|
|
|
|
|
Session 104
Poster Abstracts Adherence, Quality of Life, and Factors Related to Treatment Friday, 1:30 - 3:30 pm Hall A |
Background: Randomized clinical trials (RCT) are the “gold
standard” for assessing the efficacy and safety of antiretrovirals (ARV).
However, those included in RCT may not be representative of the general HIV-positive
population and may have an improved response to HAART. We investigated the type
of patients who start HAART as part of an RCT and the effect of participation
in an RCT on response to HAART in a complete clinic population.
Methods: We included all ARV-naïve patients in the
Royal Free cohort who had started HAART since January 1996. Virologic response
was defined as the occurrence of a viral load < 400 copies/mL 12 to 48 weeks
after starting HAART (missing=failure [M=F] and missing=excluded [M=E]). We
chose this method because it is not sensitive to the number of viral load
measurements available. We studied the factors associated with participation in
an RCT and the effect of participation in an RCT on both the virologic response
to HAART and the tolerability of HAART regimens using logistic and Cox
regression.
Results:
Of 828 patients, 227 (27%)
started HAART while participating in 1 of 30 different RCT. Risk group and
HAART regimen were associated with RCT participation (p = 0.02 and < 0.0001) with adjusted odds ratios (OR) of 0.83
(95% CI 0.58 to 1.18) and 0.23 (0.08 to 0.67) for those with heterosexual and
other risks compared to homosexual risk and of 0.86 (0.50 to 1.48) for 2
nucleoside reverse transcriptase inhibitor (NRTI)+1 non-NRTI (NNRTI), 3.47
(0.88 to 6.42) for 2 protease inhibitors (PI)+2NRTI and 5.83 (2.97 to 11.42)
for other regimens compared to 1PI+2NRTI. Race, sex, CD4 and viral load at
HAART, age, and calendar year were not associated with RCT participation. The
RCT group had a median of 7 (IQR 5 to 9) CD4 counts in the first 48 weeks of
HAART compared with 6 (4 to 8) in the non-RCT group (p < 0.0001). 198 (87%) in an RCT and 441 (73%) not in an RCT had
a virologic response to HAART (M=F; adjusted OR = 2.69; 1.64 to 4.42; p < 0.0001). Using a M=E approach,
198 of 210 (94%) in the RCT group and 441 of 513 (86%) in the non-RCT group had
a virologic response (OR = 2.83; 1.36 to 5.92; p = 0.006); 90 (40%) in an RCT and 283 (47%) not in an RCT
discontinued an ARV in the first 48 weeks of HAART (adjusted hazard ratio = 0.70;
0.59 to 0.84; p < 0.0001).
Conclusions: RCT provide unbiased comparisons between
drug regimens but absolute virologic and discontinuation responses observed in
trials may over-estimate what can be achieved in routine clinic practice. This
may be due to unmeasured differences in the type of patients entering RCT, in
the clinical care received or in the HAART regimens used in RCT.
Keywords: antiretroviral therapy; randomised clinical trials; observational studies
 |