Capacidade de 4 equações de referência para prever a força muscular respiratória de idosos neuropatas: estudo analítico

Abstract

Introduction: There are many equations to estimate respiratory muscle strength (RMS) with different predictive values. Thus, the chosen equation can decisively influence the diagnosis of the situation of the respiratory muscles and interfere with clinical decision-making. The chosen reference equation becomes as relevant as the correct technique and the measured result. Objective: To compare the predictive value of 4 reference equations with the FMR values: maximum inspiratory pressure (MIP) and maximum expiratory pressure (MEP), measured in elderly neuropaths and indicate the equations that best fits. Method: Analytical observational study. The capacities of 4 predictive equations were compared with the RMS measured in 220 elderly people, with 108 functional neuropaths and 112 healthy controls. To verify the normality of the data, the Shapiro-wilk test was used, descriptive analyzes were performed. To assess the factors that influenced the participants' health status (healthy or neuropathic), comparisons were analyzed using the Chi-square and Mann-Whitney tests. Pearson's correlation was used to correlate the variables weight and BMI among the elderly with respiratory muscle weakness. The comparison between the values of predicted by each equation and the measured values was made by adjusting a Logistic Regression model. To check the fit, the predictive capacity of the model was calculated: area under the ROC curve (AUC), Sensitivity, Specificity, Positive Predictive Value (VPP), Negative Predictive Value (VPN) and Accuracy (ACC), and the quality measures of adjustment of the equations: R², mean absolute deviation (MAD), mean absolute percentage error (MAPE) and root mean square error (RMSE), we established a ranking ranking according to the predictive capacity. The software used was R (version 3.5.1). Results: None of the equations was able to predict the PImax of the studied sample. In elderly neuropathies, the Neder equation had a better potential in predicting MIP (higher AUC) and better predictive capacity in general rancking. For MEP, the Sanchez E1 equation was the one that best fitted with the best predictive capacity when analyzing VPP, VPN, ACC and AUC. Conclusion: For the reference values of the RMS in elderly neuropaths, the Neder equation was the most adjusted to the measured values of MIP and the Sanchez E1 equation was the most adjusted to the measured values of MEP.