Clinical predictors and genetic determinants of response to warfarin therapy in Kenya patients

Abstract

Background: Warfarin is the most extensively used long-term oral anticoagulant in most parts of the world. However, sociodemographics, patients‘ clinical characteristics and individual‘s genetics contribute to inter-patient variation in warfarin dose and response. Sociodemographics have been shown to contribute to almost a fifth of inter-patient variability in warfarin dose requirements among individuals from the western communities. In addition, the most important genes that determine warfarin dose and response include Vitamin K Epoxide Reductase Complex Subunit 1(VKORC1), Cytochrome P450 2C9 (CYP 2C9) and Cytochrome P450 4F2 (CYP 4F2) which account for ~ 30%, ~ 10% and ~2 % of the inter-patient warfarin dose variation in the western countries, respectively. There is paucity of Kenyan data on the clinical predictors and genetic variability on warfarin dose and response. Study Objective: To characterize the clinical predictors and genetic determinants of response to warfarin therapy in order to develop a model for safe warfarin dosing among black Kenyan patients. Methods: The study areas were the medical records department and three anticoagulation clinics of the leading teaching and referral hospital in Kenya. Black Kenyan patients aged ≥18 years, who were on long term warfarin therapy (≥ 28 days), were recruited. Two study designs were used: Prospective longitudinal study design (180 patients) which characterized the clinical determinants of warfarin response and cross-sectional study (40 participants) which characterized the genetic variability among the patients. Sociodemographics, ethnicities, dietary habits, nutritional and herbal medicine utilization, duration of warfarin use and adverse drug reactions (ADRs) to warfarin were obtained through face-to-face patient interviews. Data on clinical indications of anticoagulation, concomitant medications and concurrent diseases were acquired through review of the medical records. Venous blood samples were drawn for the determination INRs and genetic testing. Patients were followed up for six consecutive clinic visits while determining the INRs whose therapeutic levels were set at 2-3 in accordance with the established international guidelines. DNA extraction was conducted using the Quick-DNATM Miniprep plus kit in accordance with manufacturer‘s instructions and the purity was assessed by comparing A260 and A280 ratio using the NanoDrop technology. CYP 2C9 (*2, *3, *4, *5, *6, *8, *11 and *13), VKORC1 rs9923231 as well as CYP 4F2 [*3(1347C>T) and rs2189784;G>A were analyzed for single nucleotide polymorphisms (SNPs) using MassARRAY Compact Mass Spectrometer and Agena real-time detection platform, using primers. Statistical Analyses: Data analyses were conducted onto IBM Statistical Package for Social Sciences version 23 at 95% confidence limit with values having p≤0.05 being considered statistically significant. Associations between the predictors (socio-demographics, patients‘ clinical characteristics and genetic polymorphisms) and outcomes (warfarin doses, INRs and ADRs) were computed. Regression analyses were achieved using an identity link and logit link to determine independent predictors of outcome variables. R2 and Nagelkerke‘s R2 regression models were used to estimate the relative contribution of individual predictor variables to the inter-patient variability in warfarin maintenance doses, ADRs and INRs observed. Results: There was female preponderance (77.0 %). Patients had a mean age 43.4 (±13.2) years where majority were overweight/obese (>60.0%) and married (65.6%). The median duration of warfarin therapy was 24.8 (range 1.1-375.9) months. The mean warfarin maintenance dose was 6.2±2.8mg per day. Venous thromboembolic (VTE) (56.6%) and cardioembolic (33.3%) events were the main indications of warfarin anticoagulation. Higher warfarin dose requirement was found among males, participants aged ≤ 50 years, obese/overweight and with spouses (p=0.05). Although a diagnosis of VTE was the independent predictor of higher warfarin dose requirement [β=0.72, CI: 0.07, 1.36; p=0.030], patients with cardioembolic events had better INRs than those with VTEs. Furthermore, significant proportion of males were better anticoagulated than females (51.3% vs. 30.4%; p=0.040). Gender [β=-3.38, 95% CI: -5.38 to -1.38, p=0.001] and patients‘ age (β=-0.06, 95 %CI: -50.13 to -0.00, p=0.050) independently predicted warfarin maintenance doses that would give therapeutic INRs. The most common ADR to warfarin therapy was haemorrhage (27.8%) which was significantly found among patients without spouses (p=0.008) and cardiac diseases (p=0.050). However, the independent predictor of ADRs was living without spouses (β=0.76; 95% CI: 1.13, 4.06; A.O.R=2.14, p=0.019). The patients‘ clinical data explained approximately 14.0%, 20.0% and 19.0% of the variability in inter-patient warfarin xxviii doses, INRs and ADRs, respectively with body surface area (BSA) contributing the greatest variability to warfarin dose (6.92%) and INRs (4.54%). There were two ethnolinguistic groups of Africa captured for the genetic study: Bantu and Nilotes. CYP 4F2 gene was the most polymorphic, with CYP 4F2 (rs2189784; G>A) at 60% and CYP 4F2 *3(1347C>T) (17.5 %), followed by VKORC1 rs9923231 (12.5 %). Unlike CYP 2C9*8 and *11 where *8 showed heterozygousity (GA) genotypes at 17.5% while *11 revealed heterozygousity (CT) genotypes at 2.5%, polymorphisms of CYP 2C9*2, *3, *4 *5 *6 and *13 were not detected in the population. The wild type CYP 2C9 *8 and *11, variant VKORC1 and CYP4F2 (1347C>T) (CT) genotypes required clinically higher warfarin maintenance doses. However, CYP 4F2 (rs2189784; G> A) wild type (GG) genotype [β =3.01, 95%CI: 1.36-4.65, p<0.001] and VKORC1 (rs9923231) wild type (CC) genotype [β =3.77, 95%CI: 1.51-6.03, p=0.001] were the independent genetic determinants for the warfarin maintenance doses which could achieve therapeutic INRs. The genetic data accounted for approximately 21.0 %, 38.0 % and 31.0 % of the warfarin dose variability, INR responses and ADRs among the patients in the two ethnolinguistic groups studied. CYP 4F2 polymorphisms contributed the greatest in variability (9.6%) of inter-patient warfarin dose requirements. The clinical and genetic information led to derivation of the model for safe warfarin dosing to produce INR 2-3 as: Safe Warfarin Dose (Mg) =-3.38(Male) +3.01 CYP 4F2 (rs2189784; wild type GG) + 3.77 VKORC1 (rs9923231; wild type CC). Conclusions: Higher warfarin doses among the Kenyan patients on anticoagulation may be necessitated by male gender, younger age, obesity, diagnosis of VTE, VKORC1 and CYP 4F2 polymorphisms. BSA and CYP 4F2 polymorphisms contributed the greatest variability of inter-patient warfarin doses and may guide dosing of warfarin in long-term anticoagulation. Males and patients with heart diseases were better anticoagulated; but the latter had more ADRs suggesting that intensification of warfarin anticoagulation should be done among females as well as for patients with cardioembolic disorders. Future studies should correlate health facility, prescriber-related and pharmacoeconomic factors with anticoagulation outcomes. Studies exploring the role of other genes involved in warfarin pharmacokinetics and pharmacodynamics are also recommended.