Treatment outcomes in patients who received rifampicin with nevirapine or efavirenz

Polly Clayden, HIV i-Base

In an oral presentation, Andrew Boulle from the University of Cape Town showed findings from a prospective cohort study of adults receiving rifampicin based TB treatment with either nevirapine or efavirenz containing ART.

The study was conducted in Khayelitsha, a township outside Cape Town with a 30% HIV antenatal prevalence and TB case finding above 1500/100,000. The aim was to look at outcomes, in patients initiating NNRTI-based ART, who were already receiving rifampicin containing TB treatment.

The cohort were treatment naive adults >/= 14 years of age, </=250 cells/mm3 enrolled between 2001 and June 2006. If they were receiving TB treatment there must have been 14 days overlap of TB and HIV treatment.

The analysis looked at four different treatment scenarios: people without TB starting either nevirapine or efavirenz and people receiving rifampicin-based TB treatment when starting nevirapine or efavirenz.

The investigators identified 4117 patients starting on ART during the study period. Of those, 2687 patients were without TB when they initiated HIV treatment; 1726 started on nevirapine and 961 efavirenz. The remaining 1283 started ART while receiving TB treatment, of these, 209 started with nevirapine and 1074 efavirenz.

The four treatment groups had similar baseline characteristics. There were however more women in the nevirapine vs efavirenz groups: 73.7% vs 62.1% (p<0.001) with TB treatment, and 80% vs 51.7% (p<0.001) without TB. Also in the TB treatment group, the patients receiving nevirapine had been receiving TB treatment for longer, a median of 87 days vs 73 days in the efavirenz group (<0.001). And the patients receiving nevirapine had higher median CD4 counts than those receiving efavirenz: 80 vs 61 cells/mm3 (p=0.002) in the TB treatment group and 116 vs 93 cells/mm3 (p<0.001) in the group without TB at baseline.

Viral load results were available for a subset of patients. The investigators defined a suboptimal response to HAART as failure to suppress viral load to less than 400 copies/mL. This occurred for approximately 15% at 6 months of the patients receiving nevirapine with TB treatment at 6 months vs between 6-9% for the patients receiving nevirapine (n=141) without TB treatment or efavirenz. The same analysis at 18 months showed failure to suppress in approximately 20% vs 8-13% in the nevirapine with TB treatment (n=80) and the other groups respectively.

The odds ratios for viral load >400 copies/mL at 6, 12, 18 months and all time points combined, for patients receiving TB treatment were 2.1, 1.6, 1.4 and 1.7 respectively for those receiving nevirapine and 1.2, 0.9, 1.1 and 1.1 for those receiving efavirenz.

In multivariate analysis combining all four groups the odds ratios (with efavirenz and no TB treatment as reference) were: efavirenz/TB treatment at start 1.1 (95% CI 0.7-1.7); nevirapine/no TB 1.5 (95% CI 1.0-2.1); nevirapine, TB treatment at start 2.9 (95% CI 1.8-4.7), p<0.001.

In this analysis age and gender were not significant but baseline CD4 (per 25 cell increase) 0.9 (95% CI 0.9-1.00), p<0.001; baseline weight (per 10kg increase) 1.2 (95% CI 1.1-1.3), p=0.001 and baseline viral load (per 1 log) 1.3 (95% CI 1.1 -1.6), p=0.001 were found to have an effect.

Additionally duration of ART (with 6 months as reference): 12 months 1.5 (95% CI 1.2-1.8) and 18 months 1.8 (95% CI 1.5-2.3), p<0.001 were significantly associated with viral load >400 copies/mL.

The study also looked at ART and TB treatment in patients who developed TB while receiving ART. The findings are limited by the small number of patients developing TB while on ART, but in this study there was no increased risk of viral load >400 copies/mL with TB treatment: Adjusted HR 1.00 (0.5-2.00) for nevirapine, p=0.995 and 1.2 (0.6-2.4) for efavirenz, p=0.703.

Dr Boulle summarised that receiving rifampicin-based TB treatment at the start of NVP-based ART was associated with an up to a two-fold increased risk of a viral load >400 copies/mL in the first 18 months on ART. There was no increased risk with incident TB on nevirapine-based ART, but this analysis was limited by small numbers. He said that the most likely explanation for this is lead-in dosing of nevirapine in patients with pre-existing hepatic induction due to rifampicin. He suggested to always include a 400mg/day induction arm in the nevirapine/rifampicin groups in future research.

He noted there were more nevirapine substitutions due to toxicity in patients already on TB treatment but this difference was not significant after adjustment for potential confounders.

Overall the analysis is limited by small numbers in the nevirapine/rifampicin group which he explained is “unlikely to change as local policy no longer recommends nevirapine use whilst on rifampicin.”

He described these findings as giving a “complex message” and recommended that if efavirenz is available a strong case can be made for using it for people receiving rifampicin based TB treatment in preference to nevirapine. “However, in spite of differences, 85% of patients on NVP/Rif at the start of ART achieved a viral load < 400 copies/ml at 6 months and 80% at 18 months – still highly effective in the absence of alternatives.”


Boulle A, Van Cutsem G, Cohen K et al. Antiretroviral treatment outcomes in patients who received rifampicin together with nevirapine or efavirenz. 38th World Conference on Lung Health of the Union against Tuberculosis and Lung disease. December 2007. Cape Town. Abstract TS 1893-12.

Links to other websites are current at date of posting but not maintained.