65
Performance Characteristics of Non-CD4-based Criteria for Initiation of ART in an Ambulatory Clinic Population in Phnom Penh, Cambodia
Chel Sarim*1, J Elliott1, S Huffam1, H Chenda1, P Sophea1, S Vonthanak1, J Kaldor2, D Cooper2, and M Vun1
1Natl Ctr for HIV/AIDS, Dermatology and STI, Ministry of Hlth Cambodia, Phnom Penh and 2Natl Ctr in HIV Epidemiology and Clin Res, Univ of New South Wales, Sydney, Australia
Background: Absolute peripheral CD4+ T cell
count provides independent prognostic information and guides initiation of ART.
As access to CD4 testing in resource-limited settings is restricted,
alternative initiation strategies have been recommended.
Methods: Data were prospectively collected on all ART-naïve
adult patients enrolling in an ambulatory HIV clinic in Phnom Penh. Using CD4-based ART initiation
criteria as the “gold standard,” we evaluated the performance characteristics
of clinical criteria alone, and combined with lymphocyte count, hemoglobin, or
body mass index as defined in 3 ways: by
the 2002 and 2003 World Health Organization (WHO) ART guidelines and the 2005
WHO Guideline Development Group (GDG) meeting report.
Results: Data were available on 439 patients. Median
age at enrollment was 34 years and 212 (48.3%) patients were female. Median CD4
count was 94 cells/mm3 with 280 (63.6%) counts <200 cells/mm3.
At enrolment 264 (60.1%) patients were in WHO Stage 3 and 49 (11.2%) in WHO
Stage 4. The proportions of patients eligible for ART as per the 2002 and 2003
guidelines and the 2005 GDG report were 65.8%, 73.3%, and 75.4% respectively. Compared
to CD4 criteria, the sensitivity, specificity, positive predictive value and
negative predictive value of clinical criteria alone were 0.17, 1.00, 1.00, and
0.38, respectively, as per the 2002 guidelines, 0.87, 0.71, 0.89, and 0.66,
respectively, as per the 2003 guidelines and 0.84, 0.69, 0.89, 0.59,
respectively, as per the 2005 GDG report. Addition of lymphocyte count <1200
cells/mm3 improved performance substantially as per the 2002
guidelines (0.65, 0.95, 0.96, 0.58), but marginally as per the 2003 guidelines
(0.91, 0.68, 0.89, 0.74) and 2005 GDG report (0.89, 0.65, 0.86, 0.65). Use of
alternative lymphocyte count cut-off values, hemoglobin and body mass index
either alone or in combination did not further improve predictive values. Use
of the best performing non-CD4 based criteria would have resulted in treatment
misallocation for 24.8% patients as per the 2002 guidelines, 15.0% patients as
per the 2003 guidelines and 17.1% as per the 2005 GDG report.
Conclusions: Basing initiation of ART on clinical criteria
alone would have resulted in poor treatment decision-making in this population.
Addition of lymphocyte count improved performance, but treatment misallocation
remained substantial. These data strongly support the role of CD4 testing in criteria
for initiation of ART in an ambulatory setting.
|
