Treatment completion and costs of a randomized trial of rifampin for 4 months versus isoniazid for 9 months

Study population The study population comprised patients aged 18 years or older, with documented tuberculin skin tests that met the criteria for a positive test by Canadian standards (Long R, 2000, see 'Other Publications of Related Interest' below for bibliographic details), and with 9INH as the initial recommended treatment for LTBI. Contacts of INH-resistant cases and patients with hypersensitivity to RIF were excluded. Also excluded were those who were taking therapy with potential interactions with RIF, without any acceptable alternative being available, or who refused such alternatives.


Analysis of effectiveness
The effectiveness analysis was conducted on an intention to treat basis.The primary outcome was the percentage of prescribed doses taken, as measured by an electronic device in the cap of the pill container.The patients were considered to have completed therapy if they took more than 80% of the total prescribed doses within 20 weeks for 4RIF, or 43 weeks for 9INH.The secondary outcomes included adverse events resulting in permanent discontinuation of therapy.Drug-induced hepatitis was defined as liver transaminase levels more than three times the upper limits of normal with symptoms, or transaminase levels more than five times the upper limits of normal without symptoms.The characteristics of those patients who refused and of the patients randomised to the two treatment arms were not significantly different.The characteristics of the participants in the 4RIF and 9INH groups were similar.
The proportion of doses taken each month was generally higher with RIF, although it was significantly higher only in month 2 (96.8% versus 90.8%; p=0.03).
Major adverse events, which resulted in permanent discontinuation of therapy by their treating physician, occurred in 8 (14%) patients in the 9INH group and 2 (3%) patients in the 4RIF group (relative risk 0.25, 95% confidence interval: 0.1 -1.1).Three (5%) patients in the 9INH group suffered drug-induced hepatitis versus 0 patients in the 4RIF group.
Minor symptoms, which did not result in any change of therapy, were generally similar for the two regimens.The exception was a significantly greater occurrence of fatigue with RIF (15% versus 4%; p<0.001).

Clinical conclusions
Significantly more patients in the 4RIF group completed an adequate course of therapy, with less frequent major adverse events, compared with the 9INH group.

Measure of benefits used in the economic analysis
No summary measure of health benefits was used in the economic evaluation.The study was, in effect, a costconsequences analysis.

Direct costs
Discounting was not relevant given the time horizon of the analysis.The direct costs considered in the economic analysis seem to have been those of the hospital.They included clinical visits (routine and unscheduled visits), emergency room, consultations (gastrointestinal and dermatology), drug costs, pharmacist fees, additional chest X-ray, and laboratory tests (liver function tests, sequential multiple analyser-seven, and hepatitis serology).The costs for the electronic monitors to test the number of prescribed doses taken were not included.The unit costs and the quantities of resource used were reported separately.The institutional costs were based on actual costs and the Medical Chest Institute in 2003.Professional fees were gathered from the 2003 reimbursement schedule of the Regie de l'Assurance Maladie du Quebec.The medication costs were based on pills dispensed, pharmacist fees and global TB drug facility prices.The total costs and the mean costs per patient allocated and per patient completed were reported.

Validity of estimate of measure of effectiveness
The effectiveness evidence came from an open-label randomised controlled trial, which was appropriate for the study question.The main strengths of the study were that power calculations were performed, the methods used for sample selection and randomisation were given, and the baseline comparability of the groups was stated.In addition, the numbers of patients who refused or were excluded from the initial sample were clearly reported.However, the authors reported some limitations of their study.First, the trial did not have adequate power to assess safety and efficacy.Second, the absence of blinding might have introduced bias into the outcome assessment.

Validity of estimate of measure of benefit
Not relevant since a cost-consequences analysis was performed.

Validity of estimate of costs
It appears that all the costs relevant to the perspective adopted in the study have been included.The price year and the unit costs were reported, thereby facilitating reflation exercises in other settings.The unit costs and the quantities of resource used were reported separately, thus enhancing the reproducibility of the study in other settings.Resource use was taken from this single study and a statistical comparison of quantities between the two treatments was performed.The cost estimates would appear to be specific to the study setting.Since all the costs were incurred during less than two years, discounting was appropriately not performed.