Summary

The aim of drug therapy is in general to cure diseases or reduce symptoms. However, drug therapy is ineffective in 30 to 60 percent of the patients and, on the other hand, two to four percent of all hospital admissions result from adverse drug reactions. A better prediction which patients will not respond to drug therapy or will develop adverse drug reactions may avoid these events (chapter 1). In this thesis, we analyzed both the effect of co-prescribed
drugs and genetic variation on drug response. In chapter two, we studied the exposure to and clinical consequences of drug-drug interactions (DDI). The exposure to DDIs in the elderly general population (≥70 years of age) has almost doubled between 1992 and 2005 from ten to nineteen percent (chapter 2.1). Also
the exposure to potentially life threatening DDIs almost doubled from 1.5 percent in 1992 to 2.9 percent in 2005. In the Netherlands, pharmacists are obliged to intervene prescriptions that imply a high risk for the patient. In chapter 2.2 and 2.3 we studied which factors were
involved in the dispensing of prescriptions which involved a DDI with a high risk for patient harm. In the literature, the relationship with the prescriber, the medication surveillance software and pharmacy organization were described as factors associated with these dispensings. In a subsequent study, we analyzed whether these factors were associated with the dispensing of high-risk DDIs in community pharmacies in the Netherlands. Pharmacies using the Euroned medication surveillance program and pharmacies that were part of a health care centre dispensed one high risk DDI more often. The clinical consequences of DDIs were studied in a literature review (chapter 2.4). About half a percent of all hospital admissions
were due to DDIs. In the elderly this proportion was about five percent. In chapter three we studied the effect of genetic variation on the response to drugs used in the treatment of type 2 diabetes mellitus. The antidiabetic drug tolbutamide, one of the drugs in the sulfonylurea group, is metabolized by CYP2C9. Incident tolbutamide users with a CYP2C9*3 variant allele were prescribed lower doses than users with the wildtype genotype, most likely due to a decrease in tolbutamide metabolism (chapter 3.1). In chapter 3.2 and
3.3 the antidiabetic drug metformin was studied. Metformin is not metabolized, but genetic
variation in transporters involved in the carriage of metformin may affect the glucose lowering effect. Metformin is a substrate for the organic cation transporter 1 (OCT1), encoded by the SLC22A1 gene, and the multidrug and toxin extrusion 1 (MATE1) transporter, encoded
by the SLC47A1 gene. Both OCT1 and MATE1 are located in the hepatocyte, OCT1 transports metformin into the hepatocyte and MATE1 transports metformin out of the hepatocyte into the bile. We studied whether genetic variation in these genes is associated with the change in HbA1c level in incident metformin users. The rs622342 minor C allele in the SLC22A1 gene was associated with a 0.3 % smaller reduction in HbA1c level and the rs2289669 minor A allele in the SLC47A1 gene with a 0.3 % larger reduction. Most likely, the rs622342 C allele codes for a crippled OCT1 transporter, and the rs2289669 A allele for a crippled MATE1 transporter. In chapter 3.4 we describe an interaction between these
polymorphisms. The effect of the rs2289669 polymorphism is stronger in patients with the rs622342 CC genotype than in patients with the AA or AC genotype. In patients with a crippled OCT1 influx transporter (rs622342 CC genotype) and a normal functioning MATE1 efflux transporter (rs2289669 GG genotype), the MATE1 efflux transporter will outperform the OCT1 influx transporter, resulting in low intracellular metformin levels and a hampered glucose lowering effect. In other patients, OCT1 will outperform MATE1, and the glucose Recently, the rs10494366 SNP in the NOS1AP gene was associated with an increased QTc interval. Most likely, this SNP regulates intracellular calcium levels through an effect on the inward calcium channel currents. Sulfonylurea, a group of antidiabetic drugs, indirectly trigger the opening of voltage dependent calcium channels. In view of these similarities, we studied in chapter 3.5 whether this SNP is associated with response to sulfonylurea. In
glibenclamide users, the rs10494366 TG and GG genotype are associated with a reduced response and higher mortality rates than in glibenclamide users with the TT genotype. In tolbutamide and glimepiride users these genotypes were associated with lower mortality rates. The effects of genetic variation on two groups of cardiovascular drugs, the statins and calcium channel blockers (CCB), were studied in chapter four. The statins simvastatin and atorvastatin are substrates for the P-glycoprotein (P-gp) transporter, encoded by the ABCB1 gene, and the CYP3A4 metabolizing enzyme. In chapter 4.1, the association between the 1236/2677/3435
haplotypes in the ABCB1 gene and the cholesterol lowering effect of simvastatin was studied. In men, the TTT and CGT haplotype were associated with a 0.40 to 0.53 mmol/l larger reduction in total and LDL cholesterol levels than the reference CGC haplotype. In women, no significant associations were found. In chapter 4.2 we studied whether the polymorphisms
C1236T, G2677AT and C3435T in the ABCB1 gene and the polymorphism CYP3A4*1B were associated with a decrease of the prescribed dose or a switch to another cholesterol lowering drug during simvastatin and atorvastatin therapy, possibly indicating adverse drug reactions or a too strong reduction in cholesterol level. Simvastatin and atorvastatin users with the variant CYP3A4*1B variant G allele had a two times lower risk for a dose decrease or switch to another cholesterol lowering drug. No associations were found for the ABCB1 polymorphisms or haplotypes. Women with the CYP3A4*1B variant G allele had a three times lower risk than women with the CYP3A4*1B reference A allele and in the group of ABCB1 CT or TT genotype carriers the CYP3A4*1B variant G was associated with 2.5 times As mentioned before, the rs10494366 SNP in the NOS1AP gene was associated with an increased QTc interval, most likely due to an effect on the inward calcium channel currents. CCB affect the voltage dependent calcium channels. We studied in chapter 4.3 the effect of
this polymorphism on the incidence of diabetes mellitus in calcium channel blocker users, because insulin release is triggered by an influx of calcium in the pancreatic beta-cells. CCB users with the rs10494366 TG or GG genotype had a two times lower risk of diabetes mellitus than users with the TT genotype, although small numbers preclude definitive statements and replication of these results is indicated. In chapter 4.4 we studied the effect of the rs10494366
SNP on cardiovascular mortality in calcium channel blocker users. Dihydropyridine CCB users with the TG genotype had a 3.5 times higher cardiovascular mortality risk and users with the GG genotype a 6 times higher cardiovascular mortality risk than users with the TT genotype. In the non-dihydropyridine CCB users, no associations with cardiovascular mortality were found. Also in this study, replication of the results is indicated. In chapter 3.2 we found that the rs622342 polymorphism in the SLC22A1 gene, coding for the OCT1 transporter is associated with metformin response. Also the anti-Parkinson drugs pramipexole, amantadine and, possibly, levodopa are substrates for OCT1. The rs622342 variant C allele was associated with higher prescribed doses of anti-Parkinson drugs, especially amantadine and selegiline, and a shorter survival time (chapter 5.1). After start of
levodopa therapy, patients with the CC genotype had a two times higher mortality risk and had lived on average 2.5 years shorter than patients with the AA genotype. In the general discussion (chapter 6), the results are summarized and discussed. Apart from
the identification of polymorphisms not previously associated with drug response, the most important result is that the interaction between individual polymorphisms, between polymorphisms and gender and, possibly, between polymorphism and co-prescribed drugs, do add substantially to the prediction in drug response. Whether genotyping is useful in individualizing pharmacotherapy depends on the possibility to prevent either adverse drug reactions or increased costs due to ineffective therapy, weighed against the costs of

Source: http://nvza.nl/ufc/file2/nvza_sites/karin/b18716fdaeaf5c5b10c5ef85b4bd7624/pu/Summary_Matthijs_Becker.pdf