A patient-specific approach for the IDH risk evaluation in the early stage of Haemodialysis

Introduction: Intra-Dialysis Hypotension (IDH) occurs in 25-30% of the hemodialysis (HD) sessions. An open challenge is to find an index to quantify the risk of IDH onset, in the early stage of HD, for each specific patient. A similar index (Ji), defined as the weighted sum of influent parameters has been recently described by Vito et al. Aim of this work is its optimization, validation, and evaluation in a wider set of patients. Methods: The patients have been classified in Hypotension Prone (HP) and Hypotension Resistant (HR), using an updated criterion (IDH_D) to identify IDH episodes. The statistical analysis was repeated to verify, also using IDH-D, the pertinence of the already identified predictors. Ji accuracy was verified on the 1st dataset (70 patients - 450 sessions) and evaluated on a 2nd dataset (60 patients – 360 sessions). Ji higher than 1 suggests a risk of IDH onset. The effects of the different protocols adopted to manage IDH in the four involved Dialysis Units were also studied, identifying and evaluating a center-dependent risk threshold. Results: Ji allows predicting the 77% of the IDH events when the threshold is set equal to 1. The use of center-specific thresholds allows slightly improving index specificity and sensitivity but does not substantially alter the results. Conclusions: Ji allows a reliable prediction of IDH risk, for each specific patient, at the early stage of the HD sessions. It can allow HD prescription optimization when patient-specific longitudinal data were available.