Studies regarding impact of dosing intermittency on treatment efficacy in HIV-negative tuberculosis
| Authors, year of publication | Measures of treatment efficacy | Source of study materials | Study design | Main findings† | Strengths | Weaknesses | Risk of bias |
| Chang et al, 200635 | Relapse | Multiple countries | Systematic review and meta-analysis* | A logistic risk model showed a significant dose-response relationship between dosing schedules and relapse, with the following odds (95% CI) of relapse relative to daily regimens: 1.6 (0.6 to 4.1) for daily initial phase (IP) + thrice-weekly continuation phase (CP), 2.8 (1.3 to 6.1) for daily IP + twice-weekly CP, 2.8 (1.4 to 5.7) for thrice-weekly, 5.0 (2.4 to 10.5) for daily IP + once-weekly rifapentine, and 7.1 (3.3 to 15.3) for thrice-weekly IP + once-weekly rifapentine. In the presence of cavitation, only 6 months daily or daily IP + thrice-weekly CP attained best estimated relapse risks below 5%. | The study clearly addressed a focused question with good coverage on methodology. Literature search was sufficiently rigorous. Studies were sufficiently similar to make combining them reasonable. Sensitivity analysis gave consistent findings. | Study quality was not addressed. A deterministic mathematical model was used for estimating treatment outcome in subgroups. | Low |
| Gao et al, 200936 | Relapse | USA, Canada, Hong Kong, China | Systematic review and meta-analysis* | Once- or twice-weekly rifapentine and daily rifampicin have similar efficacy for treating HIV-negative pulmonary TB. Once-weekly or less of rifapentine increases the risk of bacteriological relapse in comparison with twice- or thrice-weekly rifampicin (pooled relative risks were 2.44 (95% CI 1.15 to 5.18) and 1.71 (95% CI 1.13 to 2.58), respectively. | The study clearly addressed a focused question with good coverage on methodology and assessment of study quality. Literature search was sufficiently rigorous. | There was no evidence of heterogeneity by the chi square test of heterogeneity but degree of freedom was relatively small. | Low |
| Mwandumba et al, 2000, 200137 38 | Failure, relapse | Hong Kong | Systematic review* | With only one randomised controlled trial included, there were insufficient data for comparing treatment efficacy of fully intermittent and daily regimens. | A conventional approach was used. | Data were insufficient for drawing a conclusion. | NA |
| Cox et al, 200839 | Recurrence | Multiple countries | Systematic review* | Although several clinical trials may support the use of daily treatment regimens, studies reporting TB recurrence after intermittent regimens have been limited. Few studies under routine programmatic conditions have reported TB recurrence. Considerable heterogeneity across studies precluded attribution of TB recurrence to dosing schedules. | A conventional approach was used. | Data were insufficient for drawing a conclusion. | NA |
| Ziganshina et al, 200940 | Cure, relapse | Multiple countries | Systematic review* | It is uncertain whether intermittent treatment, taken 2 or 3 times a week, is more efficacious than daily treatment in enhancing cure rates or decreasing relapse rates of newly diagnosed TB. | A conventional approach was used. | Data were insufficient for drawing a conclusion. | NA |
| Menzies et al, 200941 | Failure, relapse | Multiple countries | Systematic review and meta-analysis* | Treatment outcomes (failure or relapse) for all intermittent schedules evaluated were similar to daily schedules, but there is insufficient evidence to support administration of treatment twice weekly throughout therapy. | The study clearly addressed a focused question with good coverage on methodology and assessment of study quality. Literature search was sufficiently rigorous. | Main potential confounders such as initial cavitation and 2-month culture status were not considered. Also, the inclusion of daily regimens with rifampicin given for <6 months might have introduced bias. | High |
| Prasad et al, 200142 | Cure | India | Controlled clinical trial | Bacteriological response was comparable and favourable for 43 culture-proven TB patients allocated to 2HREZ/4HR (n=21) and 2HREZ3/4HR3 (n=22) regardless of initial drug resistance (82.3% vs 100% without resistance, and both 100% with resistance). | The study clearly addressed a focused question. | Sample size was too small for drawing a conclusion. | NA |
| Alvarez et al, 200943 | Drug resistance | Brazil | Retrospective cohort analysis | No significant differences were found between partially intermittent treatment with daily initial phase (daily for one month, then thrice weekly) and daily treatment throughout among 5138 patients in terms of drug resistance rates. | The study clearly addressed a focused question in a well-defined cohort. Outcomes were clearly defined, with reliable assessment of exposure. | Main potential confounders were not considered. Assessment of outcome was not blind. CIs were not provided. | High |
| Chang et al, 200434 | Relapse | Hong Kong | Nested case–control study | Thrice-weekly treatment increased the risk of relapse in comparison with daily treatment (OR 3.92, 95% CI 1.78 to 8.63). For culture-proven pulmonary TB without cavitation, the 30 month relapse rate for standard thrice-weekly regimen was 1.1% (95% CI 0.6% to 2.0%). The corresponding rates in the presence of cavitation were 3.3% (95% CI 1.9% to 5.5%) for standard daily regimen, and 7.8% (95% CI 4.0% to 14.6%) for standard thrice-weekly regimen. | The study clearly addressed a focused question with cases and controls taken from the same cohort. Cases were clearly defined and differentiated from controls that were randomly selected from the cohort. Exposure status was ascertained in a standard way. Main potential confounders were considered. CIs were provided. | Assessment of outcome was not blind. | Low |
| Paramasivan et al, 199344 | Cure, acquired drug resistance | India | Non-randomised controlled clinical trial | Comparing 2HRZ2+4HR2 versus 2HRZ+4HR2; culture-positive rates at the end of treatment were 19.6% and 8.1% overall, respectively, and 13.2% and 6.9% among initially drug-sensitive cases, respectively. The corresponding acquired drug resistance rates among initially drug-sensitive cases were 81.0% and 30.2%. All differences were statistically significant. Partially intermittent treatment with daily initial phase might be better than twice-weekly treatment throughout in terms of cure rate and acquired drug resistance rate. | The study clearly addressed a focused question. | Cases and controls might not be taken from comparable populations. Only 14.7% cases treated twice-weekly throughout and 41.2% treated daily in the initial phase were included in overall assessment, less for drug-sensitive cases. No comparison was made between participants and non-participants. Confounders were not considered. No CI was provided. | High |
| Singla et al, 200945 | Failure | India | Prospective case–control analysis | The number of treatment interruptions is an independent risk factor for treatment failure. | The study clearly addressed a focused question with cases and controls taken from comparable populations. Cases were clearly defined and differentiated from controls. Exposure status was ascertained in a standard way. Main potential confounders were considered. CIs were provided. | Acid-fast bacilli smear instead of culture was used to assess treatment outcome. Only 40% of those with positive smears at month 5 were culture-negative. It is uncertain whether factors associated with failure defined by smear are the same as those of failure defined by culture. | Low |
↵* Articles identified by the literature search and included in systematic reviews identified by the current review are shown in appendix 3 in the online supplement.
↵† A number before and a suffix in subscript after a treatment regimen stands for the number of months and frequency per week, respectively.
H, isoniazid; NA, not applicable; R, rifampicin; TB, tuberculosis; Z, pyrazinamide.