High drug resistance levels could compromise the control of HIV infection in paediatric and adolescent population in Kinshasa, the Democratic Republic of Congo

Background The inadequacy of HIV viraemia and resistance monitoring in Africa leads to uncontrolled circulation of HIV strains with drug resistance mutations (DRM), compromising antiretroviral therapy (ART) effectiveness. This study describes the DRM prevalence and its therapeutic impact in HIV-infected pediatric patients from Kinshasa (Democratic Republic of Congo, DRC). Methods From 2016–2018, dried blood were collected from 71 HIV-infected children and adolescents under ART in two hospitals in Kinshasa for HIV-1 DRM pol analysis, predicted ARV-susceptibility by Stanford and phylogenetic characterization. Results HIV-1 sequences were recovered from 55 children/adolescents with 14 years of median-age. All had received nucleoside and non-nucleoside reverse transcriptase inhibitors (NRTI, NNRTI), 9.1% protease inhibitors (PI) and only one integrase inhibitor (INI). Despite the use of ART, 89.1% showed virological failure and 67.3% carried viruses with major-DRM to one (12.7%), two (47.3%), or three (5.5%) ARV-families. Most children/adolescents harbored DRM to NNRTI (73.5%) or NRTI (61.2%). Major-DRM to PI was present in 8.3% and minor-DRM to INI in 15%. Dual-class-NRTI+NNRTI resistance appeared in 53.1% of patients. Viruses presented high/intermediate resistance to nevirapine (72.9% patients), efavirenz (70.9%), emtricitabine/lamivudine (47.9%), rilpivirine (41.7%), etravirine (39.6%), doravidine (33.3%), zidovudine (22.9%), among others. Most participants were susceptible to INI and PI. Great diversity of variants was found, with a high rate (40%) of unique recombinants. Conclusion The high DRM prevalence observed among HIV-infected children and adolescents in Kinshasa could compromise the 95-95-95-UNAIDS targets in the DRC. It also reinforces the need for routine resistance monitoring for optimal rescue therapy election in this vulnerable population to control the spread of resistant HIV in the country.


Introduction
In the Democratic Republic of Congo (DRC), the coverage of people living with HIV receiving ART was 57% in 2018, but only 25% among children [9], although more than 100,000 children and adolescents are HIV-infected in the DRC. Furthermore, viraemia and HIVDR monitoring for ART optimization are absent during the clinical routine in all HIVinfected subjects. It can lead to late treatment failures identification due to resistant viruses, to empirical switches of ART regimens with recycled NRTIs, and to resistant viruses spreading. Without correct HIV monitoring, patients may spend months or even years on a failing ARV regimen, resulting in DRM accumulation and increased rates of morbidity and mortality [13][14][15][16]. Accordingly, the Joint United Nations Program on HIV/AIDS (UNAIDS) and the WHO reinforced the importance of HIVDR monitoring to control the HIV epidemic [3,4,17], mainly in key populations such as infants and adolescents.
For all these reasons, the aim of this study is to analyze HIVDR prevalence and its therapeutical impact on the children and adolescent population from Kinshasa (DRC). To find out the current situation of these populations in the DRC is particularly important due to their lifelong-treatment with more years under ART than adults.

Statistics analysis
Descriptive statistical analysis was performed calculating mean, median, and interquartile range (IQR). The percentage of virus carrying DRM was calculated with 95% confidence intervals (CI). The statistical significance was calculated using a T-Student or Mann-Whitney U tests. P-values <0.05 were considered statistically significant. Statistical analyses were conducted using GraphPad Prism version 8.0.1 (San Diego, CA, USA).

HIV-1 variant characterization
Nucleotide sequences were translated and aligned using the ClustalW algorithm implemented in MEGA6. For HIV-1 variant characterization we used reference sequences available in Los Alamos HIV Sequence Database (http://www.hiv.lanl.gov). Phylogenetic trees (PhyML tree) were reconstructed by maximum-likelihood (ML) method [21] using the general time reversible plus proportion of invariable sites plus gamma distribution parameter (GTR+I+G) evolutionary model. For estimating the bootstrap values on the inferred PhyML tree topology, Shimodaira-Hasegawa test using FastTree program was used (support >90%) [22]. Sequences not clustering with any known subtype or circulating recombinant form (CRF) were analyzed using Recombination Detection Program (RDP3v4.13) [23], identifying the subtypes involved in eventual recombination events and hypothetical recombination breakpoints. To further confirm the detected putative recombination events, new phylogenetic analyses were performed using the sequence fragments assigned to different subtypes according to the proposed breakpoint position(s) defined by RPD3. In the positive cases, the recombinant sequences were redefined as unique recombinant forms (URFs). The remaining cases were denoted as unique unclassified (U) variants.

Ethical aspects
The project was approved by the Human Subjects Review Committees at Monkole Hospital/ University of Kinshasa (DRC) and University Hospital Ramón y Cajal (Madrid, Spain). Informed consent was obtained from parents or guardians of enrolled participants. Children and adolescents also provided assent after parental consent when they could understand the meaning of participation in the study. All methods were carried out according to relevant guidelines and regulations.

Study population
DBS samples were collected between 2016 and 2018 from 71 HIV-positive subjects (34 children, 37 adolescents) under HIV care and ART in paediatric units of Monkole and Kalembelembe Hospitals. Among them, 85.5% of children/adolescents presented viraemia higher than 1,000cp/ml at sampling. HIV-1 sequences could be recovered from 55 (77.5%) patients in, at least, one genotype region for DRM analysis. ( Table 1). Table 1 describes the main features of the study population. The median age at diagnosis/ first ART of the 55 subjects with a sequence was 6 [IQR:2-12]/7 [IQR:3.5-12] years. The median age at study enrolment was 14 [IQR: [11][12][13][14][15][16] years. All of them were under ART at sampling and had received NRTI and NNRTI, 9.1% were PI-experienced, and only one adolescent (1.8%) had received INI.

HIVDR among children and adolescents on ART
Of comparing the rate of infections with resistant viruses carrying major DRM in children (0-14 years) vs. adolescents (15-21 years) under study, no significant difference was observed (S1 Table).  Table. https://doi.org/10.1371/journal.pone.0248835.g002 Infections with resistant viruses were associated with ART exposure time. DRMs to all ARV families were higher in those patients having been under ART for more than 5 years (Fig 3A). The rate of children and adolescents carrying viruses resistant to NRTI was 2 times higher in those having been under ART for more than 5 years (OR = 3.75; 95% CI [1.22-11.34], p<0.05) (Fig 3A). Fig 3B shows the steadily tend to increase of infections with resistant viruses with ART exposure time.

Predicted ARV susceptibility
Most children and adolescents with available pol genotype were infected with viruses susceptible to PI (88.9%) and INI (87.5%). A high resistance level was observed in 70.8% subjects to nevirapine, in 64.6% to efavirenz, in 45.8% to emtricitabine, or lamivudine, 37.5% to rilpivirine, and at lower rates for other ARV, except for INI (Fig 4 and S1 Fig).
When considering both high or intermediate resistance level to NNRTIs, most (72.9% and 70.9%) subjects presented high and intermediate resistance to nevirapine and efavirenz, respectively, and 41.7% to rilpivirine, 39.6% to etravirine, and 33.3% to doravidine. For NRTI, half (47.9%) of the subjects carried viruses with high and intermediate resistance to lamivudine and emtricitabine, 22.9% to zidovudine, 18.8% to abacavir, and 14.6% to tenofovir. For PI, these rates were lower than 10% and absent for INI (Fig 4 and S1 Fig).

Discussion
This study represents the first characterization of HIV drug resistance among HIV-infected children and adolescents in the DRC. Moreover, it provides the most recent data of resistance to ARV drugs in Kinshasa and the DRC. While all published studies from DRC reported resistance data in naïve [25][26][27][28] or treated adults [25,29] with samples collected from 2002 to 2014, our study had analyzed samples collected from 2016-2018. In addition, the present study is the first reporting resistance data for INI in that country, which is appropriate before the expected broad implementation of dolutegravir implementation in the DRC.
In the DRC and other Sub-Saharan Africa countries, HIV monitoring is limited due to a weak Public Health System and administration. Molecular diagnosis and VL testing are poorly accessible, mainly centralized in a national laboratory or restricted to some specific private hospitals. Moreover, resistance monitoring is not routinely available in the DRC [30]. For these reasons, ART regimen changes are determined by ARV availability, sometimes restricted by stock-outs [31], toxicity, and secondary effects in patients. No antiretroviral stock out was reported in Monkole and Kalembelembe hospitals during the study period.
According to the published data in samples from ART-naïve adults collected from 2002-2014 in DRC, the reported pre-antiretroviral-treatment drug resistance (PDR), a predictor of treatment failure, ranged from 0%-2.4% to PI, from 0-18.3% to NRTI, and from 0-9.8% to NNRTI, and unknown for INI [25][26][27][28]. Our study in children and adolescents from Kinshasa revealed that 7 out of 10 participants with available sequences harbored DRM, 4 out of 10 double resistance to NRTI+NNRTI, and 1 in 10 triple resistance to NRTI+NNRTI+PI. Of note, 7 out of 10 subjects with DRM to NNRTI had intermediate or high-level resistance to efavirenz and nevirapine, ARVs involved in first-line ART in the DRC. The observed burden of major DRM to PI in treated pediatric population was substantially higher than that previously reported in HIV-infected treated adults in the country [25,26,29]. Although adherence was promoted by medical staff, as well as in group of support with peers, accordingly to local staff, a third of 55 ART-treated children and adolescents with available sequence presented unsuppressed viraemia in the absence of resistant viruses. It would suggest that adherence support is required in this pediatric collective [3].
Since 2013, the WHO has recommended using PI-based ART regimens for children. In 2018, the WHO formally encouraged the phase-out of NNRTIs across age groups, with the introduction of dolutegravir for children with approved dose. However, in 2017, globally nearly 77% of children were still receiving nevirapine in first-line ART due to limited supplies of child-friendly drug formulations [3,32]. Previous national guidelines from ART in the DRC recommended tenofovir + lamivudine + efavirenz as the first-line regimen for infected adolescents and adults, tenofovir + lamivudine + ritonavir-boosted lopinavir for second-line, and abacavir + dolutegravir + ritonavir-boosted darunavir for the third-line ART. In infants, guidelines recommended abacavir + lamivudine + ritonavir-boosted lopinavir or efavirenz. Dolutegravir, included in 2019 in the first-line ART therapy in children, adolescents and adults on the DRC will probably improve viral load suppression, and it is expected to reduce mortality and HIV incidence compared to EFV-based regimens [3]. However, it is important to monitor their implementation and to forecast the development of DRM [33]. INI treatment would therefore be a good alternative to NNRTIs due to the high presence of DRM to this ARV family, as we observed in our study. Moreover, the absence of major DRM to INI in the study population (never exposed to this drug family, except in one case), would strongly support the implementation of new dolutegravir-based treatment in the DRC. However, the circulation of viruses with minor DRM to INI residues (observed in 15% of children/adolescents) could affect INI susceptibility in combination with other substitutions [34]. If the presence of M184I/V or K65R (DRM to NRTI) could prevent dolutegravir resistance as previously suggested [35], it should be explored further. As expected, the great diversity of HIV-1 variants in the country [36] is reflected in our study cohort, with a high rate of URF.
A limitation of this study is that samples were collected for ART treated subjects in two hospitals in Kinshasa, and it could not be representative of the situation on a city or country-wide level. However, Monkole and Kalembelembe hospitals are considered in the top 5 of clinical centers with more HIV-infected children and adolescents under ART in Kinshasa. Further analyses are required to monitor the current transmission of drug-resistant strains in ARTnaive HIV-infected children and adolescents in the DRC.
The sample size, although modest, was similar to other related resistance studies among treated patients in the DRC, analyzing from 55 [29] to 93 [25] pol sequences. Furthermore, the presented data are the first resistance information available for children and adolescents under ART in the country. Previous reports showed that HIVDR in naïve adults was very low in the DRC before the year 2007 [26,27], although some TDR in some RT residues reached prevalence higher than 10% in samples collected in Kinshasa during 2013-2014 [28]. Understanding levels of DRM prior to treatment initiation is particularly important in children because they have higher viraemias and faster disease progression compared to adults [37]. However, it would be necessary to know the TDR rate in ART-naïve children and adolescents in the city and country. Unfortunately, we cannot estimate the rate of resistant viruses which could have been transmitted at first HIV infection in our study cohort. The third limitation was the lack of complete information regarding HIV-status, treatment, and resistance data from all their mothers, which could have identified cases of vertical DRM transmission. The last limitation would be the absence of resistance data in 16 ART-treated participants with negative pol amplifications. It could be explained by the low viraemias reported in some of them due to ART control. Moreover, we cannot exclude that some negative PCR could be due to the high viral genetic variability in viral targets for primers used for pol amplication. A recent study reported the temporal trends of HIV-1 subtypes and recombinants in the DRC during a 43-year period (1976-2018) and in Kinshasa from 1983 to 2018, showing a high number of different HIV-1 variants currently circulating in that city, with an increase of complex and unique recombinants in the last years [38].
Multiple studies have demonstrated that the percentage of patients with drug resistance goes up steadily as time on treatment increases [25,39], as we have observed in the current study. Since children and adolescents are patients with a lifelong-treatment with more years under ART than adults, the study of DRM is particularly important in pediatric populations.
Thus, all generated results were communicated to clinicians in Kinshasa for better care of pediatric patients under study, providing useful data for ART regimen optimization. This work provides unique information related to these vulnerable populations in Kinshasa and in the rest of the country, where nearly 3 million children and adolescents are HIV infected without VL and resistance monitoring. We also reinforced the use of DBS as field-friendly and an useful specimen to carry out resistance analysis in low-middle income countries as the WHO recommends in the absence of plasma [3].
DRM monitoring is also crucial to control infection in these countries, where most paediatric HIV occur. Also, it is important to reinforce adherence support, and to implement routine VL quantification and resistance testing in the DRC national guidelines, strengthening the country's laboratory services. Our findings also support alternative ART regimens based on PI and INI instead of RTI in HIV-infected children and adolescent population in that country. These improvements would help to control the spread of resistant viruses among new HIV infections, getting the DRC closer to WHO's 95-95-95 targets [40].