Early treatment in the HPTN 052 study of treatment as
prevention reduced serious illness by around 40%, delegates at the Sixth
International AIDS Society conference (IAS 2011) heard – but the effect was almost entirely
accounted for by fewer cases of extrapulmonary tuberculosis.
There was no significant difference in deaths, or in the
rate of serious bacterial infections or of pulmonary tuberculosis.
The HPTN 052 study was designed to determine whether early
treatment for an HIV-infected person would reduce the risk of HIV transmission
to his or her regular partner.
The study showed that early antiretroviral treatment – started
at a CD4 count between 350 and 550 cells/mm3 – reduced the risk of HIV transmission to
uninfected partners by at least 96%.
However, the study also collected information on a number of
other endpoints: death, illness , virological and immunological responses to
treatment. These data will provide important information on the potential
benefits of early treatment in settings where bacterial infections and
tuberculosis are more prevalent than in North America and Europe.
The study will also provide information on any differences
in rates of clinical or immunological disease progression between different
regions of the world, or between men and women. Some studies, but not all, have
suggested faster rates of disease progression in sub-Saharan Africa.
The study enrolled 1763 HIV-infected individuals with
CD4 counts between 350 and 550 cells/mm3. Participants were
randomised either to receive immediate antiretroviral treatment or to defer
treatment until their CD4 cell count fell below 250 cells/mm3 on two
separate tests. This was the level at which treatment was recommended to begin
in national guidelines during the study recruitment period. The study recruited
participants in Malawi, India, Zimbabwe, Botswana, South Africa, Kenya, Thailand, the US and Brazil.
The median CD4 count at enrolment was 446 cells/mm3 and
the median viral load 4.4 log10 copies/ml (25,000 copies/ml).
Although the proportion of HIV-positive men and women
recruited to the study was equal across the study population as a whole, sites
in Africa recruited a larger proportion of women than men (58 vs 40%, p<
0.0001), reflecting the higher rates of engagement in care by women in
sub-Saharan Africa as a consequence of earlier HIV diagnosis through antenatal
African sites also recruited a larger proportion of
participants in the 18 to 24 age group (20 vs 15%, p=0.003). African participants
had slightly lower viral loads (4.4 vs 4.5 log10 HIVRNA,
Although there was no significant difference in
baseline CD4 count between regions, Africans in the deferred treatment arm were
less likely to reach the CD4 threshold that indicated they needed to start
treatment. Participants in the deferred arm outside the Africa region were 40%
more likely to start treatment, but when the hazard ratio was adjusted for
confounding factors this difference ceased to be statistically significant (aHR
1.30, 95% CI 0.9-1.8, p= 0.06).
The median time to starting treatment in the deferred
arm was 3.25 years outside Africa, compared to 4.1 years in Africa.
Overall, 21% of participants in the deferred arm needed
to start treatment during the follow-up period (n=184), an incidence rate of 12
per 100 person-years of follow-up. However, among Africans the rate of treatment
initiation was 9 per 100 person-years, compared to 15 per 100 person-years of
follow-up in participants at non-African sites.
Unsurprisingly, the factors significantly associated
with the need to start treatment in the deferred treatment arm were lower CD4
count or higher viral load at enrolment (adjusted hazard ratio per increase of
100 CD4 0.6; 95% CI,0.5-0.7) and higher log10 HIV RNA (adjusted
hazard ratio per log increase 1.5; 95% CI, 1.2-1.8).
Overall, 75% of patients in the deferred arm who
initiated treatment did so because of a decline in CD4 cell count.
Among those who initiated treatment immediately (886
patients), at a median CD4 count of 442 cells/mm3, 90% had an
undetectable viral load (below 400 copies/ml) after one year of treatment. They
experienced a mean CD4 cell increase of 158 cells/mm3, to 603
In the delayed-treatment group, the median CD4 count at
treatment initiation was 225 cells/mm3, and the median time to
treatment initiation in the deferred arm was 3.5 years. After one year, 93% had
an undetectable viral load (below 400 copies/ml) and CD4 cell counts had risen
by a mean of 191 cells/mm3, to 418 cells/mm3.
Virologic failure was rare; only 5% in the immediate-treatment arm and 2.7% in the deferred-treatment arm experienced virologic
failure. Of these, 67% in the immediate arm and 60% in the deferred arm switched
to second-line therapy.
The effect of early treatment on clinical outcomes was
Early treatment significantly reduced the risk of
clinical illness, but there was no difference in the risk of death between the
two study arms.
Primary endpoint clinical events were defined as:
- A WHO stage IV event.
- Pulmonary tuberculosis.
- Severe bacterial infection.
Around 7% of participants received cotrimoxazole
prophylaxis against bacterial infections, and around 4% received isoniazid
preventive therapy to prevent the development of active TB.
A total of 105 participants developed a primary clinical event
during 3304 person-years of follow-up. Forty primary events occurred in the
immediate arm (2.4 per 100 PY), compared to 65 in the deferred arm (4.0 per
100 PY, hazard ratio 0.59, 95% CI: 0.40 - 0.88, p=0.01.)
Early treatment did not significantly reduce the risk
of developing pulmonary tuberculosis. There were 14 cases of pulmonary TB in
the immediate-treatment arm (0.8 per 100 PY), compared with 16 in the delayed-treatment
arm (0.9 per 100 PY). Pulmonary TB cases were diagnosed at a median CD4 count
of 521 in the immediate-treatment arm and 295 in the delayed-treatment arm.
In contrast, early treatment did significantly reduce
the risk of extrapulmonary TB. Three cases were diagnosed in the immediate-treatment arm (0.2 per 100 PY), at a median CD4 count of 443, compared to 17 in
the delayed-treatment arm (1 per 100 PY), at a median CD4 count of 342.
Bacterial infections occurred somewhat more frequently
in the immediate-treatment arm (19 vs 13, 1.1 vs 0.8 per 100 PY). In each arm
the most common bacterial infection was pneumonia, and the difference between
arms was chiefly driven by four cases of sepsis in the immediate-treatment arm.
Three participants in the immediate arm each experienced more than one
There was no significant difference in the death rate
between arms: ten occurred in the immediate arm and 13 in the deferred arm.
Deaths were largely attributed to causes other than HIV in the immediate arm: 3
suicides, one stroke, and three unknown causes, with only three deaths due to
infection (leptospirosis, TB and sepsis) in the immediate arm. Almost half the
deaths in the delayed arm were of unknown cause, with two accidental deaths,
one stroke and only two deaths due to infection.
There was no significant difference in severe adverse
events between the two arms (14% in each), nor any difference in the
distribution of types of adverse event between the two arms.