Pharmacokinetic interactions between efavirenz and rifampicin in the treatment of HIV and tuberculosis: one size does not fit all

The concomitant treatment of HIV-tuberculosis co-infection is complicated by pharmacological interactions between drugs, resulting in unpredictable drug levels. We monitored efavirenz levels in all tuberculosis-HIV-treated patients over 2 years. Using 800 mg/day of efavirenz, high levels and toxicity were detected in seven out of nine patients, necessitating reduction or discontinuation. Polymorphisms in cytochrome P450 2B6 may account for this. Therapeutic drug monitoring, dose reduction or a lower starting dose may be appropriate in some patients to abrogate toxicity.

HIV-tuberculosis co-infection is a growing health burden ^[1,2], and although highly effective therapies exist for both HIV and tuberculosis, treatment is complicated by similar routes of drug metabolism, which may result in suboptimal plasma levels of either class of drug, with the risk of microbiological or virological failure and the possible emergence of drug resistance.

Current guidelines ^[3,4] recommend the use of rifampicin as part of any robust antituberculous drug regimen. Rifampicin is, however, promiscuous in its interactions with many drugs via the cytochrome P450 pathway, particularly protease inhibitors, and therefore many physicians choose to use efavirenz as the basis of an antiretroviral regime in patients receiving concurrent rifampicin-based regimes. However, recent evidence ^[5] has suggested that the concomitant use of rifampicin with efavirenz decreases the concentrations of efavirenz. Current practice is therefore to increase the dose of efavirenz from 600 to 800 mg a day in order to achieve therapeutic levels because low levels predict increased rates of treatment failure ^[6]. High levels of efavirenz are, however, associated with the increased severity and frequency of side-effects, particularly those affecting the central nervous system (CNS) ^[6].

We monitored clinical toxicity and efavirenz levels in HIV-tuberculosis-co-infected patients receiving concomitant highly active antiretroviral therapy (HAART) and antituberculous therapy over a 2-year period at this centre. During this time, 30 patients with tuberculosis-HIV co-infection were identified. Twenty-one did not receive antiretroviral treatment until the completion of antituberculous therapy. Nine patients received concomitant HAART and antituberculous therapy, consisting of standard quadruple antituberculous therapy (rifampicin, isoniazid, pyrazinamide and ethambutol) and an efavirenz-based antiretroviral regimen. Of these nine patients, seven (78%) developed significant clinical toxicity. Clinical features and efavirenz levels are summarized in Table 1.

Table 1

It can be seen that in all seven patients developing clinical toxicity, efavirenz trough levels were elevated significantly beyond the therapeutic range (median 11 680 ng/dl; 5373-19 591; therapeutic range 1200-4000 ng/dl). Six patients developed significant CNS side-effects ranging from anxiety to depression, and one patient developed hepatitis. It is likely that these side-effects were related to the high circulating levels of efavirenz, although no control group receiving efavirenz at the same dosage (800 mg a day) without antituberculous drugs was available for comparison, as such patients would routinely be started on the usual, lower dose of 600 mg a day. In one case, tuberculosis meningitis may also have contributed to the CNS symptoms reported. All but one of those developing toxicity (no. 6, who developed hepatitis) were of black ethnicity. This reflects the ethnic composition of co-infected patients at this centre, 28 out of 30 (93%) of whom were of black African origin. Three patients discontinued efavirenz; four tolerated lower doses (400-600 mg/day). All seven patients achieved a viral load of less than 50 copies/ml.

These data show that the use of efavirenz at a high dose cannot be universally endorsed in patients receiving concomitant antituberculous therapy. Therapeutic drug monitoring should be undertaken in this group of patients to obviate the development of toxicity. Some patients may benefit from a dose reduction in efavirenz or from starting with a dose of 600 mg.

Our findings add support to the previously reported variability in pharmacokinetics, therapeutic response and side-effects in HIV-infected patients from distinct ethnic origins ^[7]. All patients developing toxicity, except one, were of black African ethnicity. It has been suggested that single nucleotide polymorphisms in the cytochrome P450 2B6 genes appear to modulate efavirenz CNS toxicity: patients with a TT genotype, which is observed in 20% of the black population (cf. 3% of white individuals) have an extended clearance of efavirenz ^[8]. This may account for the increased levels of efavirenz and concomitant CNS toxicity observed in our patients.