351
The Predictors of HIV Encephalopathy and Outcome in HIV+ Patients in the Era of Effective Therapy
C Mussini1, K Bhaskaran2, S Walker2, M Dorrucci3, C Sabin4, A Phillips4, Kholoud Porter*2, and CASCADE Collaboration
1Clin of Infectious and Tropical Diseases, Modena, Italy; 2Med Res Council Clin Trials Unit, London, UK; 3Inst Superiore di Sanità, Rome, Italy; and 4Royal Free and Univ Coll Med Sch, London, UK
Background: Despite HAART success at
reducing AIDS-related morbidity, its ability to prevent the onset of HIV
encephalopathy (HIVE) remains unclear given the limited penetration into the
brain of some drugs. We investigated changes in the risk of HIVE since the
introduction of HAART, and assessed the importance of current and nadir CD4
cell count and other prognostic factors in predicting HIVE diagnosis.
Methods: We used Cox models for time
from seroconversion to HIVE diagnosis, stratifying by cohort, allowing for late
entry, in a pooled dataset of 22 seroconverter
cohorts in Europe, Australia
and Canada (CASCADE). We considered the risk of HIVE over 3 periods (to 1996,
1997-1999, and 2000-2004), adjusting for age at seroconversion, exposure
category, and sex. The prognostic roles of these cofactors, and of
(time-dependent) current and nadir CD4 count, were then further investigated.
Poisson regression was used to estimate absolute incidence rates of HIVE.
Results: Of 7923 seroconverters
included in the analysis, 152 were diagnosed with HIVE (129 pre-1997, 14 in
1997-1999, and 9 in 2000-2004). The risk of HIVE fell substantially in
1997-1999 and 2000-2004 compared with pre-1997 (RR = 0.28 [95%CI 0.15 to 0.54]
and 0.19 [0.08 to 0.45], respectively). Compared to individuals with most
recent CD4 ³350 cells/mm3, the relative
risk (95%CI) of HIVE increased to 4.5 (2.0 to 9.8), 11.9 (5.4 to 26.5), and
69.0 (34.2 to 139.1) for those with CD4 200 to 349, 100 to 199, and 0 to 99
cells/mm3, respectively. Current CD4 count was a better predictor of
HIVE risk than nadir CD4 count (p
<0.001); furthermore there was no evidence that nadir CD4 was an independent
predictor of HIVE risk after accounting for current CD4 count (p = 0.60). We found no evidence of an
effect of exposure category, sex, or age at seroconversion on the risk of HIVE
(p = 0.60, 0.60, 0.28, respectively).
The absolute incidence rate (95%CI) pre-1997 was 0.7 (0.3 to 1.5) cases per
1000 person-years at CD4³350
cells/mm3, compared with 3.2 (1.8 to 5.8), 9.6 (5.7 to 16.2), and
64.2 (52.1 to 79.1) at CD4 counts of 200 to 349, 100 to 199, and 0 to 99
cells/mm3, respectively. In 2000-2004 the equivalent incidence rates
of HIVE by current CD4 count were 0.3 (0.07 to 1.1), 1.1 (0.3 to 4.3), 3.2 (0.8
to 12.7) and 10.7 (3.5 to 33.3) cases per 1000 person-years, respectively.
Conclusions:
We found an
elevated risk of HIVE even at CD4 counts of 200 to 350
cells/mm3. The decrease in HIVE incidence at the same CD4
levels in the HAART era suggests a direct effect of HAART on the pathology of
HIVE.
|
