A one-compartment model with first-order absorption and first-order elimination with CL, V and KA as basic structural parameters, as implemented in the ADVAN2 and TRANS2 routines of NONMEM, was used to describe IDV plasma population kinetics. A basic model with the above structural parameters as well as statistical parameters describing interindividual variability was built. Results from the basic model lacking interoccasion variability are presented in Table 2.
A number of potential covariates were assessed for their effects on pharmacokinetic parameters of IDV. These covariates included body weight and age (linear), and gender, race and co-medication (categorical). Initial covariate analysis (see Methods) identified age and race as significant covariates for CL, V and KA. ΔMVOF was −31.7, −38.7 and −14.1 for age, and −23.2, −21.7 and −22.2 for race with the CL, V and KA parameters, respectively. In contrast, body weight, gender and concomitant zidovudine and lamivudine did not appear to affect IDV pharmacokinetic parameters. A full model incorporating the identified covariates was built and further refined by setting them, one by one, to their null value (zero). This step eliminated the apparent effects of age on CL and KA, as well as race on CL, V and KA (P > 0.05). A model retaining age in the V model was obtained.
The availability of multiple plasma samples collected from different visits allowed an estimate of individual variability within the same subject across different occasions for structural pharmacokinetic parameters. Variances and co-variances associated with interoccasion variability were incorporated into the refined model with age as a covariate for V. This led to a highly significant fall in MVOF of 317 from the basic model. Further refinement of the model by deleting age from the V model resulted in a statistically significant (P < 0.02) increase in the objective function value. Therefore, the initial basic model with age in the V model along with interoccasion variability was considered as the final model: CL = 54.4 l/h ;V = 117 + 0.581(age − 37) (l);KA = 2.481/h. Model estimates for the final model are detailed in Table 2. The residual variability from the final model [coefficient of variation (CV) 27.2%] was much smaller than that from the basic model (66%). A simulated plasma pharmacokinetic profile of IDV by using the population typical parameter estimates is shown in Fig. 1. A plot of weighted residuals versus final model-predicted IDV plasma levels is displayed in Fig. 2. Mean individual Bayesian estimates of CL and V, normalized by body weight, were 0.75 l/h per kg (CV 54.8%) and 1.74 l/kg (CV 82.7%), respectively. Mean values of other derived pharmacokinetic parameters were 1.50 h (CV 20.9%) for half-life, 0.87 h (CV 13.2%) for tmax, 9.51 μmol/l (CV 47.3%) for Cmax, 0.42 μmol/l (CV 57.5%) for C8h, and 29.56 μmol/l⋅h (CV 46.9%) for AUC8h As can be seen from Fig. 1, a large portion of plasma samples (331 of a total of 805; 41.1%) was obtained beyond 8 h after administration. Median sampling time for these levels was 9.5 h (range 8.1–13.5). Bayesian estimation of the median IDV concentration at this time was 0.20 μmol/l (CV 77.7%), which is in excellent agreement with the experimental median value within the same time interval: 0.19 μmol/l (CV 79.9%).
Trough levels (C8h) and AUC are generally considered as indicators of drug exposure. For patients with at least two visits, interoccasion variability of C8h and AUC8h was calculated for each individual expressed as the CV of the parameters over different occasions. As depicted in Fig. 3, while this did range from < 10% to > 60%, most subjects had interoccasion variability of 20–40% for the parameters.
Penetration of indinavir into cerebrospinal fluid
The penetration of IDV into the central nervous system was evaluated by measuring IDV levels in CSF and in plasma in a smaller group of 19 patients. Measured levels of IDV in CSF and corresponding model-predicted plasma levels of IDV are shown in Fig. 4. IDV levels in CSF in these patients were variable, ranging from 0.032 to 0.25 μmol/l, with a mean of 0.11 μmol/l (CV 49.7%). Average CSF sampling time was 1.4 h (range 0.5–3.3). Corresponding mean plasma IDV level was 10.28 μmol/l (CV 51.0%). The CSF:plasma ratio of IDV levels, an indicator of CSF penetration of the drug, ranged from 0.002 to 0.051 with a mean of 0.017 (CV 88.6%).
NONMEM analysis using sparsely collected samples has been previously applied to evaluate plasma population pharmacokinetics for several nucleoside and non-nucleoside HIV reverse transcriptase inhibitors, including zidovudine, didanosine and nevirapine, and the protease inhibitor saquinavir during combination therapy [9,10]. Implemented in large-scale studies, these population studies have provided a unique opportunity to define pharmacokinetic inter- and intraindividual variability and interoccasion variability associated with long-term antiretroviral therapy and to identify covariates that significantly affect pharmacokinetic behavior of these drugs. Results from these studies can serve to optimize anti-HIV therapy.
Pharmacokinetics of IDV was previously studied in healthy volunteers and HIV-infected patients during small-scale phase I and/or II clinical trials following single or multiple oral doses . Plasma pharmacokinetic parameters of IDV obtained from the present population study of the drug dosed at 800 mg every 8 h in 171 patients are 0.87 h, 9.51 μmol/l, 0.42 μmol/l and 29.56 μmol/l.h for tmax, Cmax, C8h and AUC8h, respectively. These results are in very good agreement with previous data as summarized in the manufacturer's package insert at the recommended dose and frequency of the drug (800 mg every 8 h):tmax, Cmax, C8h and AUC8h being 0.8 h, 12.62 μmol/l, 0.25 μmol/l and 30.69 μmol/l.h, respectively .
Clinical pharmacokinetics of IDV are associated with substantial interindividual variability, which is at least partly responsible for the variability in viral response to an IDV-containing regimen [13–15]. In addition to interindividual variability, pharmacokinetic behavior of drugs can also vary in time as a result of metabolic changes, for example in pathophysiological conditions such as renal and hepatic function. Because of the chronic nature of antiretroviral therapy, it is, therefore, of particular importance to define pharmacokinetic variability of anti-HIV drugs across different occasions within an individual (interoccasion variability). In the present study, as illustrated in Fig. 1, plasma levels of IDV varied extensively at any given time. This high variability could account for much of the variability seen among patients taking IDV with regard to both activity and toxicity. Variability (interoccasion and interindividual) of the structural parameters was successfully estimated with good precision (Table 2). Interoccasion variability (as CV) of CL and V (not normalized to body weight) were 56.7% and 60.7%, respectively, which was of comparable magnitude to their respective interindividual variability, 52.7% and 70.2%. Incorporation of interoccasion variability in the model successfully explained a portion of the residual variability, reducing this to 27.2% in the final model compared with 66.0% in the basic model.
Trough levels are defined as concentrations immediately preceding the next dose. These levels have been shown to correlate with HIV RNA response to an IDV-containing regimen [13–15]. For IDV taken every 8 h, plasma concentrations at 8 h postdosing are generally regarded as trough levels. In the present study (Fig. 1), a large number of plasma samples were obtained beyond 8 h, the indicated time to take the next dose, up to more than 13 h following the previous dose. Among the 331 samples collected after 8 h, more than two thirds were obtained later than 9 h, of which half were drawn 10 h after IDV administration. It should be pointed out that most of these late samples were drawn in the morning before the subjects being given the dose for pharmacokinetic sampling. IDV levels associated with these samples represent, therefore, the morning troughs of last night's dose. Given the short plasma half-life of the drug (1.5 h), trough levels beyond 10 h are much lower than those at 8 h. Median individual Bayesian estimates of IDV levels at 10 and 12 h post dosing were 0.13 (CV 79.1%) and 0.05 μmol/l (CV 86.1%), respectively. Despite the fact that a patient may be compliant by taking all three doses each day, it is conceivable that the time elapsing between the evening dose and the next morning dose could be much longer than 8 h, resulting in suboptimal drug exposure, which may compromise virological response over time.
Pharmacokinetics of IDV can be affected by other factors. Published data have focused primarily on drug–drug interactions between IDV, which is metabolized by hepatic CYP3A, and some co-administered drugs that influence IDV metabolism through induction or inhibition of this subfamily of enzymes [11,16]. In the present study, we examined a number of covariates that commonly affect pharmacokinetics of drugs. These factors included patient demographics and treatment regimen. The effects of these covariates were evaluated in the basic model without interoccasion variability to avoid excessive computation. IDV was administered in combination with lamivudine and zidovudine during the induction phase in all subjects and in one arm during maintenance therapy. No metabolic drug–drug interactions between IDV and these nucleoside analogs were anticipated, since they undergo different biotransformation processes. During the process of identifying covariates, incorporation of the factor ‘regimen’ (IDV alone versus in combination) into the model did not result in a better fit, demonstrating the absence of an effect of concomitant nucleoside drugs on IDV.
Body weight, a covariate that commonly correlates with CL and V, did not significantly influence IDV pharmacokinetics. Nevertheless, CL and V were still normalized to body weight for the purpose of uniformity and comparison with previously reported data. IDV is mainly cleared through hepatic CYP3A-mediated metabolism . In a previous animal study, gender was identified to affect IDV CL, with male rats exhibiting a metabolic clearance twice as high as that of females . This gender-dependent pharmacokinetics was caused by differences in the activity of CYP3A. Such a gender-dependent IDV CL, however, was not found in vivo in primates (monkey) or in vitro using human hepatic microsomes [17,18]. In the present study, gender did not affect CL when it was incorporated into the CL model, with male and female patients exhibiting comparable CL values (0.71 and 0.88 l/h per kg in males and females, respectively).
Among the patient demographics examined, age was the only covariate that significantly influenced IDV pharmacokinetics, with older subjects exhibiting a larger value for V IDV.CL remained unrelated to age in the studied age range. Since the elimination rate constant is related to CL and V, an increase in V with a stable CL results in an decrease in this value and an increase in the half-life of IDV as people age. It has been demonstrated that the aging process results in a decline in hepatic CYP3A-mediated oxidative metabolism of xenobiotics . However, metabolism of some endogenous substrates indicative of CYP3A activity, such as testosterone 6β-hydroxylation, is basically constant with increasing age [19,20]. In addition, an in vitro study that evaluated the effect of aging on human CYP3A activity also failed to show a correlation between age and CYP3A activity . These apparent paradoxes warrant further investigation of IDV pharmacokinetics in elderly subjects. The age-associated increase in the value of V for IDV may be related to its high lipophilicity. Lipophilic drugs such as diazepam were shown to exhibit an increased V in elderly individuals as a consequence of an aging-related increase in adipose tissue and a decrease in lean body mass relative to total body weight . The age-related change in IDV pharmacokinetics, as identified by the present study, should, however, have little clinical consequences since parameters underlying IDV exposure (AUC8h and trough levels) were not significantly influenced by age.
The central nervous system is an important HIV sanctuary . The effectiveness of antiretroviral therapy depends on the ability of antiretroviral agents to reach this area. IDV is the only protease inhibitor that achieves CSF levels that exceed its in vitro 95% inhibitory concentration for HIV replication (0.025–0.1 μmol/l) [23,24]. Its use, therefore, presents an opportunity to suppress viral replication in the central nervous system. HIV RNA in the CSF could not be detected in a study of combination therapy involving IDV . A recent study showed that an IDV-containing regimen led to > 1 log10 copies/ml decline in CSF HIV RNA compared with reference treatments that contained no protease inhibitors . CSF penetration of the drug has been previously documented [26–29]. For practical reasons, a single CSF sample was obtained per subject per visit in most cases. CSF IDV levels ranged from 0.05 to 0.66 μmol/l, with sampling occurring at 1–5 h after dosing [26–29]. In the present study, the mean CSF IDV level in 19 patients was 0.11 μmol/l, with an average sampling time of 1.4 h postadministration; this is consistent with other reported data. The failure observed in the IDV monotherapy arm  is, therefore, unlikely to be a result of insufficient CNS exposure of the drug. Mean CSF:plasma IDV ratio was 0.017. This ratio has so far been used as a measure of the extent of IDV CSF penetration. However, recently published data on IDV CFS penetration using levels measured up to 16 h after dosing demonstrated that IDV CSF levels remained constant over time with the standard regimen . Plasma drug levels declined rapidly, resulting in an apparent time-dependent change in the CSF:plasma ratio. The ratio was approximately 1 at 8 h postdose and 2 at 16 h . Although the CSF:plasma ratio of Cmax is generally accepted as a measure of CSF penetration, a more accurate approach would be to calculate the CSF:plasma ratio of IDV AUC through population analysis using sparse CSF samples.
In conclusion, plasma population pharmacokinetics of IDV in combination with zidovudine and lamivudine was studied in 171 HIV-infected patients. IDV pharmacokinetic parameters and associated intersubject variability were consistent with previously reported data. IDV pharmacokinetics were shown to exhibit interoccasion variability at a magnitude that is comparable to its interindividual variability. IDV CSF levels measured from 19 patients were above the 95% inhibitory concentration of HIV replication. Given that CSF is virtually free of protein, viral suppression in the central nervous system should be achievable with an IDV-containing regimen.
The authors would like to thank the patients and staff who contributed to the study, Michele Turner at the University of Alabama at Birmingham for indinavir analysis, the participating ACTU pharmacology laboratory staff and the ACTG Operations Staff.
1. Hammer SM, Squires KE, Hughes MD. et al
. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less.
:AIDS Clinical Trials Group 320 Study Team.
N Engl J Med 1997, 337: 725 –733.
2. Gulick RM, Mellors JW, Havlir D. et al
. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy.
N Engl J Med 1997, 337: 734 –739.
3. Hirsch M, Steigbigel R, Staszewski S. et al
. A randomized, controlled trial of indinavir, zidovudine, and lamivudine in adults with advanced human immunodeficiency virus type 1 infection and prior antiretroviral therapy.
J Infect Dis 1999, 180: 659 –665.
4. Havlir DV, Marschner IC, Hirsch MS. et al
. Maintenance antiretroviral therapies in HIV infected patients with undetectable plasma HIV RNA after triple-drug therapy.
:AIDS Clinical Trials Group 343 Study Team.
N Engl J Med 1998, 339: 1261 –1268.
5. Foisy ML, Sommadossi JP. Rapid quantification of indinavir in human plasma by high-performance liquid chromatography with ultraviolet detection.
J Chromatogr B 1999, 721: 239 –247.
6. Beal SL, Sheiner LB. NONMEM User's Guide.
San Francisco: University of California at San Francisco.
7. Karlsson MO, Sheiner LB. The importance of modeling interoccasion variability in population pharmacokinetic analyses.
J Pharmacokinet Biopharmaceut 1993, 21: 735 –750.
8. Ritschel WA. Pharmacokinetics of multiple dosing.
In:Handbook of Basic Pharmacokinetics
. Edited by Ritschel WA. Hamilton: Drug Intelligence Publications; 1976: 218 –234.
9. Zhou XJ, Sheiner LB, D'Aquila RT. et al
. Population pharmacokinetics of nevirapine, zidovudine, and didanosine in human immunodeficiency virus-infected patients.
Antimicrob Agents Chemother 1999, 43: 121 –128.
10. Vanhove GF, Kastrissios H, Gries JM. et al
. Pharmacokinetics of saquinavir, zidovudine, and zalcitabine in combination therapy.
Antimicrob Agents Chemother 1997, 41: 2428 –2432.
11. Plosker GL, Noble S. Indinavir – a review of its use in the management of HIV infection.
Drugs 1999, 58: 1165 –1203.
12. Merck Sharp & Dohme. Package insert for Crixivan
13. Stein DS, Fish DG, Bilello JA, Preston SL, Martineau GL, Drusano GL. A 24-week open-label phase I/II evaluation of the HIV protease inhibitor MK-639 (indinavir).
AIDS 1996, 10: 485 –492.
14. Acosta E, Henry K, Baken L, Page LM, Fletcher CV. Indinavir concentrations and antiviral effect.
Pharmacotherapy 1999, 19: 708 –712.
15. Murphy RL, Sommadossi JP, Lamson M, Hall DB, Myers M, Dusek A. Antiviral effect and pharmacokinetic interaction between nevirapine and indinavir in persons infected with human immunodeficiency virus type 1.
J Infect Dis 1999, 179: 1116 –1123.
16. Chiba M, Hensleigh M, Nishime JA, Balani SK, Lin JH. Role of cytochrome P450 3A4 in human metabolism of MK-639, a potent human immunodeficiency virus protease inhibitor.
Drug Metab Dispos 1996, 24: 307 –314.
17. Lin JH, Chiba M, Chen IW, Nishime JA, Vastag KJ. Sex-dependent pharmacokinetics of indinavir: in vivo and in vitro evidence.
Drug Metab Dispos 1996, 24: 1298 –1306.
18. Hunt CM, Westerkam WR, Stave GM. Effect of age and gender on the activity of human hepatic CYP3A.
Biochem Pharmacol 1992, 44: 275 –283.
19. Crowley JJ, Cusack B., Jue SG, Coup JR, Park BK, Vestal RE. Aging and drug interactions.
:II. Effect of phenytoin and smoking on the oxidation of theophylline and cortisol in healthy men.
J Pharmacol Exp Ther 1988, 1245: 513 –523.
20. Wrighton SA, Brian WR, Sari MA. et al
. Studies on the expression and metabolic capabilities of human liver cytochrome P-450IIIA5 (HLp3).
Mol Pharmacol 1990, 38: 207 –213.
21. Greenblatt DJ, Abernethy DR, Shader RI. Pharmacokinetic aspects of drug therapy in the elderly.
Ther Drug Monit 1986, 8: 249 –255.
22. Portegies P. HIV-1, the brain and combination therapy.
Lancet 1995, 346: 1244 –1245.
23. Emini EA, Schleif WA, Deutsch P, Condra JH. In vivo selection of HIV-1 variants with reduced susceptibility to the protease inhibitor L-735,524 and related compounds.
Adv Exp Med Biol 1996, 394: 327 –331.
24. Vacca JP, Dorsey BD, Schleif WA. et al
. L-735,524: an orally bioavailable human immunodeficiency virus type 1 protease inhibitor.
Proc Natl Acad Sci USA 1994, 91: 4096 –4100.
25. Gisslen M, Hagberg L, Svennerholm B, Norkrans G. HIV-1 RNA is not detectable in the cerebrospinal fluid during antiretroviral combination therapy.
AIDS 1997, 11: 1194. 1194.
26. Martin C, Sonnerborg A, Svensson JO, Stahle L. Indinavir-based treatment of HIV-1 infected patients: efficacy in the central nervous system.
AIDS 1999, 13: 1227 –1232.
27. Brinkman K, Kroon F, Hugen PW, Burger DM. Therapeutic concentrations of indinavir in cerebrospinal fluid of HIV-1-infected patients.
AIDS 1998, 12: 537. 537.
28. Enting RH, Hoetelmans RM, Lange JM, Burger DM, Beijnen JH, Portegies P. Antiretroviral drugs and the central nervous system.
AIDS 1998, 12: 1941 –1955.
29. Stahle L, Martin C, Svensson JO, Sonnerborg A. Indinavir in cerebrospinal fluid of HIV-1-infected patients.
Lancet 1997, 350: 1823. 1823.
Keywords:© 2000 Lippincott Williams & Wilkins, Inc.
indinavir; population pharmacokinetics; cerebrospinal fluid penetration; combination antiretroviral therapy