HIV virological non-suppression and its associated factors amongst children on antiretroviral therapy at a major paediatric treatment centre in Southern Ghana: a cross-sectional study

Adwoa K. A. Afrane (  dadwoa@yahoo.com ) Department of Child Health, Korle Bu Teaching Hospital, Accra Bamenla Q. Goka Department of Child Health, University of Ghana Medical School, Accra Lorna Renner Department of Child Health, University of Ghana Medical School, Accra Alfred Edwin Yawson Department of Community Health, University of Ghana Medical School, Accra Yakubu Alhassan Department of Health Policy Planning and Management, School of Public Health, University of Ghana, Accra Seth N. Owiafe Department of Child Health, Korle Bu Teaching Hospital, Accra Seth Agyeman Department of Immunology, Korle Bu Teaching Hospital, Accra Kwamena W.C Sagoe Department of Medical Microbiology, School of Medicine and Dentistry, University of Ghana, Legon Awewura Kwara Division of Infectious diseases and Global Medicine, Unitersity of Florida Gainseville, FL


Results
The mean (± SD) age of the study participants was 11.4 ± 2.4 years and the proportion of males was 53. 2%

Conclusion
The prevelance of virological non-suppression was high. Virological non-suppression was associated with a previous history of TB treatment, female gender, severe CD4 immune suppression status (CD4% <15% / CD4 + count < 200 cells/mm 3 3,4 In resource-limited settings, HIV viral load (HIV VL) monitoring is recommended by the WHO (World Health Organization) as a gold standard for monitoring treatment effectiveness in 2013, 5 and Ghana adopted these guidelines in 2016. According to Ghana's National Aids Control Program (NACP) guidelines, viral load testing is recommended 6 months after initiating ART and therafter annually for people who have achieved virological suppression. 6 However people with HIV VL levels > 1000 copies/ml undergo intensi ed adherence support after which the viral load is repeated three months later in order to differentiate poor adherence from treatment failure. 6 Virological non-suppression could be due to poor adherence to ART, resistance to ART (transmitted or acquired) or pharmacokinetic issues (poor absorption, under-dosing and drug interactions). 7, 8 Adherence-related factors are most often caused by the patient. 7,8 ART regimen-related factors could result in the development of resistance. 9,10 A wide range of factors are associated with virological non-suppression and these include socio-demographic factors such as younger age (less than three years), 11 male gender, 12,13 WHO advanced HIV stage, 14,15 coinfection with tuberculosis (TB), 16,17 nevirapine (NVP) based therapy, 18,19 and poor adherence to treatment. 20,21 A principle cause for virological non-suppression is non-adherence. 21,22 The paediatric population are more likely to have challenges with adherence to ART 23,24 because drug formulations for paediatric patients are often less tolerable, and require dose adjustment according to the weight as the child grows. 2324 There are also speci c challenges with the social context, such as a lack of consistent caregiver in younger children and disclosure issues, which pose a challenge to consistent adherence. 23,24 These peculiar issues to children and adolescents can result in virological non-suppression, without necessarily leading to the development of drug resistance. 8 High rates of virological non-supression suggests a higher risk of developing resistance and is common in children in Low-and middle-income countries. (LMICs) 8 , and this has implications for the spread of resistant HIV variants. Identifying patients with virological non-suppression is thus important for prompt detection of treatment failures and in identifying the characteristics of the population that might bene t from interventions such as intensi ed adherence counseling and increased frequency of follow up. 31 This would in turn prevent the development of drug resistance in the population at risk. The aim of this study was therefore to determine the prevalence of virological non-suppression and its associated factors among children living with HIV (CLWH) attending a Paediatric HIV clinic at Teaching Hospital in Ghana. This knowledge would help to target interventions for improving virological suppression that will ultimately lead to improvement in clinical care.

Study design and setting
This study used a cross-sectional design to recruit paediatric HIV positive patients attending the outpatient clinic from October 2017 to July 2018 at the Korle Bu Teaching Hospital (KBTH), Accra, Ghana. KBTH has 2500 beds and is currently the largest hospital in West Africa and the third largest on the African continent. 25 The hospital is the main tertiary referral centre in Accra and serves the majority of the southern half of Ghana. The Paediatric HIV clinic at KBTH has been providing comprehensive HIV/AIDS care and management services since 2004. Patients are referred from primary and secondary health facilities as well as from other departments within the hospital. An average of 40 patients are seen per clinic day. The National AIDS Control Program (NACP) provides ART medication free of charge. HIV VL and CD4 + counts are also paid for by the program. The clinic uses national treatment guidelines that are in line with current WHO recommendations for ART in Ghana. 26 Treatment is initiated for all patients irrespective of the CD4 + count. ART available and in use in various combinations at the clinic at the time of the study were zidovudine (AZT), lamivudine (3TC), abacvir (ABC), efavirenz (EFV), nevirapine (NVP), tenofovir (TDF), and ritonavir boosted Lopinavir (LPV/r).

Study participants
Study participants were Children living with HIV (CLHV) aged between 8 months and 15 years and who had been on ART for at least 6 months for whom caregivers/study participants had given informed written consent to participate in the study were enrolled. Children who had been transferred from referral points and did not have their previous notes available and children with HIV-2 mono infection were excluded from the study. There are currently no Food and Drug FDA-approved assays for quanti cation of HIV-2 RNA 71 and hence the exclusion of children with HIV-2 mono infection. Voluntary informed consent was obtained from parents and guardians of study participants and assent from children aged 10-15years. Before seeking informed consent/assent, study participants were screened to determine whether they were eligble or not. A questionnaire was administered to study participants and caregivers. Information was also collected retrospectively from their medical records.

Sample size determination
The Cochran's sample size formula 27 was used to calculate the sample size to determine the prevalence of nonsuppressed viremia. A minimum sample size of 250 study participants was determined using con dence level of 95% and an error margin of 5%. Consecutive cases of HIV children who met the eligibility criteria were enrolled into the study until the sample size was reached.

Data Collection
Participants had their drug doses checked for appropriateness of their current drug regimen. The dosage, (frequency and dose/kg or m 2 ) was cross checked with the recommended dosage and appropriateness was documented as 'yes' or 'no.' ART Adherence was assessed by a pharmacist on the day of recruitment using the pill count.
Pill count (the number of pills taken) was calculated based on the number of unused pills that the care giver brought back when re lling their ART medication on the day of study recruitment. 155 Total re ll was the expected number of pills to have been taken since the last visit.
Pill count was callulated as: see equation in the supplementary les.

Laboratory analysis
Laboratory investigations done on the day of recuitment were CD4 + count and HIV VL. CD4 + absolute cell count and cell percentage were quanti ed by a dual-platform ow cytometry technology using a FACS Count system (Becton-Dickinson, Franklin Lakes, NJ, USA) according to manufacturer's instructions at the Fevers Unit Laboratory of KBTH. The HIV RNA VL testing was performed at the Central Laboratory of KBTH using the COBAS ® AMPLICOR Monitor test (Roche Diagnostic Systems, Branchburg, NJ, USA), with a a lower limit of detection of 20 copies/ml. The laboratory at the Fevers unit and the Central Lab KBTH are certi ed by the South African Public Health Reference Laboratory and participates in an external quality assurance testing programme by the South African Public Health Reference Laboratory.

Operational de nitions
In accordance with WHO guidelines, study participants were categorized as having virological non-suppression if the HIV VL level was > 1000 copies/ml on the day of recruitment, after at least 6 months of using ART. Drug Adherence was determined by caregivers report and categorized according to WHO guidelines as follows: Good ≥ 95%; Fair: 85-94%: Poor < 85%. 28

Statistical analysis
The dependent variable was virological non-suppression (VL > 1000 copies/ml). The independent variables were the sociodemographic factors, clinical factors and ART Adherence factors. Pearson's chi-square test of association was used to determine strength of association between the independent categorical variables and the outcome variable (virological non-suppression). The logistic regression model was used in determining the factors in uencing HIV VL suppression amongstudy participants with statistical signi cance set at p < 0.05. The crude odds ratio (OR) and adjusted odds ratio (AOR) were determined and OR and their respective 95% con dence intervals were calculated.

Ethical considerations
Ethical appproval (KBTH-IRB/ 00060/2017) was obtained from the Institutional Review Board of Korle Bu Teaching Hospital, Accra, Ghana. Informed consent was obtained from parents or legal guardians for each minor participant prior to enrolment to participate in the study.

Baseline characteristics
The baseline characterteistics of all 250 patients are shown in Table 1. The mean duration on ART was 64 months ± 3.0 months. More than half 133 (53.2%) of the study participants were males. Majority of the study participants 148 (59.2%) were within the age range of 10 to 15 years. Overall, the mothers of 157 (62.8%) participants were HIV positive while 93 (37.2%) of mothers had unknown HIV status. The fathers of 62 (24.8%) study participants were HIV positive whilst 105 (42.0%) were HIV negative and 83 (33.2%) had unknown HIV status. Primary caregivers who were mothers were 115 (46.0%), fathers were 25 (10.0%) and 110 (44%) were guardians (grandmother, grandfather, aunt, uncle, foster caregiver). The unknown status included parents who were dead. The educational and occupation of parents is described in Table 1

Factors associated with virological non-suppression
Bivariate analysis of factors associated with virological non-suppression are shown in Table 1. Females were more likely to have virological non-suppression in comparison to males (54.2% vs 32.2%, p = 0.035). The overall CD4 + immune suppression status of the subject at study recruitment showed a signi cant association with the subjects VL. (Fischer's exact (Φ, p < 0.001). The adherence rate of subject measured by pill count percentage showed a signi cant association with the subjects VL (χ 2 = 7.99, p =0.018).
There were no signi cant differences for the following variables: age of subject, primary caregiver's relationship to subject, educational status of father, educational status of mother, occupational status of father, occupational status of mother, history of previous TB treatment, WHO stage at initiation of ART, baseline VL and baseline CD4 + count values (overall), duration on ART, current ART regimen and person responsible for child's medication (Table   1).  As antiretroviral access continues to expand in resource-limited countries like Ghana, monitoring response to ART by the use of VL measurements is critical in determining the effectiveness of ART in the population. 29,173 The Sub-Saharan African region's prevalence for virological non-suppression (>1000 copies/ml) in children who have been on ART for at least 6 months ranges from 13% to 44%. 8 The prevalence of children with detectable viral loads above 20 copies/ml but less than 1000 copies/ml (low level viraemia (LLV)) in this study was 23%. This group of patients are denoted as 'non-failures' and WHO does not make any provision for the management of LLV. These group of individuals will not receive any follow-up VL for at least six to 12 months and thus risk continuing on a failing regimen for considerable time. There is therfore the need to design algorithms for patients with LLV to have more frequent VL monitoring as literature has shown the emergence of high-level resistance in this group of individuals. 33 We observed that females were 2.5 times more likely to have virological non-suppression as compared to males.
This phenomenon was similar to the study by Muri et al,34 in Tanzanian children, whereby females were also 2.5 more times likely to have virological non-suppression. On the contrary, some studies have reported that males had increased odds of virological non-suppression, 13,35 whilst other studies however found no association between sex and virological non-suppression. 36 The role of gender in virological suppression could be biologic according to authors such as Njom et al. 37 The relationship of virological suppression and gender is therefore inconclusive and requires further studies.
A third of the study population had been previously treated for TB. Furthermore, having a history of previous TB treatment increased the odds of having virological non-suppression by as much as ve times. These ndings are in agreement with studies reported by Ahoua et al, 38 and by Rajin et al. 39 On the other hand, it has been recently reported that children who had a history of TB co-infection had better virological outcomes. 13 Reasons for this could be due to the close monitoring, frequent clinic visits and adherence support, adopted as part of TB treatment offered at the sample sites. The association of a previous history of TB and virological non-suppression in this current study could be due to the increased pill burden and drug-drug interactions between the medications that these patients are on. The signi cance of TB comorbidity on the occurrence of virological non-suppression buttresses the need for the prioritization of VL monitoring and adherence support in co-infected patients as well as patients who have a history of previous TB.
The odds of virological non-suppression was almost eight times more likely in study participants whose current drug regimen was NVP-based as compared to a study participant who had an EFV-based regimen. The ndings of this study is consistent with current literature that shows that patients on NVP-based regimens experience more virological failure than patients on EFV-based regimens. 18,19 The use of regimens containing NVP is associated with a low genetic barrier of drug resistance and a higher risk of baseline resistance in cases where NVP was used as prophylaxis in the babies for Prevention of Mother To Child Transmission (PMTCT). This current study however did not evaluate prior NVP exposure. Our ndings support the current ART guidelines being used at the HIV clinic at KBTH which is to phase out NVP based regimens and replace with LPV/r or EFV regimens for children less than 20 kg. Current guidelines recommend a Dolutegravir (DTG) based regimen as the preferred rst-line for children weighing at least 20kg. Hopefully with the introduction of DTG and its scale up, the non suppression due to certain antiretroviral drugs such as NVP will be reduced.
We observed that study participants found to have severe CD4 + immune suppression status at the time of study recruitment were 25 times more likely to have virological non-suppression. These ndings are in congruence with studies reported by Jobanputra et al, 36 among children in Swaziland and by Izudi et al, 40 among children in Northwestern Uganda where it was found that patients with low CD4 + count values at study recruitment were more likely to have virological non-suppression. This nding is expected and supports the knowledge that viral suppression leads to immune recovery and could re ect the fact that those study participants who were virologically suppressed had a chance to reconstitute their immune systems for their CD4 + counts to increase. 41 This nding also supports early initiation of ART in children and according to the current ART guideline in Ghana, all children con rmed to have HIV diagnosis after birth are started on ART regardless of the CD4+ count.
There was no association between parent-educational level, employment status of parent and virological nonsuppression in this study. This is in agreement with a Danish HIV Cohort reported by Legarth et al 42 which also showed no association between education level and virological non-suppression. This nding is however in contrast to a study reported by Mensah et al 43 in Ghana, in which a child with an unemployed caregiver was ve times more likely to have virological non-suppression. There is substantial evidence on the socioeconomic inequalities in the treatment outcomes of chronic diseases like HIV. This current study could not con rm association between employment status and virological non-suppression. This observation could be due to the fact that for almost a third of study participants, the educational and employment status of parents was not known.
Studies on the relationship between self-reported adherence to ART and virological non-suppression have shown inconsistent results. 20,44 In this study, adherence level measured by pill count was not associated with virological non-suppression. On the other hand in a clinical trial reported by Intasanet al, 45 in Cambodian children, nonadherence was associated with virological non-suppression. The measure of adherence used in the study by Intasanet al, 45 was however the 3 day self -report by caregiver. In a more recent research by Natukunda et al, 31 in adolescents, reported in 2019, more than 70% of adolescents who experienced virological non-suppression were su ciently adherent as measured by pill count ( adherence >95%). On the other hand, there are also studies whereby poorly adherent patients maintained undetectable VL. 44,46 The strength of this current study is that it did not only determine the prevalence of virological non-suppression but explored factors associated with the phenomenon in children. In the absence of drug resistance testing information, close monitoring of VL levels, multidisciplinary support and prompt clinical judgment are key in ensuring children who have failed treatment are appropriately transitioned to second line therapy. Ultimately, 'an ounce of prevention is better than a pound of cure.' Limitation: The reliance on self-reported data as a measure of adherence, which may be affected by recall and social desirability bias. Analysis of baseline VL and CD4 + count was not available for some of the subjects and hence analysis on baseline VL and CD4 + could not be done for those patients. This is because VL monitoring was started in April 2011 (as a national policy) and hence all children above eight years of age did not have the opportunity of having a baseline VL level done.

Conclusion
The high rate of virological non-suppresion is consistent with ndings in other countries in Sub-Saharan setting and emphasizes the great challenge to successfully suppressing HIV in paediatric patients and reinforces the need for regular monitoring of viral load levels in children. Patients on ART with active TB and those with a history of previous TB should be prioritized for more regular VL monitoring (6 monthly) as the national guidelines advocate for yearly monitoring of viral load for patients whose viral load levels are suppressed. Further research should focus on determining resistance patterns in this study population.

Declarations
Ethics approval and consent to participate Ethical approval (KBTH-IRB/ 00060/2017) was obtained from the Institutional Review Board of Korle Bu Teaching Hospital, Accra, Ghana. An informed consent was obtained from parents or legal guardians for each minor participant prior to enrolment to participate in the study. All procedures performed involving study participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Consent for publication
Not applicable.
Availability of data and materials The data sets used and analysed during this study are available with the corresponding author on request.

Competing interest
All authors declare that there are no competing interests. Authors' contributions AKAA contributed to conception and study design, acquisition of data, analysis and interpretation of data and drafting of the manuscript. AK, BG, LR contributed to study design, interpretation of data and substantively revised it. SO contributed to acquisition of data. SA contributed to laboratory testing. YA contributed to analysis of data and interpretation of data. AEY contributed to analysis of data, interpretation of data and substantively revised it.
KWCS contributed to conception and study design, interpretation of data and substantively revised it. All authors read and approved the nal manuscript.

Acknowledgments
The

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.