Failure to suppress viral load more likely with nevirapine than efavirenz if ART started during TB treatment

Background on TB treatment and ART

In many resource-limited settings, a large percentage of people with HIV are on TB therapy when starting ART (up to 40% in some areas of southern Africa). For instance, in Khayelitsha, a resource-constrained township near Cape Town, the HIV seroprevalence is over 30% and the TB case finding rate is above 1500/100,000 — among the highest in the world.

First-line ART is usually based on one of the non-nucleoside reverse transcriptase inhibitors (NNRTIs), either efavirenz or nevirapine. In resource-limited settings, most people go onto a nevirapine-based regimen because inexpensive generic formulations are widely available, including convenient fixed dose combinations (though not in South Africa).

Also, nevirapine is known to be safe and effective for prevention of mother to child transmission — which is important where up to two-thirds of the people on treatment are women of child-bearing potential.

The preference for nevirapine is despite its rather complex dosing schedule: after a couple of weeks on the drug, it induces its own hepatic metabolism (i.e., the body eliminates the drug more efficiently). Thus, for the first two weeks on therapy, it usually must be dosed at 200mg per day before being increased to 200mg twice daily in order to maintain therapeutic levels.

However, in patients coinfected with TB, there have been persistent concerns about giving TB treatment together with either nevirapine or efavirenz-based ART because drug interaction studies found that rifampicin can lower the serum concentrations of efavirenz by up to 20-25%, and nevirapine levels by 20-55%. There are also concerns about additive toxicity of TB therapy and ART, especially hepatitis.

As a result, over the years, experts have made different recommendations about giving TB treatment and ART together. Some recommended putting off ART until after intensive TB treatment with rifampicin has been completed — but studies have since shown that delaying ART in patients with low CD4 cell counts on TB treatment leads to more deaths in Thailand and South Africa. Others suggested using significantly less potent ART regimens containing three nucleoside analogues — which could also put patients at risk of developing cross resistance to the entire class of nucleoside analogues that are so important for constructing viable ART regimens (Pillay). Some suggested using efavirenz only at a higher dose — but data have shown that this may not be safe, particularly in some people of African origin who are more likely to carry a genetic polymorphism that slows efavirenz clearance.

Yet few studies have carefully evaluated the long-term clinical consequences of these drug interactions (particularly for viral load suppression) in a large cohort of patients.

The study in Khayelitsha

The clinics in Khayelitsha offered the perfect opportunity to follow the effect of the interactions of TB and HIV treatment over time since they have been routinely monitoring CD4 and viral load every six months in over 4,000 people put onto ART since 2001 — and the township has a very burden of TB, meaning that many patients are already receiving TB treatment when they begin ART.

The analysis included adults or adolescents (≥ 14 years old) with less than 250 CD4 cells who started ART by June 2006, but excluded those on TB treatment who did not have at least 14 days of overlapping rifampicin and HIV treatment. Patients were followed as long as they took the NNRTI which they had started on, out to 18 months or until the interruption of ART, the patient’s death, loss to follow-up or transferring out of the clinics, or the start of treatment for a new episode of TB.

2,687 people without TB started ART: 1726 on nevirapine and 961 on efavirenz; 1,283 started ART while on rifampicin: 209 people on nevirapine and 1,074 on efavirenz. Viral load results are available for about nine out of ten of the subjects reaching each time point in each of the four groups being followed.

“We think the groups are broadly comparable,” said Dr Boulle. And yet, because this was an observational cohort, and people were enrolled at different points of the programme, the baseline characteristics were not perfectly matched. For instance, a significantly higher proportion of those put on nevirapine were women, and liver enzymes were lower. Among the patients with TB, those on nevirapine had had a longer median duration on TB treatment (p<0.001), though the CD4 cell counts were higher on nevirapine (median 80 cells on nevirapine, 61 cells on efavirenz); while other variables such as age, viral load (which was only available in a subset of patients; 5.3 log in each arm), age and weight were similar. There were similar slight (but statistically significant) differences between the baseline characteristics of those without TB who were put on efavirenz and nevirapine.

And of course, people on TB treatment had more advanced disease than those who didn’t have TB — with lower CD4 cells, higher viral loads and a higher rate of AIDS diagnoses.

No effect seen when rifampicin/TB treatment is started in patients already on nevirapine/ART

In this cohort, there was a risk of about 10% per year of developing TB while on ART. For those who remained on the same ART regimen, starting TB treatment did not seem to increase the risk of viral load over 400 copies/ml, with a hazard ratio of 1.0 (0.9-2.0) on nevirapine and 1.2 (0.6-2.4) on efavirenz, though this analysis may have been limited by the relatively low numbers of patients who developed a new case of TB.

However it does raise the question of whether the drug interaction between nevirapine and rifampicin is only significant when nevirapine is started in patients who are already on rifampicin and not the other way around.

According to Dr Boulle, the most likely explanation for this is that rifampicin switches on hepatic metabolism, which means patients on rifampicin who start nevirapine “could have sub-therapeutic nevirapine levels during the lead-in dosing period.”

Indeed, a group from Chelsea and Westminster Hospital in London has previously reported that when people are switched from efavirenz (which also induces the same hepatic metabolic pathway) to nevirapine, a lead-in period of nevirapine was unnecessary. “It was found that when changing from efavirenz to nevirapine individuals should commence on 200mg twice a day, as this dose is associated with therapeutic plasma drug levels,” wrote Winston et al.

Clinical studies would need to show whether this is safe in patients in TB patients however. A recent study reported somewhat similar results in patients on rifampicin/TB treatment for the standard nevirapine dosing (with lead-in) in Thailand compared to a higher lead-in dose. However, that study also increased the dose at two weeks to 600mg a day (which was associated with a higher rate of toxicity).

Further studies should evaluate starting and staying at the 200mg bid dose — which would have the added benefit of being simpler.

Results

After six months on ART, a lower percentage, around 85%, of those who started nevirapine while on TB treatment had viral loads suppressed below 400 copies/ml compared to around 91-94% on the other three arms (these figures are the reporter’s approximations taken from a bar graph). This trend continued with 80% suppressed on nevirapine and TB treatment at 12 and 18 months, and around 86-92% for the other three arms. There was no major difference in responses between patients on efavirenz with TB treatment or those receiving efavirenz without TB treatment.
When adjusted for age, sex and baseline CD4 cell count, the odds ratio for having a viral load ≥400 copies/ml for those who started on nevirapine while taking TB treatment was highest at six months (2.1) declining to 1.4 at 18 months (1.7 for the three time points combined); while for efavirenz and TB treatment, it was 1.2 at six months and 1.1 at 18 months (1.1 for the three time points combined).

A more complete multivariate model combining the four groups, and including baseline viral load and weight, looked at the odds ratio of having a viral load over 400 copies/ml (see table). A Kaplan-Meier time to failure estimate painted a similar picture, with the nevirapine/rifampicin group having a faster time to first viral load over 400 copies/ml.

Another Kaplan-Meier analysis of time to treatment-limiting toxicity suggested a worse performance for nevirapine while it was given with TB treatment (as soon as TB treatment stopped, patients were censored from this analysis), but in the multivariate analysis this difference was lost. In the efavirenz arms there was no difference. Of note, treatment was switched in more patients on nevirapine than efavirenz due to toxicity whether or not they were on TB treatment.

Although efavirenz better in TB patients, nevirapine is not a bad option

“The study has a rather complex message,” said Dr Boulle “if efavirenz is routinely available… a strong case can be made for using it [rather] than nevirapine in TB patients starting ART. However… nevirapine [with rifampicin] is still highly effective in the absence of alternatives.”

References

Boulle, A et al. Antiretroviral treatment outcomes in patients who received rifampicin together with nevirapine or efavirenz. 38th World Lung Health Conference, Cape Town, abstract TS-71893-12, 2007.

Winston A et al. Dose escalation or immediate full dose when switching from efavirenz to nevirapine-based highly active antiretroviral therapy in HIV-1-infected individuals? AIDS;18(3): 572-574, 2004.