Paper # 108
Pre-treatment Mortality and Probability of Starting ART in Patients Enrolled in the Free State ARV Program, South Africa: Implications for Treatment Guidelines
Suzanne Ingle*1, L Fairall2, V Timmerman2, M Bachmann3, J Sterne1, M Egger4, M May1, and IeDEA Southern Africa
1Univ of Bristol, UK; 2Univ of Cape Town, South Africa; 3Univ of East Anglia, Norwich, UK; and 4Univ of Bern, Switzerland
Background: Important indicators of the performance
of ART treatment programs include the proportion of eligible patients starting
treatment and the extent of pre-treatment mortality.
Methods: All ART eligible patients (CD4<200
cells/mm3) enrolled in 36 clinics in the Free State public-sector
treatment program between May 2004 to Dec 2007 were followed until Dec 2008.
Time was measured from the first CD4 <200 (baseline) to the earlier of ART
initiation or death. We estimated the cumulative % of patients starting ART
(overall and stratified by CD4), while accounting for censoring due to the
competing risk of death. We also estimated the cumulative pre-ART mortality
accounting for the competing risk of starting ART. We used competing risks
regression to estimate the mutually adjusted associations of sex, age, CD4 and
year of enrolment with the probability of (1) starting ART and (2) pre-ART
death. We investigated patterns of CD4 count in patients with multiple pre-ART
measurements.
Results: During the study, 12963/22083 (59%)
eligible patients started ART, 5125 (23%) died before initiating treatment,
2779 (13%) were lost to follow up and 1216 (5.5%) were still waiting for ART at
the end of follow up. The table shows outcomes by baseline CD4 counts. Men (vs
women) were less likely to start ART (HR = 0.83, 95%CI 0.80 to 0.86) and had
higher mortality (HR = 1.29, 95%CI 1.22 to 1.36). Patients enrolled during 2007
(vs 2004 to 2005) were more likely to start ART (1.56, 95%CI 1.50 to 1.63) and
less likely to die (0.64, 95%CI 0.59 to 0.68). Among 2991 patients with a CD4
count prior to eligibility (median 260 (IQR 227 to 318)) the median time to
next measurement was 183 (IQR 105 to 309) days, median CD4 at eligibility was
101 (IQR 47 to 154) and the median decrease between CD4 measurements was 113 (IQR
70 to 183).
|
|
N (%)
|
Estimated %
|
Adjusted HR (95% CI)
|
|
|
Patients
|
Start ART
|
Dead pre ART
|
Start ART
|
Dead pre ART
|
|
Overall
|
22083 (100)
|
69.9
|
30.1
|
|
|
|
CD4 count
|
|
<25
|
3207 (15)
|
51.2
|
47.6
|
0.59(0.56 to 0.63)
|
3.51(3.26 to 3.78)
|
|
25-49
|
2698 (12)
|
62.0
|
36.9
|
0.74(0.70 to 0.79)
|
2.37(2.18 to 2.57)
|
|
50-99
|
5102 (23)
|
67.6
|
32.4
|
0.83(0.79 to 0.86)
|
1.87(1.74 to 2.00)
|
|
100-200
|
11076 (50)
|
78.3
|
20.2
|
1
|
1
|
Conclusions: Although survival improved between 2004
and 2007, many patients died before accessing ART. Pre-ART mortality may be
reduced by fast-tracking the most immunodeficient patients and improving access
for men. Because of the low frequency of measurements, the treatment
eligibility threshold should be increased to raise the median CD4 count at ART
initiation to 200 cells and to improve retention in care between first program
contact and ART start.
|
