Table.
GRADE evaluation of interventions for tuberculosis
| Important outcomes | Develpment of active tuberculosis, cure rates, relapse, treatment compliance rates, adverse effects | ||||||||
| Number of studies (participants) | Outcome | Comparison | Type of evidence | Quality | Consistency | Directness | Effect size | GRADE | Comment |
| What are the effects of interventions to prevent tuberculosis in people without HIV infection at high risk of developing tuberculosis? | |||||||||
| 11 (73,375) | Development of active TB | Isoniazid v placebo | 4 | 0 | 0 | 0 | 0 | High | |
| 2 (199) | Development of active TB | Isoniazid v no isoniazid in people undergoing renal transplant | 4 | –3 | 0 | 0 | 0 | Very low | Quality points deducted for sparse data and methodological flaws (uncertainty about allocation concealment, unclear randomisation) |
| 1 (20,874) | Adverse effects | Isoniazid v placebo | 4 | 0 | 0 | 0 | +1 | High | Effect-size point added for odds ratio greater than 5 |
| 1 (630) | Development of active TB | Rifampicin plus isoniazid v isoniazid alone | 4 | –2 | 0 | –1 | 0 | Very low | Quality points deducted for quasi-randomised RCT, and baseline differences. Directness point deducted for unclear population (HIV status) |
| What are the effects of different drug regimens in people with newly diagnosed pulmonary tuberculosis without HIV infection? | |||||||||
| 2 (1295) | Relapse rates | Shorter (6 months) v longer regimens (8–9 months) | 4 | –1 | 0 | 0 | 0 | Moderate | Quality point deducted for incomplete reporting of results |
| 1 (444) | Cure rates | Pyrazinamide-containing regimens v regimens not containing pyrazinamide | 4 | –1 | 0 | 0 | 0 | Moderate | Quality point deducted for incomplete reporting of results |
| 2 (1330) | Relapse rates | Pyrazinamide-containing regimens v regimens not containing pyrazinamide | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for incomplete reporting of results. Directness point deducted for inconsistent results at different lengths of follow-up |
| 1 (851) | Relapse rates | Rifampicin plus isoniazid v isoniazid | 4 | –1 | 0 | 0 | 0 | Moderate | Quality point deducted for incomplete reporting of results |
| 2 (605) | Cure rates | Intermittent chemotherapy for 6 months or longer v daily therapy | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for incomplete reporting of results. Directness point deducted for inclusion of different disease states |
| 1 (378) | Relapse rates | Intermittent chemotherapy for 6 months or longer v daily therapy | 4 | 0 | 0 | 0 | –2 | Low | Directness points deducted for short follow-up and small number of events (6 in total), suggesting it may have been too small to detect a clinically important difference |
| 4 (489) | Cure rates | Ciprofloxacin-substituted regimens v standard regimens | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for weak methods. Directness point deducted for inclusion of HIV-positive people in analysis |
| 3 (412) | Cure rates | Quinolone-substituted regimens v standard regimens | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for weak methods. Directness point deducted for inclusion of HIV-positive people in analysis |
| 3 (384) | Relapse rates | Quinolone-substituted regimens v standard regimens | 4 | –2 | 0 | –2 | +2 | Low | Quality points deducted for weak methods and unclear treatment duration. Directness points deducted for inclusion of HIV-positive people in analysis and small number of events (11 in total). Effect-size points added for RR above 5 |
| 1 (144) | Cure rates | Levofoxacin plus standard-drug regimen v standard-drug regimen alone | 4 | –2 | 0 | 0 | 0 | Low | Quality points deducted for sparse data and weak methods |
| 1 (144) | Cure rates | Levofoxacin plus standard-drug regimen v ofloxacin plus standard-drug regimen | 4 | –2 | 0 | 0 | 0 | Low | Quality points deducted for sparse data and weak methods |
| 1 (1355) | Cure rates | Ethambutol-containing regimens v rifampicin-containing regimens | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for incomplete reporting of results. Directness point deducted for composite outcome |
| 5 (2588) | Relapse rates | Short-course chemotherapy regimens (less than 6 months) v longer-course regimens | 4 | 0 | 0 | –1 | 0 | Moderate | Directness point deducted as treatments given under optimal conditions affecting generalisability (adherence likely to be lower in clinical practice) |
| What are the effects of different drug regimens in people with multidrug-resistant tuberculosis without HIV infection? | |||||||||
| 2 (unclear, at least 184 people) | Cure rates | Sparfloxacin plus standard-drug regimens v ofloxacin plus standard-drug regimens | 4 | –2 | 0 | 0 | 0 | Low | Quality points deducted for weak methods (unclear allocation concealment, unclear blinding) |
| 1 (134) | Cure rates | Adding high-dose isoniazid v adding normal-dose isoniazid or placebo to second-line therapy | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for sparse data, and lack of baseline data on sputum positivity. Directness point deducted for combined comparison group (normal-dose isoniazid and placebo groups combined in analysis) |
| 1 (134) | Adverse effects | Adding high-dose isoniazid v adding normal-dose isoniazid or placebo to second-line therapy | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for sparse data. Directness point deducted for combined comparison group (normal-dose isoniazid and placebo groups combined in analysis) |
| Which interventions improve adherence to treatment in people with tuberculosis without HIV infection? | |||||||||
| 4 (1603) | Cure rates | Direct observation treatment v self-administered treatment | 4 | 0 | 0 | –1 | 0 | Moderate | Directness point deducted for heterogeneity among RCTs |
| 1 (163) | Treatment compliance rates | Participant-chosen site v designated site | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for sparse data. Directness point deducted for restricted population (drug users) |
| 2 (1109) | Cure rates | Clinic-based support v family-member support | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for cluster randomised trial. Directness point deducted for composite outcome in 1 RCT |
| 2 (2233) | Cure rates | Community-based health-worker support v family-member-based support | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for cluster randomised trial. Directness point deducted for composite outcome (cure or completion of treatment) |
| 2 (1628) | Cure rates | Complex support interventions v usual treatment | 4 | –1 | 0 | –1 | 0 | Low | Quality point deducted for cluster randomised trial. Directness point deducted for composite outcome |
Type of evidence: 4 = RCT. Consistency: similarity of results across studies. Directness: generalisability of population or outcomes. Effect size: based on relative risk or odds ratio.