Drug interactions in primary care: impact of a new algorithm on risk determination

If managed adequately, many drug interactions do not result in clinical manifestations. Earlier studies may have overestimated the risk arising from drug interactions because they usually reported interaction frequencies and the severity of potential outcomes irrespective of their manageability.

Our objective was to estimate the risk associated with drug interactions in a large population when not only the severity of possible clinical events but also measures of their prevention (manageability, modulating factors) are considered.

We evaluated all drug pairs concurrently prescribed to 9481 adults aged 50 to 75 years who participated in a health-screening examination. Drug interactions were evaluated by use of an algorithm with 4 decision layers (severity, manageability, risk/benefit assessment, and patient-related risk factors), and this risk evaluation was compared with the conventional evaluation solely on the basis of severity.

More than 52% of the patients received combination therapy. Interaction information was available in a standard source (DRUGDEX; Thomson MICROMEDEX, Greenwood Village, Colo) for only 1029 of all 13,672 individual prescribed drug pairs. Of the drug pairs, 881 (6.4%) were identified as interacting. Of these 881 interactions, 132 (15.0%) were of major severity but 101 of 132 (76.5%) were considered manageable. Only 31 (23.5%) of 132 major interactions (ie, 31/881 [3.5% of all interacting pairs]) offered no management options and should thus be avoided.

For many commonly prescribed drug pairs, explicit drug interaction information is currently not available in DRUGDEX. For major interactions with published evidence, the overwhelming majority were manageable, and therefore risk estimates only based on the severity of potential outcomes will grossly overestimate the risk associated with such combinations.