Exposure–response modeling of the effect of glasdegib on cardiac repolarization in patients with cancer

Figures & data

Table 1. Baseline patient demographics

	Study B1371001	Study B1371002	Combined
N	47	23	70
Age, years, median (range)	69 (25–89)	61 (27–76)	67 (25–89)
Triplicate ECGs, n
ECGs within ±15 minutes of PK	589	158	747
All PK-ECG pairs	810	217	1027
Sex, n (%)
Male	28 (60)	14 (61)	42 (60)
Female	19 (40)	9 (39)	28 (40)
Race, n (%)
White	38 (81)	18 (78)	56 (80)
Asian	1 (2)	3 (13)	4 (6)
Black	4 (9)	0 (0)	4 (6)
Other	4 (9)	2 (9)	6 (9)

Due to rounding, percentages may not add up to 100.

ECG, electrocardiogram; PK, pharmacokinetic.

Table 2. Summary of ECG measurements

	QT	RR	QTcB	QTcF	QTcS
Baseline*
n	68	68	68	68	68
Range	339–463	596–1190	387–501	381–466	380–461
Median	389	813	436	419	418
Mean (SD)	391.7 (29.9)	812.2 (141.3)	437.1 (21.0)	421.1 (18.1)	419.1 (18.3)
On treatment
n	68	68	68	68	68
Range	283–520	525–1288	376–533	351–515	346–513
Median	403	830	444	430	428
Mean (SD)	404.6 (37.4)	836.3 (150.0)	444.9 (24.4)	430.6 (24.0)	428.9 (24.4)

QT, RR, QTcB, QTcF, and QTcS are means of the triplicate interval (msec) measurements.

*Two patients had baseline values that were collected after dosing and were not included in the summary of baseline measurement. One of these ‘baseline’ measurements was included in the analysis as the ECG collection was performed within 15 minutes of the PK sample.

ECG, electrocardiogram; PK, pharmacokinetics; QT, interval length between the Q and T waves; QTcB, QT interval corrected for heart rate using Bazett’s formula; QTcF, QT interval corrected for heart rate using Fridericia’s formula; QTcS, QT interval corrected for heart rate using a population-specific formula; RR, interval length between successive R waves; SD, standard deviation.

Figure 1. Relationship between glasdegib plasma concentration and (a) RR, (b) QTcF, and (c) QTcS intervals

The relationship between RR, QTcF, and QTcS with glasdegib plasma concentration is presented along with the best fit line for each measurement using a linear mixed-effects model. The CIs are provided for the glasdegib plasma concentration slope parameter in each model. The points are colored based on the study from which they were collected.CI, confidence interval; QT, interval length between the Q and T waves; QTc, QT interval corrected for heart rate; QTcF, QT interval corrected for heart rate using Fridericia’s formula; QTcS, QT interval corrected for heart rate using a population-specific formula; RR, interval length between successive R waves

Figure 2. Assessment of the correction factors for QT interval: QTcB, QTcF, and QTcS

The best fit line using a linear mixed-effects model is provided along with the CI for the slope of the RR effect. If the CI does not include 0, it indicates that the relationship with RR has not been sufficiently corrected.CI, confidence interval; QT, interval length between the Q and T waves; QTc, QT interval corrected for heart rate; QTcB, QT interval corrected for heart rate using Bazett’s formula; QTcF, QT interval corrected for heart rate using Fridericia’s formula; QTcS, QT interval corrected for heart rate using a population-specific formula; RR, interval length between successive R waves

Table 3. Final E-R models for QTcF and QTcS

	Estimate	SE	RSE	Shrinkage
QTcF
Intercept, ${\theta }_1$	423.8	2.1	0.5	–
Slope, ${\theta }_2$	4.3	0.37	8.6	–
Residual standard deviation, W	14.1	1.1	8.0	4.0
Variance of the random effect on the intercept, ${\omega }_1^2$	277.6	50.7	18.3	4.4
QTcS
${\theta }_1$, msec	422.0	2.2	0.5	–
${\theta }_2,$ msec/1000 ng/mL	4.31	0.37	8.6	–
Residual standard deviation, W	14.4	1.1	7.8	3.9
Variance of the random effect on the intercept, ${\omega }_1^2$	291.6	51.1	17.5	4.4

The estimates from the linear mixed-effects model with a random effect on the intercept are presented for both QTcF and QTcS. The estimates are reported on the microgram scale as the scaling helped the estimation procedure.

E-R, exposure–response; QT, interval length between the Q and T waves; QTcF, QT interval corrected for heart rate using Fridericia’s formula; QTcS, QT interval corrected for heart rate using a population-specific formula; RSE, relative standard error; SE, standard error.

Table 4. Median and 95% CI for the mean change in QTc at therapeutic and supratherapeutic exposures

	Mean QTcF change from baseline, msec			Mean QTcS change from baseline, msec
	Bootstrap percentiles			Bootstrap percentiles
Steady-state exposure	2.5^th	50^th	97.5^th	2.5^th	50^th	97.5^th
Therapeutic Cmax, 100 mg QD	4.40	5.30	6.24	4.32	5.36	6.25
Supratherapeutic Cmax, 40% increase	6.15	7.42	8.74	6.04	7.50	8.75
Supratherapeutic Cmax, 100% increase	10.03	12.09	14.25	9.85	12.23	14.26

The 95% CIs for the expected QTc prolongation are provided for steady-state C_max with a 100-mg QD dose (therapeutic exposure with geometric mean of 1137 ng/mL), and supratherapeutic exposures corresponding to a 40% increase in the 100-mg QD steady-state C_max, and the steady-state C_max for a 200-mg QD dose (geometric mean, 2445 ng/mL).

CI, confidence interval; C_max, maximum plasma concentration; QD, once daily; QT, interval length between the Q and T waves; QTc, QT interval corrected for heart rate; QTcF, QT interval corrected for heart rate using Fridericia’s formula; QTcS, QT interval corrected for heart rate using a population-specific formula.