PROCAARE: Trials of AZT in Africa (04)

[Hilary Curtis <network@bmaids.demon.co.uk>]

[See also 970513191334 and 970513192645 - Mod.]

Ann Smith <ASMITH@cafod.org.uk> writes:
>I refer to an article circulated in PROCAARE (24/4/97) titled "Medical Group
>condemns US AIDS Drug Tests in Africa for using Placebo".
>
>I would echo the concerns voiced by the Public Citizen Health Research Group
>regarding the medical trials of AZT use by pregnant women, which are being
>sponsored by the US governement in nine developing countries. 

Large snip, then suggestion of historical controls.

>There is abundant, sound, and well-documented epidemiological
>evidence accumulated from many African countries showing that among pregnant
>women "receiving nothing" rates of mother-to-child HIV transmission range
>from 20 to 40%. Such data provide a reliable baseline for any comparative
>study.

No they don't!  In fact if your quoted figure of 20-40% (ie a range
equal to 100% of the lower limit*) is correct then it is plain that
historical controls would _not_ work.  With such a vague baseline, one
would need an absolutely _huge_ trial to have any prospect of achieving
adequate statistical power. This still would not solve the problem of
systematic bias (see below).

Historical controls _might_ be feasible if they were based on really
precise, detailed data from _exactly_ the same populations as the trial
participants are drawn from - not just "many African countries".  This
would mean that one would need to have had a history of systematic HIV
testing of pregnant women _and_ their babies to determine accurate
transmission rates within those populations.  

I doubt that one could find a suitably well-defined population for which
good, precise historical data exists.  Even if one could, there are
still problems with using historical controls instead of randomised
contemporaneous ones, because of the possibility of systematic bias.  A
likely source of bias in this study would be due to breastfeeding rates
- which may differ markedly between the trial participants and any
historical controls.  If a trial using historical controls shows a
reduction in HIV transmission, would this be attributable to less
breastfeeding, to a change in HIV _incidence_ rates in women**, or to
the AZT intervention???

*Though this is difficult to interpret without information on what this
range  means in terms of variances and confidence estimates.

**Bearing in mind that infectiousness fluctuates with maternal viral
load and therefore is likely to peak soon after the mother first becomes
infected, then fall, then rise again as she becomes sick.  

>Along with the serious reservations I have to the inclusion of a placebo
>group, wider ethical concerns raised by the proposed trials prompt me to ask
>what are the long-term inplications for the populations involved in these
>studies? If/when the efficacy of AZT is (further) demonstrated will the US
>government ensure a continued supply of AZT for ALL pregnant HIV-infected
>women in those countries? What of the implications for those women who,
>through involvement in the trial (or its fall-out effects) discover they are
>HIV-infected, and the US government's responsibilities in this regard?

Snip re ??carcinogenicity of AZT

>It is at very least noteworthy, in this context, that the focus of these
>trials would seem to be shifting away from the litigation- and
>rights-conscious US population onto the voiceless and more vulnerable
>women and children of the Third/Developing  World.
>
>
>Ann Smith,
>CAFOD,
>London, England

My understanding is that the main feature of this study is that it uses
a shorter course of AZT than in the earlier, developed country trials.
It is entirely possible that this short course will turn out to be
(almost) as effective as the longer course, given increasing evidence
that most HIV transmission occurs at around the time of birth (or after,
if the mother breastfeeds).  If the trial does show the short course to
be effective, that might benefit both developing and developed country
women, since such a course will not only be cheaper but also presumably
_safer_ than a longer course.  In situations where women present late
(or not at all) for antenatal care, a short course has additional
benefits since it can be started when delivery is imminent. 

Apart from the issue of possible carcinogenicity, the safety profile of
AZT may be different in developing countries if, say, more women are
anaemic/borderline anaemic at study entry than in the earlier developed
country trials.  This provides a strong reason for using a placebo, to
get a reliable comparison group for possible short-term adverse effects
as well as for reduced HIV transmission.

While I do not believe that the use of placebo controls necessarily
raises ethical concerns, I think you are right to raise other issues
such as the implications for women who discover they are HIV+ as a
result of the trial, and the quality of information and consent for
trial participants.  I don't know enough details to judge whether these
issues have been properly addressed.

> Who is to provide medications once a study is complete?  

I certainly agree that this is a major issue, but one which is more
likely to be addressed successfully if the intervention is less costly,
ie a shorter course.

There is another question, which is whether AZT is the best drug to
choose for a trial of short-course therapy to reduce vertical
transmission.  I saw a paper somewhere which suggested nevirapine;
although in monotherapy it soon leads to viral resistance, it is highly
effective in suppressing viral activity in the short term, and a very
short course around the time of delivery would be fairly cheap and safe.
Resistance might not be perceived as a problem given that most
developing country women (or their babies) have essentially nil chance
of subsequently receiving combination antiviral therapy.  You may see
that remark as ethically dubious - personally I think it is just being
pragmatic.

Seriously, given the need to develop interventions for use in resource
poor settings, we should be realistic and  to acknowledge that this
_will_ sometimes mean comparing a less intensive intervention with
placebo, in circumstances when a more intensive intervention has already
shown to work in a developed country setting.  

-- 
Hilary Curtis, Executive Director, BMA Foundation for AIDS
http://www.bmaids.demon.co.uk/
Education for HIV/AIDS policy, prevention and professional practice
BMA House, Tavistock Square, London WC1H 9JP, UK

KEYWORDS: ETHICS/ AZT/ MATERNAL TRANSMISSION/ AFRICA
...............................................................gjn
--
Send mail for the `ProCAARE' conference to `procaare@usa.healthnet.org'.
Mail administrative requests to `majordomo@usa.healthnet.org'.
For additional assistance, send mail to:  `owner-procaare@usa.healthnet.org'.