Prosthesis–patient mismatch after mitral valve replacement: A pooled meta‐analysis of Kaplan–Meier‐derived individual patient data

Abstract Objective The hemodynamic effect and early and late survival impact of prosthesis–patient mismatch (PPM) after mitral valve replacement remains insufficiently explored. Methods Pubmed, Embase, Web of Science, and Cochrane Library databases were searched for English language original publications. The search yielded 791 potentially relevant studies. The final review and analysis included 19 studies compromising 11,675 patients. Results Prosthetic effective orifice area was calculated with the continuity equation method in 7 (37%), pressure half‐time method in 2 (10%), and partially or fully obtained from referenced values in 10 (53%) studies. Risk factors for PPM included gender (male), diabetes mellitus, chronic renal disease, and the use of bioprostheses. When pooling unadjusted data, PPM was associated with higher perioperative (odds ratio [OR]: 1.66; 95% confidence interval [CI]: 1.32–2.10; p < .001) and late mortality (hazard ratio [HR]: 1.46; 95% CI: 1.21–1.77; p < .001). Moreover, PPM was associated with higher late mortality when Cox proportional‐hazards regression (HR: 1.97; 95% CI: 1.57–2.47; p < .001) and propensity score (HR: 1.99; 95% CI: 1.34–2.95; p < .001) adjusted data were pooled. Contrarily, moderate (HR: 1.01; 95% CI: 0.84–1.22; p = .88) or severe (HR: 1.19; 95% CI: 0.89–1.58; p = .24) PPM were not related to higher late mortality when adjusted data were pooled individually. PPM was associated with higher systolic pulmonary pressures (mean difference: 7.88 mmHg; 95% CI: 4.72–11.05; p < .001) and less pulmonary hypertension regression (OR: 5.78; 95% CI: 3.33–10.05; p < .001) late after surgery. Conclusions Mitral valve PPM is associated with higher postoperative pulmonary artery pressure and might impair perioperative and overall survival. The relation should be further assessed in properly designed studies.

available data were included in both "any PPM" as well as "moderate PPM" and "severe PPM" pooled analyses.

| Study endpoints
Primary endpoints were perioperative mortality and overall survival.
Secondary outcomes included residual PH (defined as the absence of postoperative pulmonary artery pressure normalization, in particular, residual pulmonary artery pressure >40 mmHg, as defined in the studies included in the review) and postoperative systolic pulmonary artery pressure. Based on the timing of echocardiographic measurement, studies were stratified in early (echocardiographic assessment during the index hospitalization) and late (echocardiographic assessment at a later time point during patient follow-up) period.

| Statistical analysis
Meta-analyses were performed using Review Manager, Version 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). Fixed and random-effects models were used to obtain pooled estimates. For late mortality, study results were subgrouped by study design type: unmatched/unadjusted observational data, riskadjusted observational data, and propensity score-matched data. Studies that reported both matched or risk-adjusted and unmatched/unadjusted data were included separately for subgroup comparisons.
Heterogeneity was examined with the I 2 statistics. The degree of heterogeneity was graded as low (I 2 < 25%), moderate (I 2 = 25%-75%), and high (I 2 > 75%). Sources of heterogeneity were explored by subgroup analyses of study (method used to obtain EOA, study location, year of publication) or patient characteristics (patient age). Additionally, a metaregression was performed to assess the potential effect of clinically

| RESULTS
The database search yielded 791 potentially relevant studies (Supporting Information Data B). After removal of duplicates and titleabstract screening, 25 full-text original articles were reviewed in further detail. Four studies were additionally excluded due to no differentiation in PPM and no-PPM groups in two, use of geometric orifice area to assess PPM in one and an insufficient number of patients included in one. Sixteen retrospective single-center studies, [4][5][6][10][11][12][13][14][15][16][17][18][19][20][21][22] two retrospective multicenter studies, 7,23 and one prospective study 24 were included in the final review and meta-analysis. Two studies identified were meta-analyses. 25,26 In one study, 19 a two-tailed analysis was performed and EOA was  (Table 1). Eight studies (including 5887 patients) divided the PPM group into moderate and severe PPM subgroups.
The 0.9-cm 2 /m 2 cut-off threshold was used to define severe PPM in all of these studies. Overall, the prevalence of any PPM was 50%. In the eight studies providing data on the severity of PPM, moderate PPM was seen in 57% and severe PPM in 13%.
The EOA was measured in vivo with the CE method in all participants in 7 (37%) studies. The PHT method was used to measure the EOA in 2 (10%) studies. Other studies used either referenced values from the literature or provided by the manufacturer (n = 7; 37%) or a combination of referenced values and in vivo measurements (n = 3; 16%).

| Risk factors for PPM
The use of bioprostheses demonstrated the strongest correlation with PPM (Supporting Information Data E). Furthermore, hypertension, PH, diabetes mellitus, and chronic renal disease were all associated with PPM. In contrast, female gender was related to a lower risk of PPM. Similar results were found when the risk factors for moderate or severe PPM were explored individually. The use of bioprostheses, diabetes mellitus, and impaired left ventricular function were associated with an increased risk of moderate PPM while female gender and atrial fibrillation were associated with a lower risk of moderate PPM. Similarly, the use of bioprostheses, diabetes mellitus, and chronic renal disease was associated with an increased risk of severe PPM.

| Perioperative mortality
Any PPM was associated with increased perioperative mortality   In the article, EOA is derived from two methods; only data derived from the pressure half-time method were included. c For articles providing both the results of any PPM as well as moderate or severe PPM subgroups on the endpoints of interest, the available data were included in both "any PPM" as well as "moderate PPM" and "severe PPM" pooled analyses.

| Secondary outcomes
PPM was associated with higher pulmonary pressure both in the early

| DISCUSSION
The most important finding of our study is that PPM resulted in reduced perioperative and overall survival. The results, however, need to be interpreted with caution as the method of EOA determination varied significantly across studies and the majority of data originate from unadjusted observational data.

| Method of EOA determination
In a recent study, Cho et al. 15

| Risk factors for PPM
The use of bioprostheses rather than mechanical prostheses de-