The safe use of drugs necessitates a thorough understanding of their pharmacokinetic behavior. Older individuals have long been recognized as being more susceptible to adverse drug reactions than younger subjects. In one study, elderly individuals had a 70% higher rate of hospital admissions for adverse drug reactions than younger adults, and were more likely to be receiving multiple medications.32
Van der Hooft et al.33
observed that the frequency of hospitalizations for adverse drug reactions was related to older age. HIV protease inhibitors require stable plasma concentrations to suppress viral replication and to prevent acquisition of antiretroviral drug resistance mutations. The balancing act between insuring efficacy and minimizing toxicity may become more of a challenge as older individuals become the fastest growing demographic in the United States. As HIV disease increasingly becomes a disease of older people, an understanding of the effect of aging on antiretroviral pharmacokinetics is important for predicting virologic and immunologic outcomes in this population. Age-related decrements in renal function, medical comorbidities, and the increased number of concurrent medications in older patients can potentially affect antiretroviral drug disposition.
Our study found a positive correlation between older age and LPV trough concentration that was significant at week 24. Younger subjects tended to have lower trough concentrations of LPV, with evidence for the difference being strongest at week 24. The fact that younger subjects tended to have their trough blood drawn slightly sooner after their previous LPV dose than the older subjects could have been a source of bias. However, assuming that an earlier trough blood draw would cause LPV concentrations to be higher than if taken from a later blood draw, the bias in this study, if present, would be toward the conclusion that young subjects had higher concentrations of LPV, opposite to our findings.
In our study, we observed an effect of age on LPV pharmacokinetics independent of gender or other demographic variables. In contrast, van der Leur et al.34
in a multivariate regression analysis found that body mass index was inversely associated with lopinavir plasma concentration, but there was no effect of age. Similarly, Guillemi et al.35
found no differences in trough plasma LPV concentrations in patients greater than 60 years old receiving LPV compared to patients less than 35 years old. However, in neither of these studies was a repeated measures design employed, collecting multiple samples from each patient over a broad span of time as we did. Zhou et al.36
identified age as the primary covariate (including race, body weight, and gender) influencing indinavir pharmacokinetics, where older subjects displayed a larger volume of distribution and an increase in indinavir half-life. Nevertheless, they found no effect of aging on indinavir trough plasma concentrations or AUC8h
, suggesting that the decline in clearance with age might balance the effect of Vd. They further observed an age-associated decrease in clearance in a univariate analysis. We also evaluated body weight as a covariate influencing LPV pharmacokinetics and found weight negatively associated with volume. Similarly, Bouillon-Pichault et al.37
found that body weight was significantly associated with the probability of achieving adequate LPV exposure. They also found that differences in body weight accounted for much of the variability in LPV clearance, an observation that may help explain the marked variability in protease inhibitor plasma concentrations reported by other investigators.38,39
Their study had a number of differences from ours, including a larger sample size, broader age range, use of different LPV doses, a significantly higher proportion of women, and use of drugs in the combination known to affect LPV pharmacokinetics (i.e., NNRTIs).
Our observations may help explain other age-related differences observed in patients on HAART. Studies have suggested that virologic response to HAART is greater in older patients than in younger patients, but the immunologic response (recovery of CD4+
cells) is blunted. Although some investigators observed no differences across age groups in virologic suppression in HAART-treated patients,38
a number of researchers have observed better virologic responses including a higher proportion of virologically suppressed patients,40,41
a shorter time to becoming suppressed,42
greater virologic suppression,43
and greater durability of viral suppression40
in older adult patients. Interestingly, we observed no age-related differences in the occurrence of grade 3–4 toxicities. This is a potentially important finding supporting the safety of lopinavir/ritonavir in older patients. Improved medication adherence in older patients, as observed in our study, is consistent with the results reported by others.41
Better adherence among older patients could contribute to higher plasma concentrations and greater virologic responses, although Goodkin et al.44
observed better virologic responses in older patients independent of the effect of medication adherence.
In contrast to the enhanced virologic responses seen in older patients, several studies have found the recovery of CD4+
lymphocytes in response to HAART to be blunted in older patients.45
These effects include a lower absolute CD4+
lymphocyte count increase in response to HAART46
and slower rates of CD4+
a decreased proportion of naive CD4+
cells in untreated individuals, and diminished naive CD4+
although some investigators report no age-related changes in these parameters.38,49
Several factors could explain age-related changes in LPV pharmacokinetics. LPV is mainly a cytochrome (CYP) 3A4 substrate, and changes in the expression and activity of the subclass have been reported at various stages from infancy to adulthood.54,55
However, decreases in CYP3A activity in elderly individuals have not been consistently demonstrated.56,57
This may reflect the biological importance of CYP3A and the large capacity for CYP3A metabolism in the liver. In some studies of phenotyping using CYP3A4 probes, gender differences in metabolism are observed, which persist at older ages.56–59
However, in population studies of calcium channel blocker pharmacokinetics, drugs that are also CYP3A substrates, observed gender differences showed no effect of age.60–62
has suggested that coadministered medications may play a more important role than age or gender in older individuals because they are more likely to be on multiple drugs.
Coadministration of LPV with ritonavir, a highly potent CYP3A inhibitor, may increase the sensitivity of LPV as a probe for age-related changes in metabolism of CYP3A substrates. Combining LPV with ritonavir results in a 13-fold increase in steady-state LPV concentrations,63
and CYP3A4 is wholly responsible for the metabolism of LPV. Because of the profound effect of ritonavir in boosting LPV concentrations, even a modest increase in ritonavir concentrations could translate into a significant pharmacokinetic effect. Unfortunately, plasma sample volumes were not adequate to assay for ritonavir concentrations in our study.
Changes in liver size and liver blood flow with age seem to be well supported in the medical literature.16
Between young adulthood and old age, liver size decreases by 24–35% and liver blood flow decreases by 35%,64–66
effects that can result in diminished clearance of drugs with a high first-pass metabolic extraction, such as LPV and ritonavir.62
Recognizing that HIV protease inhibitors are substrates of MRPs and MDR-1/p-glycoprotein, important studies examining age-related changes in transporter expression come from research in oncology. For example, Plasschaert et al.67
reported higher activity and expression of p-glycoprotein in older patients with T cell acute lymphoblastic leukemia. Ritonavir is both a p-glycoprotein inhibitor and substrate.68
The effects of drug transporters on pharmacokinetics are difficult to predict as changes in transporter function on drug absorption compared with drug elimination could produce opposite effects on plasma concentrations.
The binding of protease inhibitors to plasma proteins may also be an important interaction that modulates the disposition of these drugs. Some HIV protease inhibitors are highly bound to orosomucoid or α1
-acid glycoprotein (AAG), and the concentration of this serum protein influences free concentrations of these antiretrovirals and their pharmacologic effects.69,70
Plasma concentrations of AAG were strongly associated with indinavir concentrations but less so with ritonavir concentrations.71
Concentration-dependent binding of lopinavir to orosomucoid appears to occur in vivo
, an interaction that influences the level of unbound drug and may be important in lopinavir pharmacokinetics.72
Concentrations of AAG have been reported to be affected by age and disease states,73–75
but were not measured in this study.
Our studies point to a decrease in the clearance of LPV as a likely contributor to the increased trough concentrations seen in older subjects. Clearance was calculated using data from all 77 of these subjects. In older patients, hepatic drug clearance may be reduced by up to 30% with aging, and renal elimination decreased by up to 50%.76
identified reduced hepatic and renal clearance as the most significant changes influencing pharmacokinetics with normal aging, and suggested that changes in oral bioavailability in aging result from reduced first-pass hepatic metabolism for high extraction drugs, such as LPV and ritonavir. She suggests that changes in volume of distribution are smaller than changes in clearance and contribute less significantly to other pharmacokinetic parameters. The association between age and trough LPV concentration was significant only at 24 weeks. A trend toward higher plasma concentrations in older individuals was also observed at weeks 48 and 96, but failed to reach statistical significance most likely because there were fewer data points at these times. Even though older subjects had a higher median number of nonantiretroviral medications than younger subjects, it is not likely that drug interactions explains the differences in trough concentration, as drugs known to interact with lopinavir/ritonavir were not allowed in the study.
We have demonstrated modest age-related differences in the concentrations of LPV. Although these are unlikely to affect LPV efficacy or toxicity, given its broad therapeutic index, more attention should be paid to age-related changes in concentrations of other drugs used in this patient population as the epidemic matures and new classes of antiretroviral drugs become available. Recent unpublished studies have found increased concentrations of darunavir78
in older subjects. Future studies should consider the effects of aging on concentrations of other antiretrovirals, given the potential impact on long-term efficacy and safety.