Virological response and resistances over 12 months among HIV-infected children less than two years receiving first-line lopinavir/ritonavir-based antiretroviral therapy in Cote d’Ivoire and Burkina Faso: the MONOD ANRS 12206 cohort

Abstract Introduction: Lopinavir/ritonavir-based antiretroviral therapy (ART) is recommended for all HIV-infected children less than three years. However, little is known about its field implementation and effectiveness in West Africa. We assessed the 12-month response to lopinavir/ritonavir-based antiretroviral therapy in a cohort of West African children treated before the age of two years. Methods: HIV-1-infected, ART-naive except for a prevention of mother-to-child transmission (PMTCT), tuberculosis-free, and less than two years of age children with parent’s consent were enrolled in a 12-month prospective therapeutic cohort with lopinavir/ritonavir ART and cotrimoxazole prophylaxis in Ouagadougou and Abidjan. Virological suppression (VS) at 12 months (viral load [VL] <500 copies/mL) and its correlates were assessed. Results: Between May 2011 and January 2013, 156 children initiated ART at a median age of 13.9 months (interquartile range: 7.8–18.4); 63% were from Abidjan; 53% were girls; 37% were not exposed to any PMTCT intervention or maternal ART; mother was the main caregiver in 81%; 61% were classified World Health Organization Stage 3 to 4. After 12 months on ART, 11 children had died (7%), 5 were lost-to-follow-up/withdrew (3%), and VS was achieved in 109: 70% of children enrolled and 78% of those followed-up. When adjusting for country and gender, the access to tap water at home versus none (adjusted odds ratio (aOR): 2.75, 95% confidence interval (CI): 1.09–6.94), the mother as the main caregiver versus the father (aOR: 2.82, 95% CI: 1.03–7.71), and the increase of CD4 percentage greater than 10% between inclusion and 6 months versus <10% (aOR: 2.55, 95% CI: 1.05–6.18) were significantly associated with a higher rate of VS. At 12 months, 28 out of 29 children with VL ≥1000 copies/mL had a resistance genotype test: 21 (75%) had ≥1 antiretroviral (ARV) resistance (61% to lamivudine, 29% to efavirenz, and 4% to zidovudine and lopinavir/ritonavir), of which 11 (52%) existed before ART initiation. Conclusions: Twelve-month VS rate on lopinavir/ritonavir-based ART was high, comparable to those in Africa or high-income countries. The father as the main child caregiver and lack of access to tap water are risk factors for viral failure and justify a special caution to improve adherence in these easy-to-identify situations before ART initiation. Public health challenges remain to optimize outcomes in children with earlier ART initiation in West Africa.


Introduction
By the end of 2013, there were 3.2 million (2.9-3.5 million) children younger than 15 years living with HIV globally. Among these HIV-infected children, 2.9 million (2.6-3.2 million) were living in Sub-Saharan Africa [1]. Without access to antiretroviral therapy (ART), more than 50% will die by the age of two years [2].
In 2008, the children with HIV early antiretroviral therapy (CHER) trial demonstrated the efficacy of early ART in reducing mortality in HIV-infected infants treated before 12 weeks of age compared to those treated according to the 2006 World Health Organization (WHO) guidelines: 4% versus 16%, respectively [3]. Hence, the WHO recommended the initiation of ART in all HIV-infected children before the age of 12 months, regardless of their disease progression [4]; these recommendations were extended to the age of two years in 2010 [5] and then five years in 2013 [6]. Furthermore, systematic early infant diagnosis of HIV was recommended for all HIV-exposed children from the age of six weeks to initiate early ART in those diagnosed as HIV infected [7]. Despite these recommendations, early ART initiation in HIV-infected children remains challenging particularly in West Africa [8,9], and this is mainly due to a late infant diagnosis among HIV-exposed infants [10][11][12]. In 2013, in Sub-Sahara Africa, only 24% (22-26%) of eligible HIV-infected children were receiving ART [1].
Regardless of previous exposure to prevention of motherto-child transmission (PMTCT) of HIV, studies have shown that treatment failure at 24 weeks is more likely in children starting nevirapine-based ART than in those starting lopinavir/ritonavir (LPV/r)-based ART [3,13]. Thus, the WHO 2013 guidelines recommended that all children below the age of three years start LPV/r-based ART when feasible [6]. However, the availability of LPV/r remains a challenge in low-income countries [14]. The available oral syrup formulation of LPV/r in infants has cold chain requirements. The syrup is unpalatable, with the potential for suboptimal adherence. Since 2015, oral pellets received approval for use in children and do not require a cold chain; however, their effectiveness still needs to be assessed in Africa [15]. However, little is known about the field implementation and effectiveness of starting a LPV/r-based regimen in eligible HIV-infected children, especially in West Africa. In 2010, we implemented a randomized clinical trial to assess a simplification strategy of a LPV/r-based ART started in children younger than two years for at least 12 months [9]. In this study, we assessed the 12-month clinical, immunological, and virological response to LPV/r-based ART over the pre-randomization prospective cohort of children treated before the age of two years in Ouagadougou, Burkina Faso, and Abidjan, Côte d'Ivoire, and investigated its correlates.

Study design
The MONOD-ANRS-12206 study is an international, noninferiority, open-label, phase 3, randomized clinical trial conducted in Ouagadougou, Burkina-Faso, and Abidjan, Côte d'Ivoire (ClinicalTrial.gov registry number: NCT01127204). Study sites were the Abobo-Avocatier urban health clinic, the CePReF-enfant and the Yopougon and Cocody University Hospitals in Abidjan, and the Yalgado Ouédraogo and the Charles de Gaulle University Hospitals in Ouagadougou [9,16]. The present study refers to the initial 12-month ART therapeutical cohort conducted in the pre-randomization phase. The protocol was approved by the "Comité d'Ethique pour la Recherche en Santé du Burkina Faso", and the "Comité National d'Ethique et de la Recherche en Côte d'Ivoire".

Participants
All HIV-1-infected children (confirmed by a positive HIV-1DNA polymerase chain reaction (PCR) on blood sample), less than two years old, ART naive except for PMTCT exposure, and whose parents agreed to participate in the MONOD ANRS-12206 project were enrolled in a therapeutic cohort between May 2011 and February 2013 in the participating healthcare facilities. All children enrolled started ART with abacavir (ABC, 20 mg/mL, 8 mg/kg every 12 h) or zidovudine (ZDV, 10 mg/mL, 4 mg/kg every 12 h), lamivudine (3TC, syrup 10 mg/mL, 4 mg/kg every 12 h), and LPV/r (syrup 80/20 mg/mL, 12 mg/kg every 12 h) with cotrimoxazole (sulfamethoxazole, 20 mg/kg + trimethoprim, 4 mg/kg once-daily) as prophylaxis against opportunistic infections. Comprehensive counselling regarding treatment adherence was provided to caregivers. Children with either known intolerance to one of the drugs, who co-infected with HIV-2, with haemoglobin less than 7 g/dL, with neutrophils less than 750/mm 3 , with creatinine above five times the normal threshold, or with aspartate aminotransferase or alanine aminotransferase above five times the normal threshold were excluded. Children with tuberculosis infection at inclusion received once-daily efavirenz (EFV)-based ART and were excluded from this study. The drugs were provided by the national AIDS programmes under the responsibility of the country coordinating centres in charge of supplies and qualification of the batches.

Sample size
Based on our sample size estimation to conduct the MONOD trial [16], we expected to recruit 146 children in virological success. As virological data were not available from the field yet at the time of our protocol, we expected a 90% suppression rate on LPV/r, according to the Yeni report 2008 [17]. Thus, assuming a 90% virological suppression rate among survivors at 12 months on LPV/r, we planned to recruit 162 children in the initial LPV/r-based cohort.

Procedures
Patient's data were collected prospectively using standardized questionnaires at their first contact, their pre-inclusion visit (four weeks prior to ART initiation), inclusion visit (at ART initiation), and monthly visits until 12 months. The pre-inclusion visit included informed consent, an interview to assess medical history (PMTCT exposure), a complete clinical examination (weight, height, and WHO clinical staging), tuberculosis screening (chest x-ray), a standard blood test (haematology, creatinine, urea, aspartate contre-transfert (AST), alanine-amino-transferase (ALT), total bilirubin, alkaline phosphatase (ALP), glucose, alkaline phosphatases, amylase, lipase, lipid assessment (total cholesterol, triglycerides, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and very-low-density lipoprotein (VLDL)), a CD4 lymphocyte subpopulation count (percentage and absolute count), and a qualitative DNA PCR to confirm the child's HIV status. In case of symptoms suggesting tuberculosis (prolonged fever, chronic cough, recent malnutrition, or failure of classic antibiotics for an infectious syndrome), the diagnosis was completed with a tuberculin skin test, gastric lavage on three consecutive days, and stool examination for Mycobacterium tuberculosis. CD4 count was performed using FACScan flow cytometer. HIV RNA viral loads were measured with real-time PCR using a commercial assay (Generic HIV Charge Virale, Biocentric, France). Blood samples were repeated at 3, 6, 9, and 12 months post-ART initiation and tested for HIV RNA, and a complete blood count was performed. CD4 and standard blood tests were repeated at 6 and 12 months post-ART initiation. Children were examined by a paediatrician monthly and whenever medically warranted. At each visit, when the drugs were given to families, therapeutic education was conducted systematically first by the assistant pharmacist and then by the social worker.

Statistical analysis
The main outcome measured was viral suppression (VS) defined as one single HIV-1 RNA <500 copies/mL after 12 months on ART. The 500 copies/mL threshold was chosen as this measurement was the validated threshold in both country laboratories in 2011. This single viral load measurement per patient was aimed to be operational and in line with the 2016 WHO recommendations [18]. We compared baseline demographic, clinical, and biological characteristics between countries where study took place using Student t-test for continuous or a Chi-square or Fisher's exact test for categorical variables. Using the Kaplan-Meier method, we estimated the cumulative probability of reaching the first viral load <500 copies/mL by 12 months on ART. Follow-up time was censored at 12 months of follow-up or at time of last visit for children lost-to-follow-up or date of death for children who died. Independent variables were characteristics at pre-enrolment and at ART initiation: study country, age groups, gender, first-line nucleoside reverse-transcriptase inhibitor (NRTI) backbone, caregiver, access to electricity, access to tap water and fridge, history of ART drug exposure, anthropometric z-scores (analysed with WHO Anthro Software, version 3.2.2, January 2011), WHO clinical stage, viral load, and the difference in CD4 percentage between ART initiation and 6 month post-ART (referred to as CD4 delta).
We analysed the correlates of VS at twelth month, using a multivariate logistic regression including the variables above which were associated in the univariate analysis with p < 0.25. Then, we conducted a stepwise descendant analysis. Variables were retained in the final model if significantly associated with VS (p < 0.05) or their exclusion led to changes in the estimated odds ratio of other covariates by more than 10%. Final results were adjusted for country and gender to account for differences between study countries. All analyses were performed using SAS 9.1.3 (Cary, NC, USA).

Results
Cohort profile and baseline characteristics Between May 2011 and January 2013, 226 children less than two years, first diagnosed as HIV-1 infected by HIV DNA PCR performed on dried blood spot, were referred to MONOD project clinics. Among these children, 65 (28%) did not meet inclusion criteria; 5 (2%) were co-infected with tuberculosis and initiated EFV -based therapy and thus excluded; and 156 (69%) initiated a LPV/r-based therapy and were included ( Figure 1). The full inclusion process is described elsewhere [9].
Characteristics of the 156 children who initiated LPV/rbased therapy are shown in Table 1. Overall, 53% were girls; the median age at ART initiation was 13.9 months (interquartile range (IQR): 7.8-18.4). Children in Ouagadougou were significantly older by two months than those in Abidjan. The primary caregiver was the biological mother for 81% of the children; the biological father was the primary caregiver for 10% and the legal guardian for the remaining 9%. The proportion cared for by the biological father or legal guardian was significantly higher in Ouagadougou compared to Abidjan. In Abidjan, children's families were significantly more likely to have tap water, electricity, and a fridge at home compared to those in Ouagadougou. Overall, 37% of the children were not exposed to any kind of PMTCT intervention or maternal ART, and this proportion was significantly higher in Ouagadougou compared to in Abidjan. At ART initiation, children from Ouagadougou had significantly more advanced HIV disease compared to those from Abidjan: they were more likely to be classified WHO clinical Stage 3 or 4, had lower anthropometric parameter z-scores, had lower CD4 percentage, and had higher viral load. The firstline NRTI backbone was ZDV-3TC for 91% of the children.
Among the 156 children included, the Kaplan-Meier cumulative probability of reaching a first VL measurement <500 copies/mL was 0.47 at 3 months, 0.81 at 6 months, 0.88 at 9 months, and 0.92 at 12 months. After 12 months on ART, the proportion of VS was 0.70 (95% confidence interval (CI): 0.62-0.77) of all children enrolled and 0.78 (95% CI: 0.71-0.85) of those followed-up. There was no statistically significant difference in VS between the two countries ( Table 2).
During the first 12 months on ART, 37 (22.4%) children had at least one hospitalization or clinical severe adverse event. These were more common in children in Ouagadougou compared to those in Abidjan (p = 0.04). Ten (6.4%) children changed antiretroviral drug (substituting zidovudine with abacavir) because of severe adverse effects linked to anaemia. There were no significant differences in daytime or night-time sleeping disorders, declared by caregivers, between the two study countries (p = 0.98). Neutropenia was the most commonly reported grade 3 or 4 biological adverse event and was significantly more frequent in Abidjan than in Ouagadougou (p = 0.02). Hyperbilirubinaemia was more frequent in Ouagadougou than in Abidjan (p = <0.01). There were no significant differences between both countries for other grade 3 or 4 biological adverse events ( Table 3).

Correlates of viral suppression
In univariate analyses, study country, sex, first-line NRTI backbone, access to tap water, having electricity and a fridge at home, the main caregiver, having disclosed the child's HIV status to the father, baseline, WHO clinical stage, WAZ, and the CD4 percentage delta between inclusion and six months on ART were all associated with VS at the 12month visit with a p-value of <0.25 (Table 4).
In the final analysis, adjusted for study country and sex, access to tap water at home (adjusted odds ratio (aOR): 2.75, 95% CI: 1.09-6.94), having the biological mother as the main caregiver (aOR: 2.82, 95% CI: 1.03-7.71), and a ≥10% increase of CD4 percentage within the first six months on ART (aOR: 2.55, 95% CI: 1.05-6.18) were all significantly associated with a higher rate of VS at 12 months on ART (Table 5).

Drug resistance
Among the 29 children with VL ≥1000 copies/mL at 12 months, 28 had genotypic resistance testing: 21 (75%) had a virus resistant to at least one drug and 11 (52%) existed prior to ART initiation and were classified as transmitted resistant virus (3 children were exposed to maternal PMTCT alone, 3 exposed to maternal ART initiation during breastfeeding, 2 exposed to both perinatal and postnatal ART, 2 unexposed, and one unknown). The most frequent resistance was to lamivudine (61%); 29% had resistance mutations to non-nucleoside reverse-transcriptase  inhibitors (NNRTIs) and 4% had resistance mutations to both zidovudine and LPV/r (Table 6).

Discussion
Our study reports for the first time in West African settings the 12-month response and its correlates to a LPV/r-based ART in a cohort of ART-naive HIV-1-infected children treated before two years of age. Despite a relatively late access to ART initiation, at an already advanced stage of HIV disease, high rates of VS can be achieved among young children who initiate protease-inhibitor-based ART. After 12 months on ART, 78% of those followed-up achieved VS. However, we also observed a high rate of HIV-related deaths reaching 7%, and occurring early after ART initiation, which we explain by the late access to ART. Access to tap water, having their biological mother as the primary caregiver, and an increase of CD4 percentage between ART initiation and six-month visit equal or greater than 10% were identified as correlates to the 12-month VS after ART initiation in this cohort. History of exposure to a PMTCT intervention was not associated with 12-month VS in this study. First, we observed high early mortality after ART initiation that has already been described in other cohorts of HIV-infected children initiating ART [19][20][21][22][23]. In a South African cohort, mortality was higher, reaching 14% at 39 weeks, mainly related to tuberculosis, while in our study children with tuberculosis at inclusion were excluded [20]. This high rate of mortality is related to the delay in ART initiation [8], which is mainly due to the difficulty of early infant diagnosis, as already observed in other studies conducted in low-income countries [24,25]. Young HIVinfected children are particularly vulnerable to rapid HIV disease progression and early mortality [2]. In our cohort, median age was 13.9 months, with advanced HIV disease, low CD4 percentage, and a high mean viral load level of 6.2 log/mL at ART initiation. This emphasizes the need to reduce more significantly the delay to ART initiation among HIV-infected children to optimize their survival outcome. Second, despite delayed access to ART, children experienced a high rate of VS reaching 78% among those surviving that was comparable to other settings, in Africa, as well as in high-income countries [26][27][28][29]. This rate of VS is also similar to the rate of 84% recorded in an LPV/r-treated cohort follow-up in South Africa [21] and of 73% recorded in an NNRTI-treated cohort in Uganda and Zimbabwe [23,27]. As reported in previous adult studies, rates of VS are expected to be higher in the "on-treatment analyses" than in the "intention-to-treat analyses" and also to increase as the viral RNA thresholds used to define suppression increases [30]. Again, earlier access to ART would probably have optimized this initial rate of VS. This highlights the fact that access to early infant diagnosis and early ART before the age of two years still remains challenging in real life in West African settings [31]. The scaling-up of recommended early ART remains a major public health challenge in resource-limited settings [10].
Third, at ART initiation, access to tap water at home was a determinant for VS, with a 2.75 higher rate at 12 months in our study. Lack of access to tap water at home reflects deprived socio-economic conditions, which could probably be the marker of a poor adherence leading to virological failure. We also showed the key role of the caregiver in the success of HIV-infected children treatment: the fact that the father or another person than the biological mother was the main child caregiver was associated with a significant lower VS rate compared to the mother. This result underlines that the mother has a leading role in child survival as already reported elsewhere [32]. Before ART initiation, access to tap water and not having the biological mother as the principal caregiver could therefore be two indicators of first-line ART failure, easy to verify, to strengthen their support to adherence and to monitor closely treatment response. We noticed in our study that the father is also more frequently in charge of the child in Ouagadougou than in Abidjan, and declared as the main caregiver, even if the mother is alive because in Burkina Faso the father is empowered to take the decisions about the children's healthcare seeking and treatments. Consequently, they also accompanied more frequently their child to the medical visit, as already reported in the inclusion process [9].
Fourth, our results also show good immune recovery, with a 10% increase in CD4 percentage, from 20% at baseline to 30% in median at 6 months. This result is comparable to those of other cohorts of children receiving ART in Africa [23,33,34]. We did not find an association between VS and factors found in other studies such as age, gender, previous exposures to PMTCT or to maternal ART, baseline WHO stage, CD4 percentage, or baseline viral load [23,[35][36][37]. But, we found that an increase of CD4 percentage between baseline and six months equal or greater than 10% doubled the odds of VS 12 months after ART initiation. This could be used as an early indicator to monitor the risk of virological failure in low-income countries where viral load is not widely available. We also found that LPV/r-based ART was well tolerated with few severe adverse events, most of which were clinical adverse events and hospitalizations mostly HIV related rather than drug related. We will further analyse adherence data collected using a four-day recall questionnaire of missed doses correlated to pharmacokinetic data to study the relationship with VS. Among    WAZ: weight-for-age z-score; HAZ: height-for-age z-score; ZDV: zidovudine; 3TC: lamivudine; ABC: abacavir; LPV/r: lopinavir/ritonavir; ART: antiretroviral therapy; PMTCT: prevention of mother-to-child transmission. children with virological failure, we recorded a high rate of resistance to at least one antiretroviral drug (75%), of whom 52% were acquired after ART initiation. This high rate is of particular concern in our setting where ART options are limited. This proportion is comparable with the pooled proportion of 90% of at least one detectable resistance mutation found in a systematic review of antiretroviral resistance data in children from developing countries [38]. As previously showed by other studies, we recorded a high rate of drug resistance to lamivudine and to NNRTIs [38][39][40][41]. As the preferred regimens for second line recommended by WHO for children older than 3 years is efavirenz-based ART [18,42], these children with resistance to NNRTIs will be at higher risk of treatment failure when switching to second-line regimens. We did not find any differences in VS between children with and without previous PMTCT exposure, and drug resistance to LPV/r was rare in our study. This reflects the high genetic barriers of LPV/r and its interest for first-line therapy in young children, consistent with others studies [29,43] and with WHO consolidated guidelines and recommendations [18]. Finally, the occurrence of triple-class resistance was exceptional in our study. Our study highlights the interest of a first-line LPV/r-based regimen in the context of multi-exposure to PMTCT and postnatal antiretroviral drugs in West African children that will be further investigated. In the context of early treatment for all HIV-infected children and increased risk of drug resistance with the duration of therapy [38,44,45], there is an urgent need to identify children not responding to ART as early as possible, to promote optimum adherence to ART in these children and their caregivers, to develop ART simplification strategies, and to develop new drugs and formulations to optimize the treatment of children with virological failure. Our study however presents some limitations. First, we included clinics from urban settings only, with probably a ART: antiretroviral therapy; OR: odds ratio; aOR: adjusted odds ratio; CI: confidence interval; Forced variables: country and gender; ZDV: zidovudine; 3TC: lamivudine; ABC: abacavir; LPV/r: lopinavir/ritonavir; PMTCT: prevention of mother-to-child transmission; WHO: World Health Organization; WAZ: weight-for-age z-score; HAZ: height-for-age z-score. better access to early infant diagnosis and ART compared to rural settings, where the delay in care could be even longer than that reported in our study. Second, the high rate of VS in our study was obtained in the context of clinical research involving all guarantees for optimal monitoring of the family, leading to an overestimation of the rate of VS compared to real-life routine ART programmes. Third, we were not able to report here the role of adherence in relation with the virological response. In the MONOD project, we used an adherence assessment survey based on the drug doses missed over the last four days before each visit; however, this method was found to be unreliable. Pharmacokinetic studies are ongoing and these data will be considered in, separate, subsequent analyses. Fourth, although the relatively small size of our cohort leads to a lack of statistical power in comparisons, we achieved very good follow-up quality with only 3% of children lost-tofollow-up or withdrawing. Despite these limitations, our findings are original and described for the first time a West African cohort of children treated with LPV/r-based therapy before the age of two years and documents the 12month response.
In conclusion, in this West African cohort, we show that initiating a first-line LPV/r-based ART before the age of two is feasible and appropriate and a VS rate of 78% is achieved by 12 months. We identified two correlates to low VS, probably linked to poor adherence when the mother was not the main caregiver or if the family had no access to tap water. These markers could be easily notified at ART initiation, and reinforced monitoring could be offered to improve adherence and VS. We also report that over the 2011-2013 period, challenges still remain for improving early access to ART in HIV-infected children in West Africa. An earlier access combined with targeted interventions for those at risk for treatment failure will be helpful in optimizing the VS rate in West Africa and tend towards the 90% expected. Nevertheless, the development of a protease-inhibitor formulation easier to handle and more baby-friendly combined formulations remain also critical to achieve this 90% VS rate goal in this population [46].