The quality of the information included in the selected studies of the meta-analysis was assessed through a combined strategy. First, potential publication bias was assessed by using a funnel plot (figure 1 of the supplementary data). Although Egger's test suggested a lack of asymmetry, its power was limited by the small number of studies (less than 10). Second, a review protocol was followed and assessment of alternative specific bias was analyzed demonstrating limited quality in terms of cointerventions but a low risk of bias for the rest of the evaluated items (figure 2 of the supplementary data). Finally, meta-regression analysis was performed according to the study design, as shown in figure 3 of the supplementary data, demonstrating that there was a clear trend favoring the use of RASi in the matched studies; indeed, only 1 study did not favor its use and it was the only study that was not matched. However, the P value (.110) was not statistically significant due to the small number of studies that could be included in this study. Other variables such as left ventricular ejection fraction could not be assessed through meta-regression analysis due to heterogeneity in data reporting.

After the assessment of 22 full-text articles, the final studies included in the meta-analysis were the 6 that met the inclusion criteria, as explained in methods, including 3 focused on transcatheter7–9 and 3 on SAVR patients.10–12 The flowchart summarizing study selection is depicted in figure 1. The main baseline characteristics and the total number of patients included in each study are reported in table 1. The pooled analysis was performed in 21 390 patients including 17 846 receiving TAVI, and 3544 treated with SAVR. Patients harboring transcatheter aortic valve replacement had a higher mean age (82.5 years)7–9 than those from surgical studies (74, 72, and 65.6 years, respectively).10–12 The distribution of this age difference is graphically presented in figure 2. No other baseline variable showed a clear pattern favoring any of the subgroups despite certain differences across the specific studies, namely a higher rate of hypertension reported by Inohara et al. in the transcatheter cohort,7 and a higher diabetes mellitus rate in the surgical cohort described by Goel et al.11

Figure 1.

PRISMA diagram for search strategy and study selection.

(0.24MB).

Figure 2.

Meta-regression for the variable of age in the surgical (red) and transcatheter (blue) therapeutic groups.

(0.05MB).

The mortality reported by each of the 6 studies is summarized in figure 3. The overall mortality for patients treated with TAVI, SAVR, and for the global population is also reported. The pooled analysis of the patients from the 6 studies included in this meta-analysis demonstrated a 36% reduction in the risk of death at 1 year of follow up (HR, 0.64; 95%CI, 0.47-0.88; P <.001); however, the subgroup analysis could not confirm this finding for the SAVR group (HR, 0.61; 95%CI, 0.29-1.30). Only patients who underwent TAVI had a significant mortality benefit at mid-term (HR, 0.67; 95%CI, 0.49-0.93). Lack of asymmetry was demonstrated (figure 1 of the supplementary data) with pooled HR, 0.779, 95%CI, 0.574-1.057; P=.11; I2=89.89% and, as shown in the Egger graph, the test did not provide evidence of the presence of small-study effects; in addition, sensitivity analyses sequentially eliminating dissimilar studies did not substantially alter the primary result favoring RASi prescription (figure 4).

Figure 3.

Forest plot showing reduction in all-cause mortality with renin-angiotensin system inhibitors in the global population treated with aortic valve replacement (yellow) and in the subgroup analysis for the transcatheter aortic valve recipients (TAVI, in blue), but not for the surgically treated patients (SAVR, in red). 95%CI, 95% confidence interval; D+L, DerSimonian and Laird; I-V, inverse variation.

(0.19MB).

Figure 4.

Sensitivity analyses showing no substantial alterations to the primary result favoring RASi prescription after sequential elimination of dissimilar studies. HR, hazard ratio.

(0.17MB).

Left ventricular mass regression, which seems to occur specially in patients with paravalvular regurgitation after TAVI has been independently related to RASi prescription.10 Indeed, persistent hypertrophy after SAVR has also been associated with worse clinical outcomes including mortality,15,16 whereas hypertrophy regression after SAVR following RASi prescription17 is associated with a lower incidence of myocardial infarction and stroke.16,18 However, the present meta-analysis could not confirm the association of RASi prescription after SAVR with better survival and potential left ventricular hypertrophy regression was not investigated.

RASi prescription in patients undergoing TAVI could also be associated with regression of myocardial fibrosis—probably more advanced in TAVI than in SAVR candidates given their advanced age—which may bring about better clinical outcomes including survival.8 Remarkably, in the propensity score matched study by Inohara et al., 7 RASi prescription was associated with a lower 1-year incidence of readmission for heart failure (absolute risk difference, −1.8%; 95%CI,−2.8% to−0.7%) after TAVI, but there was no difference in 1-year mortality between the RASi and no RASi groups (HR, 0.95; 95%CI, 0.81-1.12) among patients with reduced left ventricular ejection fraction (≤ 40%), despite lower 1-year mortality in the RASi group (HR, 0.78; 95%CI, 0.71-0.86) among patients with preserved left ventricular ejection fraction (> 40%), which is inconsistent with the recommendation in the current guidelines19,20 of RASi prescription only for patients with reduced left ventricular ejection fraction in the setting of heart failure. These contradictory findings, however, may be explained by a number of factors, such as smaller baseline left ventricular systolic dimensions and lower grades of postprocedural aortic and/or mitral regurgitation, which are suggested to be associated with greater left ventricular mass regression after SAVR21 and TAVI.7

Sympathetic modulation and antiarrhythmic effect have gained interest in the treatment of heart failure with RASi.22,23 The reduction in sympathetic activity as a mechanisms of modulating heart failure24 might explain the better outcomes of TAVI patients treated with RASi and could specifically benefit patients with residual aortic paravalvular regurgitation given the positive effect shown in the reduction of left ventricular volumes.7 On the other hand, patients with aortic stenosis often have concomitant coronary artery disease and myocardial ischemia; the electrolytic modulation induced by RASi can translate into an antiarrhythmic effect that might also improve outcomes.25

The present findings should be interpreted with caution because the results were drawn from nonrandomized studies (observational cohort studies), in which attrition is often worse and inadequately reported. In addition, the lower number of patients in the surgical group might explain the nonsignificant effect of RASi on mortality in that group. Additionally, prescription of RASi was defined at baseline but continuous treatment during follow up is unknown. Finally, publication bias might favor “positive” selection of those studies reporting good outcomes following RASi prescription; often, nonsignificant covariates in the univariate analysis are not entered into the multivariate analysis and are seldom reported; however, asymmetry was not detected with the present statistical assessment, suggesting minimal publication bias.