Data collection
The data were collected from the medical records of enrolled patients. Age, gender, body mass index (BMI), past medical history, length of hospital stay, intensive care unit (ICU) admission (directly or after admission to the ward), patients’ outcome (discharge or death), and the medications prescribed during 24-48 h of hospital admission were gathered. Patients who stayed in the hospital for less then 48 h were excluded. The Lexi-Interact database (Lexicomp®, Wolters Kluwer, Hudson, Ohio, United States, available on UpToDate, 2020) was applied to assess pDDIs. Based on the severity, the interactions were categorized into five categories including A (unknown), B (minor), C (moderate), D (major), and X (contraindicated) [11]. The reliability of the interactions was scaled as excellent (E = the interaction has been clearly demonstrated in well-controlled studies), good (G = the studies strongly suggest that interaction exists; however, the proof of well-controlled studies is lacking), and fair (F = available evidence is poor, but clinicians suspect interaction on the basis of pharmacologic considerations; or, evidence is good for an interaction of pharmacologically similar drug) [12, 13].