Search

Diuretics have been one of the well-known nephrotoxic drugs which can lead to acute kidney injury (AKI). However, there are few real-world studies on the incidence of AKI in hospitalized patients received diuretics. In the present study, a single-center retrospective study was conducted in our center. The clinical data of hospitalized patients received diuretics from January 2018 to December 2020 was retrospectively analyzed. Among the 18,148 hospitalized patients included in the study, 2,589 patients (14.26%) were judged as incidence with AKI, while only 252 patients were diagnosed with AKI in the medical record. Among diuretics drugs in the study, the incidence rate of AKI with torasemide was the highest with 21.62%, and hydrochlorothiazide had the lowest incidence rate (6.80%). The multiple logistic regression analysis suggested that complicated with hypertension, anemia, pneumonia, shock, sepsis, heart failure, neoplastic diseases, combined use of proton pump inhibitors (PPI) were independent risk factors for AKI related to diuretics. The logic regression models for diuretics related AKI were developed based on the included data. The model for diuretics-AKI achieved the area under the receiver operating characteristic curves (AUC) with 0.79 on 10-fold cross validation. It is urgent to improve the understanding and attention of AKI in patients received diuretics for medical workers, and the assessment of risk factors before the use of diuretics should be contributed to the early prevention, diagnosis and treatment of AKI, and ultimately reducing morbidity and improving prognosis.

Drug induced nephrotoxicity is a major clinical challenge, and it is always associated with higher costs for the pharmaceutical industry and due to detection during the late stages of drug development. It is desirable for improving the health outcomes for patients to distinguish nephrotoxic structures at an early stage of drug development. In this study, we focused on in silico prediction and insights into the structural basis of drug induced nephrotoxicity, based on reliable data on human nephrotoxicity. We collected 565 diverse chemical structures, including 287 nephrotoxic drugs on humans in the real world, and 278 non-nephrotoxic approved drugs. Several different machine learning and deep learning algorithms were employed for in silico model building. Then, a consensus model was developed based on three best individual models (RFR_QNPR, XGBOOST_QNPR, and CNF). The consensus model performed much better than individual models on internal validation and it achieved prediction accuracy of 86.24% external validation. The results of analysis of molecular properties differences between nephrotoxic and non-nephrotoxic structures indicated that several key molecular properties differ significantly, including molecular weight (MW), molecular polar surface area (MPSA), AlogP, number of hydrogen bond acceptors (nHBA), molecular solubility (LogS), the number of rotatable bonds (nRotB), and the number of aromatic rings (nAR). These molecular properties may be able to play an important part in the identification of nephrotoxic chemicals. Finally, 87 structural alerts for chemical nephrotoxicity were mined with f-score and positive rate analysis of substructures from Klekota-Roth fingerprint (KRFP). These structural alerts can well identify nephrotoxic drug structures in the data set. The in silico models and the structural alerts could be freely accessed via https://ochem.eu/article/140251 and http://www.sapredictor.cn, respectively. We hope the results should provide useful tools for early nephrotoxicity estimation in drug development.