TITAN & DUET Studies COMMENTARY
No patient left behind-better treatments for resistant HIV infection
The Lancet July 7, 2007
Bernard Hirschel a and Thomas Perneger ba. Division of Infectious Diseases, Geneva University Hospital, Geneva 1211, Switzerlandb. Clinical Epidemiology Service, Geneva University Hospital, Geneva, Switzerland
Darunavir (TMC114) is a protease inhibitor specifically designed for “resistance to resistance”1 and for effectiveness on HIV resistant to earlier protease inhibitors, such as indinavir or lopinavir. At 600 mg combined with ritonavir 100 mg, both twice daily, darunavir proved more efficacious than an investigator-selected protease inhibitor when added to an optimised background regimen in patients with drug-resistant HIV.2 The drug is now licensed in many countries worldwide.
The TITAN randomised study3 in today's Lancet explored ritonavir-boosted darunavir compared with the combination of lopinavir and ritonavir (Kaletra). All patients also received nucleotide or nucleoside reverse-transcriptase inhibitors chosen after resistance testing. Their previous treatment experience was moderate in length of time (median 5 years) and exposure to drugs; about two-thirds had received protease inhibitors before the trial, but none had been exposed to lopinavir or darunavir. Darunavir performed impressively: at week 48, significantly more patients on darunavir-ritonavir than on lopinavir-ritonavir achieved HIV RNA to below 50 copies per mL (71% vs 60%; intention-to-treat difference 11%, 95% CI 3-19%, p=0·005).
Side-effects were generally mild in both groups, with an excess of rashes in the darunavir group (16% vs 7%) and an excess of diarrhoea in the lopinavir group (15% vs 8%). Most side-effects were low grade, as shown by the proportion that led to discontinuation of treatment (7% in both groups).The comparator treatment in TITAN was not exactly the current gold standard, because the Kaletra to which darunavir was compared is not the Kaletra stocked today in most hospital or community pharmacies. TITAN patients ingested Kaletra capsules, which have since been replaced by tablets. The introduction of the new formulation is acknowledged by the TITAN investigators, who are endeavouring to switch the participants to tablets as soon as circumstances permitted, but at the time of analysis, only 18% in the lopinavir-ritonavir group had switched to the tablets. Compared with capsules, tablets need not be refrigerated, produce similar (though less variable) plasma concentrations of lopinavir, and less nausea and diarrhoea.4 What would have been TITAN's results if tablets had been used throughout? We can venture a guess: the difference in efficacy may have been the same, but the difference in gastrointestinal side-effects might be absent or smaller.Do the TITAN results presage a large-scale switch from lopinavir to darunavir? Probably not, for two reasons. First, darunavir's superior efficacy may be difficult to show in drug-naive patients, in whom pre-existing resistance to lopinavir is not the rule. Second, darunavir-ritonavir is more expensive, particularly in the USA, than lopinavir-ritonavir.
In patients similar to those in TITAN, resistance testing (before treatment with lopinavir-ritonavir in the majority with no resistance mutations) would represent a cost-effective alternative to darunavir-ritonavir, with little risk of failure.
Also in today's Lancet are the DUET-1 and DUET-2 randomised studies,5,6 which examined the efficacy of etravirine in advanced treatment-resistant HIV infection. Etravirine (TMC125) is a non-nucleoside reverse-transcriptase inhibitor (NNRTI), of the same class as nevirapine and efavirenz. Very small doses are effective in vitro,7 even against strains showing resistance to nevirapine and efavirenz, but it has been a challenge to produce a reliably bioavailable formulation. Currently, the 400 mg daily dose is administered in four pills. In a recent study, etravirine rescue was not efficacious after failure of a first-line NNRTI regimen.8 Resistance to the nucleoside reverse-transcriptase inhibitors given with etravirine, and lack of bioavailability of an etravirine formulation that is no longer in use, may have contributed to this failure. By contrast, etravirine lowered the viral load, compared with a standard-of-care control group, in patients with HIV that was resistant to a NNRTI.9 Interest in the combination of darunavir and etravirine arose when a small-scale study to investigate pharmacokinetic interactions (results were reassuring on that point) also showed surprising antiviral activity.10 These results stimulated recruitment for the DUET trials, which were finished in record time.
The typical DUET patients had already received a dozen antiretroviral drugs with selection of extensively resistant HIV, had had one or more AIDS-defining opportunistic diseases, and started the trials with a CD4 count of only 100/μL. They were randomised to darunavir plus etravirine, or to darunavir plus placebo, in addition to background treatment chosen by investigators on the basis of resistance tests. Etravirine's effect was measured by the proportion of patients reaching a viral load of less than 50 copies per mL after 24 weeks. That proportion was higher, by 17% in DUET-1 (95% CI 9-25%) and 18% in DUET-2 (11-26%), in patients on etravirine than in those in the group on the placebo combination. The difference was particularly striking when genotypic tests predicted that etravirine would be the only active drug (response rates of 44% with etravirine, 7% with placebo, difference 37%). No side-effects, except generally mild rashes, could plausibly be assigned to etravirine.
The two DUET papers leave important questions open. These questions include quality-of-life measurements, and detailed correlations between resistance genotype and treatment success which may help to gauge etravirine's prospects in individual patients. Instead of answers, which have been presented at a scientific meeting,11 we get two identical trials published as two papers, a type of duplicate publication in our view. We object to space and readers' time and attention thus being wasted, especially because combined analysis yields additional results.For instance, consider the issue of etravirine's effect on the occurrence of opportunistic infection and death. DUET-1 and DUET-2 describe the difference between the etravirine and control groups as not statistically significant. However, a pooled analysis of clinical events, with 22 of 599 patients on etravirine (3·7%) and 41 of 604 patients in the control group (6·8%) shows a statistically significant difference (by our calculation, Fisher's exact test, p=0·02; effect difference 3·1%, 95% CI 0·3-6·8%), and establishes that etravirine reduces by half the risk of clinical progression over 24 weeks (pooled odds ratio 0·52, 95% CI 0·30-0·89).
People care whether they get sick and die, and rather less whether their laboratory results are normal. Therefore it matters whether a new drug improves outcomes of clinical importance, in addition to surrogate measures, such as viral HIV RNA concentrations and CD4 counts. Antiviral therapies prevent clinical events, making it more difficult to show that new treatments reduce morbidity and mortality. A combined analysis of the DUET studies would have allowed the investigators to claim as of today a clinically important benefit for etravirine. It is a shame to see this opportunity delayed.Furthermore, a pooled analysis would have more clearly established that the efficacy of etravirine, in terms of viral load, differs by enfuvirtide status. Results of both studies suggested that etravirine benefited only patients who did not use or reused enfuvirtide. In new users of enfuvirtide, etravirine failed to confer any added benefit. The difference between new users of enfuvirtide and others, was of borderline statistical significance in DUET-1 (test of interaction in the paper: asymptotic p=0·046, exact p=0·055) and not significant in DUET-2 (by our calculation, asymptotic p=0·08, exact p=0·10). Again, a pooled analysis would have established that etravirine is significantly more effective in the absence of new treatment with enfuvirtide (odds ratio of viral load >50 copies per mL, 0·39, 95% CI 0·30-0·51) than in its presence (0·81, 0·51-1·30; test for the difference between these two odds ratios, p=0·008).
Purists may insist that separate studies always need to be reported separately, but that is not true when the only difference between them is the place of recruitment. Otherwise, results of different centres in multicentre studies could not be combined either. Some may fall prey to the erroneous perception that the regulatory requirement for “at least two adequate and well-controlled studies, each convincing on its own, to establish effectiveness”12 require two separate publications. Separate publication offers two spots instead of one for the prestigious position of first author and may allow the sponsor to honour a commitment made when the study started. Journals may play along because they compete for high-profile drug trials. Additionally, when prestigious duplicate studies are quoted together, journals get two citations instead of one, and doubled revenue from reprint sales. None of these reasons are compelling, and The Lancet and similar journals owe it to their readers to stop the practice. The high-ranking journals ought to publish the combined analysis of studies with similar protocols.But let us accentuate the positive. Occasionally one hears that the days of innovation in HIV therapy are over and that there is neither the scientific nor economic incentive for further progress. The TITAN and DUET studies, and those with the CCR5 entry inhibitor maraviroc,13,14 the non-nucleoside reverse-transcriptase inhibitor rilpivirine (TMC278), and the integrase inhibitor raltegravir,15,16 show that such pessimism is not justified. Not only are the new drugs effective, but they are also well tolerated. Raltegravir has hardly any drug interactions, which is welcome news when concomitant opportunistic infections, such as tuberculosis, must be treated at the same time.
The day will come when suppression of viral load to undetectable levels can be attained by all. We look forward to reading about it in The Lancet-three times over if need be.
In 2005-06, BH was a member of the data safety and monitoring committee (DMSB) of another study of etravirine, now completed; he is also a member of the DSMB of an ongoing study with darunavir. He has accepted travel grants and speakers' honoraria from Tibotec, Merck, Roche, GlaxoSmithKline, and Pfizer, and has participated in advisory boards for Tibotec, Pfizer, and Merck. TP declares that he has no conflict of interest.