INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Directly observed treatment, short-course (DOTS) is the strategy advocated by the World Health Organization for global control of tuberculosis (TB). This strategy emphasizes the direct observation of drug administration to patients by a designated health-care worker or another person as one of its crucial elements (1). Although DOTS has proven highly effective in program settings in both developed and developing countries, its implementation poses significant operational difficulties. If all the drug components, instead of being administered daily or three times weekly, could be administered no more than once weekly while maintaining the efficacy of the treatment, the implementation of DOTS would be far less operationally demanding.

In our previous mouse experiments, we have demonstrated that rifapentine (P) gave a significantly higher serum peak level (Cmax) and had a longer half-life (t_1/2) than did either rifampin (R) or rifabutin (5). The activity of R alone for prevention of TB was significantly diminished when the frequency of administration was reduced from six to three times weekly, whereas significant bactericidal activity was still observed in mice (5, 6) treated with P alone at a frequency of up to once fortnightly (1 in 14 d). Furthermore, the addition of isoniazid (H), given either once weekly (1 in 7 d) or 1 in 14 d, significantly enhanced the bactericidal activity of P, to the extent that the combination PH given once weekly (1 in 7 d) achieved the same level of activity as daily (6 in 7 d) treatment with the combination R plus pyrazinamide (Z) in both immunocompetent (normal) and nude mice (6). These results suggested that PH given once weekly may be the core component of fully intermittent regimens for both the prevention and treatment of TB (6).

The objectives of the current experiment were to compare, during the initial phase of treatment in normal mice infected with a larger inoculum of Mycobacterium tuberculosis: (1) the bactericidal activities of several once weekly P-containing regimens with that of the standard daily RHZ (6 in 7 d) regimen; and (2) the ability of once-weekly (1 in 7 d) P-containing combination regimens to prevent the selection of R-resistant mutants of M. tuberculosis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice

Female Swiss normal mice, 6 wk old, were purchased from the Janvier Breeding Center, le Genest Saint-Isle, France.

M. tuberculosis Strain

The virulent H37Rv strain of M. tuberculosis was grown on Löwenstein-Jensen (L-J) medium. Colonies were subcultured in Dubos broth (Diagnostics Pasteur, Paris, France) for 7 d at 37° C. The turbidity of the resulting suspension was adjusted with normal saline to match that of a standard suspension of M. bovis bacille Calmette- Guerin (BCG) (1 mg/ml), and was further diluted 5-fold for mouse inoculation. The minimum inhibitory concentrations (MICs) (µg/ml) for strain H37Rv were 0.25 for R, 0.06 for P, 0.1 for H, and 2.0 for streptomycin (S) on 7H11 medium, and 2.0 for ethambutol (E), and 10 for Z on L-J medium.

Infection of Mice

Three hundred twenty mice were infected intravenously with 0.5 ml of a bacterial suspension containing 1.3 × 10⁷ cfu of M. tuberculosis strain H37Rv.

Chemotherapy

Ten mice each were killed on Days 1 and 14, respectively, after infection for providing the baseline values of spleen weights and the number of cfu of M. tuberculosis in the animals' spleens and lungs. On Day 14, the remaining mice were allocated to an untreated control group and to nine treatment groups with 30 mice each, and treatments were begun on the same day. Each treatment group received one of the nine regimens for 8 wk (Table 1), to mimic the duration of the initial phase of chemotherapy for TB.

Dosages of the drugs were selected to provide areas under the concentration-time curve (AUC) in mice that were comparable with those achievable in humans (5). All drugs, except S, which was given by subcutaneous injection, were administered through an esophageal cannula (gavage). All surviving mice were killed on Day 70 after infection.

Assessment of Infection and Treatment

The severity of infection and the effectiveness of treatments were assessed from the survival rate, change in body weights, spleen weights, and mean numbers of cfu in the treated animals' spleens and lungs. The total cfu and the cfu of R-resistant mutants were enumerated, respectively, on plain L-J medium and L-J medium containing R at a concentration of 40 µg/ml (7). A bactericidal effect of the treatment was defined as a significant decrease in the mean number of cfu for the treated group from the pretreatment value (i.e., the values obtained from mice killed on Day 14).

Statistical Analysis

Multiple comparisons among pairs of group means were made with Bonferroni's method (8). Because there were 10 groups for comparison, the differences were considered significant at a level of 0.05/ [10(10  1)/2] = 0.0011 (8).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Survival Rate

As shown in Figure 1, untreated control mice began to die from 2 wk after infection; almost 80% of mice died by the end of Week 4, and the remaining mice died within 8 wk after infection. At Day 70, the survival rate for the untreated control group was significantly lower than for any of the treated groups (p < 0.01); the survival rates for mice treated with S-containing regimens (i.e., Groups F and H) were significantly lower than those of the remaining treated groups (p < 0.01), but did not differ significantly in Groups F versus H. There were 12 deaths in Group F and 13 in Group H; all occurred exclusively after the first weekly injection of S in Group F or after the first (three deaths) or second (10 deaths) daily injection of S in Group H. Because no further death was observed in the two groups despite continuation of treatment, and because mortality of mice treated with daily injection of S at the same dosage had never been a problem in our previous experiments, the deaths observed in the present experiment were most likely due to an accident caused by the injection of S rather than by the toxicity of S at the dosage selected.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Survival rates of mice during 70 d after intravenous infection with 1.3 × 107 cfu of M. tuberculosis H37Rv. When treatments began at Day 14, there were 30 mice in each group. The total duration of treatment was 8 wk. Group A was treated with R alone 6 in 7 d; B with P alone 1 in 7 d; C with RHZ 6 in 7 d; D with (P 1 in 7 d) + (HZ 6 in 7 d); E with PHZ 1 in 7 d; F with PHZS 1 in 7 d; G with PHZE 1 in 7 d; H with [2 wk (P 1 in 7 d + HZS 6 in 7 d) + 6 wk (PHZS) 1 in 7 d]; and I with [2 wk (P 1 in 7 d + HZE 6 in 7 d) + 6 wk (PHZE) 1 in 7 d]. Curve of Group C duplicates that of I from Days 14 to 21, and that of B from Days 28 to 70; curve of Group G duplicates that of D from Days 35 to 70; and curve of Group E duplicates that of Group I from Days 35 to 70.

Mean Body Weights

The mean body weight of mice began to decrease after Day 7, and steeply declined in untreated control mice between Days 14 and 21, in coincidence with a significant mortality. Among the few untreated mice that survived after Day 28, the mean body weight fluctuated, with wider individual variations. On the other hand, the mean body weights of treated mice began to increase when the treatments began, were significantly greater than the untreated control at Day 21 (p < 0.0001), and continued the same tendency until the end of the experiment (i.e., Day 70). The body weights did not differ significantly among the treated groups.

Spleen Weights

As shown in Figure 2, the mean spleen weight of the untreated control mice had increased significantly at Day 14 after infection over that of mice killed at Day 1. At the end of 8-wk of treatment, the mean spleen weights of Groups A and B, (i.e., mice that had been treated with R or P alone) did not differ significantly from the pretreatment values, but were significantly greater than those of the remaining treated groups (i.e., Groups C to I). These results indicate that the treatment effectively prevented the further development of splenomegaly caused by TB, and that the combination regimens were more effective than R or P monotherapy in reversing the splenomegaly.

View larger version (14K): [in this window] [in a new window]   Figure 2.   Mean spleen weights of mice surviving at 70 d after infection with H37Rv, except that the Control D1 and Control D14 bars represent, respectively, the mean spleen weight of mice killed on the next day and 14 d after infection. The error bars represent the standard deviations. For treatment regimens of Groups A to I see the legend to Figure 1.

Enumeration of cfu (log₁₀) in the Organs

M. tuberculosis was isolated from all the spleens and lungs of mice killed at Day 1 after infection. The mean number of cfu increased significantly between Days 1 and 14 (p < 0.0001), and reached 6.62 ± 0.16 log₁₀ per spleen and 7.58 ± 0.33 log₁₀ per lung. As shown in Figure 3, at Day 70 the mean number of cfu in all treated groups was reduced by 3 log₁₀ to 6 log₁₀ from the pretreatment values (p < 0.0001), indicating that all the regimens had various degrees of bactericidal effects.

View larger version (30K): [in this window] [in a new window]   Figure 3.   Enumerations of cfu in the spleens (A) and lungs (B) of mice. Mice were infected intravenously with 1.3 × 107 cfu of M. tuberculosis H37Rv; treatments began 14 d after infection and continued for 8 wk. The Control D1 and Control D14 bars represent, respectively, the mean numbers of cfu in mice killed on the next day and 14 d after infection. Each column represents the mean numbers of cfu for 17 to 30 mice. The error bars represent the standard deviations. For treatment regimens of Groups A to I see the legend to Figure 1.

Among mice treated with daily R alone (Group A) or once-weekly P alone (Group B), the reduction in cfu counts was different in the spleens and the lungs. In the spleens, the mean number of cfu was 2.23 ± 0.38 log₁₀ in Group A and 2.66 ± 1.08 log₁₀ in Group B, and the reductions, of about 4 log₁₀ from the pretreatment value, were not statistically different (p = 0.049). However, among the 28 surviving mice of Group B, one spleen was culture negative and three yielded R-resistant mutants. In two of these spleens the entire bacterial population consisted of R-resistant organisms, and resistant organisms represented 13% of the bacterial population in the third (Table 2). Among the 30 mice of Group A, all the spleens were culture positive, and from no spleen were resistant organisms isolated. In the lungs, the mean number of cfu was 3.04 ± 0.56 log₁₀ in mice treated daily with R alone, and 3.77 ± 0.22 log₁₀ in mice treated once weekly with P alone, indicating that the reduction of the cfu in the lungs was greater in the former group (p < 0.0001). Moreover, R-resistant mutants were isolated from 14 (46.7%) of 30 lungs in Group A and 17 (60.7%) of 28 lungs in Group B. In a majority of cases the proportion of resistant mutants was well above 1% of the bacterial population (Table 3).

Among mice treated with any of the combination regimens, not a single R-resistant organism was isolated from either the spleens or the lungs, suggesting that all the tested combination regimens, including the regimen of PHZ given once weekly, were able to prevent the selection of R-resistant mutants.

However, the bactericidal activities varied widely among the combination regimens. As shown in Figure 3, the smallest cfu counts (1.37 ± 0.59 log₁₀ in the lungs and 1.84 ± 0.62 log₁₀ in the spleens), or the greatest reductions from pretreatment values, were observed in mice treated with the daily standard regimen RHZ (Group C). These values were significantly smaller than the 2.16 ± 0.85 log₁₀ cfu count in the lungs and 3.00 ± 0.32 log₁₀ cfu count in the spleens of mice of Group D (p < 0.0001), in which daily R had been replaced by once-weekly P while H and Z were still given daily. The magnitude of the difference in cfu counts, about 1 log₁₀, between Groups C and D was similar to that between Groups A and B.

No significant difference was observed in the cfu counts for Groups D and E (PHZ 1 in 7 d), nor for Groups D versus E on one hand and Groups G versus I on the other hand, in which E was added as part of the combination.

Additionally, the inclusion of S in the regimen (Groups F and H) seemed to be beneficial. The mean numbers of cfu were slightly lower in Group F (PHZS 1 in 7 d) than in Groups D, E, G, or I, but the differences did not attain statistical significance. However, the cfu counts (1.45 ± 0.76 log₁₀ in the lungs and 2.34 ± 0.37 log₁₀ in the spleens) in Group H (6-wk PHZS 1 in 7 d preceded by 2-wk P 1 in 7 d plus HZS 6 in 7 d) were significantly smaller than those in Group E (p < 0.0001), and did not differ significantly from those in mice treated with RHZ 6 in 7 d (Group C) (p = 0.697).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As in normal mice infected with a smaller inoculum of M. tuberculosis (5, 6), the current experiment showed that 8-wk treatment with once-weekly P alone had very promising bactericidal activity, achieving about 4 log₁₀ killing of the bacilli in mice infected with a larger bacterial inoculum. It confirmed the assumption that once-weekly administration of P may be a key component of regimens for the initial phase of chemotherapy for TB. Furthermore, the invariably stronger bactericidal effects shown in the lungs of mice treated with the fully intermittent, once-weekly combination regimens, e.g., P combined with HZ [Group E], HZS [Group F] or HZE [Group G] as compared with R alone were very encouraging.

As is the case with other combination regimens, the fully intermittent, once-weekly regimens were able to prevent the selection of R-resistant mutants in the lungs of mice, whereas monotherapy with R or P selected resistant mutants in approximately 50% of the animals. That monotherapy selected R-resistant mutants in the lungs but not in the spleens was probably due to the greater bacterial population size in the lungs (7.58 ± 0.33 log₁₀ cfu) than in the spleens (6.62 ± 0.16 log₁₀ cfu) at the time the treatment was begun. Because the frequency of spontaneously occurring R-resistant mutants among M. tuberculosis is about 10⁷ (9, 10), the probability of finding resistant mutants in the lungs was much greater than the probability of finding them in the spleens.

P given alone once weekly was less bactericidal than R given alone daily. After 8 wk of treatment, the number of bacilli killed by the former regimen was about 1 log₁₀ less than that killed by the latter regimen. Likewise, except for regimen H, which included S given daily for 2 wk and then once weekly, the once-weekly P-containing combination regimens were approximately 1 log₁₀ less bactericidal than the combination of RHZ 6 in 7 d, chosen as the standard positive-control regimen for the experiment. However, the differences represent only a fraction of the overall 4 log₁₀ to 5 log₁₀ magnitude of killing effects in the groups.

The weaker activities of the P-containing combination regimens were unrelated to the frequency of administration of H and Z, because the bactericidal activity of the combination PHZ 1 in 7 d did not increase significantly when H and Z were given daily (Group E versus Group D). The difference in the bactericidal activity was probably due to the pharmacokinetic properties of P, whose AUC and t_1/2 are only three to four times greater than those of R (5, 11, 12). Therefore, the superior pharmacokinetic properties P may not be sufficient to compensate for the decreased bactericidal effect found when the administration of P is reduced from daily to once weekly.

A number of possible options would permit intermittent P-containing regimens to achieve a degree of activity comparable to those found with daily R-containing regimens. One option would be to give P twice-weekly. However, from an operational viewpoint, twice-weekly treatment with P-containing regimens would not represent a meaningful improvement over what can be achieved with twice-weekly administration of RH (13, 14).

Because P has very long t_1/2, another option would be to increase the weekly dosage of the drug (e.g., giving 20 mg/kg instead of 10 mg/kg). However, the finding of serious immunoallergic side effects in 35% to 57% of patients treated with R at a once-weekly dose of 1,200 to 1,800 mg (15) suggests that intermittent therapy with a higher dosage of P might be associated with unacceptable toxicity.

A third option would be to prolong the duration of treatment with the once-weekly, P-containing combination regimens. After 8 wk of treatment, the bactericidal activity of the once-weekly P-containing combination regimens was about one order of magnitude (out of four to five orders of magnitude of killing) smaller than that of the standard daily regimen. Thus, if the total duration of treatment were prolonged from 6 mo to 8 mo, the bactericidal effect of the once-weekly P-containing combination regimens might achieve the same level of activity as that of a 6-mo standard daily regimen. It is unlikely that such a prolongation of treatment would pose significant operational difficulties, because currently the duration of antituberculosis therapy in many developing countries is 8 mo (3).

An important observation is that the addition of S improved the bactericidal effect of PHZ 1 in 7 d and was more beneficial than the addition of E. Furthermore, the bactericidal effect of the PHZS 1 in 7 d was significantly strengthened when the regimen was preceded by an initial 2-wk phase of HZS given daily (Regimen H), such that the overall killing effect at 8 wk was similar to that of RHZ 6 in 7 d. In the clinical trials that showed the promising results of intermittent R-containing regimens, the treatment often began with 2 wk of daily therapy with either RHS (13) or RHZS (14). Considering the provision of DOTS, adding 2 wk of daily therapy is operationally more difficult than fully intermittent therapy, but is still far less operationally demanding than a daily or three-times-weekly regimen given throughout the entire period of treatment. Furthermore, throughout the world many TB patients are initially hospitalized, so that providing for daily DOTS during the first week or two of treatment is quite feasible. A short period of daily treatment may also facilitate patient education and improve adherence.

One might expect an objection to the use of S, which is increasingly being replaced by E in both developed and developing countries. Although S is a very active drug, it has to be given by injection, with the related constraints and risks. On the other hand, it has been suggested that patients are more likely to comply with DOTS if an injectable drug is included in the treatment regimen. Another drug with a promising postantibiotic effect (16), such as a fluoroquinolone, might be quite effective as a replacement for S, improving the activity of once-weekly PHZ. It might also be possible that with such a drug, the duration of treatment would be further shortened.

A final and most important issue, which has been addressed in part by this study, is the number of drugs to be combined with P for preventing the selection of R-resistant organisms in all cases, and the duration for which these drugs have to be given (e.g., during the initial phase only or throughout the entire duration of treatment). All of these issues will be addressed in future animal experiments designed to provide information about the optimal use of P as an important new drug for treating TB.