Page 255 - HIVMED_v21_i1.indb

Page 2 of 4 Scientific Letter

management for patients with low CD4 counts to mitigate more than 3 months after ART where the CD4 count was not
higher morbidity and mortality risks. a duplicate of the baseline count. The proportion of patients
with post-baseline CD4 counts ≤ 200 and ≤ 350 cells/mm
3
The South African ART programme is routinely managed was calculated and compared between different ART
using an electronic database known as Three Interlinked durations using a chi-squared test.
Electronic Registers.Net (TIER.Net). This study analysed
routine programmatic TIER.Net data to assess implementation Immunological non-responders were defined as patients
of post-ART CD4 testing and occurrence of poor CD4 who had been on ART for more than 4 years and were
recovery in order to highlight areas for intervention to virally suppressed (VL < 1000 copies/mL) with a CD4 count
improve CD4 monitoring and subsequent patient ≤ 350 cell/mm based on CD4 testing performed 1 year
3
management. Specifically, this study aimed to: (1) calculate before to 2 years after the VL test. Time on ART for all
the proportion of patients on ART with CD4 tests subsequent patients, including those receiving second- or third-line
to ART initiation, (2) describe the proportion of patients with ART regimens, was calculated as overall time on treatment.
3
subsequent CD4 counts ≤ 200 and ≤ 350 cells/mm by ART Mixed effects logistic regression was used to assess
duration and (3) assess the proportion of INRs and describe characteristics of INRs, adjusting for random effects at the
their characteristics. district level. Fixed effects covariates included ART regimen
(first-line, second-line or third-line), age at ART start,
Methods gender, baseline CD4 and tuberculosis (TB) status. p < 0.05
Data source and study population was considered significant.
Three Interlinked Electronic Registers.Net data were Ethical consideration
extracted in March 2020 for Johannesburg and Sedibeng
districts in Gauteng province and Capricorn and Mopani The study was approved by the Human Sciences Research
districts in Limpopo province. These districts were chosen Council Research Ethics Committee (REC 3/22/08/18).
as a convenience sample, as Anova Health Institute is the Individual patient consent was not required, as no data were
designated United States Agency for International collected for the purposes of this study. Anonymised TIER.
Development support partner in these districts. Urban Net data that were routinely collected at healthcare facilities
Johannesburg district is very densely populated, with 3162.1 for monitoring purposes were used.
persons/km , compared with 236.0, 61.7 and 61.2 persons/
2
km in Sedibeng, Capricorn and Mopani, respectively. Results
9
2
Medical scheme coverage is highest in Johannesburg Implementation of CD4 testing
(22.2%), followed by Sedibeng (20.8%), Capricorn (8.3%)
and Mopani (6.8%). Antenatal HIV prevalence, an indicator Baseline CD4 count was recorded for 80.9% of the 869 571
9
of overall population prevalence, was 34.1% in Sedibeng, patients newly initiating ART. Amongst all 1 178 190 adults
on ART, only 46.5% had a CD4 count recorded subsequent to
30.9% in Johannesburg, 26.6% in Mopani and 22.5% in baseline. Amongst patients who had been on ART for 12–18
Capricorn in 2017. 10
months (n = 56 181), only 21.9% had a post-baseline CD4 test
on record.
Records from TIER.Net were included in the analysis for
adult PLHIV aged 15–80 years who had initiated ART from
2004 onwards and had been on ART for a minimum of 12 Low CD4 counts
months (n = 1 224 366). Records were excluded where there Amongst all patients with a baseline CD4 (n = 703 869), 50.3%
were data quality concerns regarding CD4 testing, namely had counts ≤ 200 cells/mm and amongst those starting ART
3
counts ≤ 0 or ≥ 2000 cells/mm and nonsensical testing dates since 2017, 37.2% had a baseline count ≤ 200 cells/mm . CD4
3
3
(n = 722). Records with nonsensical VL testing dates were count decreased after baseline in 11.0% of the 443 443 patients
also excluded (n = 9), as were records from facilities which with both a baseline and subsequent CD4. Amongst all
had not exclusively used TIER.Net as their ART management patients with a CD4 test performed after ART start, 14.3%
3
tool, resulting in incomplete TIER.Net data (n = 45 445). The (n = 78 494) had a CD4 count ≤ 200 cells/mm . This proportion
final data set comprised 1 178 190 records – 673 606 from was highest amongst patients on ART ≤ 2 years (19.7%)
Johannesburg, 160 607 from Sedibeng, 162 020 from Capricorn compared with longer treatment durations (14.8%, 14.2% and
and 181 957 from Mopani. 13.8% amongst patients on ART 2–4, 4–6 and > 6 years,
respectively, p < 0.001).
Statistical analysis Just over one-third of patients with a post-baseline CD4 test

The proportion of patients with a baseline CD4 test and CD4 had a CD4 count ≤ 350 cells/mm (35.5%, n = 194 140). This
3
test subsequent to baseline was calculated, the former was proportion was highest amongst patients on ART ≤ 2 years
for patients newly initiating ART and the latter was for all (43.9%) compared with longer durations (36.1%, 35.2% and
patients in the cohort (both new initiators and re-initiators). 34.7% amongst those on ART 2–4, 4–6, and > 6 years,
Post-baseline CD4 tests were defined as CD4 tests performed respectively, p < 0.001).

http://www.sajhivmed.org.za 247 Open Access