Immunologic Response of HIV-Infected Children to Different Regimens of Antiretroviral Therapy: A Retrospective Observational Study

Background Both abacavir- (ABC-) based and zidovudine- (AZT-) based regimens are widely utilized for managing HIV infection in children. Unfortunately, there is a lack of data regarding their immunological response and associated risk factors in Ethiopia. Methods A retrospective hospital-based cohort study was conducted on HIV-infected children in Jimma Medical Center (JMC). A total of 179 records were reviewed by including data from November 2015 to April 2017. Data were collected on sociodemographic, clinical characteristics of patients and drug-related variables. Data analysis was done using STATA 13.1. Mixed-effect linear regression was performed to assess the difference in CD4+ changes between groups adjusting for baseline characteristics. The change in predicted CD4 count attributed to each regimen was also assessed by marginal analysis. P < 0.05 for slope of the random-effect linear regression was used as an indicator for the presence of association. Result Of 179 patients, 98 (54.7%) were females. The mean (±SD) duration of follow-up was 939.8 ± 478.3 and 984.92 ± 453.1 days for ABC and AZT groups, respectively. AZT group had a significant CD4+ count gain per visit compared with their ABC counterparts ((β = 20.51, 95% CI [6.37–34.65]), P = 0.004) over time. The regimen AZT + 3TC + LPV/r tended to have an excellent predicted CD4+ lymphocyte count change relative to all other regimens, while ABC + 3TC + LPV/r had the least immunologic recovery (margins 338.0 cells/mm3 versus 249.13 cells/mm3 (P < 0.001)). Baseline CD4+ lymphocyte count, ART group, WHO clinical stages, and viral load were independent predictors for CD4+ change overtime. Conclusion AZT-based regimens seem to have better immunological response compared to ABC-based regimens. Immunologic response was described worse in patients with a viral load of >1000copies/ml, low baseline CD4+ count, advanced WHO clinical stages, and ABC-containing regimens. Further study is needed to clarify these aspects.


Background
Currently, 1.7 million children are living with human immunodeficiency virus (HIV) and over 90% of them live in sub-Saharan Africa [1,2]. Ethiopia has the largest population of HIV-infected children in the region. According to an estimate by the Federal HIV/AIDS Prevention and Control Office (FHAPCO), there are over 738,976 people living with HIV/AIDS in Ethiopia [3]. Of these, 178,500 are children younger than 15 years of age [4]. e initiation of potent antiretroviral therapy (ART) in the mid-1990s has dramatically reduced HIV-associated morbidity and mortality [5][6][7][8]. Antiretroviral therapy reduces HIV-associated morbidity and mortality by suppressing HIV replication to undetectable levels and providing a consistent increase in the number of CD4+ T lymphocytes [9,10]. e resulting effect is the recovery of the immune system, and thus, immune reconstitution is an important outcome measure of ART [11]. According to the recent report, 59% of children living with HIV were on treatment in Ethiopia [12]. e earlier World Health Organization (WHO) guidelines recommended using a nucleoside reverse transcriptase inhibitor (NRTI) backbone including a thymidine analogue stavudine (d4T) or zidovudine (AZT) together with lamivudine (3TC) for paediatric HIV treatment [13]. But in 2010, the WHO guideline discouraged the use of d4T because of high lipodystrophy rates in adults and adolescents [14]. As a result, the current guidelines were changed to recommend substitution of the thymidine analogue with abacavir (ABC) depending on studies which reported fewer side effects and improved virological responses with ABC compared to d4T or AZT [15,16].
In Ethiopia, the initial ART regimen included the use of d4T as a preferred first-line NRTI for paediatric HIV treatment. In 2012, Ethiopia implemented the d4T phase-out program, and ABC became routinely used as a standard of care [17]. Currently, both ABC-and AZT-based regimens are routinely utilized. However, there is a lack of study regarding the immunological response of ABC-and AZTcontaining regimens among HIV-infected children in Ethiopia. Studies conducted in some African children [18][19][20][21] showed an encouraging and comparable immunologic response to those obtained among children in wellresourced countries [11,22,23]. In this study, we compared the immunological response and associated factors in a cohort of HIV-infected children receiving ABC-and AZTcontaining regimens.

Study Design and Setting.
A retrospective hospital-based cohort study was conducted on HIV-infected children in Jimma Medical Center (JMC). JMC is located in Jimma town, 355 km from Addis Ababa. It is currently the only teaching and specialized hospital in the Southwest region of Ethiopia. e hospital serves as a referral site and provides specialized care for Southwest Ethiopia with a catchment population of about 15 million. e study was conducted from April 10 to May 10, 2017, by including data from November 2015 to April 2017.

Study Population and Variables.
We included HIV-infected children (<15 years) who were on ABC-and AZTbased regimens that fulfill the eligibility criteria. Patients on ABC-and AZT-based first-line regimens, having at least six months of follow-up with good adherence, whose records were legible and complete, who have CD4 count at least at the baseline and six months, and younger than 15 years were included in the study.
ose transferred out within <6 months of follow-up and patients with incomplete records were excluded. e study was conducted by dividing the total sample into two major classes as ABC group and AZT group.

Sample Size Determination.
e number of patients who fulfilled the eligibility criteria for the ABC group was 87 and all of them were included. From those on AZT-based regimen, 92 eligible patients were included, making ABC-to-AZT group ratio of 1 : 1.05. Finally, a total of 179 subjects, with 87 charts of patients from the ABC group and 92 patients from the AZT group were reviewed.

Sampling Technique.
Since the number of patients on ABC-based regimen was limited, we included all eligible patients. A simple random sampling technique was used to select 92 patient charts for AZT-based regimens using a computer-generated random number.

Data Collection Procedure and Quality Assurance.
Data on socio-demographic, clinical, laboratory, and drugs administered were collected by record review using English version checklists. e data collection tool was carefully prepared after reviewing relevant literature studies to enable the data collectors to gather all the information required to address the study objectives. A 2-day training was provided on the data collection tool and general procedures for data collectors, i.e., 2 pharmacists (B. Pharm) and 2 nurses. Data from antiretroviral drugs and patient information sheet were collected by pharmacists, and data from ART clinic intake form, HIV care/ART follow-up, and patient sheet were collected by the nurses. e baseline body mass index of the subjects was later calculated after collection of baseline height and weight of the patient from the patient chart. Pretest was conducted on 5% of the eligible records.

Operational Definition of Terms
Good adherence: estimated adherence level of >95% [24] as recorded by ART physicians/nurses Child: age <15 years [25] Lost to follow-up: refer to a patient who has missed clinical or drug pickup appointments permanently [26].

Ethical Consideration.
e study was approved by the Institutional Review Board (IRB) of Jimma University. It has designated with an IRB number of IHRPGB/112/2017. e need for informed consent was waived because of the retrospective, anonymous nature of the study. During data collection, confidentiality was ensured, and for this reason, the name and address of the patient were not recorded in the data collection checklist.

Statistical
Analysis. Data were double-entered into Epi-Data and exported to STATA 13.1 for cleaning and analysis. Descriptive analysis was performed, and results were presented in text, tables, and charts. Bivariate and multivariate mixed-effects linear regressions for repeated measurements were performed to assess the adjusted effect of the ART regimen and identify additional predictors of CD4+ recovery. e coefficient of mean CD4+ count with 95% confidence intervals was used as a measure of the strength of association, and P < 0.05 was considered to declare a statistical significance. Marginal analysis was also conducted to see the difference among the specific regimen category.

Overview of the Study Participants.
During the study period, a total of 367 patients started antiretroviral therapy (ART) and were treated for at least 6 months. Of these, 108 from ABC group and 212 from AZT group have a complete CD4+ count at the 6 th month of treatment. irty-two patients were excluded initially from either regimen due to missed CD4+ count at 6 months and 30 (21 and 9 from ABC and AZT, respectively) because of the adherence problem, and 179 patients were included in the analysis ( Figure 1). e mean ± standard deviation (SD) duration of followup was 939.8 + 478.3 and 984.92 + 453.1 days for the ABC and AZT groups, respectively. During the study period, a total of 4 patients (1 (1.14%) patient from the ABC group and 3 (3.3%) from the AZT group; P � 0.339) died.

Baseline Socio-Demographic and Clinical Characteristics.
We included 179 patients, of whom 98 (54.75) were females. One hundred forty (78.2%) of the patients had a BMI of less than fifth centile, and there was a statistically significant difference between groups (P � 0.03). Baseline socio-demographic and clinical characteristics of the study participants are depicted in Table 1.

Immunological Response.
e overall change in the mean standard deviation (±SD) of CD4+ lymphocyte count was 139.3 (±55.7) cells/mm 3 for the ABC group and 150.5 (±42.1) cells/mm 3 for the AZT group. e mean change in CD4+ lymphocyte count from baseline for the past 18 months was presented comparatively in Figure 2. e SDs are 55.3 vs. 41.6, 66.3 vs. 63.1, and 87.7 vs. 65.1 at 6 th , 12 th , and 18 th months for ABC-and AZT-based regimens, respectively.
As shown in the figure, the CD4+ lymphocyte count trajectory belonging to AZT-based regimen looks steeper after the first 6 months of therapy.

Predictors of CD4+ Lymphocyte Count Change.
We conducted a mixed-effect linear regression using a consecutive CD4+ lymphocyte count measured over the past 18 months as an outcome variable. e slope of random-effect linear regression was used to report the overtime change in CD4+ count attributed to the predictor variables. e overall average gain in CD4+ count every six months was 53.0 cells/mm 3 (P < 0.001 and 95% CI [46.86-59.20]) with an interclass correlation coefficient of 65.5% (P < 0.001). is implies a significant proportion of variability in CD4+ lymphocyte count change due to patient-specific factors. Among the groups, 36.73% of the variation in CD4+ lymphocyte change variability was explained by differences in ART regimens (P < 0.001).
Body mass index, weight for height, ART group, baseline, CD4+ lymphocyte count, occurrence of opportunistic infections (OIs), WHO clinical staging, viral load, nutritional status, and exposure to antitubercular drugs were associated with CD4+ lymphocyte count change on binary linear regression. After adjusting for all confounders, ART group, baseline CD4+ lymphocyte count, WHO clinical stages, and viral load remained independent predictors for CD4+ lymphocyte count change.
erefore, patients who commenced on AZT-based regimens had a significant change in CD4+ lymphocyte count at each visit. Accordingly, patients exposed to AZTbased regimens had 20.51 cells/mm 3

Postestimation Treatment Effects.
We also conducted a marginal analysis to predict the changes in the mean CD4+ lymphocyte counts associated with each regime and to identify the regimen with the lowest performance. erefore, except ABC + 3TC + NVP, the other two ABC-based regimens, namely, ABC + 3TC + EFV and ABC + 3TC + LPV/r, deemed to be the lowest performing regimens as compared to their AZT counterparts. Accordingly, the predicted mean CD4+ count for paediatrics treated with AZT + 3TC + NVP had 318.3 cells/mm 3 of CD4+ lymphocyte count change, and it was 296.33 cells/mm 3 CD4+ for those treated with ABC + 3TC + EFV. e change is statistically significant in either case (P < 0.001).

Discussion
is study was the first ever to compare the immunological response of ABC-and AZT-based regimens among HIVinfected children in Ethiopia. e biannual analysis of mean AIDS Research and Treatment    Figure 2: Comparative mean CD4+ count gains for paediatric patients exposed to ABC versus AZT from April 10 to May 10, 2017.  CD4+ gain showed that the maximum gain in mean CD4+ count was attained with AZT + 3TC + LPV/r, while ABC + 3TC + LPV/r had the least immunologic recovery over the entire treatment course. e CD4+ lymphocyte count trajectory showed a linear trend. Baseline CD4+ lymphocyte count, ART group, WHO clinical stages, and viral load were independent predictors for CD4+ changes over time.
Our finding of mean CD4+ recovery was inconsistent with a randomized open-label study by Mulenga et al. [27], where a comparable immunologic response was achieved between AZT-and ABC-containing regimens. A study by Cassim et al. [28] also failed to show a significant difference in terms of immunologic response between ART regimens. e higher CD4+ recovery with AZT-based regimens in this study might be due to a higher baseline CD4+ count of those patients who were on AZT-containing regimens (Table 1). In addition, the difference in the study design might contribute to the variation.
is is thus an additional evidence to advocate to the test and treat strategy as opposed to waiting for dropping of CD4 counts at lower thresholds [24]. e finding from the marginal analysis also indicated that the change in mean CD4+ count is significantly higher in AZT-based regimens. Adult immunologic studies indicated that a linear trend in CD4+ increment at the early phase of therapy is expected and it became flat with minimal CD4+ gain later after 18 th month of treatment course [29]. However, the linear trend of immunologic recovery may indicate our study was ended prematurely. e overall random-effect linear regression analysis had pointed out that baseline CD4+ lymphocyte count, ART group, WHO clinical stages, and viral load were independent predictors for CD4+ change overtime. e marginal effects of each regimen confirmed that the immunologic outcome associated with AZT+3TC + LPV/r (margins � 338.0 cells/mm3/, P < 0.001) was the most favorable followed by ABC+3TC + NVP (m � 337.8 cells/ mm3, P < 0.001) and AZT + 3TC + EFV (m � 320.3, P < 0.001). However, ABC+3TC + LPV/r had the lowest predicted change in CD4+ count (m � 249.13, P < 0.001), implying minimal immunologic response, and further study is needed to clarify this finding.
Patients who commenced on AZT-based regimens had greater CD4+ improvement over time. Accordingly, those patients who had exposed to AZT-based regimens were associated with an average CD4+ count advantage of 20.51 cells/mm 3 (β = 20.51, 95% CI [6.37-34.65]). is finding is inconsistent with the Paediatric European Network for the Treatment of AIDS (PENTA 5) study, where children on ABC-based regimens had better immunologic response as compared to AZT-containing regimens [30].
is variation might be due to differences in study design (RCT vs. retrospective), study setup, and sample size.
In the cohort, baseline CD4+ count was another positive predictor for successful immunologic recovery. erefore, each unit increment in the baseline CD4+ count will contribute to 0.55 cells/mm 3 CD4 gain every half year (β = 0.55, 95% CI [0.49-0.69]). is is concurrent with previous studies [31][32][33][34], in which a higher baseline CD4+ lymphocyte count was associated with better immunologic response. is could be due to less extensively depleted immune system, which will be boosted easily after initiation of ART. It is also in agreement with the current WHO ART guidelines [35], which recommends the initiation of ART at diagnosis (test and treat) regardless of CD4+ count and WHO stage.
On the contrary, patients with WHO stages III and IV were in a precarious situation in terms of their CD4+ count recovery. . is is in agreement with studies conducted in Nigeria [36] and Tanzania [37], in which advanced clinical stage was associated with worse CD4+ recovery. Similar observation was also reported by Kaufmann et al. [38], in which patients with advanced HIV stage had poor immune recovery compared to those in the early stage of the disease. is finding suggests that patients with advanced HIV/AIDS and profound immune suppression need special attention to improve their outcome. Similarly, patients with a viral load of >1000 copies/ ml had a worse immunologic response (β � −27.68[−47.75− (−7.61)]). is finding is consistent with the study conducted by Zhou et al. [39].
is study had several possible limitations. Firstly, because of the retrospective nature of the study, we were not able to capture some important data such as adverse drug effects of ART. Second, as the sample size was relatively small, the power to detect definitive differences may have been limited. It was a single-site hospital-based study, and therefore, the findings may not be generalizable to the general population, and measures of adherence by health professionals that may not fit the reality are some of the limitations.

Conclusion
In the current study, patients with AZT + 3TC + LPV/r had a better immunologic recovery. Immunologic response was described worse in patients with a viral load of >1000 copies/

Data Availability
e data used to support the findings of this study are available from the corresponding author on request.
Ethical Approval e study was approved by the Institutional Review Board (IRB) of the Institute of Health Sciences, Jimma University. e need for informed consent was waived because of the retrospective, anonymous nature of the study.

Conflicts of Interest
e authors of this paper have no competing interests.

Authors' Contributions
TA conceived the study, collected and analysed the data, and drafted the manuscript. FB conceived the study and collected and analysed the data. GZ conceived the study, analysed the data, and reviewed the manuscript. All authors have read and approved the manuscript.