Abstract
Background:
Introduction of pediatric-friendly dolutegravir (DTG) formulations to treat children living with HIV (CLHIV) has presented new opportunities to improve viral suppression; however, there are limited data on emerging patterns of HIV drug resistance (HIVDR) in this population.
Methods:
We identified CLHIV aged 0–15 years experiencing virologic failure on first-line antiretroviral therapy in Uganda between January 2021 and November 2021 from the Central Public Health Laboratory database. Remnant viral load samples were tested for HIVDR using Sanger sequencing. Drug resistance mutations (DRMs) were identified, and antiretroviral susceptibility was interpreted using the Stanford University HIVDR Database algorithms. The prevalence of DRMs and HIVDR is summarized using descriptive statistics, and associations with demographic and clinical factors were determined using χ2 tests.
Results:
Among 333 successfully genotyped patient samples, 122 (36.6%) CLHIV were on a DTG-based regimen and 136 (40.8%) on a lopinavir-based regimen. Among all CLHIV, 271 (81.4%) had DRMs and 262 (78.7%) had any intermediate-to-high level HIVDR [including DTG (1.8%), lopinavir (9.6%), abacavir (ABC) (25.5%), and zidovudine (19.2%)]. Among CLHIV on a DTG-based regimen, 6.6% had any integrase inhibitor DRMs and 4.1% had intermediate-to-high resistance to DTG. Children on DTG-based regimens were less likely to have DRMs (P = 0.007) or HIVDR (P = 0.02) compared with all other regimens. CLHIV on DTG-based regimens with an ABC backbone were no more likely to have DRMs (P = 0.9) or HIVDR (P = 0.5) compared with those with a backbone not containing ABC.
Conclusions:
Low rates of DTG resistance among Ugandan CLHIV failing first-line DTG-based antiretroviral therapy regimens underscore the durability of DTG. Lack of difference in ABC resistance across all regimens supports maintaining an ABC backbone for CLHIV <30 kg when transitioning to a DTG-based regimen despite virologic failure.
Keywords: virological failure, viraemia, HIV drug resistance, children living with HIV, dolutegravir
INTRODUCTION
Globally, in 2023, there were an estimated 1.4 million children living with HIV (CLHIV) 0–14 years, with only 57% receiving antiretroviral therapy (ART), and only 84% of these being virally suppressed, compared with 94% of adults.1 In Uganda in 2023, there were an estimated 72,000 CLHIV this age, with only 83% diagnosed, 92% of those on treatment, and only 82% of those being virally suppressed.2
Historically, inadequate viral suppression (VS) rates among CLHIV have been driven by the widespread use of suboptimal antiretrovirals (ARVs), including nonnucleoside reverse transcriptase inhibitors (NNRTIs), poor palatability of pediatric formulations, limited access to optimal ART, a lack of pediatric-trained clinicians, and notably high pretreatment drug resistance and acquired drug resistance in this population.3–6 For instance, a 2017 systematic review found 42.7% of CLHIV exposed to ARVs in utero or infancy for the prevention of mother-to-child transmission had pretreatment drug resistance, consistent with smaller more recent analysis.4–6 In Uganda, acquired drug resistance data from 2019, when NNRTI-based ART was most prevalent, revealed that 70.6% of CLHIV on first-line ART with a viral load (VL) ≥1000 copies/mL had any nucleoside reverse transcriptase inhibitors (NRTI) resistance and 78.4% had resistance to the NNRTIs efavirenz (EFV) or nevirapine (NVP).7
In 2018, the World Health Organization (WHO) updated global guidance, and dolutegravir (DTG)–based regimens became the preferred first-line regimen for all CLHIV. At the time of that update, however, there were no pediatric DTG formulations approved for CLHIV weighing <20 kg. Subsequently, dispersible pediatric DTG tablets were approved in 2020 and introduced globally in 2021.3 In resource-limited settings, national HIV treatment guidelines for selecting first-line or second-line regimens are typically based on WHO’s global treatment recommendations rather than individual HIV drug resistance (HIVDR) results due in part to the cost and availability of this testing and historically limited availability of pediatric-friendly formulations. Although HIVDR data are limited among CLHIV in the DTG treatment era, it is imperative to understand existing and emerging HIVDR to ensure appropriate management of CLHIV.3 We report HIVDR outcomes among Ugandan CLHIV failing on first-line ART in 2021 who received HIVDR testing after virologic failure (VF).
METHODS
Data for this analysis were obtained following the 2020 revision of the Uganda HIV Prevention, Care, and Treatment Guidelines that recommended HIVDR testing for all children and adolescents living with HIV failing ART. In 2021, the Uganda Ministry of Health conducted standard-of-care HIVDR testing for CLHIV failing first-line ART to optimize treatment outcomes, through the support of the CDC-funded Faith-based Action for Scaling up Testing and Treatment for the Epidemic Response (FASTER) project. FASTER was implemented between 2019 and 2021 by Catholic Relief Services, in partnership with the Clinton Health Access Initiative, to support ministries of health in 4 countries (Uganda, Nigeria, Tanzania, and Zambia) to improve testing and treatment of CLHIV.8
Data for this analysis came from previously collected VL samples from children 15 years or younger on first-line ART enrolled in any ART-accredited health facility in Uganda whose VL sample satisfied the following criteria: (1) sample tested for VL between January and November 2021, (2) sample type was plasma, and (3) VF defined as VL ≥1000 copies/mL for children on EFV or NVP or repeat VL >1000 copies/mL for children on lopinavir/ritonavir (LPV/r) or DTG. Repeat VL refers to a VL test taken after a child with an initial unsuppressed VL has undergone intensified adherence support for at least 3 months with 3 consecutive good adherence scores. Samples from CLHIV whose prior regimen changes were made for treatment simplification or due to policy changes and not VF were classified as being on first-line ART.
A paper-based VL and HIVDR request form completed at sample collection contained details such as demographics, ART history, clinical condition, sample type, and collection date for each sample. At the time of this evaluation, all VL testing was performed centrally at the Central Public Health Laboratory. The Central Public Health Laboratory had an electronic VL laboratory information system (LIMS) where the details from the VL and HIVDR request form were entered along with VL testing results, once available. LIMS data were used to filter remnant VL samples eligible for HIVDR testing based on the above inclusion criteria.
The eligible VL samples were then automatically identified by the HIV Drug Resistance Database (HIVDR Database). The HIVDR Database acts as the relay database for cross-laboratory systems interoperability. The database received eligible sample profiles from the VL LIMS and provided a user interface for referring these samples to the HIVDR laboratories. Once a sample fulfilling the inclusion criteria above was identified, it was retrieved and sent to the HIVDR laboratory at Joint Clinical Research Centre to be tested using Sanger sequencing technology if there was sufficient sample volume. Once a sample was referred to the HIVDR laboratory, clinicians received email notifications from the HIVDR Database to upload their clients’ social, psychosocial, and medical profiles into the database. When the HIVDR results were available, the HIVDR laboratory systems automatically pushed the results into the HIVDR Database, from which clinicians and experts could review the results and make ART recommendations. Once individual treatment decisions were made and documented, these clinical recommendations were also uploaded into the HIVDR database.
The data presented in this work were therefore a combination of the routine data inputs by clinicians, VL and HIVDR laboratories, and drug resistance review committees, all aggregated within the HIVDR Database. These data were stored in a relational database and were extracted, cleaned, and transformed using the Structured Query Language and exported as a Comma Separated Value database for further cleaning and analysis.
The analytic data set of demographic data and HIVDR information for sampled specimens were retrieved from the HIVDR Database in March 2022. HIV drug resistance mutations (DRMs) and drug susceptibility profiles were generated using Stanford University’s HIVdb algorithm (version 9.0) (https://hivdb.stanford.edu/hivdb). All analyses were conducted using Stata 18 statistical software (StataCorp LLC, College Station, TX). Frequencies of DRMs were calculated overall and stratified by drug class of the mutation, including mutations against NRTIs, NNRTIs, protease inhibitors (PIs), and integrase inhibitors (INSTIs). The frequencies of individual DRMs were also calculated overall and stratified by regimen at sample collection. Mutations and resistance were classified as non-NNRTI if they were identified in any of the following drug classes: NRTIs, PIs, and/or INSTIs. Individuals were considered to have clinically significant HIVDR if their drug susceptibility profile revealed intermediate-to-high level drug resistance to one or more of the following individual ARVs: DTG, abacavir (ABC), zidovudine (AZT), tenofovir disoproxil fumarate (TDF), lamivudine (3TC), LPV/r, darunavir (DRV), EFV, or NVP. All calculations were estimated overall, and stratified by regimen at blood sample collection, and by previous exposure to the child’s known historical treatment regimens.
To determine possible correlates of DRMs and HIVDR, associations with various demographic and clinical factors were determined using χ2 tests. P-values < 0.05 were considered statistically significant. The outcomes of interest were the presence of any DRMs and the presence of intermediate-to-high level drug resistance to any non-NNRTI ARVs. Demographic and clinical variables explored included sex, age category, duration of time on ART, reported level of adherence to ART (good, fair, or poor), VL category at sample collection (defined as VL <10,000, VL 10,000–100,000, and VL >100,000), and regimen at sample collection. A standardized behavioral questionnaire was completed for participants as part of their clinical care; correlates of DRMs and drug susceptibility were also explored for a variety of social susceptibility factors. To better characterize children on optimized ART, demographic factors were described for children on DTG-based regimens by their specific NRTI backbone. Among these children, the presence of DRMs and non-NNRTI drug resistance was compared between children on DTG + ABC regimens compared with other DTG-based regimens using χ2 tests.
Frequencies were also calculated for the clinical recommendations reported after HIVDR testing, both overall and stratified by drug class of DRMs. The potential clinical recommendations included maintaining on current regimen, substituting with another ARV within the first-line, switching to another drug class or second-line regimen, or pending if clinical recommendations were not noted.
This activity was reviewed by CDC, deemed not research, and was conducted consistent with applicable federal law and CDC policy (see e.g., 45 C.F.R. part 46, 21 C.F.R. part 56; 42 U.S.C. §241(d); 5 U.S.C. §552a; and 44 U.S.C. §3501 et seq.).
RESULTS
Between January and November 2021, 648 VL samples from 164 sites met the inclusion criteria. A total of 428 (66.0%) samples were retrieved from the sample biorepository and sent to the HIVDR laboratory for testing; 128 (19.8%) did not meet the Sanger testing criteria due to low sample volume or other factors. Of the 428 samples referred for HIVDR testing, 333 (77.8%) samples were successfully tested while 95 (22.2%) samples failed to amplify.
Among the 333 CLHIV with successful sample amplification, 61 (18.3%) were aged 1–4 years, 79 (23.7%) aged 5–9 years, 89 (26.7%) aged 10–12 years, and 104 (31.2%) aged 13–15 years (Table 1). There were 101 (30.3%) CLHIV taking ART for <5 years, 104 (31.2%) taking ART 5–9 years, and 128 (38.4%) taking ART 10 or more years. At the time of sample collection, 122 (36.6%) CLHIV were taking a DTG-based regimen, 136 (40.8%) were taking an LPV/r-based regimen, and 74 (22.2%) were taking either an EFV-based or NVP-based regimen. Among the 122 CLHIV on DTG-based regimens, the median age was 13 years (interquartile range [IQR] = 11–14); 54 (44.2%) were on regimens with an ABC/3TC backbone, 65 (53.3%) had a TDF/3TC backbone, and 3 (2.5%) had an AZT/3TC backbone. The median age in years for these subgroups was 12 for ABC (IQR = 10–13), 13 for TDF (IQR = 12–14), and 9 for AZT (IQR = 2–10).
TABLE 1.
Demographic and Clinical Characteristics of Children in Uganda Aged 0–15 years with Virologic Failure on First-Line ART with an Eligible HIV Drug Resistance Profile, 2021
| Characteristic | Total (%) |
|---|---|
| Male | 172 (51.6) |
| Female | 156 (46.9) |
| Missing sex | 5 (1.5) |
| Age 1–4 | 61 (18.3) |
| Age 5–9 | 79 (23.7) |
| Age 10–12 | 89 (26.7) |
| Age 13–15 | 104 (31.2) |
| On ART<2 yrs | 11 (3.3) |
| On ART 2–4 yrs | 90 (27.1) |
| On ART 5–9 yrs | 104 (31.2) |
| On ART 10+ yrs | 128 (38.4) |
| Good adherence* | 81 (24.3) |
| Fair adherence* | 60 (18.0) |
| Poor adherence* | 47 (14.1) |
| Missing adherence | 145 (43.5) |
| Viral load 1–10K† | 111 (33.3) |
| Viral load >10–100K† | 124 (37.2) |
| Viral load >100K† | 98 (29.4) |
| DTG regimen | 122 (36.6) |
| LPV regimen | 136 (40.8) |
| NVP regimen | 33 (9.9) |
| EFV regimen | 41 (12.3) |
| Other regimen | 1 (0.4) |
| Total | 333 |
As determined by a clinical questionnaire administered to caregivers.
At time of sample collection.
Art, Antiretroviral treatment; DTG, Dolutegravir; LPV, Lopinavir; NVP, Nevirapine; EFV, Efaviren
Of the 333 CLHIV, 271 (81.4%) had any DRMs (Table 2), 10 (3.0%) had any INSTI DRMs, 39 (11.7%) had any PI DRMs, 217 (65.2%) had any NRTI DRMs, and 225 (67.6%) had any NNRTI DRMs. Many CLHIV had DRMs against multiple classes of ARTs: 184 (55.3%) had DRMs against 2 classes and 36 (10.8%) had DRMs against 3 classes. These frequencies varied by regimen at sample collection (Table 2) and history of ARV exposure (Table 3). Of note, children on DTG at sample collection had lower frequencies of DRMs for all drug classes except INSTIs. Rates of DRMs against INSTIs among children on DTG were low (6.6%, 8/122), although were twice as common in these children when compared with the overall sample (3.0%, 10/333) and about 6 times more common when compared with CLHIV not on DTG at sample collection (0.95%, 10/333). The most common individual DRMs are presented in Table 4. The top 12 most common DRMs were NRTI or NNRTI mutations; 203 (61.0%) CLHIV had the most common NRTI mutation (M184V), and 143 (42.9%) had the most common NNRTI mutation (K103N). The two most common PI mutations were each found in 23 (6.9%) CLHIV (M46, V82), and 4 (1.2%) CLHIV had the most common INSTI mutation (R263).
TABLE 2.
Presence of Drug Resistance Mutations and Intermediate-to-High Level Drug Resistance by Antiretroviral Drug Class and Regimen at Sample Collection Among Children Aged 0–15 in Uganda with Virologic Failure on First-Line ART, 2021
| Total (%) | ARV Present in Regimen at Sample Collection | ||||||
|---|---|---|---|---|---|---|---|
| DTG (%) | ABC (%) | AZT (%) | TDF (%) | LPV (%) | Any NNRTI (%) | ||
| All participants | 333 | 122 | 198 | 53 | 77 | 136 | 74 |
| Presence of mutations | |||||||
| Any drug resistance mutations | 271 (81.4) | 90 (73.8) | 161 (81.3) | 46 (86.8) | 59 (76.6) | 118 (86.8) | 62 (83.8) |
| NRTI mutations | 217 (65.2) | 68 (55.7) | 132 (66.7) | 40 (75.5) | 42 (54.6) | 100 (73.5) | 48 (64.9) |
| NNRTI mutations | 225 (67.6) | 74 (60.7) | 133 (67.2) | 41 (77.4) | 46 (59.7) | 97 (71.3) | 53 (71.6) |
| PI mutations | 39 (11.7) | 8 (6.6) | 23 (11.6) | 7 (13.2) | 7 (9.1) | 20 (14.7) | 10 (13.5) |
| INSTI mutations | 10 (3.0) | 8 (6.6) | 5 (2.5) | 0 (0.0) | 5 (6.5) | 1 (0.7) | 1 (1.4) |
| Mutations against 2 classes | 184 (55.3) | 58 (47.5) | 111 (56.1) | 36 (67.9) | 34 (44.2) | 83 (61.0) | 42 (56.8) |
| Mutations against 2 non-NNRTI classes | 39 (11.7) | 10 (8.2) | 24 (12.1) | 6 (11.3) | 7 (9.1) | 20 (14.7) | 8 (10.8) |
| Mutations against 3 classes | 36 (10.8) | 10 (8.2) | 21 (10.6) | 6 (11.3) | 7 (9.1) | 17 (12.5) | 8 (10.8) |
| Intermediate-high ARV resistance | |||||||
| Any intermediate-high ARV resistance | 262 (78.7) | 85 (69.7) | 156 (78.8) | 45 (84.9) | 56 (72.7) | 115 (84.6) | 61 (82.4) |
| Any non-NNRTI resistance | 217 (65.2) | 70 (57.4) | 134 (67.7) | 39 (73.6) | 41 (53.3) | 97 (71.3) | 49 (66.2) |
| Resistance to DTG | 6 (1.8) | 5 (4.1) | 3 (1.5) | 0 (0.0) | 3 (3.9) | 1 (0.7) | 0 (0.0) |
| Resistance to ABC | 85 (25.5) | 26 (21.3) | 49 (24.8) | 18 (34.0) | 16 (20.8) | 37 (27.2) | 21 (28.4) |
| Resistance to AZT | 64 (19.2) | 22 (18.0) | 29 (14.7) | 17 (32.1) | 15 (19.5) | 19 (14.0) | 22 (29.7) |
| Resistance to TDF | 29 (8.7) | 8 (6.6) | 17 (8.6) | 7 (13.2) | 5 (6.5) | 12 (8.8) | 8 (10.8) |
| Resistance to 3TC | 203 (61.0) | 62 (50.8) | 128 (64.7) | 36 (67.9) | 37 (48.1) | 96 (70.6) | 44 (59.5) |
| Resistance to LPV | 32 (9.6) | 5 (4.1) | 22 (11.1) | 5 (9.4) | 3 (3.9) | 19 (14.0) | 7 (9.5) |
| Resistance to DRV | 6 (1.8) | 1 (0.8) | 2 (1.0) | 2 (3.8) | 2 (2.6) | 2 (1.5) | 3 (4.1) |
| Resistance to EFV | 220 (66.1) | 71 (58.2) | 129 (65.2) | 40 (75.5) | 46 (59.7) | 96 (70.6) | 52 (70.3) |
| Resistance to NVP | 223 (67.0) | 73 (59.8) | 129 (65.2) | 40 (75.5) | 49 (63.6) | 96 (70.6) | 53 (71.6) |
| Resistance against 2 classes | 180 (54.1) | 58 (47.5) | 109 (55.1) | 34 (64.2) | 34 (44.2) | 80 (58.8) | 41 (55.4) |
| Resistance against 2 non-NNRTI classes | 35 (10.5) | 8 (6.6) | 23 (11.6) | 5 (9.4) | 5 (6.5) | 19 (14.0) | 7 (9.5) |
| Resistance against 3 classes | 33 (9.9) | 8 (6.6) | 21 (10.6) | 5 (9.4) | 5 (6.5) | 17 (12.5) | 7 (9.5) |
TABLE 3.
Presence of Drug Resistance Mutations and Intermediate-to-High Level Drug Resistance by Antiretroviral Drug Class and History of Any Known Exposure to Each ARV Among Children Aged 0–15 in Uganda with Virologic Failure on First-Line ART, 2021
| Total (%) | Any Known Exposure to ARV in Current or Prior Regimens | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| DTG (%) | ABC (%) | AZT (%) | TDF (%) | LPV (%) | Any INSTI (%) | Any PI (%) | Any NNRTI (%) | ||
| All participants | 333 | 129 | 252 | 135 | 80 | 195 | 130 | 200 | 175 |
| Presence of mutations | |||||||||
| Any drug resistance mutations | 271 (81.4) | 95 (73.6) | 202 (80.2) | 112 (83.0) | 61 (76.3) | 168 (86.2) | 96 (73.9) | 173 (86.5) | 135 (77.1) |
| NRTI mutations | 217 (65.2) | 73 (56.6) | 161 (63.9) | 96 (71.1) | 44 (55.0) | 142 (72.8) | 74 (56.9) | 146 (73.0) | 107 (61.1) |
| NNRTI mutations | 225 (67.6) | 79 (61.2) | 167 (66.3) | 98 (72.6) | 48 (60.0) | 139 (71.3) | 80 (61.5) | 144 (72.0) | 115 (65.7) |
| PI mutations | 39 (11.7) | 8 (6.2) | 32 (12.7) | 17 (12.6) | 7 (8.8) | 28 (14.4) | 9 (6.9) | 29 (14.5) | 20 (11.4) |
| INSTI mutations | 10 (3.0) | 8 (6.2) | 7 (2.8) | 4 (3.0) | 5 (6.3) | 5 (2.6) | 8 (6.2) | 5 (2.5) | 6 (3.4) |
| Intermediate-high ARV resistance | |||||||||
| Any intermediate-high ARV resistance | 262 (78.7) | 90 (69.8) | 195 (77.4) | 109 (80.7) | 58 (72.5) | 164 (84.1) | 91 (70.0) | 169 (84.5) | 129 (73.7) |
| Any non-NNRTI resistance | 217 (65.2) | 75 (58.1) | 160 (63.5) | 94 (69.6) | 43 (53.8) | 140 (71.8) | 76 (58.4) | 144 (72.0) | 105 (60.0) |
| Resistance to DTG | 6 (1.8) | 5 (3.9) | 4 (1.6) | 2 (1.5) | 3 (3.8) | 3 (1.5) | 5 (3.9) | 3 (1.5) | 3 (1.7) |
| Resistance to ABC | 85 (25.5) | 27 (20.9) | 62 (24.6) | 40 (29.6) | 17 (21.3) | 58 (29.7) | 28 (21.5) | 59 (29.5) | 46 (26.3) |
| Resistance to AZT | 64 (19.2) | 23 (17.8) | 41 (16.3) | 39 (29.9) | 16 (20.0) | 38 (19.5) | 24 (18.5) | 38 (19.0) | 44 (25.1) |
| Resistance to TDF | 29 (8.7) | 9 (7.0) | 21 (8.3) | 15 (11.1) | 6 (7.5) | 19 (9.7) | 9 (6.9) | 19 (9.5) | 18 (10.3) |
| Resistance to 3TC | 203 (61.0) | 67 (51.9) | 153 (60.7) | 87 (64.4) | 39 (48.8) | 137 (70.3) | 68 (52.3) | 141 (70.5) | 96 (54.9) |
| Resistance to LPV | 32 (9.6) | 5 (3.9) | 27 (10.7) | 13 (9.6) | 3 (3.8) | 26 (13.3) | 6 (4.6) | 27 (13.5) | 14 (8.0) |
| Resistance to DRV | 6 (1.8) | 1 (0.8) | 5 (2.0) | 5 (3.7) | 2 (2.5) | 5 (2.6) | 1 (0.8) | 5 (2.5) | 5 (2.9) |
| Resistance to ATV | 25 (7.5) | 5 (3.9) | 21 (8.3) | 11 (8.2) | 3 (3.8) | 20 (10.3) | 6 (4.6) | 21 (10.5) | 12 (6.9) |
| Resistance to EFV | 220 (66.1) | 76 (58.9) | 164 (65.1) | 97 (71.9) | 48 (60.0) | 138 (70.8) | 77 (59.2) | 143 (71.5) | 113 (64.6) |
| Resistance to NVP | 223 (67.0) | 78 (60.5) | 164 (65.1) | 97 (71.9) | 51 (63.8) | 129 (73.7) | 79 (60.8) | 143 (71.5) | 114 (65.1) |
ARV, Antiretroviral; DTG, Dolutegravir; ABC, Abacavir; AZT, Zidovudine; TDF, Tenofovir; NNRTI, Non Nucleoside Reverse Transcriptase Inhibitor; NRTI, Nucleoside Reverse Transciptase Inhibitor; PI, Protease Inhibitor; INSTI, Integrase Stand Transfer Inhibitor; 3TC Lamivudine; DRV, Darunavir; ERV, Efavirenz; NVP, Nevirapine.
TABLE 4.
Frequencies of HIV Drug Mutations Overall and by Antiretroviral Drug Present in Regimen at Sample Collection Among Children Aged 0–15 in Uganda with Virologic Failure on First-Line ART, 2021 (n = 333)
| Mutation drug class | Mutation | Overall | ARV Present in Regimen at Sample Collection | |||
|---|---|---|---|---|---|---|
| DTG | ABC | AZT | LPV | |||
| NRTI | M184 | 203 | 62 | 128 | 36 | 137 |
| T215 | 57 | 21 | 25 | 15 | 34 | |
| M41 | 40 | 14 | 21 | 10 | 29 | |
| L74 | 34 | 11 | 25 | 4 | 25 | |
| K219 | 34 | 10 | 17 | 11 | 19 | |
| D67 | 28 | 10 | 16 | 8 | 15 | |
| K70 | 25 | 7 | 14 | 7 | 13 | |
| L210 | 22 | 9 | 10 | 4 | 14 | |
| Y115 | 19 | 6 | 14 | 2 | 14 | |
| V75 | 10 | 5 | 7 | 1 | 6 | |
| T69 | 7 | 2 | 2 | 1 | 3 | |
| K65 | 4 | 3 | 3 | |||
| F77 | 3 | 2 | 3 | 2 | ||
| F116 | 1 | 1 | 1 | |||
| Q151 | 1 | 1 | 1 | |||
| NNRTI | K103 | 143 | 43 | 93 | 23 | 92 |
| G190 | 55 | 13 | 33 | 11 | 35 | |
| Y181 | 47 | 16 | 21 | 12 | 27 | |
| V179 | 33 | 9 | 24 | 4 | 21 | |
| K101 | 29 | 8 | 14 | 5 | 15 | |
| P225 | 18 | 4 | 13 | 3 | 12 | |
| V106 | 12 | 5 | 6 | 3 | 7 | |
| Y188 | 12 | 5 | 8 | 10 | ||
| L100 | 7 | 2 | 5 | 5 | ||
| M230 | 6 | 1 | 2 | 4 | 5 | |
| PI | M46 | 23 | 6 | 12 | 5 | 15 |
| V82 | 23 | 4 | 18 | 2 | 19 | |
| I54 | 19 | 5 | 13 | 3 | 15 | |
| L76 | 12 | 4 | 6 | 3 | 7 | |
| I47 | 6 | 5 | 1 | 5 | ||
| I50 | 3 | 1 | 2 | 2 | ||
| I84 | 3 | 1 | 1 | 3 | ||
| L90 | 3 | 1 | 1 | 1 | ||
| V32 | 2 | 2 | 2 | |||
| N88 | 1 | 1 | ||||
| INSTI | R263 | 4 | 3 | 3 | 2 | |
| T66 | 2 | 2 | 2 | |||
| G118 | 2 | 2 | 1 | |||
| E138 | 2 | 2 | 1 | |||
| Y143 | 1 | 1 | ||||
| Q148 | 1 | 1 | 1 | |||
| N155 | 1 | 1 | 1 | |||
| Total | 333 | 122 | 198 | 53 | 136 | |
DTG, Dolutegravir; ABC, Abacavir; AZT, Zidovudine; LPV, Lopinavir; NNRTI, Non Nucleoside Reverse Transcriptase Inhibitor; NRTI, Nucleoside Reverse Transciptase Inhibitor; INSTI, Integrase Strand Transfer Inhibitor; PI, Protease Inhibitor.
Drug susceptibility analysis revealed that most CLHIV (78.7%, 262/333) had clinically significant resistance (intermediate-to-high level drug resistance) to at least 1 ARV, including DTG (1.8%), LPV (9.6%), DRV (1.8%), NVP (67.0%), EFV (66.1%), 3TC (61.0%), ABC (25.5%), and AZT (19.2%) (Table 2). Similar to DRMs, individual drug susceptibility varied by regimen at sample collection (Table 2) and history of ARV exposure (Table 3). CLHIV on DTG-based regimens at the time of sample collection had the lowest prevalence of any intermediate-to-high level drug resistance at 69.7%, with resistance to EFV (58.2%) and NVP (59.8%) most frequent (Table 2). Of the 122 CLHIV failing on DTG, intermediate-to-high-level resistance was lower for other ARVs, including for DTG (4.1%), LPV (4.1%), DRV (0.8%), ABC (21.3%), AZT (18%), and TDF (6.6%) (Table 2). Among 136 children failing on LPV/r, intermediate-to-high level resistance included LPV/r (14.0%), DTG (0.7%), DRV (1.5%), ABC (27.2%), AZT (14.0%), and TDF (8.8%) (Table 2). All those with AZT resistance were also resistant to ABC leaving only 18 (13.2%) of those failing LPV/r with intermediate-high level ABC resistance and without similar AZT resistance. On average, intermediate-to-high level resistance to a particular drug tended to be higher in CLHIV on a regimen containing that drug at sample collection compared with the level of resistance to that drug among all sampled CLHIV. For instance, resistance to DTG was approximately 8 times as high among CLHIV currently taking DTG (4.1%) compared with CLHIV not taking DTG (0.5%) [χ2 P = 0.02] and approximately 2 times higher to AZT (32.1% compared with 16.8%, P = 0.01) and LPV (14.0% compared with 6.6%, P = 0.02) for those on regimens containing these drugs, respectively. However, notably, this pattern was not seen for ABC (Table 2).
In bivariate analyses, similar patterns were seen for the presence of DRMs and the presence of non-NNRTI intermediate-to-high level drug resistance (Table 5). Overall, the youngest CLHIV were more likely to have DRMs (P = 0.02) and were more likely to have any non-NNRTI resistance (P = 0.01). CLHIV on DTG-based regimens at sample collection were less likely to have both DRMs (P = 0.007) or non-NNRTI resistance (P = 0.02) compared with those on non-DTG-based regimens, and CLHIV on LPV/r-based regimens were more likely to have both DRMs (P = 0.04) and non-NNRTI resistance (P = 0.05) compared with those on non-LPV/r-based regimens. Among CLHIV on a DTG-based regimen at sample collection, the 54 (44.3%) children with an ABC-containing backbone were no more likely to have clinically relevant mutations (P = 0.9) or non-NNRTI resistance (P = 0.5) compared with children on a DTG-based regimen with a different backbone. No relationship was seen for either DRMs or intermediate-to-high level resistance by sex, duration on ART, adherence score, VL at sample collection (Table 5), or most social susceptibility factors (see Table, Supplemental Digital Content, http://links.lww.com/QAI/C590).
TABLE 5.
Bivariate Correlates of HIV Drug Resistance (HIVDR) Mutations and NNRTI Intermediate-to-High Drug Resistance Among Children Aged 0–15 in Uganda with Virologic Failure on First-Line Antiretroviral Treatment, 2021
| Characteristic | Any HIVDR Mutation (%) | No HIVDR Mutation (%) | Chi P* | Any Non-NNRTI Resistance (%) | No Non-NNRTI Resistance (%) | Chi P* | Total |
|---|---|---|---|---|---|---|---|
| Male | 138 (80.2) | 34 (19.8) | 0.4 | 118 (68.6) | 54 (31.4) | 0.2 | 172 |
| Female | 131 (84.0) | 25 (16.0) | 97 (62.2) | 59 (37.8) | 156 | ||
| Age 1–4 | 57 (93.4) | 4 (6.6) | 0.02 | 45 (73.8) | 16 (26.2) | 0.01 | 61 |
| Age 5–9 | 65 (82.3) | 14 (17.7) | 57 (72.2) | 22 (27.8) | 79 | ||
| Age 10–12 | 72 (80.9) | 17 (19.1) | 60 (67.4) | 29 (32.6) | 89 | ||
| Age 13–15 | 77 (74.0) | 27 (26.0) | 55 (52.9) | 49 (47.1) | 104 | ||
| On ART <2 yrs | 9 (81.8) | 2 (18.2) | 0.7 | 6 (54.5) | 5 (45.5) | 0.4 | 11 |
| On ART 2–4 yrs | 75 (83.3) | 15 (16.7) | 60 (66.7) | 30 (33.3) | 90 | ||
| On ART 5–9 yrs | 81 (77.9) | 23 (22.1) | 62 (59.6) | 42 (40.4) | 104 | ||
| On ART 10+ yrs | 106 (82.8) | 22 (17.2) | 89 (69.5) | 39 (30.5) | 128 | ||
| Good adherence† | 66 (81.5) | 15 (18.5) | 0.8 | 39 (48.1) | 21 (25.9) | 0.6 | 81 |
| Fair adherence† | 49 (81.7) | 11 (18.3) | 55 (91.7) | 26 (43.3) | 60 | ||
| Poor adherence† | 36 (76.6) | 11 (23.4) | 28 (59.6) | 19 (40.4) | 47 | ||
| Viral load 1–10K‡ | 88 (79.3) | 23 (20.7) | 0.8 | 74 (66.7) | 37 (33.3) | 0.3 | 111 |
| Viral load 10–100K‡ | 103 (83.1) | 21 (16.9) | 85 (68.5) | 39 (31.5) | 124 | ||
| Viral load >100K‡ | 80 (82.6) | 18 (18.4) | 58 (59.2) | 40 (40.8) | 98 | ||
| DTG regimen‡ | 90 (73.8) | 32 (26.2) | 0.007 | 70 (57.4) | 52 (42.6) | 0.02 | 122 |
| Non-DTG regimen‡ | 181 (85.8) | 30 (14.2) | 147 (69.7) | 64 (30.3) | 211 | ||
| LPV regimen‡ | 118 (86.8) | 18 (13.2) | 0.04 | 97 (71.3) | 39 (28.7) | 0.05 | 136 |
| Non-LPV regimen‡ | 153 (77.7) | 44 (22.3) | 120 (60.9) | 77 (39.1) | 197 | ||
| NVP regimen‡ | 28 (84.8) | 5 (15.2) | 0.6 | 26 (78.8) | 7 (21.2) | 0.08 | 33 |
| Non-NVP regimen‡ | 243 (81.0) | 57 (19.0) | 191 (63.7) | 109 (36.3) | 300 | ||
| EFV regimen‡ | 34 (82.9) | 7 (17.1) | 0.8 | 23 (56.1) | 18 (43.9) | 0.2 | 41 |
| Non-EFV regimen‡ | 237 (81.2) | 55 (18.8) | 194 (66.4) | 98 (33.6) | 292 | ||
| Total | 271 (81.4) | 62 (18.6) | 217 (65.2) | 116 (34.8) | 333 |
P values calculated by χ2 tests.
As determined by a clinical questionnaire administered to caregivers.
At the time of sample collection.
NNRTI, Non Nucleoside Reverse Transciptase Inihibitor; ART, Antiretroviral Therapy; DTG, Dolutegravir; ABC, Abacavir; LPV, Lopinavir; NVP, Nevirapine; EFV, Efavirenz
At the time of data collection, almost half of CLHIV in the database (N = 145, 43%) had no information on clinical recommendations after HIVDR testing (see Table, Supplemental Digital Content, http://links.lww.com/QAI/C590). However, among those with clinical recommendations, 94 (28.2%) were maintained on their current regimen, 55 (16.5%) were switched to another drug class, and 39 (11.7%) had a single drug substitution.
DISCUSSION
This analysis demonstrated low levels of INSTI and PI HIVDR among Ugandan CLHIV experiencing VF on first-line ART and represents the largest examination of HIVDR among CLHIV on DTG-based regimens in Sub-Saharan Africa to date. Although overall rates of DRMs were high, most DRMs were to NNRTIs with low rates to INSTIs and PIs. We found that CLHIV on a DTG-based regimen were least likely to have any DRMs (including non-NNRTI DRMs) compared with all other regimens, with >95% susceptible to DTG, LPV/r, and DRV/r. The finding of limited DTG resistance among children failing on DTG indicates that most children could be maintained on their DTG-based ART with more emphasis on intensified, effective adherence support to help them achieve VS. Low rates of DRV/r resistance further support the use of DRV as an effective second-line ARV, even among those CLHIV failing on LPV/r. We also found that the prevalence of intermediate-to-high level resistance to ABC and AZT was relatively low (≤25%) and was similar across all CLHIV, regardless of regimen at sample collection, suggesting the preferred maintenance of an ABC backbone after first-line failure. These novel HIVDR findings are important as countries develop and implement guidelines to optimize all CLHIV to first-line DTG-based regimens and determine optimal management of VF in the setting of DTG.
Consistent with existing literature is the finding that most (81.4%) CLHIV with VF have at least 1 DRM, with a similar prevalence of NNRTI (67.6%) and NRTI (65.2%). DRMs and the most common mutations seen by class were K103N and M184V, respectively.7,9–11 Consistent with emerging literature is the finding that the proportion of CLHIV with NNRTI DRMs does not vary between those on a PI-based or NNRTI-based regimen,12 although this study found lower rates of NNRTI DRMs among CLHIV on DTG-based regimens. Our results also showed lower rates of clinically significant resistance to 3TC among those on DTG compared with those on LPV/r, which is consistent with a small but growing body of research that suggests DTG-based regimens may be associated with lower rates of emerging HIVDR.13,14 Similar to other smaller HIVDR studies among CLHIV on DTG-based regimens, we saw low rates of INSTI DRMs (6.6% in our study, range 5.8%–18% in other studies) and even lower rates of intermediate-to-high level resistance to DTG (4.1% in our study, range 8.9%–18% in other studies).14–16 Although analysis methodologies differ across previous studies due to differences in the sample population, ART at sample collection, and sampling approaches, all findings continue to demonstrate DTG efficacy and support DTG-based regimens as preferred first-line among CLHIV.
The novel findings in our study included the relatively large number of CLHIV on a DTG-based regimen with a WHO-recommended ABC + 3TC backbone for CLHIV <30 kg compared with other HIVDR studies among CLHIV on DTG-based regimens.14,15,17 Although results from the Nucleosides and Darunavir/Dolutegravir in Africa (NADIA) trial have demonstrated the durability and non-inferiority of DTG when paired with a 3TC + TDF backbone even when no NRTI activity is predicted, questions remain as to whether these findings apply to CLHIV, especially those who would be optimized to DTG using a single drug substitution while maintaining their ABC + 3TC backbone.18 Our findings demonstrate that among this highly treatment experienced population of CLHIV, those with a DTG + ABC containing regimen were no more likely to have DRMs or intermediate-to-high level non-NNRTI resistance than those on DTG with a non-ABC containing backbone. In addition, similar rates of clinically significant resistance to ABC and AZT were seen among CLHIV on a DTG-based regimen, despite higher known exposure to ABC than AZT and few children currently on AZT. While further data are needed, our findings also support the maintenance of an ABC-containing backbone among children failing LPV/r-based regimens, as we demonstrated that only 13.2% of our population failing LPV/r may potentially benefit from switching to an AZT-containing backbone. The low rates of intermediate-to-high level resistance to TDF seen in both CLHIV on DTG and LPV/r along with results from the NADIA trial underscore the importance of transitioning all CLHIV on DTG to a TDF-containing regimen once they reach 30 kg, such as tenofovir/lamivudine/dolutegravir, which has been shown to improve VS regardless of underlying NRTI resistance.
Our study found higher rates of LPV/r resistance than previously reported among children on LPV/r-based regimens. Our findings of 14% intermediate-to-high level resistance to LPV/r were slightly higher than the 1%–10.7% reported in other studies.12,15,19 However, our study also contributes to a small but increasingly important body of evidence which demonstrates very low rates of intermediate-to-high level resistance to DRV/r amongst children on LPV/r and DTG, indicating that DRV may be the more promising PI for second-line regimens, even in the presence of LPV/r resistance.12,15,19
Consistent with limited available data, our study found that younger children (younger than 5 years) are more likely to have DRMs and intermediate-to-high level resistance.7,19 This has important clinical implications, as these young children will need ART throughout their lives, and it will be important to prevent additional development of resistance to currently available, optimal regimens. Sample size limited our ability to further analyze resistance by regimen in these young children, but this would be an important focus of future HIVDR surveys. This analysis took place before full introduction of DTG 10 mg in Uganda, so further investigation into this age band is warranted as most CLHIV <5 years will have now fully transitioned to DTG.
At the time of this analysis only 56.5% (188/333) of CLHIV had a clinical outcome reported in their chart. Half of those with a reported outcome were maintained on their current regimen. These data suggest moderate durability of current regimens, despite high levels of HIVDR; however, minimal clinical outcome data limited our ability to draw meaningful conclusions about the impact of HIVDR testing on clinical outcomes.
An important strength of this analysis is that it is one of the first to result from HIVDR testing conducted as part of standard of care HIV treatment for all CLHIV. However, there were also several limitations, including those originating from identified gaps in programmatic HIVDR testing. Primarily, samples were only obtained from those CLHIV on ART who returned for follow-up VL testing, none of whom were eligible CLHIV 16–19 years and only about half of the samples identified as eligible for inclusion were successfully traced, processed, and amplified. This limits the study’s full representativeness of the population and prevents interpretation for older adolescents. To improve data quality in the future, improvements can be made in tracing and processing for samples eligible for HIVDR testing, as well as to better understand the factors that contributed to unsuccessful amplification for a number of the samples. These challenges and resulting lower sample size consequently limited our ability to perform more robust analyses between and within treatment groups. In addition, owing to delays in HIVDR turnaround time and review of clinical HIVDR results, as well as gaps in documentation of clinical management using HIVDR results, we had limited ability to analyze treatment decisions’ outcomes and therefore could not robustly evaluate the clinical impact of standard of care HIVDR testing. Further evaluations will be important once measures to improve turnaround time and documentation are implemented.
In conclusion, among this cohort of Ugandan CLHIV failing their first-line ART regimen, the findings of low rates of INSTI mutations and overall lower rates of clinically significant HIVDR among those on a DTG-based regimen underscores the durability of DTG-based regimens, highlighting the importance of a rapid transition to DTG-based ART for all CLHIV, and the value of effective adherence support for those CLHIV on DTG with VF. The rates of cross-resistance seen between ABC and AZT, regardless of ART backbone at sample collection, support maintaining an ABC backbone for CLHIV <30 kg, at the time of DTG transition even without VL results. Further data on clinical outcomes of CLHIV with VF maintained on DTG-based regimens, especially in the presence of NRTI and INSTI resistance, could better inform guidance on the timing and necessity of a switch to a second-line regimen.
Supplementary Material
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.jaids.com).
ACKNOWLEDGMENTS
The authors acknowledge the contribution of CLHIV whose data were analyzed, health care workers at ART-Accredited facilities and at the HIV Viral load and Drug Resistance testing laboratory, and the national and regional HIVDR committees that supported the review of the HIVDR results. The authors also acknowledge CDC that funded this work through the FASTER project. Finally, the authors acknowledge the team at the Ministry of Health STD/AIDS Control Program for their stewardship.
Supported by the US President’s Emergency Plan for AIDS Relief through the United States Centers for Disease Control and Prevention under the terms of Cooperative Agreement NU2GGH001463.
Footnotes
The authors have no conflicts of interest to disclose.
Presented at the 30th International Workshop on HIV Drug Resistance and Treatment Strategies; September 20–23, 2023; Cape Town, South Africa.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the funding agencies.
REFERENCES
- 1.Joint United Nations Programme on HIV/AIDS. The Path that Ends AIDS: 2023 UNAIDS Global AIDS Update. Geneva Switzerland: UNAIDS; 2023. [Google Scholar]
- 2.UNAIDS. AIDSinfo. Available at: https://aidsinfo.unaids.org/. Accessed December 10, 2024.
- 3.World Health Organization. Consolidated Guidelines on HIV Prevention, Testing, Treatment, Service Delivery and Monitoring: Recommendations for a Public Health Approach; 2021. Available at: https://www.who.int/publications/i/item/9789240031593 Accessed September 23, 2024. [PubMed]
- 4.Boerma RS, Sigaloff KC, Akanmu AS, et al. Alarming increase in pretreatment HIV drug resistance in children living in sub-Saharan Africa: a systematic review and meta-analysis. J Antimicrob Chemother. 2017;72:365–371. [DOI] [PubMed] [Google Scholar]
- 5.Soeria-Atmadja S, Amuge P, Nanzigu S, et al. Pretreatment HIV drug resistance predicts accumulation of new mutations in ART-naïve Ugandan children. Acta Paediatr. 2020;109:2706–2716. [DOI] [PubMed] [Google Scholar]
- 6.Takem EN, Coox C, Shang J, et al. The association between HIV pretreatment drug resistance and virological outcomes in children and adults in sub-Saharan Africa: a systematic review and meta-analysis. PLoS One. 2024;19:e0300456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.World Health Organization. HIV Drug Resistance Report 2021; 2021. Available at: https://www.who.int/publications/i/item/9789240038608 Accessed October 31, 2024.
- 8.Oliver D, Mabirizi D, Hast M, et al. A model for accelerating access to care and treatment for children and adolescents living with HIV in Nigeria, Tanzania, Uganda, and Zambia: the faith-based action for scaling-up testing and treatment for the epidemic response (FASTER) initiative. J Int Assoc Provid AIDS Care. 2023;22:23259582231186701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Pillay S, Bland RM, Lessells RJ, et al. Drug resistance in children at virological failure in a rural KwaZulu-Natal, South Africa, cohort. AIDS Res Ther. 2014;11:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Vaz P, Buck WC, Bhatt N, et al. Compromise of second-line antiretroviral therapy due to high rates of human immunodeficiency virus drug resistance in mozambican treatment-experienced children with virologic failure. J Pediatr Infect Dis Soc. 2020;9:6–13. [DOI] [PubMed] [Google Scholar]
- 11.Muri L, Gamell A, Ntamatungiro AJ, et al. Development of HIV drug resistance and therapeutic failure in children and adolescents in rural Tanzania: an emerging public health concern. AIDS. 2017;31:61–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hackett S, Teasdale CA, Pals S, et al. Drug resistance mutations among South African children living with HIV on WHO-recommended ART regimens. Clin Infect Dis. 2021;73:e2217–e2225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Frange P, Blanche S, Veber F, et al. Dolutegravir in the long term in children and adolescents: frequent virological failure but rare acquired genotypic resistance. HIV Med. 2021;22:958–964. [DOI] [PubMed] [Google Scholar]
- 14.Kityo C, White E, Turkova A, et al. Virological failures and genotypic resistance in children and adolescents randomised to dolutegravir-based ART vs. standard-of-care in the ODYSSEY trial. Presented at: International Pediatric HIV Workshop; 2021. Session Ab 10. Virtual Abs PEBLB. [Google Scholar]
- 15.Kamori D, Barabona G, Rugemalila J, et al. Emerging integrase strand transfer inhibitor drug resistance mutations among children and adults on ART in Tanzania: findings from a national representative HIV drug resistance survey. J Antimicrob Chemother. 2023;78:779–787. [DOI] [PubMed] [Google Scholar]
- 16.Kingwara L, Inzaule SC, Momanyi L, et al. Impact of nucleos(t)ide reverse transcriptase inhibitor resistance on dolutegravir and protease-inhibitor-based regimens in children and adolescents in Kenya. AIDS. 2022;36:611–613. [DOI] [PubMed] [Google Scholar]
- 17.Townsend CL, O’Rourke J, Milanzi E, et al. Effectiveness and safety of dolutegravir and raltegravir for treating children and adolescents living with HIV: a systematic review. J Int AIDS Soc. 2022;25:e25970. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Paton NI, Musaazi J, Kityo C, et al. Efficacy and safety of dolutegravir or darunavir in combination with lamivudine plus either zidovudine or tenofovir for second-line treatment of HIV infection (NADIA): week 96 results from a prospective, multicentre, open-label, factorial, randomised, non-inferiority trial. Lancet HIV. 2022;9:e381–e393. [DOI] [PubMed] [Google Scholar]
- 19.Meyers T, Sawry S, Wong JY, et al. Virologic failure among children taking lopinavir/ritonavir-containing first-line antiretroviral therapy in South Africa. Pediatr Infect Dis J. 2015;34:175–179. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.