Table 1.

Study Characteristics in the Randomized Trials of Genotype-Guided Warfarin Dosing

Author, year	Study arm	N Randomized	N Losses	Mean follow-up, days	Mean daily starting dose, mg	Gene(s) tested	PG Algorithm	C Algorithm
Hillman47	PG	18	0	28	4.6	CYP2C9	Hillman model70‡	Marshfield algorithm71
2005	C	20	0	28	5
Caraco48	PG	142	47	22	8.6	CYP2C9	Algorithm constructed de novo48	DAWN AC computer algorithm72
2007	C	141	45	40	6.7
Anderson57	PG	101*	N/R*	46	5.1†	CYP2C9, VKORC1	Carlquist regression equation73‡	10 mg × 2 days, then 5 mg74
2007	C	99	N/R	46	5

PG: Pharmacogenetic arm; C: control dosing arm; N: number; N/R: not reported

*Data were not presented by study group. Of the patients, 206 were randomized; 6 dropped out: 3 because surgery was canceled, and 3 stopped warfarin before the first INR was drawn

†Derived from weekly starting dose. Actual starting dose was double the algorithm PG: 10.2 mg, C: 10 mg × 2 days followed by regular dose per study by Kovacs et al.⁷⁴

‡Multiple regression algorithms. In addition to genotype, Hillman et al.⁴⁷ accounted for age, sex, body surface area, concomitant medications, co-morbidities and clinical indication; Anderson et al.⁵⁷ accounted for age, sex, weight