Ritonavir has shown a potent antiretroviral activity[1,2]. However, a large number of patients is unable to tolerate this drug because of neurological and gastrointestinal disturbances occurring during the early weeks of treatment. There is evidence that such side-effects may be dependent on ritonavir plasma concentrations. For example, dose escalation during the early weeks of treatment improves ritonavir tolerance by avoiding excessive plasma concentrations. Furthermore, it has recently been reported that a longer escalation dosing scheme is better tolerated than a shorter one[3,4]. Side-effects also have a tendency to occur more frequently in correspondence with the peak concentration of ritonavir in the plasma, i.e. from 2 to 6h after administration.
It should be noted that, even though the strategy of escalating the dose during the first weeks of treatment significantly improves ritonavir tolerance, the percentage of patients unable to tolerate it remains quite high. It may be hypothesized that the patients who experience side-effects are those with higher plasma concentrations, possibly because of greater absorption or lower metabolism.
The purpose of this study was to verify whether neurological or gastrointestinal side-effects occurring during the early weeks of treatment with ritonavir could be predicted on the basis of ritonavir concentrations in the plasma.
Patients and methods
Ambulatory HIV-positive individuals with indications for treatment with a protease inhibitor based on current guidelines  were enrolled in the study. The study was approved by the Institutional Review Board and a written informed consent was obtained from each patient before enrolment. A complete medical history, a physical examination, and a panel of laboratory tests consisting of a chemistry screen and a complete blood cell count with differential and platelet count were available within 2 weeks before the initiation of treatment with ritonavir. Viral load in the plasma was determined with the branch DNA assay (Chiron, Emeryville, CA, USA) with a lower limit of quantification of 400 copies/ml.
Two groups of patients were observed prospectively: patients experiencing neurological or gastrointestinal side-effects during the early weeks of treatment (group A), and patients who did not experience such side-effects (group B). Patients were included in group A if they seriously complained of side-effects and asked for intervention because the side-effects were barely tolerable or were intolerable.
All patients were nucleoside analogue experienced and protease inhibitor naive. Four patients in each group had experienced a previous AIDS-defining event. All the other patients were asymptomatic or mildly symptomatic. None of the patients had signs or symptoms of acute illness at the time of the pharmacokinetic study. None of the female patients was pregnant. The Karnofsky performance score was over 70 and the red blood cell counts were over 3.5×106 cells/μl.
All the patients received ritonavir in combination with two nucleoside analogues. Concomitant nucleoside analogues (number of patients) were: lamivudine (3TC)-stavudine (D4T) (five), 3TC-didanosine (DDI) (three), 3TC-zidovudine (ZDV) (two), and DDI-D4T (one) in group A versus 3TC-D4T (four), 3TC-DDI (one), 3TC-ZDV (four), and DDI-ZDV (one). Other concomitant medications were trimethoprim- sulphamethoxazole (TMP-SMX) (five), pentamidine (one) and methadone (one) in group A versus TMP-SMX (six), pentamidine (two), acyclovir (one), and methadone (one) in group B.
Patients were initially treated with ritonavir (capsule formulation) using the following escalation dosing scheme: 300mg twice a day for 3 days, 400mg twice a day for 4 days, 500mg twice a day for 5 days, then the full dose of 600mg twice a day. In the patients who experienced side-effects blood sampling for pharmacokinetic analysis was performed no later than 24h after the occurrence of side-effects. All the patients in group A experienced side-effects within 4 weeks of the initiation of treatment with ritonavir. Blood samples were scheduled to be drawn at: 0 (pre-dose), 1, 2, 3, 4, and 6h after a morning dose. The patients were asked to take their evening dose at exactly 12h before the scheduled time for the predose blood sample. On the day of blood sampling all patients received ritonavir after a light continental breakfast.
In the patients without side-effects, blood sampling was performed at the full dose of 600mg twice a day after at least 3 days of the administration of such a dosage regimen, at the same time schedule as in group A. Blood samples were drawn in pre-heparinized tubes, centrifuged and the plasma was stored at -20°C until assayed.
Ritonavir concentrations in the plasma were determined by a newly developed high-pressure liquid chromatography assay. Briefly, 2ml of phosphate buffer (Na2HPO4/NOH, pH12) were added to 1ml of patient plasma or calibration standard in culture tubes and mixed thoroughly. Standard curves and quality controls were prepared using ritonavir powder kindly donated by Abbott Laboratories, IL, USA. Standard curves were linear in the range 0.5-40mg/l. Patient plasma, standards and quality controls were applied to the Extrelut 3 column (Darmstadt, Germany) and eluted with diethylether. The organic phase containing the lipophilic substances was evaporated to dryness and reconstituted in 200μl of mobile phase. Ritonavir was separated with a Supecosil (Bellefonte, PA, USA) LC-ABZ column (5μm, 15cm × 4.6mm) using a mixture of phosphate buffer:acetonitrile (55:45) as mobile phase at a flow rate of 1ml/min. The detector was set at a wavelength of 205nm. Precision (intra and interday variabilities) of the assay was less than 10% at each of the quality control concentrations of 0.8, 3, and 15mg/l. The average recovery was over 75% and the lower limit of quantitation was 0.5mg/l.
Pharmacokinetic and statistical analysis
The peak concentration (Cmax) was defined as the highest concentration observed in the plasma, time-to-peak (Tmax) was the time of Cmax observation, trough concentration (Cmin) and the actual trough (actual Cmin) were the observed predose concentration, and the lowest concentration observed after ritonavir administration, respectively.
The area under the concentration versus time curve from time 0 to the end of the sampling time (AUC 0-6) was computed using the log trapezoidal rule with the program Siphar (Simed, Creteil, France).
Cmax, Cmin, actual Cmin, and AUC 0-6 were compared in the two study groups using the Mann-Whitney U test. The Pearson correlation coefficient, obtained by regression analysis, was also used to evaluate the relationship between pharmacokinetic and demographic parameters, after pooling the data obtained in the two groups of patients.
The software Statview II (Abacus Concepts Inc., Berkeley, CA, USA) was used for statistical analysis.
Patient demographics are shown in Table 1. The two study groups were balanced in terms of patient age, body mass index (calculated as weight/height2), serum creatinine [0.99±0.30 (mean±SD) in group A versus 0.89±0.10 mg/dl in group B], alanine aminotransferase (44±33 versus 36±28IU/l), bilirubin (0.58±0.43 versus 0.61±0.26mg/dl), alcoholism (four out of 11 versus two out of 10), detectable HCV RNA (six out of 11 versus five out of 10), tobacco smoking (seven out of 11 versus six out of 10), and baseline viral load. Patients with side-effects had significantly lower weight and higher baseline CD4 cell counts. There was a higher number of women patients (82%) in the group with adverse effects compared with the other group (40%). Only one patient in group B had an undetectable trough concentration, which was entered into the analysis as 0.5mg/l. The lowest concentration in the remaining 20 patients was 4.8mg/l, also observed in the group without side-effects (Fig. 1).
Pharmacokinetic data are shown in Table 2. Patients with side-effects had higher Cmax, Cmin, actual Cmin, and AUC 0-6. Linear regression revealed a significant positive association of Cmax with baseline CD4 cell counts (r=0.59, P=0.005) (Fig. 2a), and a negative association with weight (r=0.57, P=0.007) (Fig. 2b) and baseline log viral load (r=0.58, P=0.014). Baseline CD4 cell counts and viral load were also significantly associated with Cmin and actual Cmin whereas patient weights did not correlate. The peak concentration was significantly higher in women patients [26.7 (17.8-33.4) versus 16.7 (15.8-18.6) mg/l, P=0.04], whereas Cmin and actual Cmin did not differ.
A subanalysis was conducted using the data obtained in the 13 women enrolled in the study. The association of Cmax with side-effects was still evident. Cmax in the nine women with side-effects was higher than in the four women without side-effects [28.0 (23.4-34.3) versus 15.0 (12.7-25.0) mg/l, P=0.06] as well as Cmin [12.6 (8.6-14.4) versus 6.0 (2.7-8.3), P=0.02]. By pooling all the women patients in the two study groups, the relationship between ritonavir concentrations and baseline CD4 cell counts was as follows: Cmax versus CD4 cells (r=0.49, P=0.09), Cmax versus viral load (r=0.70, P=0.01), and Cmin versus CD4 cells (r=0.67, P=0.01).
Six patients in group A underwent pharmacokinetic sampling while taking a 500mg dose. Therefore, a subanalysis was done for the patients studied at the 600mg twice a day dosage regimen. Cmax was 26.7 (24.5-32.6) mg/l in patients with side-effects (n=5) versus 16.2 (13.4-17.0) mg/l (P=0.01) in patients without side-effects (n=10). Cmin was 10.3 (7.7-12.8) versus 7.5 (4.9-8.6) (P=0.07), respectively. The association with baseline CD4 cell counts was still significant for both Cmax (r=0.53, P=0.04) and Cmin (r=0.69, P=0.005).
Of note is the fact that patients with neurological and gastrointestinal side-effects were pooled in a single group (group A) for the purpose of this analysis. Ten of the 11 patients in group A complained primarily of neurological side-effects even though complaints of gastrointestinal disturbances (mild to severe) were reported by six of them. Three of these six patients plus one patient who did not have neurological side-effects complained of serious gastrointestinal side-effects.
Our study shows that ritonavir neurological or gastrointestinal side-effects may be predicted by higher Cmax, Cmin, actual Cmin, and AUC 0-6. On the basis of a generally accepted concept, and in the absence of any data proving the contrary, it may be assumed that Cmax is the true parameter of ritonavir exposure associated with side-effects, whereas Cmin and actual Cmin may also be significantly correlated, as they were in our data set, because of their relationship with Cmax.
Of note is the fact that plasma concentrations were not the only variables significantly associated with adverse events. Higher baseline CD4 cell counts appeared to be significantly associated with side-effects. This may reflect higher plasma concentrations of ritonavir in patients with high CD4 cell counts. In fact, it has been reported that plasma levels of most of the drugs are lower in HIV-infected patients at an advanced stage of disease[7-10], possibly as a consequence of impaired absorption, although this is not true for all the drugs[11-13]. A lower absorption in patients at an advanced stage of disease, using the CD4 cell count as a marker of disease stage, may be the case for ritonavir. In fact, when pooling the data obtained in the two study groups, ritonavir concentrations showed a significant positive association with baseline CD4 cell counts and a negative association with baseline viral load.
Women patients appeared to be at a higher risk of side-effects in our study. There is a lack of data regarding ritonavir tolerance in women patients[1,2]. In a previous report, sex was not a risk factor for side-effects, even though only a minority of the patients studied were women. A higher incidence of side-effects in women patients may be due to higher ritonavir concentrations as a consequence of a lower volume of distribution. In fact, Cmax but not Cmin or actual Cmin were significantly higher in women than in men, and Cmax was also associated with patient weight, which was lower in women patients (59.0 versus 65.7kg, P=0.02).
In order to gain a further insight into the relevance of our findings regarding the correlation between ritonavir plasma levels and side-effects, a subanalysis was done using only the data obtained in the women patients. The association of Cmax and Cmin with side-effects as well as with baseline CD4 cell counts was confirmed, even though, as expected from the smaller sample size, it had a lower level of significance.
An aspect of our data that should be pointed out is the relatively high interindividual variability in plasma concentrations compared with previously reported data obtained in healthy volunteers  or in HIV-infected patients in a less advanced stage of disease. The peak plasma concentration showed a frequency distribution skewed to the right (Fig. 3), with values in agreement with a recent study in a patient population comparable to ours. The high variability of ritonavir concentrations may be an effect of saturable kinetics as well as interindividual variability in the activity of the P450 isoenzymes metabolizing ritonavir and their extent of saturability.
It is of interest that because ritonavir induces its own metabolism and such induction is completed by 2 weeks of administration, it may be argued that the plasma concentrations in some of the patients observed at the 500mg twice a day dosage were higher than those observed in the other patients because induction was not yet completed. This should not be of concern, because the analysis of the subset of patients observed at the 600mg twice a day regimen confirmed the results obtained with the whole data set.
The results of our study may have a relevant clinical application. Our study suggests that individualization of the ritonavir dosage regimen, e.g. a downward titration of the ritonavir dose in patients with side-effects, guided by plasma level monitoring, may result in a substantial increase in the percentage of patients tolerating ritonavir without increasing the risk of treatment failure caused by the selection of resistant viral mutants as a consequence of suboptimal systemic exposure to ritonavir. According to the current guidelines, if a patient treated with a protease inhibitor experiences a side-effect incompatible with further administration, a reduction of the dose should be avoided, and the patient should be treated with an alternative drug. This recommendation is based on the evidence that higher doses of protease inhibitors are associated with a better outcome[1,2,18]. This is certainly true when comparing response to therapy in patients treated with different doses of protease inhibitors. However, because of the interindividual variability in pharmacokinetic processes[18,19], it is believed that parameters of drug exposure, such as Cmax, Cmin or AUC may be more precise in predicting outcome than the dose administered[18,20,21]. In particular, it is thought that a Cmin higher than the mean IC90 would be a reasonable parameter to target for. Therefore, in case certain side-effects correlate with high plasma concentrations, as shown in our study, it may be possible to reduce the dosage regimen as long as Cmin is maintained above IC90. Individualization of the dosage regimen on the basis of peak and trough concentrations is a well known procedure known as therapeutic drug monitoring (TDM).
Our study was designed to address some of the issues that may arise in the hypothesis of applying TDM principles to ritonavir, such as the optimal sampling times for peak and trough levels as well as the evaluation of such levels as a guide to dose adjustment. The peak concentration of ritonavir occurs on average at approximately 4h after administration. This was confirmed in our study. Therefore, in routine clinical practice this may be an appropriate time to sample for peak, even because a ritonavir concentration determined within ±1h from Tmax may be considered a very close approximation of the true peak, due to Michaelis-Menten kinetics. In a case when the peak concentration in a patient with side-effects is higher than a target value, e.g. 1SD above the mean peak concentration observed in patients without side-effects, the patient may be considered eligible for dose reduction. Clearly, it is necessary to predict whether dose reduction will result in a sub-inhibitory trough concentration, e.g. less than the mean IC90 of 2.1mg/l. In our data set, the lowest trough concentration observed in patients with side-effects was 7.9mg/l, a value that leaves plenty of room for dose reduction. Clearly, the trough concentration should be re-evaluated after a dose reduction in order to verify that the concentration is indeed above the desired threshold. Michaelis- Menten kinetics may, in fact, cause lower-than-expected concentrations after dose reduction. In a case when the trough concentration is, or is predicted to be, unacceptably low after dose reduction, a more reasonable approach may be to administer ritonavir three times a day. Such an approach may result in a small percentage of patients undergoing dose adjustment.
It should be noted that although the lowest plasma concentration is usually the pre-dose concentration, defined as ‚trough‚, this may not be the case for all patients treated with ritonavir. In fact, in some patients plasma concentrations of protease inhibitors continue to decay after administration due to delayed absorption. In our study, the lowest concentration occurred at time 0 in 11 patients and between 1 and 3h in the remaining patients, with a coefficient of correlation of Cmin versus actual Cmin of r=0.87. The actual Cmin was up to 49% lower than the pre-dose concentration. A further complication may be that ritonavir presents kinetics sensitive to circadian rhythm, so that the trough before the dose in the afternoon is 32% lower than the trough in the morning. A conservative approach may be to consider the morning trough concentration as an overestimation of the actual Cmin in the afternoon, i.e. the actual lowest daily concentration, which may be up to approximately 80% lower than the morning pre-dose concentration, in the worst scenario.
Although the application of TDM was not the purpose of this study, we adopted it in some of the 11 patients who had side-effects. We considered a peak higher than 1SD above the mean observed in patients without side-effects as an indication for dosage adjustment. One patient who had side-effects at 600mg twice a day was titrated downwards to 500mg twice a day, and another patient who also had side-effects at 600mg twice a day was titrated to 500mg twice a day, and then again to 400mg twice a day. A patient who had side-effects at 500mg twice a day was kept at this dosage regimen, because the side-effects decreased after approximately 10 days from their manifestation. Another patient with side-effects at 500mg twice a day was kept at this dosage regimen for a few weeks, then subsequent to a reduction in side-effects, the dosage regimen was increased to 600mg twice a day, because her trough concentration was only barely above IC90. In all such patients who were treated at a dose different from the recommended full dose of 600mg twice a day, trough concentrations remained above twice the IC90, and viral load remained under the limit of detection (two patients) or kept on decaying (two patients) at 4 months of follow-up. Of the remaining patients with side-effects, two patients did not have an indication for TDM based on Cmax, three refused dose reduction and preferred to be treated with an alternative drug, a patient in denial refused any further treatment, and the remaining patient was lost at follow-up as a result of relocation.
This preliminary experience is in agreement with a recent observation in which patients receiving ritonavir in combination with saquinavir had their ritonavir dose titrated downwards with improved tolerance and no lack of efficacy. Other preliminary reports of TDM application to protease inhibitors other than ritonavir are encouraging[22,23].
A number of issues should be addressed for the optimal application of TDM to ritonavir. It is necessary to know the percentage of patients who may ultimately benefit from such intervention and, therefore, its cost-effect ratio. It is possible that counselling before the initiation of treatment, with an assurance to patients that, in case of the occurrence of side-effects, TDM has a high probability of being effective, may increase the percentage of patients compliant with such an intervention.
Clearly, the usage of a mean IC90 as a reference value to be used in TDM application is a rather inaccurate approach. The knowledge of the actual IC90 in a single patient would be of greater help in fine tuning the trough concentration in that particular patient. In this respect, the introduction of an antivirogram in clinical practice is urgently needed.
We have found a significant relationship between ritonavir plasma concentrations and the side-effects that occur early after the initiation of treatment with this drug. Dosage adjustment based on plasma concentrations may result in a higher percentage of patients capable of tolerating ritonavir without a reduction of antiviral efficacy.
1. Markowitz M, Saag M, Powderly WG, et al. A preliminary study of ritonavir, an inhibitor of HIV-1 protease, to treat HIV-1 infection. N Engl J Med 1995, 333:1534-1539.
2. Danner SA, Carr A, Leonard JM, et al. A short term study of the safety, pharmacokinetics, and efficacy of ritonavir, an inhibitor of HIV-1 protease. N Engl J Med 1995, 333:1528-1533.
3. Arribas JR, Lazzarin A, Martin T, Moroni M, Vanhove G, MacLeod C. Effect of dose escalation on the safety, tolerability and efficacy of ritonavir. 4th International Congress on Drug Therapy in HIV Infection. Glasgow, 8-12 November 1998 [Abstract P41].
4. Gatti G, Arribas JR, Lazzarin A, Martin T, Moroni M, Bassetti D. Ritonavir systemic exposure using a 14-day vs. a 7-day dose escalation scheme. 4th International Congress on Drug Therapy in HIV Infection. Glasgow, 8-12 November 1998 [Abstract P48].
5. Merry C, Barry M, Gibbons S, Mulcahy F, Back D. Improved tolerability of ritonavir derived from pharmacokinetic principles. Br J Clin Pharmacol 1996, 42:787.
6. Carpenter C, Fischl M, Hammer, et al. Updated recommendations of the International AIDS Society Panel - USA Panel. JAMA 1997, 277:1962-1969.
7. Gordon SM, Horsburgh CR, Peloquin CA, et al. Low serum levels of oral antimycobacterial agents in patients with disseminated Mycobacterium avium complex disease. J Infect Dis 1993, 168:1559-1562.
8. Lake-Bakaar G, Winston T, Lake-Bakaar D, et al. Gastropathy and ketoconazole malabsorption in the acquired immunodeficiency syndrome (AIDS). Ann Intern Med 1988, 109:471-473.
9. Macnab KA, Gill MJ, Sutherland LR, De Boer Visser N, Church D. Erratic zidovudine bioavailability in HIV seropositive patients. J Antimicrob Chemother 1993, 31:421-428.
10. Gatti G, Guerra M, Miletich F, Gatti C, Cruciani M, Bassetti D. Comparative study of dapsone plasma levels in HIV+ patients and healthy volunteers using a population approach. 36th Interscience Conference on Antimicrobial Agents and Chemotherapy. New Orleans, 15-18 September 1996.
11. Crivat M, Vittecoq D, Kodjo A, et al. Pharmacokinetic of saquinavir and study of intestinal function in HIV patients. 37th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, 28 September 28-1 October 1997 [Abstract A9].
12. Gatti G, Flaherty J, Bubp J, White J, Borin M, Gambertoglio JG. Comparative study of bioavailabilities and pharmacokinetics of clindamycin in healthy volunteers and patients with AIDS. Antimicrob Agents Chemother 1993, 37:1137-1143.
13. Gatti G, Papa P, Torre D, Andreoni M, et al. Population pharmacokinetics of rifabutin in human immunodeficiency virus-infected patients. Antimicrob Agents Chemother 1998, 42:2017-2023.
14. Casado JL, Perez-Elias MJ, Antela A, et al. Risk factors for ritonavir intolerance and outcome after change to indinavir [letter]. AIDS 1998, 12:335-336.
15. Norvir. Abbott Laboratories, Abbott Park, North Chicago, IL, USA. Prescribing information, June 1999.
16. Hsu A, Granneman GR, Witt G, et al. Multiple-dose pharmacokinetics of ritonavir in human immunodeficiency virus-infected subjects. Antimicrob Agents Chemother 1997, 41:898-905.
17. Taylor S, Dedicoat M, Hackforth S, et al. The use of ritonavir and saquinavir plasma level monitoring as an aid to dosage adjustment. 4th International Congress on Drug Therapy in HIV Infection. Glasgow, 8-12 November 1998 [Abstract P50].
18. Schapiro JM, Winters MA, Stewart F, et al. The effect of high dose saquinavir on viral load and CD4+ T-cell counts in HIV-infected patients. Ann Intern Med 1996, 124:1039-1050.
19. Gatti G, Di Biagio A, De Pascalis C, et al. High variability of indinavir parameters of systemic exposure in HIV+ patients observed prospectively in everyday clinical practice. 38th Interscience Conference on Antimicrobial Agents and Chemotherapy. San Diego, 24-27 September 1998 [Abstract A63].
20. Acosta EP, Henry K, Weller D, et al. Indinavir pharmacokinetics and relationship between exposure and antiviral effect. 37th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, 28 September-1 October 1997 [Abstract A15].
21. Molla A, Korneyeva M, Gao Q, et al. Ordered accumulation of mutations in HIV protease confers resistance to ritonavir. Nature Med 1996, 2:760-766.
22. Dieleman J, Gyssens IC, Van Der Ende MEM, De Marie S, Burger DM. Urologic complaints in relation to indinavir plasma levels in HIV-infected patients. 12th World AIDS Conference. Geneva, 28 June-3 July 1998 [Abstract 12372].
23. Perello L, Goujard C, Delfraissy JF, Taburet AM. Therapeutic drug monitoring in HIV-infected patients receiving indinavir. 12th World AIDS Conference. Geneva, 28 June-3 July 1998 [Abstract 42272].
© 1999 Lippincott Williams & Wilkins, Inc.