Table 3.

Performance of different study designs in identifying regulatory QT prolongation risk at C_max. Shown for each chemical and study design is the percentage of subsampling iterations that resulted in a positive classification of QT prolongation risk at C_max of 10 msec prolongation at 95% confidence. For comparison, the reference clinical classification based on in vivo human data and the classification based on the full cohort (n = 43) are also shown. Parenthesis around a clinical classification indicates that the classification is ambiguous (some studies above and some below the regulatory threshold).

	Percent subsamples “+”
Chemical: clinical classification	Std. Donor	n = 5	n = 10	n = 20	n = 43
Cisapride: +	100	100	100	100	+
Citalopram: +	100	100	100	100	+
N-acetylprocainamide: +	100	100	100	100	+
Quinidine: +	100	100	100	100	+
Sematilide: +	100	100	100	100	+
Vernacalant: +	100	100	100	100	+
Sotalol: +	98	100	100	100	+
Disopyramide: +	84	100	100	100	+
Dofetilide: +	98	92	100	100	+
Moxifloxacin: (+)	100	94	91	96	+
Cabazitaxel: –	12	6	0	0	–
Mifepristone: –	2	16	3	0	–
Lamotrigine: (–)	100	76	31	8	–

+: Positive for QTc prolongation; –: Negative for QTc prolongation; (+) or (–): mixed or ambiguous clinical results in terms of 10 msec regulatory threshold at 95% confidence.