Protease inhibitor not superior to NNRTI in sub-Saharan Africa

Protease inhibitor-based treatment did not result in superior rates of viorologic suppression after two years when compared to NNRTI-based treatment when tested in a randomised study in the Democratic Republic of the Congo, researchers reported at the 19^th Conference on Retroviruses and Opportunistic Infections in Seattle last week.

Virologic outcomes by intent to treat analysis for patients in the Lubumbashi trial on a first line protease-inhibitor (PI)-based regimen were almost identical to those on a non-nucleoside reverse transcriptase inhibitor (NNRTI)-based regimen at 96 weeks (63% and 64% with viral load below 50 copies/ml respectively, p=0.4393) 

Presenting the results of this two-year prospective randomised trial Nathan Clumeck of the Saint-Pierre University Hospital, Brussels, noted that those on the NNRTI-based regimen had statistically significantly higher rates of virological failure (27 (12.9%) compared to 13 (6%) for LPV/r), as well as a higher incidence of resistance mutations in both RTI classes. This may hinder the effectiveness of second-line treatment with a PI as well as having the potential for increased resistance transmission.

However this difference in virologic outcomes was offset by a higher rater of loss to follow-up in the protease inhibitor group.

The World Health Organization (WHO) recommends using NNRTIS (nevirapine or efavirenz) with tenofovir/emtricitabine (TDF/FTC) or zidovudine/lamivudine (AZT/3TC) as the preferred first-line therapy in resource-poor settings. Reasons include low cost, ease of use of fixed-dose combinations and no need to refrigerate. However the advantages come with the primary disadvantage of a low genetic barrier to the development of resistance in cases of treatment interruption. This is particularly important with low adherence or stock-outs.

The rationale and advantages of a PI-based regimen for first-line include a high genetic barrier; the possibility for more treatment options at the time of failure; LPV/r is in tablet form so refrigeration is unnecessary. Nonetheless disadvantages include higher cost, more pills and metabolic changes, as well as the potential for drug-drug interactions.

The authors decided that in spite of the higher cost and other concerns they would test the hypothesis that PI-based regimens would be better than NNRTI-based regimens in resource-poor settings.  Therapeutic failures at 96 weeks were defined as:

·        WHO clinical stage 3 or 4

·        Confirmed virological failure (viral load over 1000 copies/mL)

·        Stopping or changing treatment because of toxicity

Eligibility for enrolment included being over 18 and ART-naïve; women who had taken single-dose nevirapine were included as long as they had taken the dose more than a year previously; WHO criteria were used for starting ART (WHO clinical stage 4, or stage 3 and CD4 under 350 cells/mm³ or CD4 under 200 cells/mm³ regardless of clinical stage).

The primary reason for exclusion was tuberculosis (TB) at screening. Dr. Clumeck noted this is important given that the high death rates in all developing countries within the first six months of treatment are due primarily to TB. Other criteria for exclusion included being pregnant or breastfeeding and abnormal liver enzymes and haemoglobin levels.

425 patients were first randomised to LPV/r (n=216) or NVP (n=209) and then secondly to TDF/FTC or AZT/3TC. All got cotrimoxazole prophylaxis.

There were no differences between the groups at baseline. Approximately 70% were women; the majority were WHO clinical stage 3. One woman within the LPV/r arm and two women within the NVP arm had previous exposure to NVP. Median CD4 cell counts at baseline were in the region of 170 cells/mm³. Approximately 30% had CD4 counts under 100 cells/mm³ and approximately 50% had a viral load above 100,000 copies/ml.

At week 96 there was no statistically significant difference in the proportions with viral load below 50 copies/ml between the two arms but differences in virological failure were statistically significant.  

Changes in CD4 cell counts were not statistically significant, but there was a trend toward a higher response in the PI arm.

Of the 40 patients with virological failure (13 in the PI arm and 27 in the NVP arm) 32 had genotyping done at the time of failure.

NNRTI mutations were found in 86% of those in the NVP arm and none in the PI arm. There were no major PI mutations in either arm.

NRTI mutations, namely M184V and K56R, were more frequent in the NVP arm compared to the LPV/r arm, occurring in 73% (16/22) and 20% (2/10), respectively. 

Baseline resistance mutations were available for 36 of the 40 with virological failure. Three out of the 25 in the NVP arm had drug resistance mutations for NNRTI and all failed at week 24. None of the 11 in the LPV/r had drug resistance.

Of those who failed 13 switched (1 in the LPV/r arm and 12 in the NVP arm) and 9 succeeded in re-suppressing viral load to under 50 copies/mL, one with a NVP-based regimen and eight with LPV/r-based regimen.

Overall both regimens and their backbones were equally well tolerated.

However, Dr. Clumeck concluded the significantly higher rate of virological failure and higher incidence of resistance mutations in both RTI classes of those on an NNRTI-NRTI first-line regimen in resource-poor settings warrants the evaluation over the long-term of the cost-effectiveness of a PI-based regimen for first-line use in these settings.