Differential Diagnostic Models Between Vasovagal Syncope and Psychogenic Pseudosyncope in Children

Abstract

Objective: We aimed to establish useful models for the clinical differential diagnosis between vasovagal syncope (VVS) and psychogenic pseudosyncope (PPS). Methods: This bicentric study included 176 patients (150 VVS and 26 PPS cases) for model development. Based on the results of univariate and multivariate analyses, a logistic regression model and a scoring model were established and their abilities to differentiate VVS from PPS were tested. Another 78 patients (53 VVS and 25 PPS) were used for external validation. Results: In the logistic regression model, the outcome indicated that the QT-dispersion (QTd) (P < 0.001), syncope duration (P < 0.001), and upright posture (P < 0.001) acted as independent factors for the differentiation of VVS from PPS, which generated an area under the curve (AUC) of 0.892. A cutoff value of 0.234 yielded a sensitivity and specificity of 89.3 and 80.8%, respectively, for the differentiation between VVS and PPS in the logistic regression model. In the scoring model which consists of three variables, a cutoff score of three points yielded a sensitivity and specificity of 91.3 and 76.9%, respectively, with an AUC of 0.909. The external validation test indicated that the negative and positive predictive values of the scoring model were 78.8 and 91.7%, respectively, and the accuracy was 80.8%. Conclusion: The scoring model consisting of three variables is an easy-to-perform, inexpensive, and non-invasive measure for initial differential diagnosis between VVS and PPS.

Keywords: differential diagnosis; logistic regression model; psychogenic pseudosyncope; scoring model; vasovagal syncope.

Figure 1

Flow chart of enrollment of study population in the first part of the study. PPS, psychogenic pseudosyncope; VVS, vasovagal syncope.

Figure 2

ROC curve of the diagnostic value of the logistic regression model and the scoring model between VVS and PPS. The y-axis represents sensitivity. The x-axis represents false positive rate (1-specificity). The 45° straight line stands for the reference line indicating sensitivity being equal to false positive rate. In the logistic regression model, the AUC was 0.892 (95% CI: 0.815–0.969; P < 0.001) and in the scoring model, the AUC of was 0.909 (95% CI: 0.845–0.973; P < 0.001). AUC, area under the curve; CI, confidence interval; PPS, psychogenic pseudosyncope; ROC, operation characteristic curve; VVS, vasovagal syncope.