Impact of ongoing side effects on the efficacy of initial antiretroviral therapy

Proposal
1456

Title of Proposed Research

Lead Researcher

Andrew Carr

Affiliation

St Vincent's Hospital, Sydney

Funding Source

This work does not require funding - the investigators will undertake the project within their allocated research time.

Potential Conflicts of Interest

AC has received research grant support from BMS, Gilead, MSD, Pfizer and ViiV Healthcare; participated in advisory boards for Gilead, MSD, and ViiV Healthcare; and received honoraria from Gilead, MSD, and ViiV Healthcare.
SK has no conflict of interest.

Data Sharing Agreement Date

7 July 2016

Lay Summary

HIV medications need to be safe, effective and well tolerated to be able to control HIV indefinitely. The current way of estimating treatment effectiveness requires that a person take the treatment without change and for full control of HIV. However, if a person has side effects from the HIv treatment, but continues to take it, the treatment is regarded as successful. This would mean that permanent side effects are regarded as unimportant. We propose to analyse data from two studies using the existing method, and compare it to a new method in which treatment is regarded as successful if is not only controls HIV and is taken continuously, but also does not cause side effects, and does not require of other medications to control drug side effects.

Study Data Provided

[{ "PostingID": 1387, "Title": "VIIV-ING113086", "Description": "A randomized, double blind study of the safety and efficacy of GSK1349572 50mg once daily to raltegravir 400mg twice daily both administered with fixed-dose dual nucleoside reverse transcriptase inhibitor therapy over 96 weeks in HIV-1 infected antiretroviral therapy naive adult subjects" },{ "PostingID": 1388, "Title": "VIIV-ING114467", "Description": "A Randomized, Double-Blind Study of the Safety and Efficacy of GSK1349572 Plus Abacavir/Lamivudine Fixed-Dose Combination Therapy Administered Once Daily Compared to Atripla over 96 Weeks in HIV-1 Infected Antiretroviral Therapy Naive Adult Subjects" },{ "PostingID": 2000, "Title": "VIIV-ING114915", "Description": "A Phase IIIb, randomized, open-label study of the safety and efficacy of GSK1349572 (dolutegravir, DTG) 50 mg once daily compared to darunavir/ritonavir (DRV/r) 800 mg/100 mg once daily each administered with fixed-dose dual nucleoside reverse transcriptase inhibitor therapy over 96 weeks in HIV-1 infected antiretroviral naïve adult subjects" }]

Statistical Analysis Plan

Statistical analysisData will be summarised for the following endpoint components of the two comparison algorithms, by study arm, within trial:• The proportion of study participants with viral load <50 copies/mL will be summarised at weeks 4,12, 24, 48, 72, 96, 120 and 144. • The cumulative proportion of subjects switching from their randomised study regimen (excluding changes in the backbone regimen) will be summarised over the duration of follow-up. A line listing of the change and reason for change will be made.• The proportion of subjects with a study medication-related AE will be described over all follow-up, and the proportion of patients with adverse events found post-study to be significantly more common in one treatment group, even if not reported to be study drug related will be described, and the cumulative proportion over the duration of follow-up will be summarised.Comparisons of proportion of subjects meeting the definition of failure in the FDA snapshot and Snapshot-plus analyses will be made using binary endpoints. Consistent with FDA snapshot methodology, failure is based on the following events:1) A detectable viral load (>50 copies/mL), after achieving virological suppression, up and including the week in which the snapshot comparison is made 2) Study participants who die or who are lost to follow-up, and those missing a viral load assessment at the study visit in question, and those who have ceased their randomised regimen before the week of the snapshot assessment (except for reasons permitted in the protocol) will be imputed as failures. The Snapshot-plus algorithm would define failure as 1 and 2 described above, in addition to the following:3) No adverse events, up to and including the week of the snapshot assessment The difference in proportion of patients with a successful responses (95%CI) will be compared between the dolutegravir and comparator arm, using the snapshot and snapshot plus algorithm. The comparison will be made separately using the FLAMINGO, SPRING-2 and SINGLE datasets, at weeks 48 and 144 of follow-up. A test of interaction will be made by baseline viral load strata =5 or >5 log10 copies/mL. The difference in the proportion of patients with successful responses between the snapshot and snapshot plus algorithm, and 95%CI will be calculated and a p-value for the difference in proportions will be calculated using a z-test of independent proportions. If a significant interaction is found in the responses by viral load strata in the snapshot or snapshot plus algorithms, a sensitivity analysis will compare the difference in proportion of patients with successful outcomes (95%CI) by viral load strata.

Publication Citation

http://www.croiconference.org/sessions/fda-snapshot-algorithm-may-overestimate-efficacy-initial-art