Legemiddel­interaksjoner hos sykehuspasienter – forekomst og klinisk betydning

Abstract

Introduction: Little is known about the occurrence of drug interactions of hospitalised patients. We aimed to identify the frequency of potential drug interactions among hospitalised patients and, furthermore, to evaluate how many of these were problematic and clinically significant. Methods: We investigated drug interactions by two methods: by applying a computerised drug-drug interaction programme and by prospective clinical evaluation. The study was carried out at departments of internal medicine and rheumatology in five Norwegian hospitals in 2002. Patient characteristics and information on drug use were recorded for 827 patients consecutively included. Medication reviews were carried out by clinical pharmacists and drug-related problems (DRPs), among which drug interactions is one category, were identified and discussed in the hospital multidisciplinary team, chaired by a physician. Retrospectively, the patients’ drug regimens were screened by using a computer programme for drug-drug interactions, DRUID. This programme is universally used in Norway and classifies DDIs into four categories according to assumed severity: A, avoid; B, avoid/take precautions; C, take precautions; D, no action needed. Results: A total of 1513 DDIs were found in 544 patients (66% of the total sample of 827 patients) by computer screening. Many of these were related to drugs started in the hospital, that is they were new drug interactions. By bedside evaluation, 99 DDIs were found in 73 patients (9%). Of these, 89 were also identified by computer screening. Thus, only 6% of the computer identified drug interactions were assessed to be clinically problematic. Interactions of all degrees of severity, according to the computer programme, were identified as problematic by bedside screening. Drugs most often causing new drug interactions were warfarin, acetylsalicylic acid, digitoxin and combinations of codeine and paracetamol. Conclusions: The majority of hospitalised patients have potential drug interactions, however, less than one in ten have interactions of clinical importance. Computer graded severity has limitations when intended to be used as an indicator for clinical importance. Another way of picking up new, possibly serious drug interactions would be to select indicator drugs, which are drugs frequently known to be involved in interactions, and then undertake the main interaction search on these drugs.