Race-Specific Dosing Guidelines Urged for Warfarin

Since the National Institutes of Health issued a call for broader racial representation in clinical research, clinicians and researchers have met that challenge with the development of clinical trials that take race-specific factors into consideration. But in the case of anticoagulation – and the significant inter-patient variability in optimal anticoagulant dosing – there are several other race-specific pharmacogenetic factors that still need to be accounted for.

To that end, investigators, led by Nita Limdi, PharmD, PhD, from the University of Alabama at Birmingham, compared race-combined and race-specific dosing algorithms to determine whether incorporating certain clinical and genetic factors into specific algorithms could help clinicians better calculate warfarin dosing for both African-American and European-American patients.

Previous interventional studies, such as the Clarification of Optimal Anticoagulation through Genetics (COAG) trial that evaluated genotype-guided dosing, have delivered incongruous percent-time-in-therapeutic-range (PTTR) data for the two racial groups – fueling discussions about the utility of this type of guided therapy. “We propose that differences in the proportional African representation can explain PTTR differences among COAG participants,” stated Dr. Limdi and colleagues. “As the algorithms adjust (rather than stratify) for race, the heterogeneity introduced by race sheds light on the divergent findings in COAG.”

In their prospective study published in Blood, Dr. Limdi and colleagues analyzed the genetic and clinical factors of 1,357 patients (44% African American, 56% European American) treated with warfarin, then calculated and compared their recommended dose according to both race-adjusted dosing models (e.g., COAG) and race-specific dosing models. With such a large African-American population in the study, the authors noted, they were able to conduct a robust assessment of the impact of clinical and genetic factors on warfarin dose by race.

“As the outcomes of disease can vary by race, so can response to medications,” said Dr. Limdi. “Therefore, warfarin dosing equations that combine race groups for analysis (race-adjusted analysis) assume that the effect of variables – such as age and genetics – are the same across race groups, which may compromise dose prediction among patients of both races.”

The investigators identified several factors that affected warfarin metabolism and warfarin dose requirements, including clinical factors (such as age, body surface area [BSA], chronic kidney disease [CKD], and amiodarone use) and genetic factors (CYP2C9 *2, *3, *5, *6,*11, rs12777823, VKORC1, and CYP4F2).

As seen in the TABLE, warfarin dose in both racial groups was influenced by all identified clinical factors, as well as by the CYP2C9*3 and VKORC1 variants. However, while the CYP2C9*2 and CYP4F2 variants primarily influenced warfarin dose variability in European Americans, the rs12777823 variant accounted for the larger variability in dosing requirements among African Americans.

Overall, clinical predictors accounted for greater dose variability among European Americans compared with African Americans, while genetic predictors accounted for a greater portion of variability in African Americans compared with European Americans.

“Race-specific pharmacogenetic algorithms allow incorporation of factors – genetic and clinical – that influence [warfarin] dose in African Americans, and should be included in [warfarin] dosing decisions,” Dr. Limdi told ASH Clinical News. For example, she said, dosing in African Americans should account for kidney function and genetic variants such as CYP2C9*5, *6, *11, and rs12777823. For European Americans, addition of the CYP4F2 variant and kidney function improves dose prediction. “Dosing for both race groups should include a measure of kidney function,” she stressed.

While they were able to identify several factors to help better predict optimal warfarin dosing, the study authors noted some limitations. First, their results may not be generalizable to Hispanics and Asians, and, second, the incorporation of other genetic and clinical factors may improve dose prediction and alter predictor effect sizes even further. Moreover, warfarin dose variations can be influenced by a variety of other factors, such as fluctuations in diet and concurrent illness.

“Our findings highlight the need for adequate racial representation in warfarin dosing studies to improve our understanding of how the factors that influence warfarin dose differ according to race,” said Dr. Limdi. “This is the first step to developing race-specific algorithms to personalize therapy.”


Reference

Limdi N, Brown T, Yan Q, et al. Race influences warfarin dose changes associated with genetic factors. Blood. 2015 May 29. [Epub ahead of print.]

TABLE. Effect of Clinical and Genetic Factors on Warfarin Dose Requirements by Race

Race-Combined

Race-StratifiedEuropean Americans

Race-Stratified

African Americans

% Dose change (95% CI)

% Dose change (95% CI)

% Dose change (95% CI)

Age

–0.68

(–0.81 to –0.56; p<0.001)

–0.62(–0.79 to –0.45; p<0.001)

–0.71

(–8.69 to –5.00; p<0.001)

BSA, per m2

51.09

(40.10 to 62.94; p<0.001)

49.27(35.69 to 64.21; p<0.001)

54.89

(8.26 to 14.95; p<0.001)

Chronic kidney disease

-8.24

(-10.94 to -5.47; p<0.001)

–10.92(–14.54 to –7.15; p<0.001)

–6.46

(–10.37 to –2.38; p=0.002)

Amiodarone use

–19.99

(–24.85 to –14.81; p<0.001)

–17.52(–23.37 to –11.23; p<0.001)

–25.75

(–33.81 to –16.7; p<0.001)

CYP2C9*2

–18.47

(–22.40 to –14.33; p<0.001)

–20.64(–24.66 to –16.41; p<0.001)

2.98

(–10.52 to 18.53; p=0.68)

CYP2C9*3

–34.83

(–39.08 to -30.29; p<0.001)

–34.05(–38.52 to –29.24; p<0.001)

–37.83

(–48.92 to –24.33; p<0.001)

VKORC1

-26.64

(–29.02 to –24.18; p<0.001)

–28.94(–31.49 to –26.29; p<0.001)

–19.99

(–25.53 to –14.05; p<0.001)

CYP4F2 4.58

(0.97 to 8.31; p=0.01)

5.89(1.9 to 10.05; p=0.004)

1.23

(–6.32 to 9.38;p=0.76)

rs12777823 –7.59

(–11.32 to –3.70; p=0.0002)

–2.26(–7.44 to 3.21;p=0.41)

–12.26

(–17.55 to –6.62; p<0.001)

SHARE