ABC | Volume 114, Nº1, January 2019

Original Article Otto et al. Aortic prosthesis mismatch in public health system Arq Bras Cardiol. 2020; 114(1):12-22 The objective of this study was to assess the frequency of PPM in a representative population treated in the Brazilian Public Health System and to identify the preoperative factors that are associated with the occurrence of PPM. Methods In this cross-sectional retrospective study, performed from January 2011 to July 2016, we included patients older than 18 years who underwent AVR. Patients who died prior the first postoperative echocardiography or with incomplete clinical and echocardiographic data were excluded. Informed consent was obtained from each patient and the study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the institution’s human research committee. All subjects underwent surgical AVR and transthoracic echocardiogram (ETT) within 30 days after surgery. Three hundred and sixteen (316) patients met the inclusion criteria. However, data of indexed EOA (iEOA) to determine the degree of PPM was available only in 176 patients. These data were not found in the echocardiogram report, nor were the values for calculation in images available at the hospital imaging server in 140 patients. After the publication of the European prosthetics guidelines 1 there was mandatory standardization for calculation of iEOA in our Echocardiogram Laboratory. The echocardiographic evaluation was performed following the recommendations of the American Society of Echocardiography Guideline, obtaining two-dimensional, pulsed and continuous Doppler and M mode images with Philips HDI 5000, HD 7, iE33 or GE E9 ultrasound systems with 2-5 Hz multifrequency transducer. The left atrial volume and ejection fraction (EF) were measured by Simpson's method (for LVEF < 53%) or Teicholz (for LVEF ≥ 53%). LV mass was obtained by the Devereux formula (measured from M or 2-dimensional mode) and indexed to the BSA. 13 LV diameters were obtained by Mor bi-dimensional mode. 13 LV outflow tract (LVOT) was evaluated at the plane before the aortic valve, 1,10,14 the peak and mean gradients, the velocity time integrals (VTI) ratio of the LVOT and aortic prosthesis and the calculation of EOA were performed according to the ASE recommendations. EOA = (LVOT area × LVOT VTI)/Aortic flow VTI). 1,10 The calculation of EOA was indexed to BSA estimated by the Dubois and Dubois formula: BSA = (Weight 0.425 × Height 0.725 ) × 0.007184 and was used to identify the degree of PPM. 1,5 Definitions of PPM Definition #1: PPM was defined as severe if iEOA was < 0.65 cm 2 /m 2 , moderate if iEAO was between 0.65 cm 2 /m 2 and 0.85 cm²/m² and absent if iEOA > 0.85 cm 2 /m 2 . Definition #2: We also used the definition of PPM adjusted for high BMI as recommended by European recommendations. 1 For BMI < 30 mg/kg, moderate PPD is considered if iEOA is < 0.70 cm 2 /m 2 and severe if iEOA < 0.55 cm 2 /m 2 . Definition #3: Severe PPM was also defined on the basis of the mean transprosthetic gradient > 20 mmHg. We tested three different definitions for PPM in this study population to check which of them would identify better variables associated with mismatch. Statistical analysis Continuous variables with normal distribution were presented as mean and standard deviation and categorical variables in absolute numbers and percentages with confidence intervals, when necessary. Means of the three PPM groups were compared with one-way ANOVA after the Shapiro‑Wilk normality test and Tukey's post hoc test. For categorical variables, a Chi-square test was use to compare proportions and frequencies. The association between preoperative variables and occurrence of severe PPM was assessed using the Poisson regression with robust variancemodel. In the univariate analysis, the association between each independent variable and the occurrence of PPM was assessed, and those that presented p < 0.1 were selected for entry into the multivariable analysis. The multivariable models were built by the consecutive exclusion of one variable from each complete model that presented the highest value of p of the Wald test, as described by Hosmer and Lemeshow. Data for multivariable models were complete for 148 patients. A receiver operating curve (ROC) analysis was performed to assess the predictive value of the multivariable model for the prediction of severe PPM. ROC analysis was performed only for the PPM definition with more independent variables in the study, i.e., Definition #2. The analyses were conducted using the SAS 9.4 software and p < 0.05 was considered significant. Results Frequency and Comparison of PPM Groups Severe and moderate PPM occurred in 33.4% and 36.2% of patients, respectively. Tables 1 and 2 compare baseline clinical and echocardiographic characteristics of the 3 PPM groups. Even though 19% of the patients (34 patients out of 176 with PPD data) had rheumatic etiology, few presented significant mitral valve disease and underwent concomitant valve surgery (Table 1). There was loss of iEOA data in 140 patients with an average gradient of 18.7 ± 7 mmHg and a peak of 32.1 ± 5 mmHg. Patients with severe PPM were younger and had larger BSA and BMI, smaller LVOT diameter, and higher prevalence of rheumatic heart disease. There was low incidence of aortic root enlargement at the time of surgery in all groups. Table 3 shows the types and numbers of implanted prostheses and was presented in a descriptive way according to the type, number and category of PPD. There was a wide range of types and sizes of prostheses used in the AVR, which makes it impossible to analyze the association between prosthesis type, prosthesis number and degree of PPD. The data in Table 3 is too sparse to allow for any statistical model. Saint Jude bioprosthesis was implanted in 58% of patients (it is the most frequent), but cannot be tested as a determinant of PPM. 13