UK-CAB 25 – Meeting Report

Reading study results — D:A:D study — BHIVA treatment guidelines — GSK

11 April 2008

Summary

The morning session was mainly a presentation on key concepts in clinical trials by Simon Collins, HIV i-Base and Caroline Sabin, Royal Free Hospital designed to work through key recent studies to look at: what is an event, timelines for events, sample size to capture an event, determining causal factors for an event ire comparing “chance” occurrence to ARV-related occurrence of heart attack, and the statistical concepts of odds ratios, confidence intervals, absolute and relative risk.

Simon Collins provided feedback on key studies from CROI, including data on abacavir, ddI and increased risk of heart attack in the D:A:D study and biomarkers of inflammation in the SMART study.

The afternoon was a company meeting with GlaxoSmithKline, who presented their well-known data on abacavir safety. Strictly speaking this is not comparable with the data from the D:A:D study, since the information is drawn from pre-marketing approval safety and efficacy studies not designed to capture long-term adverse events like heart attack (as noted in the recent Lancet editorial).

There was also a discussion on the draft BHIVA treatment guidelines. Key points from the meeting were posted on the website and summarized in a letter provided to Garry Brough and Elijah Amooti for circulation to BHIVA.

Contents

1. Conference feedback: key studies from CROI 2008.
2. D:A:D study: long-term cardiovascular risks of ARVs, reading study results – some applied statistics.
3. Company meeting: GlaxoSmithKline
4. BHIVA treatment guidelines: new draft

1. Conference feedback: key studies from CROI 2008

CROI is one of the most important scientific meetings of the year, always held in Boston.

The summary of key studies is available online as a PDF (4 Mb) and webpage.

D:A:D study: long-term cardiovascular risks of ARVs, reading study results – some applied statistics

Dr. Caroline Sabin, Professor of Medical Statistics and Epidemiology, Research Department of Infection and Population Health, Division of Population Health, Royal Free and UC Medical School and Simon Collins, HIV i-Base

Dr. Sabin (CS) discussed the concept of an event and how to identify risk factors contributing to an event happening. To illustrate and explain these terms and topic, she used the D:A:D study. The D:A:D study is a long-term international cohort study to look at the relationship between exposure to ARVs and heart attack.

CS started by explaining why one would want to study the risk of an event. The reasons may be:

  • determine the probability that some event will occur in a population
  • to examine the rate at which an event occurs in a population or the time taken for an event to occur
  • to compare risks in different populations eg old and young people, people using or not using ARVs.

Other reasons are possible.

Prevalence measures how common something is at any one time, eg total number of people with HIV in any one year. To determine how often something will occur in a population you need to look at incidence, eg how many new HIV infections are there in a year. Determining incidence needs a longitudinal approach. For a rare event, eg a heart attack, you need to look at a large number of people over a long time, especially if you want to compare possible causes.

Q: During the trial design stage, do they allow for drop out percentage?
A: If it is done properly, yes. They would say that took into consideration that patients might drop out.

Q: What is a two-tailed study?
A: Most studies have two possible outcomes, eg does a drug make you better or get worse? Researches need to be prepared for any result. It is very rare they know in advance which way the study will go.

Q: Is the increase in risk of heart attack from ARVs in the D:A:D study more than you would expect?
A: Yes it is more than you would expect. PI exposure in the study is on average 10 years. Enough people have been looked at to say with reasonable confidence that PIs increase the risk of heart attack in HIV-positive people. The same applies to abacavir. We have stated the relative risk ie the risk compared to people not using these drugs. The actual risk for an individual depends on other risk factors, eg smoking. Abacavir seems to double the risk of heart attack. For someone with a low baseline risk of heart attack, using abacavir may not make a significant difference, but for someone with a high baseline risk, this may be important to know.

Q: If the relative risk of an event is two does it mean I am twice as likely to experience an event?
A: Yes, but this cannot tell you definitively why. Patients’ risk factors vary a lot. Poisson regression allows us to calculate the relative risk adjusted for any difference in patient characteristics. Confidence interval shows how much uncertainty there is the calculation. Good studies have confidence levels that are very close. If the confidence interval is wide then the study was too small. Width of the confidence interval also depends on what you choose to study.

Simon Collins (SC) talked about relative risk of illness or death in the SMART study.

SMART was a randomized trial. The aim was to study the hypothesis that there may be an advantage in using intermittent ARV therapy guided by CD4 level ie less toxicity, compared to continuous treatment. But in the drug conservation (treatment break) arm, people had more severe illness. SMART was a big study of 6,000 patients planned for 9 years. It was stopped after 2 years because data showed the relative risk of serious illness or death for people taking breaks was 2 times that of people in the continuous treatment arm.

Q: How soon after stopping treatment they could they see any difference in health in the drug conservation arm?
A: Within 4 weeks they could see the difference?

Q: Did people go back on same treatment, when they had the break?
A: No, they went on a different combination of therapy.

Q: did people recover their health on restarting treatment?
A: Not completely. This will be looked at further, but it seems their CD4 did not return to the level it was before people first stopped treatment.

3. GlaxoSmithKline

Dr. Grahame Taylor (GT) presented GSK’s data on abacavir safety from pre-marketing approval studies. This data does not show a high or increased risk of heart attack from abacavir. However, these trials were designed to look at anti-viral effect, not the risk of heart attack.

GT then outlined the ACTG 5202 study – phase IIIb randomized study, comparing two nuke backbones: TDF/FTC v ABC/3TC plus either EFV or ATV/r. 1,600-person study that began enrolling in 2005. Study appears to show ABC backbone not as effective for people with pre-treatment viral load of 100,000. GSK has conducted other studies which do not show this result, and the factors for people on the ABC backbone not achieving viral suppression need further examination.

Q: What does the GSK data actually show?
A: It shows that there is no increase risk of heart attack from abacavir. However, D:A:D has raised the importance of looking at cardiovascular risk factors when using ARVs. It is recommended that all patients should be assessed for cardiovascular risk, using Framingham Risk scores and actively managed to reduce their risk. Factors like diabetes; age, smoking etc place an important role in all of these factors. The D:A:D study findings are unexpected. It is important to take note of the study’s findings, because it is difficult to look at the total picture until you have complete information. We believe the picture is incomplete because we have no comparable data on tenofovir.

Q: But doesn’t D:A:D show abacavir is very harmful and dangerous drug?
A: Really difficult to answer. There are many risk factors for heart disease in HIV-positive people. If it is dangerous it should be withdrawn.

Q: When will final analysis become available?
A: Do not know. Hopefully as soon as possible.

GSK research: 3 integrase inhibitors in development processes. Looking at vaccines.

Q: What is the time frame of this new drug?
A: Started 10 years ago. Various companies have designed the technologies so far; it is still years away from the final product. Intergrase inhibitors are in phase 2.

Q: Pre-registration trials – it advisable to keep a follow up on the people that used the drug during the trial period?
A: Certainly sound logical but it is difficult to put it into a process. But in principle it sounds like a great idea.

Q: What are GSK doing on data on abacavir risks?
A: We are aiming to present it at the next national conference.

Q: Do you hear of people that modify treatment because of the risks?
A: If someone has a severe cardiovascular risk they should stop using the abacavir if the drug affects him or her in that way. Smoking is an independent risk factor. Stress factors are much higher amongst HIV positive people.

4. BHIVA treatment guidelines: new draft

With Gary Brough, UKCAB treatment guidelines writing group representative

Key points arising from this discussion were included in the letter drafted after the meeting (members only, sign in required)

Attendees

Zhana Books, Zhana Productions, London
Paul Clift, UK CAB, London
Simon Collins, HIV i-Base, London
Ben Cromarty , North Yorkshire AIDS Action, Northallerton
Emma Hudson, The Brunswick Centre, Halifax
Rupert Jones, West Yorkshire African Group, Leeds
Mohamade Jowata, Brent PCT, London
Mary Kawomera , W O M M, Doncaster
Angeline Marang, HIV i-Base, London
Michael Marr , Waverley Care Solas, Edinburgh
Nakamba Ng’ambi , Leeds Skyline Services, Leeds
Roger Pebody, THT, London
Jo Robinson, THT , London
Winnie Sseruma, HIV i-Base, London
Jack Summerside, Independent, London
Anthony Tukai, Oxfordshire County Council, Oxford
Brian West, HIV Scotland, Edinbugh
Matthew Williams, Monument Trust, London

Speakers

Simon Collins, HIV i-Base
Caroline Sabin, Royal Free Hospital
GrahamTaylor, GlaxoSmithKline
Kevita Sud, GlaxoSmithKline

Views expressed in this report are personal and do not represent the views of any organisation involved in the meeting. Reasonable steps have been taken to ensure this report is accurate but it is for information only and not a substitue for professional advice.

Edit
Published: April 11, 2008
Last edited: January 5, 2011