 |
 
2nd
International Workshop on Clinical Pharmacology of HIV Therapy
Noordwijk, The Netherlands April 2 - 4, 2001
Poster: Evaluation
of the Inhibitory Quotient as a Pharmacodynamic Predictor of the
Virologic Response to Protease Inhibitor Therapy # 7.1
Dale Kempf,
Ann Hsu, Jeffrey Isaacson, Ping Jiang, Scott Brun, Cheryl Renz, G.
Richard Granneman and Eugene Sun; Abbott Laboratories, Abbott Park, IL
USA
|
|
Share this Poster with a Colleague
ABSTRACT
Background: Pharmacodynamic
(PD) models that incorporate information on both drug pharmacokinetics and viral
susceptibility are important for the understanding of the in
vivo potency of antiretroviral agents. The
inhibitory quotient (IQ), calculated as the ratio of Ctrough and
human serum-adjusted IC50, is a PD model that has been used to estimate the
activity of protease inhibitors (PIs). We have assessed the predictive ability
of the IQ model in two clinical studies in PI-experienced subjects, where
differences in baseline drug susceptibility due to pre-existing resistance allow
for a large range in IQ values.
Methods: Study
957 evaluated the activity of lopinavir/r (LPV/r) + efavirenz + NRTIs in 57
multiple PI-experienced, NNRTI-naïve subjects with median baseline HIV RNA 4.5
log copies/mL. Study 985 evaluated the intensification of indinavir (IDV)
therapy in subjects failing IDV + NRTIs by dose modification with ritonavir in
37 subjects with median baseline HIV RNA 3.3 log copies/mL. The IQ for each
subject was calculated as Ctrough/(baseline fold IC50 x serum-adjusted IC50 for
wt HIV) . In Study 985, baseline fold IC50 was defined as the
virtual phenotype based on the baseline genotype.
Results: In
Study 957, response (400 at Week 24) was observed in 67% (14/21), 80%
(12/15) and 100% (16/16) of subjects with LPV IQ of <4, 4-15 and >15,
respectively (p <0.026). In Study 985, response (0.5 log10 decline or <50
c/mL at Week 24), was observed in 75% (12/16), of subjects with IDV IQ >2 and
0% (0/9) of subjects with IQ <2 (p <0.001).
Conclusion: In
two clinical studies evaluating the activity of two different PIs with very
different serum free fractions in PI-experienced subjects, IQ calculated for
individual subjects was statistically significantly associated with virologic
response.
INTRODUCTION
The inhibitory quotient (IQ) is a
model for understanding and evaluating drug pharmacology that encompasses both
pharmacokinetics and drug susceptibility (Fig. 1). Originally articulated for
the effects of antibiotics (1), the IQ can be applied to HIV protease inhibitors
(PIs) as the ratio of the Ctrough and the IC50. In order for a
pharmacodynamic model such as IQ to be useful, it should be (a) standardized and
(b) predictive of the in vivo efficacy of more than one drug. In this
study, we examined the association of the IQ of lopinavir (LPV) and indinavir (IDV)
in two separate studies in PI-experienced subjects experiencing virologic
failure.
METHODS
Study Designs
- Study 985 examined the effect
of intensification of ongoing IDV (800 mg TID) triple therapy in patients
with HIV RNA >50 copies/mL by addition of RTV to a final dose of RTV/IDV
of 400/400 mg BID (Fig. 3). After Week 3, changes in the NRTI portion of the
regimen were allowed.
- Virologic response at Weeks 3
and 24 was defined as having plasma HIV RNA <50 copies/mL and/or having
plasma HIV RNA >0.5 log copies/mL below the baseline value. Response was
defined using a ‘dropouts as censored’ definition (2).
Virology
- Baseline phenotype in Study
957 was measured by the PhenoSense™ method.
- Baseline genotype in Study 985
was determined by population sequencing and used to determine the baseline
virtual phenotype (3). In a subset of patients for whom phenotype was
determined, there was a high correlation between determined phenotype and
virtual phenotype.
- Ctrough was determined at
steady state. In Study 985, predose concentrations were used directly as Ctrough,
and four patients for whom the Cpredose sample was taken close to
the Tmax from the previous dose were censored from the analysis.
- Inhibitory quotient
values for each patient were calculated by the following equation:
- Serum-adjusted
IC50 values were taken from Molla et al. (4) and represent the mean IC50 for
three wild-type laboratory strains (IIIB, HXB2 and pNL4-3) in the presence
of 50% human serum plus 10% fetal calf serum. Values for LPV and IDV are
0.059 and 0.053 µg/mL, respectively. The free fraction of LPV in 50% human
serum plus 10% fetal calf serum has been shown to approximate the free
fraction in 100% human serum (5).
Statistical Analysis
- The univariate association of
virologic response (categorical) with IQ was examined using Fisher’s exact
test. Multivariate analyses were conducted using stepwise logistic
regression with an entry and exit p-value of 0.15.
RESULTS
Characterization of baseline
inhibitory quotient
- In
both studies, a broad range of IQ values were observed. In Study 957, LPV IQ
values ranged from 0.05 to 279, with a median of 9.9 (Fig. 4).
- In Study 985, IDV virtual IQ (vIQ)
values ranged from 0.33 to 59, with a median of 5.3. Themedian RTV vIQ in
Study 985 was <5% of the median IDV vIQ (Fig. 5).
Virologic response with respect
to inhibitory quotient
- In both studies, virologic
response (dropouts as censored) was predicted by the inhibitory quotient.
- In Study 957, plasma HIV RNA
<400 copies/mL at Week 24 was observed in 100%, 80% and 67% of patients
with IQ at baseline of >15, 4 to 15 and <4, respectively. Decline to
<50 copies/mL was observed in 94%, 67% and 48% of the above patients,
respectively.
- In Study 985, virologic
response (defined by the study protocol as plasma HIV RNA <50 copies/mL
and/or a 0.5 log copies/mL decline from baseline) was significantly higher
in those patients with IDV vIQ >2 than in those with vIQ <2.
Relative
predictive value of IDV virtual IQ vs. virtual phenotype or IDV trough
- In Study 985, the
relationships of IDV vIQ to baseline IDV virtual phenotype and IDV Ctrough
were examined.
- All subjects with <6-fold
baseline IDV virtual phenotype had a vIQ >2, and each subject with a
baseline virtual phenotype of >12 had a vIQ <2. The intermediate range
of virtual phenotype (between 6- and 12-fold) represented the range within
which differences in Ctrough had a major effect on virologic
response (Fig. 8).
- Six subjects had a
Ctrough < 0.75 µg/mL (median Ctrough) but had a vIQ >2. At
Week 3, 5/6 of these subjects experienced virologic response, confirming
that the phenotypic susceptibility to IDV was an important contributor to
the rate of response, particularly in those subjects with relatively modest
Ctrough (Fig. 9).
- Since IDV and RTV vIQ values
were highly correlated, two stepwise logistic regression analyses of the
virologic response in Study 985 were performed using the parameters listed
in Table 1, with an entry and exit p-value of 0.15.
- In model 1, only log10
IDV vIQ entered the model and remained significantly associated with
virologic response at Weeks 3 (p=0.012) and 24 (p=0.009).
- In model 2, log10 RTV vIQ was
substituted for log10 IDV vIQ. In this case, the log10
IDV virtual phenotype entered the model and remained significantly
associated with response at Week 3 (p=0.015).The Week 24 model included
log10 IDV virtual phenotype (p=0.094), log10 IDV Ctrough at Week
3 (p=0.084) and number of PI mutations p=(0.143).
CONCLUSIONS
- In two studies of PI therapy
in PI-experienced patients, the inhibitory quotient, considering both viral
and pharmacokinetic parameters, was a significant predictor of
response.
- Similar IQ breakpoints were
observed with two PIs with substantially different serum protein binding.
- In Study 985, IDV vIQ appeared
to be a better predictor of virologic response than IDV virtualphenotype,
IDV Ctrough or RTV vIQ.
- The use of inhibitory
quotients to assess the antiretroviral activity of protease inhibitors in
vivo merits additional study.
ACKNOWLEDGMENTS
Abbott Laboratories
J Sylte, S Boller, M
King, J Swerdlow, R Rode, O Jasinsky, R Bertz K Real, B Bernstein, T Japour, K
Rynkiewicz, J Ryan, J Alfano J Lamm, J Franczyk, B Richards
Study 957 Investigators
N Clumeck, D Ho, S Becker, A
Lazzarin, F Bergmann, G Pantaleo, G Carosi, J Rockstroh, S Danner ,R Tubiana,
P-M Girard
Study 985 Investigators
A Zolopa, J Gallant, N Shulman,
E Race, D Havlir,
Virco
Brendan Larder
Virologic
Nick Hellmann
TOP
Poster: Evaluation
of the Inhibitory Quotient as a Pharmacodynamic Predictor of the Virologic
Response to Protease Inhibitor Therapy
© 2001 Medical Advocates
for Social Justice
225 West Washington Street, Suite 2200, Chicago, IL 60606
Telephone (312) 419-7289 Facsimile: (312) 419-7151
Email: info@medadvocates.org
