Predictors of virological rebound in HIV-1-infected patients initiating a
protease inhibitor-containing regimen
AIDS 2002; 12:21-29. Jan 4
Vincent Le Moinga,b; Genevieve Cheneb; Maria P. Carrieric; Ahmadou Alioumb;
Francoise Brun-Vezinetd; Lionel Pirothe; Jill P. Cassutof; Jean-Paul Moattic;
Francois Raffig; Catherine Leporta; and the Aproco Study Group*
Objective: To study the predictors of virological rebound in patients having
early virological response to protease inhibitor (PI)-containing regimen.
The authors recommend using more potent regimen when viral load is high;
single PI regimen may not be adequate. They found more durable responses when
therapy was started at CD4 counts above 500, when immunity was preserved.
Design and methods: APROCO cohort study prospectively enrolled 1283
HIV-infected patients starting a PI-containing regimen in 1997-1999.
Adherence to therapy was measured with self-administered questionnaires after
4 months of therapy (M4). Virological rebound was defined as a viral load
(VL) > 500 copies/ml in patients having early virological response, defined
as a VL < 500 copies/ml at M4. Predictors of time to virological rebound were
studied with multivariate proportional hazards model.
Results: During a median follow-up of 20 months, virological rebound was
observed in 32% of the 830 patients with early virological response.
Virological rebound was more frequent when patients had received previous
antiretroviral treatment [adjusted hazards ratio (HR) = 2.4; P < 0.0001],
were younger (HR = 1.4 per each 10 years younger; P < 0.0001), had baseline
CD4 cell count < 500 (HR = 2.3; P < 0.001), had higher baseline VL (HR = 1.4
per each log10 copies/ml higher; P < 0.001), reported low adherence to
therapy at M4 (HR = 2.1; P < 0.001) or had stopped PI at M4 (HR = 1.7; P =
Conclusion: Initiation of treatment at a stage of preserved immunity is
associated with a more durable virological response under protease inhibitor.
Every effort should be made to monitor and strengthen adherence to therapy,
even in patients having early virological response.
At first, patients were classified as reporting high adherence if in the
detailed table they reported that they had taken 100% of their prescribed
doses. Among these patients, those who however declared that they had skipped
a dose during the last week-end, that they had 'almost totally' followed
their HAART regimen, or they had to modify several times the prescribed
schedule or that they took their regimen through a unique intake were
reclassified as reporting
moderate adherence. Patients were also classified as reporting moderate
adherence if they declared having missed no more than 20% of their prescribed
doses of HAART (80-99% adherence). Finally, patients who declared that they
had taken less than 80% of pills, and/or that they had 'partially' or 'not
at all' followed their HAART prescriptions, were considered as reporting low
Almost half of them were antiretroviral naive; the, median baseline VL was
4.5 log10 copies/ml [interquartile range (IQR), 3.6-5.1] and median baseline
CD4+ cell count was 291 (IQR, 133-436). PI prescribed at baseline was
indinavir in 45.8% of patients, nelfinavir in 29.9%, ritonavir in 13.5%,
saquinavir-hgc in 6.8% and a combination of PIs in 4.0%. At M4, 8.6% of
patients had changed
PI and 5.4% had stopped it, either temporarily or definitively. These
modifications of the PI component of the regimen were due to intolerance in
69 of 116 (59.5%) patients.
Risk of rebound was higher, but not significantly so, in patients having
moderate adherence compared to those reporting high adherence.
The effect of baseline CD4+ cell counts on the risk of virological rebound
did not appear to be
log-linear: the risk of rebound was quite uniform for the lowest counts and
was significantly lower solely in the patients having a CD4+ cell count 500.
In our study however, the association between baseline CD4+ cell counts and
sustained virological response was not linear and was significant only when
baseline CD4+ cell count was above 500.