Published ahead of print on March 30, 2006, doi:10.1164/rccm.200602-280OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200602-280OC
Potent Twice-Weekly Rifapentine-containing Regimens in Murine TuberculosisCenter for Tuberculosis Research, Department of Medicine, Johns Hopkins University School of Medicine; Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland; Infectious Diseases Pharmacokinetics Laboratory, National Jewish Medical and Research Center; Departments of Pharmacy and Medicine, University of Colorado, Denver, Colorado; and Division of Tuberculosis Elimination, Centers for Disease Control and Prevention, Atlanta, Georgia Correspondence and requests for reprints should be addressed to Eric L. Nuermberger, M.D., 1503 E. Jefferson Street, Baltimore, MD 21231-1002. E-mail: enuermb{at}jhmi.edu
Rationale: Recent studies have demonstrated that intermittent administration of rifamycin-based regimens results in higher rates of tuberculosis relapse and treatment failure compared with daily therapy. Twice-weekly treatment with rifampin, isoniazid, and pyrazinamide may be improved by increasing Mycobacterium tuberculosis exposure to rifamycin by substituting rifapentine for rifampin. Methods: To test this hypothesis, we compared the activities of standard daily and twice-weekly rifampin plus isoniazid-based regimens to those of twice-weekly rifapentine plus isoniazid- or moxifloxacin-containing regimens in the murine model of tuberculosis. Relapse rates were assessed after 4, 5, and 6 mo of treatment to assess stable cure. Single- and multiple-dose pharmacokinetics of rifampin and rifapentine were also determined. Results: After 2 mo of treatment, twice-weekly therapy with rifapentine (15 or 20 mg/kg), moxifloxacin, and pyrazinamide was significantly more active than standard daily or twice-weekly therapy with rifampin, isoniazid, and pyrazinamide. Stable cure was achieved after 4 mo of twice-weekly rifapentine plus isoniazid- or moxifloxacin-containing therapy, but only after 6 mo of standard daily therapy. Twice-weekly rifapentine (15 mg/kg) displayed more favorable pharmacodynamics than did daily rifampin (10 mg/kg). Conclusions: By virtue of the enhanced rifamycin exposure, twice-weekly regimens containing rifapentine (15 or 20 mg/kg) may permit shortening the current treatment duration by 2 mo. Such regimens warrant clinical investigation.
Key Words: moxifloxacin • rifampin • rifapentine • tuberculosis • treatment In the treatment of drug-susceptible pulmonary tuberculosis (TB), daily administration of a standardized 6-mo regimen including rifampin (R), isoniazid (H), and pyrazinamide (Z; collectively, RHZ) is highly effective in achieving cure (1). But, because supervision of daily treatment imposes a substantial burden on both patients and treatment programs (2, 3), intermittent regimens with predominantly twice- or thrice-weekly drug administration are recommended for supervised therapy (4, 5). Although at least one randomized clinical trial has shown that daily and thrice-weekly RHZ-based regimens have similar efficacy, more recent observational studies suggest intermittent therapy is not as effective as daily therapy, particularly in patients at high risk for relapse (6–8). Recent experience in the murine model of TB supports the latter observations. Daily administration of the 6-mo RHZ-based regimen is significantly more effective than the same regimen administered according to a predominantly twice-weekly schedule (9). The rifamycins are the key drugs in modern short-course therapy. Rifapentine (P) is a rifamycin with a lower minimum inhibitory concentration (MIC) against Mycobacterium tuberculosis and a much longer serum half-life compared with R (10). While P is approved for twice-weekly administration at a dose of 10 mg/kg during the initial phase of therapy (11), several influential clinical trials have focused attention on the use of once-weekly administration of P in combination with H during the continuation phase. In Tuberculosis Trials Consortium Study 22, once-weekly PH was less effective than twice-weekly RH, and was associated with rifamycin-resistant relapse among HIV-infected patients (12–15). As a result, once-weekly PH is recommended only for HIV-seronegative patients at low risk for relapse (4). Studies in the murine model have demonstrated that increasing the dose of P from 10 to 15 mg/kg increases the anti-TB activity (16), a finding consistent with the concentration-dependent activity of the rifamycins (17–19). Once-weekly continuation phase regimens based on a 15-mg/kg dose of P are at least as active as twice-weekly RH, but are not as active as daily RH (9). This led us to hypothesize that using P twice weekly rather than once weekly would significantly increase the rifamycin exposure and result in activity similar to that of standard daily therapy. Prior experience in the murine model suggested that the substitution of moxifloxacin (M) for H would lead to additional gains in activity (9, 20, 21). In the current series of studies, the bactericidal and sterilizing activities of modified twice-weekly regimens were compared with those of standard daily therapy in the murine model of TB. In Study 1, we used the predominantly twice-weekly RHZ-based regimen as the foundation to assess the effect of three modifications: substitution of M for H, substitution of P for R during the intermittent phase of therapy, and increasing the dose of P from 10 to 15 and 20 mg/kg. In Study 2, we systematically determined the contribution of individual drugs or drug combinations to the activity of twice-weekly P-based regimens. Finally, single- and multiple-dose pharmacokinetic studies were conducted to determine whether twice-weekly P regimens provide increased effective rifamycin exposure compared with daily R.
Antimicrobials Drugs were obtained and stock solutions prepared as previously described (9).
M. tuberculosis Strain
Aerosol Infection
Chemotherapy
Study 1
Study 2 The scheme of Study 2 is presented in Table 2. The overall duration of Study 2 was 2 mo. Mice in Positive Control Groups 1 and 2 received 2 mo of daily RHZ and RM2Z, respectively. To assess the role of Z, H, and M, mice in Test Groups 3–7 received 2 wk of daily RHZ followed by 6 wk of twice-weekly P15 either alone or in combination with Z, HZ, M2Z, or M2HZ, respectively; mice in Test Groups 8 and 9 received 2 wk of daily RM2ZH, followed by 6 wk of twice-weekly P15M2Z, or P15M2ZH, respectively. To assess the effect of duration of the initial daily phase, mice in Groups 10 and 11 received a daily initial phase of 2 and 1 wk of RM2Z, followed by 6 and 7 wk of twice-weekly P15M2Z, respectively, whereas mice in Group 12 received 8 wk of twice-weekly P15M2Z.
Infected, untreated negative control mice were kept to assess mortality from infection for each study.
Assessment of Treatment Efficacy
Pharmacokinetic and Pharmacodynamic Analysis
Statistical Analysis
Body Weights In both Studies 1 and 2, no differences were observed in mean body weights between any of the treatment groups, suggesting that none of the tested drug regimens caused significant adverse effects.
Treatment Efficacy in Study 1 After the initial 2 wk of daily therapy, the lung cfu counts were 6.13 ± 0.05, 5.88 ± 0.04, and 5.83 ± 0.08 log10 in mice treated with RHZ, RM1Z, and RM2Z, respectively (p > 0.05). After 8 wk of therapy, as shown in Figure 1, the lung cfu counts were more than 1 log10 lower (p < 0.01) in Group 1 mice treated with daily RHZ, (2.56 log10) compared with Group 2 mice treated with 2 wk of daily RHZ, followed by 6 wk of twice-weekly RHZ (3.61 log10). This 1-log10 difference in the lung cfu count is consistent with previous studies, and confirms the superior activity of RHZ when administered daily in the mouse (9). In mice of Groups 3 and 4 that were treated with twice-weekly RMZ, the cfu counts were lower than those in mice treated with twice-weekly RHZ (Group 2), and were lower in mice receiving M2 than in mice receiving M1. However, in all of these groups, the cfu counts were higher than in control mice treated with daily RHZ. In mice of Group 5 that received twice-weekly P10HZ after an initial 2 wk of daily RHZ, the cfu counts were also higher than in Group 1 control mice treated with daily RHZ for 8 wk, but were lower than in Group 2 control mice that received twice-weekly RHZ. In Group 6 mice that received twice-weekly P15HZ, and especially in Group 7 mice that received twice-weekly P20HZ, the cfu counts were lower than in control mice treated with daily RHZ. These results provide support for the dose-dependent activity of P, and are in agreement with what has previously been observed with R (22). Finally, the lung cfu counts were lowest in mice that received M2 instead of H and twice-weekly P15 (Group 8; 0.84 log10) or P20 (Group 9; 0.35 log10) instead of R (p < 0.01 vs. daily RHZ). The substitution of M for H when administered with either P15 or P20 (Group 6 vs. 8 and Group 7 vs. 9, respectively) clearly demonstrates the added benefit in activity obtained by substituting M for H.
Mice in control Group 1 (daily treatment throughout), Test Group 6 (twice-weekly P15H during the continuation phase) and Test Group 8 (twice-weekly P15M2 during the continuation phase) received treatment for a total of 4, 5, or 6 mo. At the 4-mo time point, 1 of the 5 mice killed from Group 1 harbored a single cfu in the lung, whereas all mice in Groups 6 and 8 were culture negative (Table 3). In Group 1, the proportions of mice that relapsed 3 mo after completing 4, 5, and 6 mo of therapy were 9 of 12 (75%), 2 of 12 (17%), and 0 of 12 (0%), respectively, whereas in Groups 6 and 8, the proportion was uniformly 0 of 12 (0%) at each time point.
Treatment Efficacy in Study 2 The day after infection (Day –11), the mean cfu count in the lungs was 3.39 ± 0.16 log10. On Day 0, the mean cfu count had increased to 6.56 ± 0.08 log10. No difference in cfu counts was detected between mice infected by Runs 1 through 3. Infected, untreated control mice died 3–4 wk after infection. After 2 wk of daily treatment with RHZ (Group 1), RM2Z (Group 2), and RM2ZH (Group 8), the lung cfu counts were 5.65 ± 0.25, 5.42 ± 0.48, and 5.38 ± 0.40, respectively (p = 0.47). At Weeks 4 and 8 of therapy (Table 4), all drug regimens displayed potent bactericidal activity. At the 8-wk time point, mice treated with daily RM2Z had approximately 1 log10 fewer organisms in the lungs compared with mice treated with daily RHZ; all twice-weekly P15-containing regimens (Groups 3–12) were more active than the daily RHZ regimen. In mice initially treated for 2 wk with daily RHZ, the cfu counts were 1.67 ± 0.13 log10 when the subsequent twice-weekly phase contained P15 alone, and were substantially lower (0.53 ± 0.22 log10 cfu) when Z was added to P15 (Group 4). However, the addition of H, M, or MH to P15Z did not result in an additional reduction in the cfu counts.
The potent bactericidal activity of 2 wk of daily RM2Z plus 6 wk of twice-weekly P15M2Z was decreased when H was administered concomitantly during the first 2 wk of therapy, and further decreased when H was administered for the entire 8-wk period (p = 0.001; Figure 2). These findings suggest that administration of H may negatively impact the efficacy of PMZ-containing therapy in an exposure-dependent manner.
The three regimens based on twice-weekly P15M2Z that did not contain H (Groups 10–12) were all significantly more active than the daily RM2Z regimen (p < 0.01), and not significantly different from each other after 2 mo of therapy, irrespective of whether the twice-weekly P15M2Z component was preceded by 2 wk of daily RM2Z, 1 wk of daily RM2Z, or no daily therapy at all. These results imply that twice-weekly P15 is at least as active as daily R, even during the first 2 wk of treatment.
Pharmacokinetics and Pharmacodynamics of Rifampin and Rifapentine
After a single oral dose, the total drug mean peak serum concentration (Cmax) values were 12.1 ± 1.3 µg/ml for 10 mg/kg of rifampin and 19.2 ± 1.0 µg/ml for 15 mg/kg of P. The time to peak serum concentration (Tmax) values were 2 and 2.7 h for R and P, respectively. The serum half-life (t1/2) of P was, on average, five times longer than that of R. Accordingly, the mean area under the serum concentration–time curve (AUC0–24 h) for total drug was 116 ± 1.0 µg h/ml for 10 mg/kg of R, whereas the mean AUC0–72 h was 637 ± 40 µg h/ml for 15 mg/kg of P. When adjusted for protein binding to reflect estimated free-drug concentrations (97.5% for P and 82.5% for R) (11, 19), these AUC values decreased to 20.3 ± 3.6 µg h/ml and 15.9 ± 0.40 µg h/ml, respectively. At the end of 1 wk (i.e., five daily doses of R and 2 doses of P administered 3 d apart), no significant differences were detected in the Cmax, Tmax, t1/2, or AUC as compared with single-dose values. At the end of 5 wk, however, the serum drug concentrations were significantly decreased as compared with values obtained after a single dose. Although there was no change in the Tmax and Cmax, there were 72 and 44% reductions in the t1/2 for R and P, respectively, and a 25% reduction in the AUC for both drugs (Table 5). The primary pharmacodynamic indices obtained on Weeks 1 and 5 with daily R and twice-weekly P are presented in Table 6. During Week 1, total drug Cmax:MIC ratio for twice-weekly P at 15 mg/kg (322) was approximately seven times higher than that for daily R at 10 mg/kg (48), but after adjusting for protein binding, the estimated free-drug Cmax:MIC ratios were quite similar for P (8.03) and R (8.50). On the other hand, the estimated free-drug weekly AUC:MIC ratios were approximately 30% higher for twice-weekly P (531) compared with daily R (405). Finally, although both total and estimated free-drug concentrations remained above the MIC throughout the entire week (168 h) in mice treated with twice-weekly P, the time above MIC per week (T > MIC) in mice receiving daily R was approximately 144 h for total drug, and fell to 80 h for free-drug concentrations.
During the fifth week of treatment, the free-drug Cmax:MIC ratio remained unchanged, whereas the free-drug AUC/MIC ratio fell by 25% for both regimens. Likewise, the weekly T > MIC was reduced from 168 to 108 h for P and from 80 to 60 h for R when free drug was considered.
The current series of experiments yielded three important findings. First, the substitution of rifapentine for rifampin in the twice-weekly rifampin, isoniazid, and pyrazinamide regimen significantly improves the anti-TB activity, especially when the dose of rifapentine is increased from 10 to 15 and from 15 to 20 mg/kg. Second, the substitution of moxifloxacin for isoniazid in the twice-weekly regimen is also beneficial. Finally, there is autoinduction of rifampin and rifapentine metabolism in the mouse that results in reduced rifamycin exposures after the first month of therapy. Each of these findings may have important clinical applications, because intermittently administered rifampin-containing regimens have lower intrinsic activity than the same regimens administered daily in mice (9) and in humans. A recent nested case-control study in Hong Kong indicated that thrice-weekly treatment with a rifampin-containing regimen was associated with a nearly fourfold increased risk of relapse compared with daily therapy (7). In another study conducted in New York City, patients infected with HIV and treated with twice- or thrice-weekly rifampin-containing regimens were 6.7 times more likely to relapse and 6.4 times more likely to develop acquired rifampin resistance when intermittent therapy was commenced during the initial phase of therapy rather than during the continuation phase (8). Thus, despite their obvious advantages for the directly observed therapy of TB, intermittent rifampin-containing regimens may be less effective than their daily counterparts. One way to improve the activity of twice-weekly therapy containing rifampin, isoniazid, and pyrazinamide is to increase the activity of the rifamycin component by replacing rifampin with rifapentine. Rifapentine offers significant advantages over rifampin, including a lower MIC and a longer serum half-life (25, 26). Moreover, rifapentine is well tolerated in humans when doses of 15 mg/kg, and even 20 mg/kg, are administered once weekly (27). In contrast, there is significant risk of serious adverse events, such as the flulike syndrome when rifampin is administered intermittently at doses above 10 mg/kg (28). When we assessed the impact of substituting rifapentine for rifampin during the initial 2 mo of treatment in the murine model, we found that substitution of rifapentine at 10 mg/kg twice-weekly resulted in bactericidal activity that was intermediate between that obtained with daily and that obtained with twice-weekly rifampin-based therapy. Increasing the rifapentine dose to 15 or 20 mg/kg, however, resulted in improved bactericidal activity that was equivalent to or better than, respectively, that of standard daily rifampin-based therapy. The benefit of substituting rifapentine for rifampin was still more apparent during the continuation phase of treatment, when the sterilizing activity was assessed using culture conversion and relapse as outcomes. The predominantly twice-weekly regimen, including rifapentine 15 mg/kg, isoniazid, and pyrazinamide, resulted in complete culture conversion and no relapse (stable cure) in all mice after just 4 mo of therapy, whereas the standard daily rifampin-based regimen required 6 mo of therapy to reach the same endpoint. To better understand the reasons for the remarkable potency of twice-weekly rifapentine-containing regimens, we examined the single-dose and multiple-dose pharmacokinetics of rifapentine (15 mg/kg) and rifampin (10 mg/kg) administered twice-weekly and daily, respectively. The single-dose pharmacokinetics corroborated previous single-dose mouse studies, and are consistent with the single-dose human pharmacokinetics of rifampin and rifapentine (9, 12, 26, 29, 30). At the end of 5 wk of treatment, the weekly AUC was reduced by 25% for both rifampin and rifapentine, with a concomitant reduction in half-life. This finding strongly suggests that both rifampin and rifapentine undergo significant autoinduction of metabolism in the mouse. The reduction in rifampin exposure was consistent with the effects of autoinduction observed in humans (31–33). However, it is unclear whether rifapentine undergoes a similar increase in elimination in humans (11, 30, 34). Further clinical studies are needed to quantify the degree of autoinduction after chronic rifapentine administration. The pharmacodynamic parameters of rifampin and rifapentine provide a ready explanation for the superior activity of twice-weekly rifapentine-containing regimens over that of standard daily therapy. Overall, the rifamycin exposure, as defined by the weekly free-drug AUC:MIC or T > MIC, was substantially higher after administration of twice-weekly rifapentine (15 mg/kg) as compared with daily rifampin (10 mg/kg). In other words, the tubercle bacilli infecting mice treated with twice-weekly rifapentine were exposed to more free drug for longer periods of time than those in mice treated with daily rifampin, and this difference was reflected in the enhanced sterilizing activity of the twice-weekly rifapentine-containing regimens. It is tempting to speculate that the reduced rifampin exposure caused by autoinduction after the first few weeks of treatment significantly diminishes the activity of rifampin, as there is evidence in humans and in murine models that doses lower than 600 mg (10 mg/kg) are less effective (22, 35, 36). The greater drug exposures obtained with twice-weekly rifapentine-containing regimens may limit the reduction in activity that occurs with autoinduction. Unfortunately, the pharmacodynamic parameter that best predicts the sterilizing activity of rifamycins, and the threshold value of that parameter associated with optimal activity, remain unknown. Existing data suggest that twice-weekly rifapentine-based regimens would be safe and well tolerated in humans. In our first mouse study, mice that were treated with twice-weekly rifapentine at 10, 15, and 20 mg/kg experienced no detectable adverse events nor any differences in body weight compared with control animals. The human safety and tolerability of rifapentine at 10, 15, and 20 mg/kg administered once weekly during the continuation phase of therapy was assessed in Tuberculosis Trials Consortium Study 25 (27). The authors concluded that drug-associated adverse effects were possibly associated with increases in the dose of rifapentine (p = 0.051), but that doses of 10 and 15 mg/kg of rifapentine were safe and well tolerated. When the authors conducted a secondary analysis after removing a first-trimester spontaneous abortion that was "possibly" associated with the 20 mg/kg dose, there was no association between dose and drug-associated adverse events (p = 0.14). Furthermore, a second study conducted in South Africa demonstrated that two doses of 15 mg/kg of rifapentine administered 4 d apart were safe and well tolerated in patients with pulmonary TB (34). Finally, in a third clinical study, rifapentine (10 mg/kg) was administered twice weekly in combination with daily isoniazid and pyrazinamide plus ethambutol during the initial 2 mo phase of therapy (11). No differences in treatment-related adverse events were noted, except for a 4% increase in hyperuricemia in the twice-weekly rifapentine arm over that of the daily rifampin-containing control arm. Although the recent clinical experience with high-dose rifapentine and twice-weekly dosing is encouraging, careful investigation will be required before twice-weekly rifapentine (15 or 20 mg/kg)-based regimens can be considered safe and tolerable. Based on our prior experience in the mouse model (20, 21, 37), we also hypothesized that substitution of the fluoroquinolone, moxifloxacin, for isoniazid might further increase the activity of the most potent twice-weekly rifapentine-based regimens. Such increased activity was clearly evident during the initial 2 mo of treatment, during which the substitution of moxifloxacin for isoniazid in rifapentine-containing regimens resulted in an additional reduction in lung cfu counts of approximately 1.5 log10, similar to the impact of substituting moxifloxacin for isoniazid in the standard daily regimen of rifampin, isoniazid, and pyrazinamide (20). The impact of substituting moxifloxacin for isoniazid on sterilizing activity was difficult to define, due to the unanticipated potency of the rifapentine-containing regimens. All mice treated with either the twice-weekly rifapentine (15 mg/kg) plus isoniazid-based regimen or the rifapentine (15 mg/kg) plus moxifloxacin-based regimen achieved stable cure after just 4 mo, as compared with 6 mo with standard daily rifampin plus isoniazid-based therapy. Whether the activity of these two regimens would have been different after shorter durations of therapy (e.g., 3 mo) is currently under investigation, but in one prior report, daily administration of rifapentine (20 mg/kg), isoniazid, and pyrazinamide for 3 mo failed to achieve stable cure in all mice (38). Prior observations in the murine model have demonstrated that isoniazid antagonizes the activity of combinations containing rifampin and pyrazinamide, with or without a fluoroquinolone (20, 37, 39). In our second experiment described herein, we observed that isoniazid also antagonizes the activity of the predominantly twice-weekly rifapentine, moxifloxacin, and pyrazinamide regimen in an exposure-dependent fashion during the initial 2 mo of treatment. The longer isoniazid was administered together with this regimen, the greater the apparent antagonism (Figure 2). Whether such antagonism occurs in the treatment of human TB is debated, but there are currently no clinical data available to address the issue. Phase II studies to evaluate the benefit of substituting moxifloxacin for isoniazid in the standard daily regimen are now underway. Beyond the potential benefit of avoiding isoniazid-related antagonism, the substitution of moxifloxacin for isoniazid in twice-weekly rifamycin regimens also has the theoretical advantage of reducing the risk of acquired rifamycin resistance. Selection of rifamycin resistance has been observed after therapy with twice-weekly combinations of rifampin or rifabutin with isoniazid, and with once-weekly therapy with rifapentine and isoniazid, chiefly among HIV-infected patients with CD4 counts below 100 cells/mm3 (13, 40). Because the half-life of moxifloxacin is substantially longer than that of isoniazid, and similar to that of rifapentine, and because twice-weekly rifapentine is substantially more potent than twice-weekly rifampin, the twice-weekly administration of rifapentine and moxifloxacin should be more effective in preventing the selection of rifamycin-resistant mutants compared with twice-weekly regimens combining a rifamycin with isoniazid. This hypothesis has not been investigated in the mouse, however, primarily because the clearances of isoniazid and, especially, moxifloxacin are greatly increased in the mouse compared with humans, making it difficult to precisely model the pharmacokinetics observed in humans with intermittent dosing. In conclusion, this study reports on the development of two highly efficacious twice-weekly rifapentine (15 mg/kg)-containing regimens capable of producing stable cure in the murine model of TB after just 4 mo of administration, rather than the 6 mo required to reach the same endpoint with the standard daily rifampin-based regimen. Whether twice-weekly regimens containing rifapentine (15 or 20 mg/kg) and moxifloxacin can produce the same results after even shorter durations of therapy is currently under investigation. However, even a 4-mo regimen that remains highly efficacious, despite twice-weekly administration, would significantly advance the treatment of TB. Thus, although there is justifiable excitement over several new anti-TB drugs in development, more rational use of existing drugs remains a promising strategy for improving the treatment of TB.
Supported by the National Institute of Allergy and Infectious Diseases TB Drug Development (contract N01-AI-40007), National Institutes of Health (grants AI58993 and EB005094, and supplement to grant AI43846), and the Potts Memorial Foundation. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200602-280OC on March 30, 2006 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form February 25, 2006; accepted in final form March 30, 2006
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
