Page 275 - HIVMED_v21_i1.indb

Page 3 of 10 Original Research

Services (NHLS). Some data that were more than 1 year old, Antiretroviral therapy treatment,
were excluded as these did not represent the patient’s status clinical and laboratory
at the time of the study census. These data included laboratory
values of creatinine, haemoglobin and total cholesterol blood characteristics
results.
Out of 157 participants for whom the ART regimens and the
Data analysis latest serum creatinine blood results were available, some
12/150 participants (8%) had an eGFR of ≤ 50 mL/min/1.73m³
A list of study definitions can be found in Appendix 1. Stata/ whilst on first-line ART. Those on second-line ART showed a
IC 16.0 software (STATA Corporation, LLC, TX, US 2019) was low eGFR in 1/7 participants (14.3%). Not a single participant
used to analyse data. For descriptive statistics numbers, with a serum creatinine of ≤ 100 µmol/L had renal dysfunction
percentages, medians, minimum and maximum values and as per the chronic kidney disease (CKD)- Epidemiology
interquartile ranges (IQR) were used. For evaluating Collaboration Equation (EPI) equation, whereas 13/36
associations between demographic, clinical and laboratory (36.1%) of those with a creatinine of ≥ 100 µmol/L had renal
characteristics with four treatment outcomes viz. the most dysfunction, according to this equation.
recent VL and CD4 cell count, LTFU and death, Pearson’s chi-
square, Fisher’s exact, Mann–Whitney (Wilcoxon rank-sum), Twenty-nine out of 34 participants (85.3%) with an
Kruskall-Wallis or Spearman’s correlation tests were used unsuppressed VL (≥ 400 copies/mL) remained on a non-
where appropriate. Most values for the continuous variables nucleoside reverse transcriptase inhibitor (NNRTI)-based
were shown to be non-normally distributed as per the ART regimen, whereas 5/34 (14.7%) were on a second-line
Shapiro–Wilk normality test. Even after logarithmic regimen with a protease inhibitor (PI) boosted with ritonavir.
transformation, the data were non-normally distributed. All The PI was lopinavir in all cases. Of the participants with a
variables were initially assessed individually against the VL of 400 copies/mL – 999 copies/mL, 14/16 (87.5%) were
treatment outcomes. If p values were ≤ 0.05 or there were > 0.5 on a NNRTI first-line regimen, whereas for those with a VL of
correlations for measures of association, these predictors were ≥ 1000 copies/mL, 15/18 (83.3%) were on a NNRTI first-line
entered into a linear or logistic regression model where regimen.
appropriate. A level of ≤ 0.05 was considered statistically
significant. During the initial analysis, the VL and latest CD4 The median CD4 cell count improved from a baseline
cell count outcomes were assessed as continuous variables as of 279.5 cells/mm³ (IQR 167–433) to 536 cells/mm³
well as clinical categories. Whenever there was a statistically (IQR 337.5–703.5) at the most recent CD4 cell count
significant relationship between any continuous or categorical monitoring visit, resulting in a median improvement of
outcome in the initial analysis, it was used in the binomial 256.5 cells/mm³. Moreover, some 33% of the cohort had a
logistic regression model provided that the number of CD4 cell count of ≥ 350 cells/mm³ at baseline compared to
observations was ≥ 10 per cell. The binomial logistic regression 70.7% at the most recent CD4 cell count monitoring visit.
model categorised the VL as suppressed (< 400 copies/mL) or Close to 10% of the cohort still had a CD4 cell count of
unsuppressed (≥ 400 copies/mL) and the latest CD4 cell as < 200 cells/mm³ at the latest CD4 cell count monitoring
3
high (≥ 350 cells/mm ) or low (< 350 cells/mm³). visit. For this group, 10/19 (52.3%) participants had an
unsuppressed VL of ≥ 400 cells/mL. This group also
Ethical consideration showed that 12/19 (63.2%) suffered from a defined
co-morbidity. The majority of this group (n = 16/19, 84.2%)
Ethical clearance was granted by both the Human Research were on first-line ART.
Ethics Committee (Medical) of the University of the
Witwatersrand, Johannesburg (M180304) and the North West Recent haemoglobin blood results were available for
Health Research Committee (NW_2018_006). The Tlokwe 24 participants. Of these, 4/24 (16.7%) had a haemoglobin of
sub-district Office of the Primary Health Care Manager also < 8g/dL, whilst 3/24 (12.5%) had a haemoglobin of 8 g/dL –
granted permission for the research. Owing to the 12 g/dL and 17/24 (70.8%) had a haemoglobin of ≥ 12g/dL.
retrospective nature of the study, patient’s anonymised data Total cholesterol blood results were available for 75 participants.
were evaluated. Consequently, informed consent was waived Of these, 40/74 (53.3%) had readings of < 5 mmol/L and 35/74
for this study. (46.7%) had readings of ≥ 5 mmol/L. For participants who
were classified with hypercholesterolaemia, 9/14 (64.3%)
Results had recent total cholesterol readings of ≥ 5 mmol/L whereas
26/61 (42.6%) participants who were not classified to have
Data from a total of 191 clinic files of OPLWH were hypercholesterolaemia had recent total cholesterol readings of
examined. Clinic A provided n = 38/191 (19.9%), Clinic B, ≥ 5 mmol/L.
n = 96/191 (50.2%) and Clinic C, n = 57/191 (29.8%). One
patient file from Clinic A had to be excluded as this was a Co-morbidities
duplicate file. Eleven patient files were missing from Clinic B,
as were 12 files from Clinic C. An overview of the demographic, One or more co-morbidity was found in 123/199 (64.4%)
treatment, clinical and laboratory characteristics is provided participants. Eighty-four participants (44%) had one chronic
in Table 1. All participants were pensioners. condition, a number of 34 participants (17.8%) had two

http://www.sajhivmed.org.za 267 Open Access