Objective: To describe the use of second line protease-inhibitor (PI) regim
ens across Europe and to determine factors associated with virological and
immunological response.
Design: Analysis of data from 984 patients with a median follow-up of 21 mo
nths enrolled in EuroSIDA. Patients started their second PI-containing regi
men at least 16 weeks after starting the first PI-containing regimen and wi
th viral load > 1000 copies/ ml.
Methods: Virological response was defined as a viral load < 500 copies/ml a
nd immunological response as an increase of 50 x 10(6)/l or more in CD4 lym
phocyte count.
Results: The median CD4 cell count at starting the second PI was 171 x 10(6
) cells/l; viral load was 4.45 log copies/ml. As a second PI regimen, 45% w
ere using a dual PI, while of those on one PI, indinavir (42%) and nelfinav
ir (34%) were most common. In multivariate Cox models, a higher viral load
at starting the second PI relative hazard (RH), 0.67 per 1 log higher; 95%
confidence interval (CI), 0.58-0.77; P < 0.0001) and a lower CD4 cell count
(RH, 1.15 per 50% higher; 95% CI, 1.06-1.26; P = 0.0014) were associated w
ith a reduced probability of virological response. Those who had achieved v
iral suppression on the first PI-regimen were more likely to respond to the
second (RH, 1.65; 95% CI, 1.30-2.10; P < 0.0001) as were those who added o
ne or two new nucleosides to their second PI.
Conclusions: Patients who initiate a second PI regimen at lower viral load,
higher CD4 cell count or who added new nucleosides tended to be more likel
y to achieve a viral load < 500 copies/ml. The roles of cross-resistance an
d adherence in response to second-line regimens needs further investigation
. (C) 2001 Lippincott Williams & Wilkins.