TABLE 5.

Analysis by multivariate ANCOVA of the effects on rifampin pharmacokinetic parameters by transporter gene polymorphism, demographic and clinical factors, and rifampin dosage

Main effects	Adjusted mean rifampin AUC0-24 (μg·h/ml)	Adjusted mean of rifampin peak concn	ANCOVA P valuea	False discovery rate P value
Rifampin AUC0-24 (n = 86)
Rifampin dosage vs ln AUC0-24, slope = 1.041b			0.002	0.02
SLCO1B1 c.463C>A
CC (n = 71)	46.7		0.001	0.02
CA (n = 15)	29.8
Groupsd			0.02d
(1) Healthy controls (n = 16)	49.7		(1) vs (2) = .007	(1) vs (2) = .03
(2) Subjects with TB, Africa (n = 37)	29.0		(2) vs (3) = 0.07	(2) vs (3) = 0.23
(3) Subjects with TB, non-Africa (n = 33)	36.0		(3) vs (1) = 0.06	(3) vs (1) = 0.23
Rifampin Cmax (n = 88)
Rifampin dosage vs ln Cmax, slope = 1.53c			0.0003	0.002
Groupsd			0.0006d
(1) Healthy controls (n = 16)		14.5	(1) vs (2) = 0.0002	(1) vs (2) = 0.002
(2) Subjects with TB, Africa (n = 37)		5.5	(2) vs (3) = 0.24	(2) vs (3) = 0.56
(3) Subjects with TB, non-Africa (n = 35)		6.5	(3) vs (1) = 0.0004	(3) vs (1) = 0.002

^aMain effects retained in the ANCOVA model, with P values of <0.05.

^bln AUC_0-24 = 1.041 times (ln rifampin dosage in mg/kg).

^cln C_max = 1.53 times (ln rifampin dosage in mg/kg).

^dComparisons in ANCOVA using post hoc least significant difference (LSD) method.