS may be a critical parameter for evaluation of occult RV dysfunction, suggestive of the increased susceptibility for damage within the thinner RV wall. Overall, among the studies that evaluated RV parameters, evaluation of strain may provide the most utility in early prediction of the effect of chemotherapy on the right ventricle. The discrepancies among the studies are most probably due to echocardiographic imaging techniques, as image acquisition and interobserver variability which may compromise the overall assessment. Additionally, the study population consisted of varying cardiovascular comorbidities, follow up periods and small sample sizes.

Despite the heterogeneity among the study population and the various parameters studied, there appears to be substantial evidence to suggest chemotherapy induced RV dysfunction. Larger studies with longer follow-up is necessary to confirm and establish the clinical impact and long term effects of these findings.

There is a growing body of evidence emphasizing the predictive significance of right ventricular structure and pathophysiology in different cardiac conditions (5, 6, 44–49). RV function is adversely affected even in untreated cancer patients, in part due to the abnormal physiology of increased pro-inflammatory markers, neurohumoral changes and circulating reactive oxidative species (47–49). However, there exists less evidence and guidance regarding the evaluation of right ventricle in patients undergoing chemotherapy. The implications of right ventricular dysfunction in cancer patients treated with Anthracycline and/or Trastuzumab have been reported in literature. Oliviera et al. (50) reported significant differences in the need for right ventricular assist devices in patients with chemotherapy associated cardiomyopathy, the majority of whom were treated with Anthracycline compared to all other causes of cardiomyopathy (19% vs. 9.3%, p < 0.0001), suggesting that RV dysfunction likely occurs in parallel with the more defined left ventricular dysfunction in this patient population (50).

Milano et al. (51) demonstrated a decrease in RV and LV thickness among mice infused with Doxorubicin or a combination of Doxorubicin and Trastuzumab. Both groups expressed increased RV free wall fibrosis in contrast to placebo or mice treated with Trastuzumab alone (51). Thus, it is intuitive that RV structure and function might be compromised following chemotherapy in human beings.

The present meta-analysis has several limitations. First, this is a study level meta-analysis. Study level meta-analyses are limited in their ability to identify inter study heterogeneity. The dose of Anthracyclines and/or Trastuzumab used varied across studies. However, this has not been accounted for in the present analysis. All studies included in the present analysis had a follow-up period of less than 12 months. This was planned to avoid selection bias of survivors and because Anthracyclines and/or Trastuzumab induced cardiotoxicity is reversible. The study by Cottin et al., included both breath cancer and lymphoma patients. Finally, the present analysis did not assume any correlation between the pre and post treatment mean, so that a conservative pooled estimate with a wide confidence interval is obtained.