TABLE 2

Summary of pioglitazone, rosiglitazone, raloxifene, and erythromycin kinetic parameter estimates for two- (corrected and uncorrected k_obs) and three-parameter models, Spearman's rank correlation coefficient (r_s), and a test for a positive linear regression of k_obs versus ln[I]

Ninety-five percent profile-likelihood confidence intervals for each kinetic estimate are listed in parentheses. The rank correlation and linear regression tests do not depend on the nonlinear model fit and are reported on the same row as the three-parameter model estimates. k_{obs[0 μM]} is not estimated in the 2p model.

Compound		k_{obs[0 μM]}	K_I	k_inact	r_s	k_obs − ln[I] Slope
		min⁻¹	μM	min⁻¹
Pioglitazone	3p	NC	NC	NC	0.96	Yes
					(p = 0.0002)	(p = 0.0033)
	2p_C	NC	NC	NC
	2p_U	NC	NC	NC
Rosiglitazone	3p	0.0043	6.0	0.0103	1	Yes
		(0.0034, 0.0052)	(4.1, 9.1)	(0.0092, 0.0115)	(p = 0)	(p < 0.0001)
	2p_C	NC	6.1	0.0103
			(4.7, 7.9)	(0.0096, 0.0110)
	2p_U	NC	2.8	0.0142
			(0.2, 9.1)	(0.0108, 0.0183)
Raloxifene	3p	−0.001	3.2	0.2389	0.96	Yes
		(−0.0373, 0.0336)	(1.6,6.9)	(0.1895, 0.2901)	(p = 0.0002)	(p = 0.0013)
	2p_C	NC	3.4	0.2318
			(2.0, 5.8)	(0.1976, 0.2739)
	2p_U	NC	3.3	0.2381
			(0.7, 2.0)	(0.2059, 0.2769)
Erythromycin	3p	0.0037	12.5	0.0289	1	Yes
		(0.0018, 0.0055)	(8.8, 18.4)	(0.0262, 0.0316)	(p = 0)	(p < 0.0001)
	2p_C	NC	11.9	0.0291
			(9.5, 14.9)	(0.0272,0.0312)
	2p_U	NC	8.2	0.0311
			(4.5, 14.6)	(0.0266, 0.0366)

3p, three-parameter model used to estimate k_{obs[0 μM]}, K_I, and k_inact; NC, not calculated; 2p_U, two-parameter model used to estimate K_I and k_inact, wherein corrections were not made for k_{obs[0 μM]}; 2p_C, two-parameter model used to estimate K_I and k_inact, wherein k_obs values were generated using data corrected for loss of activity occurring at [I] = 0 μM.