We restricted inclusion in this meta-analysis to studies that met all the following inclusion criteria: (1) RCTs, (2) Patients undergoing chemotherapy; (3) Receiving ARNI comparing with control being standard care, placebo or no treatment; (4) Reporting any outcomes of interest; and (5) studies available for review in English and in full text. We excluded studies (1) with no control arm; (2) overlapping population; (3) studies that evaluated patients with a history of cancer, but not in treatment at the moment; and (4) Studies reported only as abstracts or conference presentations were excluded.

We conducted a systematic search of PubMed, EMBASE, and the Cochrane Central Register of Controlled Trials through November 2025 for studies published in English, with no temporal restrictions. The following search terms in all 3 databases: (“Chemotherapeutic agents” OR “Antitumor drugs” OR “Cancer treatment OR “Cardiotoxicity”) AND (“Angiotensin-Neprilysin Inhibition” OR sacubitril OR Entresto OR “Sacubitril-Valsartan” OR ARNI). The exact search strategy applied to each database is reported in the Supplementary Material . Furthermore, references from all included studies, previous systematic reviews, and meta-analyses were manually searched for any additional studies. Two authors (J.P.O and C.O.F.B) independently extracted data following predefined search criteria and quality assessment methods. Disagreements between the 2 authors were resolved by consensus, with the assistance of a third author (R.H).

The outcomes of interest were CTRCD events, left ventricular ejection fraction (LVEF), GLS, blood pressure, hypotension, and dyspnea events. CTRCD was defined as reported in each study ( Supplementary Material 1 ).

Risk ratios (RR) with 95% confidence intervals (CIs) were calculated to compare the incidence of binary outcomes between ARNI and control groups, while continuous outcomes were assessed using mean differences (MDs) with 95% CIs. For binary end points, we applied the Mantel–Haenszel random-effects model, and for continuous outcomes, we used the inverse-variance approach. A restricted maximum likelihood random-effects model was employed for all outcomes to address methodological and demographic variability across the included studies. Heterogeneity was evaluated using Cochran’s Q test, the I² statistic, and τ², and was categorized as low (I² = 0%–25%), moderate (I² = 26%–50%), or high (I² > 50%).

For 2 continuous outcomes (LVEF and GLS), we planned to extract the mean change from baseline. In studies that reported only baseline and final mean ± SD values, the mean change was calculated by subtracting the final mean value from the baseline mean value. The SD of Change was estimated using the formula recommended by the Cochrane Handbook. ( Supplementary Material 2 ).

All statistical analyses were performed using R statistical software using the meta package (version 4.4.2). We estimated missing means and SD from medians and interquartile range using the method by Wan et al, as per Cochrane recommendations.

A leave-one-out sensitivity approach was conducted for all outcomes. In this analysis, each study was sequentially excluded, and the pooled RR or MD was recalculated based on the remaining studies.

The risk of bias in each study was evaluated using the tool recommended by the Cochrane Collaboration Handbook. We evaluated the risk of bias in RCTs using the Risk of Bias 2 (RoB-2) tool. The assessment was performed by 2 independent investigators (MEM and MYS), and any disagreements were resolved by consensus, with the involvement of a third author (RH) when necessary.

In addition, the certainty of the evidence was appraised by 2 other independent reviewers (LAL and RH) using the Grading of Recommendations Assessment, Development and Evaluation (GRADE). In accordance with Cochrane recommendations, funnel plots and Egger’s test were not conducted because fewer than 10 studies were included for each outcome (n < 10).

The systematic search yielded 1,680 records. After the removal of duplicate records and studies not related to the topic of this meta-analysis, a total of 17 remaining studies were fully reviewed, and 4 studies. ^,,, met the inclusion criteria and were incorporated into the quantitative synthesis ( Figure 1 ).

PRISMA flow diagram of study screening and selection.

In total, 412 participants were included, of whom 176 (42.7%) received ARNI versus standard of care, placebo or no treatment. Follow-up durations ranged from 6 to 18 months. The mean age across studies was 51.8 years, and most participants were female. Baseline LVEF values were preserved in all RCTs (ranging from 57% to 64%), and GLS was nearly identical across the studies (−19% to −21%), as summarized in Table 1 . Notably, exposure to anthracyclines was substantial across trials, with doxorubicin-equivalent doses ranging from 167 to 276 mg/m². Three studies included HER2-directed monoclonal antibodies, with trastuzumab used in 5% to 40% of participants and pertuzumab reported only in the Hsu study. Taxane-based regimens were frequently employed in PRADA II and SARAH (62%–96%). Other concomitant therapies, particularly cyclosporine and corticosteroids, were commonly used in the SARAH and Katogiannis studies, as detailed in Supplementary Table S2 .

In patients undergoing chemotherapy, ARNI therapy was associated with measurable improvements in cardiac function, including improvement in GLS (MD −0.93%; 95% CI −1.49 to −0.38; p < 0.01; I² = 0%; Figure 2 ) and increases in LVEF (MD 1.47%; 95% CI 0.59 to 2.34; p < 0.01; I² = 0%; Figure 2 ). However, ARNI therapy was not associated with a significant reduction in chemotherapy-induced cardiac dysfunction, as reflected by CTRCD incidence (RR 0.40; 95% CI 0.08–1.97; p = 0.258; I² = 29.3%; Figure 3 ). Similarly, no differences were observed in all-cause mortality (RR 0.63; 95% CI 0.08–5.01; p = 0.659; I² = 0%; Figure 3 ).

The effect of Sacubitril/valsartan on echocardiographic parameters. ( A ) GLS was significantly improved in patients receiving Sacubitril/valsartan compared with control. ( B ) LVEF was also significantly increased in the Sacubitril/valsartan group. ARNI = angiotensin receptor–neprilysin inhibitor; GLS = global longitudinal strain; IV = inverse variance; LVEF = left ventricular ejection fraction; MD = mean difference.

The effect of Sacubitril/valsartan on clinical outcomes. ( A ) The incidence of CTRCD was numerically lower in the Sacubitril/valsartan group, although no statistically significant difference was observed. ( B ) All-cause mortality did not differ between the Sacubitril/valsartan and control groups. ( C ) Hypotension was significantly more frequent among patients receiving Sacubitril/valsartan. ARNI = angiotensin receptor–neprilysin inhibitor; CI = confidence interval; CTRCD = chemotherapy-related cardiac dysfunction; MH = Mantel–Haenszel; RR = risk ratio.

Moreover, ARNI therapy increased the risk of hypotension (RR 4.58; 95% CI 1.55–13.48; p = 0.006; I² = 0%; Figure 3 ), consistent with significant reductions in systolic blood pressure (SBP) (MD −9.98; 95% CI −15.51 to −4.45; p < 0.01; I² = 77.3%; Supplementary Figure S1a ) and diastolic blood pressure (DBP) (MD −5.72; 95% CI −8.71 to −2.73; p < 0.01; I² = 62%; Supplementary Figure S1b ). No significant between-group differences were observed in NT-proBNP levels (SMD −0.19; 95% CI −0.45 to 0.07; p = 0.25; I² = 16.9%; Supplementary Figure S2a ) or dyspnea rates (RR 0.64; 95% CI 0.27–1.47; p = 0.289; I² = 0%; Supplementary Figure S2b ).