Objective The aim was to develop a drug-drug interaction database (SFINX) to be integrated into decision support systems or to be used in website solutions for clinical evaluation of interactions. Methods Key elements such as substance properties and names, drug formulations, text structures and references were defined before development of the database. Standard operating procedures for literature searches, text writing rules and a classification system for clinical relevance and documentation level were determined. ATC codes, CAS numbers and country-specific codes for substances were identified and quality assured to ensure safe integration of SFINX into other data systems. Much effort was put into giving short and practical advice regarding clinically relevant drug-drug interactions.Results SFINX includes over 8,000 interaction pairs and is integrated into Swedish and Finnish computerised decision support systems. Over 31,000 physicians and pharmacists are receiving interaction alerts through SFINX. User feedback is collected for continuous improvement of the content. Conclusion SFINX is a potentially valuable tool delivering instant information on drug interactions during prescribing and dispensing.

Previous studies have rarely been powered to determine the quantitative influence of specific CYP2C9 genotypes on warfarin dose requirements. The results from our pooled analysis are likely to be the most accurate to date and the methodology could serve as a model for future pharmacogenetic meta-analyses.

ABSTRACT:The most important enzyme in hepatic drug metabolism is cytochrome P450 3A4. Published in vitro data indicate that vitamin D may up-regulate the expression of the CYP3A4 gene. Individual vitamin D levels are highly dependent on sunlight exposure and show great seasonal variability in northern countries. The aim of the present study was to investigate whether plasma concentrations of CYP3A4 drug substrates exhibit seasonal changes compatible with a stimulatory effect of vitamin D on drug metabolism. Three immunosuppressants (tacrolimus, sirolimus, and cyclosporine) were analyzed, because these CYP3A4 drug substrates are subject to long-term use and repeated concentration determinations. In addition, mycophenolic acid was included in the analysis as a control drug independent of CYP3A4 metabolism. Concentration-to-dose ratios were extracted from the Karolinska Therapeutic Drug Monitoring database and compared between the 3-month periods of lowest and highest vitamin D levels. Sirolimus and tacrolimus levels showed seasonal variability that was highly consistent with changes in vitamin D; for example, significantly lower drug concentrations in July to September than in January to March. As expected, no significant difference was evident for mycophenolic acid, but this result was also the case with cyclosporine, possibly due to cross-reactivity of CYP3A4-mediated metabolites with the immunoassay used for quantification. In conclusion, there is cyclic variation in blood levels of important immunosuppressants throughout the year that correlates with UV light-dependent changes in vitamin D levels. Even though a causal relationship remains to be established, it is suggested that individual differences in vitamin D may contribute to variability in drug metabolism and disposition.

Drug usage has increased steadily, and the more drugs used, the higher the risk for adverse effects or loss of effect due to drug-drug interactions. For drug prescribers it is difficult to know what drugs a patient is taking and whether they interact. Computerizing of health care records has made it possible to connect patients' drug lists to clinical decision support systems giving the prescriber information about e.g. drug-drug interactions, duplicated prescriptions and drugs in pregnancy. The aim of this thesis is to create a knowledge base suitable for usage in decision support systems, to evaluate the database in clinical practice, and to use existing clinical databases to create new knowledge about possible drug-drug interactions and their mechanisms. LIST OF PUBLICATIONS This thesis is based on the following publications. The articles will be referred to in the text by their roman numerals.