Abstract
Background
Dolutegravir is being rolled out globally as part of preferred antiretroviral therapy (ART) regimens, including among treatment-experienced patients. The role of viral load (VL) testing before switching patients already on ART to a dolutegravir-containing regimen is less clear in real-world settings.
Methods
We included patients from the International epidemiology Databases to Evaluate AIDS consortium who switched from a nevirapine- or efavirenz-containing regimen to one with dolutegravir. We used multivariable cause-specific hazards regression to estimate the association of the most recent VL test in the 12 months before switching with subsequent outcomes.
Results
We included 36 393 patients at 37 sites in 5 countries (Democratic Republic of the Congo, Kenya, Rwanda, Tanzania, Uganda) who switched to dolutegravir from July 2017 through February 2020, with a median follow-up of approximately 11 months. Compared with those who switched with a VL <200 copies/mL, patients without a recent VL test or with a preswitch VL ≥1000 copies/mL had significantly increased hazards of an incident VL ≥1000 copies/mL (adjusted hazard ratio [aHR], 2.89; 95% confidence interval [CI], 1.99–4.19 and aHR, 6.60; 95% CI, 4.36–9.99, respectively) and pulmonary tuberculosis or a World Health Organization clinical stage 4 event (aHR, 4.78; 95% CI, 2.77–8.24 and aHR, 13.97; 95% CI, 6.62–29.50, respectively).
Conclusions
A VL test before switching to dolutegravir may help identify patients who need additional clinical monitoring and/or adherence support. Further surveillance of patients who switched to dolutegravir with an unknown or unsuppressed VL is needed.
Keywords: antiretroviral agents, clinical decision-making, HIV integrase inhibitors, viral load, prognosis
Patients who switched to dolutegravir with a viral load ≥1000 copies/mL or without a recent test had worse subsequent outcomes than those known to be suppressed. These findings have implications for clinical care and surveillance as dolutegravir is implemented.
Dolutegravir, an integrase strand transfer inhibitor, with 2 nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs), is now the preferred antiretroviral therapy (ART) regimen globally for people living with human immunodeficiency virus (HIV) [1, 2]. In low- and middle-income countries (LMICs), dolutegravir is being scaled up most commonly as a first-line regimen of a once-daily fixed-dose combination tablet that also contains tenofovir disoproxil fumarate and lamivudine [3]. In addition to starting patients newly initiating ART on dolutegravir, patients already on regimens that contain non-nucleoside reverse transcriptase inhibitors (NNRTIs), such as nevirapine and efavirenz, are being switched to dolutegravir [4, 5] because of its superior efficacy and tolerability and high barrier to HIV drug resistance [6]. By 2025, dolutegravir-containing regimens are expected to be the most widely used ART among people living with HIV in LMICs [7].
The World Health Organization (WHO) guidelines recommend at least yearly testing of viral load (VL) for patients established on ART and encourage assessing VL before switching from an NNRTI to dolutegravir [1, 2]. A VL <1000 copies/mL is the recommended cutoff to switch to a first-line regimen of dolutegravir with a tenofovir NRTI backbone. For patients with a VL ≥1000 copies/mL that remains elevated despite adherence support, the WHO currently recommends a second-line dolutegravir regimen that substitutes tenofovir with zidovudine. Resource-constrained settings often have limited access to genotypic resistance testing [8], so a persistently unsuppressed VL may indicate underlying resistance mutations [9]. Because resistance to both tenofovir and lamivudine/emtricitabine is prevalent [10, 11], some ART-experienced patients with an unknown or unsuppressed VL might be switched to regimens in which dolutegravir is the only fully active antiretroviral drug [12, 13]. Exposure to dolutegravir monotherapy may potentially increase the risk for treatment failure and integrase strand transfer inhibitor resistance mutations [14–16]. Despite the WHO guidance, some patients may not have a VL test before switching to dolutegravir. Use of routine VL monitoring remains limited in some resource-constrained settings [8, 17], and national ART guidelines differ in their recommendation for VL testing before regimen changes [18, 19]. In addition, VL testing coverage is sensitive to disruptions of public health systems, whereby VL testing may only be done or available at irregular time intervals, as evidenced during the coronavirus disease 2019 pandemic [20].
Routine clinical care data from the initial stages of the global dolutegravir rollout may provide evidence for how VL monitoring can inform clinical care for patients in HIV treatment programs transitioning to dolutegravir and ultimately preserve it as a treatment option. In this study, we sought to describe VL testing status among patients who switched to dolutegravir in sub-Saharan Africa and then examine associations with subsequent HIV treatment outcomes while on dolutegravir.
METHODS
Data Sources
Data came from the International epidemiology Databases to Evaluate AIDS research consortium, which collects and harmonizes HIV treatment and care data across 7 geographical regions. This analysis included data from the Central Africa (CA) and East Africa (EA) regions. Sites contributing to CA and EA are mostly public facilities with diverse levels of care and comprehensiveness of clinical services offered. Patient data are collected at clinical encounters and are deidentified at the site level before being transmitted to a regional data management center. Site-level data are collected through site surveys that ascertain level of care, availability of services, and other service-related and contextual information [21]. Research ethics committees at sites and regional data management centers provided ethical oversight and approved the use of deidentified data for this analysis.
Inclusion and Exclusion Criteria
We included patients who were aged ≥16 years at the time of HIV care enrollment who switched to a dolutegravir-containing regimen from a nevirapine- or efavirenz-containing regimen from July 2017 through February 2020, started ART ≥6 months before switching, and had ≥6 months of possible follow-up after switching. We excluded patients who had a conflicting dolutegravir start date.
Measures
We examined the following 5 outcomes: an incident VL ≥1000 copies/mL, with a subsequent VL test ≥1000 copies/mL or no subsequent VL test during follow-up; new or recurrent pulmonary tuberculosis or WHO clinical stage 4 event, based on documentation of a clinical diagnosis, with pulmonary tuberculosis additionally imputed based on documentation of initiating a tuberculosis treatment regimen; switch to a protease inhibitor (PI)–containing regimen; switch back to an NNRTI-containing regimen; and death from any cause or loss to program. Death was based on a site’s standard practices for ascertainment (ie, active or passive), including estimation of date of death if unknown. Loss to program was defined as either lost to follow-up (ie, no record for ≥7 months immediately preceding the date of site database closure) or known to leave care. The last recorded date of contact was used to determine when patients were lost to program. Death or lost to program was a competing event for the first 4 outcomes. Switching to a non-dolutegravir–containing regimen (other than that related to the outcome, if applicable) was a competing event for all outcomes. Patients were followed from the date of switching to dolutegravir until the outcome of interest, a competing event, date of site database closure, or administrative censoring 18 months after switching. As some clinical events (eg, incident VL ≥1000 copies/mL) were documented on the same day as switching regimens, the clinical event always took precedence.
The main exposure variable was preswitch VL status, that is, a VL measured in the 12 months before switching to dolutegravir (inclusive of the day of switching) and VL result among those with a test (ie, <200 copies/mL, 200–399 copies/mL, 400–999 copies/mL, and ≥1000 copies/mL). VL testing was conducted in accordance with sites’ local clinical protocols and was reported as a numeric variable or as undetectable, using the lower limit for detection of the assay used. For patients who had multiple VL tests prior to switching to dolutegravir, we used the VL test closest to the date of switching.
Other variables examined were age group, sex, initial NRTI backbone when starting dolutegravir, year of starting ART, preswitch NNRTI, prior AIDS diagnosis (ie, CD4 count <200 cells/mm3 or CD4+ percentage <14%, WHO clinical stage 4 event for EA, or documentation of AIDS diagnosis for CA), history of disengagement from care (ie, gap with no record for ≥7 months in up to 5 years before switching), site urbanicity, site level of care, and country. The time interval used to define a history of disengagement from care and lost to follow-up extended the previously validated 6-month interval [22] by 1 month to reduce scheduling-related misclassification among patients with differentiated service delivery with visits every 6 months.
Statistical Analyses
We computed frequencies and proportions of the sample characteristics, medians and interquartile ranges (IQRs) for time from most recent VL test until switching to dolutegravir and time from switching until an event (among those with the outcome of interest), and incidence rates for outcomes, overall and stratified by preswitch VL status. We used cause-specific hazards regression to determine the association between preswitch VL status and each outcome, with a random effect for site. We first ran crude models and then multivariable models adjusted for age group, sex, history of disengagement from care, and prior AIDS diagnosis as potential confounders. Patients with a history of disengagement from care may be less likely to have a VL test because of missing visits and be at higher risk for worse HIV treatment outcomes [23–26]. Prior AIDS diagnosis was included because patients with advanced HIV disease are at higher risk for poor outcomes [27] and may be subject to more intensive monitoring [28], inclusive of VL testing, but may also have more difficulty in achieving viral suppression [26]. Analyses for the VL outcome were limited to patients who had VL testing after switching to dolutegravir (Supplementary Table 1). Analyses for the pulmonary tuberculosis and WHO clinical stage 4 event outcome were limited to EA, as these data were not available for CA. We conducted sensitivity analyses of the multivariable models by limiting to countries outside of Kenya, as most data were from Kenya, and stratifying by median time from preswitch VL until initiating dolutegravir. Analyses were conducted using SAS 9.4 (SAS Institute).
RESULTS
Sample Disposition and Characteristics
From July 2017 through February 2020, 37 765 patients who switched to dolutegravir were assessed for eligibility, and 36 393 individuals were included (Supplementary Figure 1) from 37 sites in 5 countries: Kenya (82%), Uganda (9%), Rwanda (5%), Tanzania (3%), and Democratic Republic of the Congo (0.5%). Less than half (42%) of patients received care at rural sites, and 33% received care at sites with a primary level of care. Overall, 84% of patients were aged 30–59 years, and 52% were female. The majority (82%) started ART before 2016, 54% had a prior AIDS diagnosis, and 12% had a history of disengagement from care. Over half (55%) of patients switched from efavirenz, and 96% had an initial NRTI backbone that contained tenofovir after switching to dolutegravir (Table 1).
Table 1.
Viral Load Status in the 12 Months Before Switching to Dolutegravir and Characteristics of Patients and Sites
| Variables | Overall Sample | Most Recent VL in the 12 Months Before Switching to Dolutegravir | ||||
|---|---|---|---|---|---|---|
| <200 Copies/mL | 200–399 Copies/mL | 400–999 Copies/mL | ≥1000 Copies/mL | No VL Test | ||
| Preswitch VL status, N (row%) | 36 393 (100) | 31 848 (87.5) | 1720 (4.7) | 993 (2.7) | 400 (1.1) | 1432 (3.9) |
| Preswitch VL test to dolutegravir switch, median (interquartile range), months | 1.8 (0.9–4.6) | 1.8 (0.9–4.6) | 1.8 (0.9–4.6) | 2.4 (0.9–5.5) | 2.1 (0.7–4.6) | Not applicable |
| Patient characteristics | N = 36 393 | N = 31 848 | N = 1720 | N = 993 | N = 400 | N = 1432 |
| Age group, N (column%), y | ||||||
| 17–30 | 1172 (3.2) | 945 (3.0) | 42 (2.4) | 19 (1.9) | 39 (9.8) | 127 (8.9) |
| 30–39 | 5817 (16.0) | 4981 (15.6) | 239 (13.9) | 135 (13.6) | 109 (27.3) | 353 (24.7) |
| 40–49 | 13 169 (36.2) | 11 551 (36.3) | 624 (36.3) | 381 (38.4) | 135 (33.8) | 478 (33.4) |
| 50–59 | 11 514 (31.6) | 10 195 (32.0) | 604 (35.1) | 300 (30.2) | 78 (19.5) | 337(23.5) |
| ≥60 y | 4721 (13.0) | 4176 (13.1) | 211 (12.3) | 158 (15.9) | 39 (9.8) | 137 (9.6) |
| Sex, N (column%) | ||||||
| Female | 18 907 (52.0) | 16 626 (52.2) | 877 (51.0) | 530 (53.4) | 199 (49.8) | 675 (47.1) |
| Male | 17 486 (48.1) | 15 222 (47.8) | 843 (49.0) | 463 (46.6) | 201 (50.3) | 757 (52.9) |
| Year of starting antiretroviral therapy, N (column%) | ||||||
| ≤2005 | 2647 (7.3) | 2354 (7.4) | 138 (8.0) | 83 (8.4) | 18 (4.5) | 54 (3.8) |
| 2006–2010 | 13 419 (36.9) | 11 901 (37.4) | 680 (39.5) | 395 (39.8) | 92 (23.0) | 351 (24.5) |
| 2011–2015 | 13 719 (37.7) | 12 009 (37.7) | 625 (36.3) | 369 (37.2) | 155 (38.8) | 561 (39.2) |
| ≥2016 | 6608 (18.2) | 5584 (17.5) | 277 (16.1) | 146 (14.7) | 135 (33.8) | 466 (32.5) |
| Preswitch non-nucleoside reverse transcriptase inhibitor, N (column%) | ||||||
| Efavirenz | 20 118 (55.3) | 17 597 (55.3) | 854 (49.7) | 380 (38.3) | 251 (62.8) | 1036 (72.4) |
| Nevirapine | 16 275 (44.7) | 14 251 (44.8) | 866 (50.4) | 613 (61.7) | 149 (37.3) | 396 (27.7) |
| Nucleoside/nucleotide reverse transcriptase inhibitor backbone when starting dolutegravir, N (column%) | ||||||
| Tenofovir + lamivudine/emtricitabine | 34 683 (95.5) | 30 407 (95.7) | 1643 (95.6) | 916 (92.7) | 352 (88.2) | 1365 (95.5) |
| Zidovudine + lamivudine/emtricitabine | 1193 (3.3) | 998 (3.1) | 70 (4.1) | 64 (6.5) | 28 (7.0) | 33 (2.3) |
| Othera | 439 (1.2) | 374 (1.2) | 6 (0.4) | 8 (0.8) | 19 (4.8) | 32 (2.2) |
| Missing | 78 | 69 | 1 | 5 | 1 | 2 |
| Prior AIDS diagnosis, N (column%) | 19 740 (54.2) | 17 187 (54.0) | 1034 (60.1) | 649 (65.4) | 210 (52.5) | 660 (46.1) |
| History of disengagement from care, N (column%) | 4399 (12.1) | 3357 (10.5) | 208 (12.1) | 133 (13.4) | 97 (24.3) | 604 (42.2) |
| Country, N (column%) | ||||||
| Democratic Republic of the Congo | 165 (0.5) | 10 (0.03) | 3 (0.2) | 1 (0.1) | 9 (2.3) | 142 (9.9) |
| Kenya | 29 923 (82.2) | 26 360 (82.8) | 1645 (95.6) | 923 (93.0) | 243 (60.8) | 752 (52.5) |
| Rwanda | 1669 (4.6) | 1341 (4.2) | 27 (1.6) | 14 (1.4) | 61 (15.3) | 226 (15.8) |
| Tanzania | 1221 (3.4) | 1034 (3.3) | 11 (0.6) | 11 (1.1) | 54 (13.5) | 111 (7.8) |
| Uganda | 3415 (9.4) | 3103 (9.7) | 34 (2.0) | 44 (4.4) | 33 (8.3) | 201 (14.0) |
| Site urbanicity, N (column%) | ||||||
| Rural | 15 379 (42.3) | 13 425 (42.2) | 613 (35.6) | 398 (40.1) | 185 (46.3) | 758 (52.9) |
| Urban | 21 014 (57.7) | 18 423 (57.9) | 1107 (64.4) | 595 (59.9) | 215 (53.8) | 674(47.1) |
| Site level of care, N (column%) | ||||||
| Primary | 12 122 (33.3) | 10 532 (33.1) | 532 (30.9) | 263 (26.5) | 180 (45.0) | 615 (43.0) |
| Secondary | 13 602 (37.4) | 11 886 (37.3) | 778 (45.2) | 450 (45.3) | 123 (30.8) | 365 (25.5) |
| Tertiary | 10 669 (29.3) | 9430 (29.6) | 410 (23.8) | 280 (28.2) | 97 (24.3) | 452 (31.6) |
Abbreviation: VL, viral load.
Other primarily consisted of abacavir-containing backbones (91%; 399 of 439).
Viral Load Testing Before Switching to Dolutegravir and Differences by Sample Characteristics
In the 12 months before switching to dolutegravir, 96% of patients had a VL test and 4% did not. Of those with a VL test, 91% had a VL <200 copies/mL, 5% had a VL 200–399 copies/mL, 3% had a VL 400–999 copies/mL, and 1% had a VL ≥1000 copies/mL (Table 1). Median time from VL test to switch was 1.8 months (IQR, 0.9–4.6). Compared with patients with a preswitch VL <200 copies/mL, patients without a preswitch VL test or with a preswitch VL ≥1000 copies/mL had a younger age distribution, started ART later, more often switched from efavirenz, more often had a history of disengagement from care, and more often received care at sites that were outside of Kenya, rural sites, and sites with a primary level of care.
HIV Treatment Outcomes After Switching to Dolutegravir
For the incident VL ≥1000 copies/mL outcome (N = 30 459), the median follow-up time was 11.5 months (IQR, 9.8–13.9) since switching to dolutegravir until the outcome (n = 444), switching to a PI- or NNRTI-containing regimen (n = 2183), death or loss to program (n = 544), or database closure or administrative censoring (n = 27 288). For the pulmonary tuberculosis and WHO clinical stage 4 event outcome (N = 34 559), the median follow-up time was 11.0 months (IQR, 9.0–12.9) since switching to dolutegravir until the outcome (n = 109), switching to a PI- or NNRTI-containing regimen (n = 2886), death or loss to program (n = 1296), or database closure or administrative censoring (n = 30 268). For switch to a PI- or NNRTI-containing regimen and death or lost to program outcomes (N = 36 393), the median follow-up time was 11.1 months (IQR, 9.0–13.3) since switching to dolutegravir until death or loss to program (n = 1392), switching to a PI-containing regimen (n = 115) or NNRTI-containing regimen (n = 2790), or database closure or administrative censoring (n = 32 096). Among patients who switched to an NNRTI-containing regimen, 69% were female (Supplementary Table 2). The descending incidence of HIV treatment outcomes was as follows: switch to an NNRTI-containing regimen (8.3 per 100 person-years), death or loss to program (4.1 per 100 person-years), incident VL ≥1000 copies/mL (1.5 per 100 person-years), pulmonary tuberculosis or WHO clinical stage 4 event (0.3 per 100 person-years), and switch to a PI-containing regimen (0.3 per 100 person-years; Table 2). Among patients with each outcome of interest, median time until event varied by preswitch VL status.
Table 2.
Description of Human Immunodeficiency Virus Treatment Outcomes During Follow-up and Associations With Viral Load Status Before Switching to Dolutegravir
| Outcomes,Total and By Preswitch VL Status | Number of Events | Number of Person-Years | Crude Incidence Rate per 100 Person-Years (95% CI) | Median (Interquartile Range) Months Until Event | Unadjusted Hazard Ratioa (95% CI); P Value | Adjusted Hazard Ratiob (95% CI); P Value |
|---|---|---|---|---|---|---|
| Incident VL ≥1000 copies/mL | N = 30 459 | N = 30 459 | ||||
| Total | 444c | 29 739 | 1.5 (1.4–1.6) | 5.3 (3.2–9.2) | ||
| Preswitch VL, copies/mL | ||||||
| <200 | 322 | 26 297 | 1.2 (1.1–1.4) | 5.5 (3.3–9.2) | Reference | Reference |
| 200–399 | 34 | 1547 | 2.2 (1.6–3.1) | 5.6 (3.3–11.2) | 1.63 (1.14–2.33); P = .007 | 1.61 (1.13–2.30); P = .008 |
| 400–999 | 28 | 1036 | 2.7 (1.9–3.9) | 6.9 (3.5–14.9) | 1.79 (1.21–2.65); P = .004 | 1.75 (1.18–2.59); P = .005 |
| ≥1000 | 26 | 245 | 10.6 (7.2–15.6) | 4.0 (2.8–6.6) | 7.23 (4.79–10.91); P < .001 | 6.60 (4.36–9.99); P < .001 |
| No VL test | 34 | 614 | 5.5 (4.0–7.8) | 2.5 (1.0–7.0) | 3.57 (2.48–5.15); P < .001 | 2.89 (1.99–4.19); P < .001 |
| Pulmonary tuberculosis or World Health Organization clinical stage 4 event | N = 34 559 | N = 34 559 | ||||
| Total | 109d | 31 760 | 0.3 (0.3–0.4) | 3.2 (1.2–7.1) | ||
| Preswitch VL, copies/mL | ||||||
| <200 | 78 | 28 066 | 0.3 (0.2–0.3) | 3.8 (2.0–7.4) | Reference | Reference |
| 200–399 | 4 | 1637 | 0.2 (0.1–0.7) | 7.7 (4.1–11.1) | 1.49 (0.54–4.12); P = .441 | 1.46 (0.53–4.04); P = .465 |
| 400–999 | 2 | 1077 | 0.2 (0.05–0.7) | 4.8 (4.1–5.6) | 1.06 (0.26–4.35); P = .933 | 1.00 (0.25–4.11); P = .996 |
| ≥1000 | 8 | 250 | 3.2 (1.6–6.4) | 2.9 (1.0–5.8) | 15.98 (7.65–33.42); P < .001 | 13.97 (6.62–29.50); P < .001 |
| No VL test | 17 | 732 | 2.3 (1.4–3.7) | 0.9 (0.5–1.8) | 5.49 (3.22–9.36); P < .001 | 4.78 (2.77–8.24); P < .001 |
| Switch to a protease inhibitor–containing regimen | N = 36 393 | N = 36 393 | ||||
| Total | 115e | 33 799 | 0.3 (0.3–0.4) | 3.3 (1.8–7.6) | ||
| Preswitch VL, copies/mL | ||||||
| <200 | 79 | 29 647 | 0.3 (0.2–0.3) | 3.7 (1.8–7.0) | Reference | Reference |
| 200–399 | 5 | 1671 | 0.3 (0.1–0.7) | 7.6 (2.8–9.9) | 1.35 (0.54–3.35); P = .519 | 1.35 (0.54–3.36); P = .516 |
| 400–999 | 6 | 1094 | 0.5 (0.2–1.2) | 9.5 (7.6–10.2) | 2.80 (1.20–6.52); P = .017 | 2.72 (1.17–6.35); P = .020 |
| ≥1000 | 19 | 323 | 5.9 (3.7–9.2) | 1.8 (1.0–5.3) | 31.14 (18.36–52.80); P < .001 | 30.53 (17.87–52.18); P < .001 |
| No VL test | 6 | 1064 | 0.6 (0.3–1.3) | 2.8 (2.1–12.5) | 2.14 (0.91–5.00); P = .080 | 2.13 (0.90–5.04); P = .086 |
| Switch to a non-nucleoside reverse transcriptase inhibitor–containing regimen | N = 36 393 | N = 36 393 | ||||
| Total | 2790f | 33 799 | 8.3 (8.0–8.6) | 3.1 (1.9–6.4) | ||
| Preswitch VL, copies/mL | ||||||
| <200 | 2352 | 29 647 | 7.9 (7.6–8.3) | 3.2 (2.0–6.4) | Reference | Reference |
| 200–399 | 116 | 1671 | 6.9 (5.8–8.3) | 3.7 (2.5–7.3) | 0.98 (0.81–1.18); P = .798 | 0.99 (0.82–1.20); P = .940 |
| 400–999 | 90 | 1094 | 8.2 (6.7–10.1) | 3.6 (1.7–7.4) | 1.24 (1.00–1.53); P = .048 | 1.20 (0.97–1.49); P = .092 |
| ≥1000 | 69 | 323 | 21.3 (16.8–27.0) | 1.4 (0.9–2.8) | 3.51 (2.75–4.48); P < .001 | 2.83 (2.22–3.61); P < .001 |
| No VL test | 163 | 1064 | 15.3 (13.1–17.9) | 2.3 (1.0–4.6) | 1.80 (1.53–2.12); P < .001 | 1.55 (1.31–1.83); P < .001 |
| Death or loss to program | N = 36 393 | N = 36 393 | ||||
| Total | 1392g | 33 799 | 4.1 (3.9–4.3) | 4.0 (1.0–8.1) | ||
| Preswitch VL, copies/mL | ||||||
| <200 | 1050 | 29 647 | 3.5 (3.3–3.8) | 4.4 (1.4–8.4) | Reference | Reference |
| 200–399 | 69 | 1671 | 4.1 (3.3–5.2) | 4.9 (1.8–8.8) | 1.14 (0.89–1.46); P = .286 | 1.14 (0.90–1.46); P = .283 |
| 400–999 | 43 | 1094 | 3.9 (2.9–5.3) | 7.2 (3.1–11.1) | 1.12 (0.82–1.52); P = .478 | 1.12 (0.82–1.52); P = .486 |
| ≥1000 | 34 | 323 | 10.5 (7.5–14.7) | 4.4 (1.0–7.2) | 3.03 (2.15–4.28); P < .001 | 2.56 (1.81–3.63); P < .001 |
| No VL test | 196 | 1064 | 18.4 (16.0–21.2) | 1.4 (0.03–3.9) | 5.38 (4.57–6.33); P < .001 | 3.77 (3.19–4.46); P < .001 |
Abbreviations: CI, confidence interval; VL, viral load.
Models included a random effect for site but were not adjusted for covariables.
Models included a random effect for site and were adjusted for age group, sex, prior AIDS diagnosis, and history of disengagement from care.
18% (79 of 444) had a subsequent VL test ≥1000 copies/mL, and 82% (365 of 444) had a single VL ≥1000 copies/mL with no subsequent VL test. Patients who had a single VL ≥1000 copies/mL and later resuppressed on a subsequent VL test were not included.
In nonmutually exclusive categories, 83% (90 of 109) had pulmonary and/or extrapulmonary tuberculosis, 9% (10 of 109) had extrapulmonary cryptococcosis, 6% (6 of 109) had Kaposi sarcoma, and 5% (5/109) had documentation of another World Health Organization clinical stage 4 event.
72% (83 of 115) switched to a ritonavir-boosted atazanavir-containing regimen, and 28% (32 of 115) switched to a ritonavir-boosted lopinavir-containing regimen.
86% (2403 of 2790) switched to an efavirenz-containing regimen, and 14% (387 of 2790) switched to a nevirapine-containing regimen.
15% (213 of 1392) died, 36% (508 of 1392) were lost to follow-up, and 48% (671 of 1392) were known to leave care.
Association Between Preswitch Viral Load and HIV Treatment Outcomes
Patients with a preswitch VL ≥1000 copies/mL had significantly greater hazards of an incident VL ≥1000 copies/mL (adjusted hazard ratio [aHR], 6.60; 95% confidence interval [CI], 4.36–9.99), pulmonary tuberculosis or a WHO clinical stage 4 event (aHR, 13.97; 95% CI, 6.62–29.50), switch to a PI-containing regimen (aHR, 30.53; 95% CI, 17.87–52.18), switch to an NNRTI-containing regimen (aHR, 2.83; 95% CI, 2.22–3.61), and death or loss to program (aHR, 2.56; 95% CI, 1.81–3.63) compared with those who switched with a VL <200 copies/mL (Table 2). Patients without a preswitch VL load test had significantly greater hazards of an incident VL ≥1000 copies/mL (aHR, 2.89; 95% CI, 1.99–4.19, pulmonary tuberculosis or a WHO clinical stage 4 event (aHR, 4.78; 95% CI, 2.77–8.24), switch to an NNRTI-containing regimen (aHR, 1.55; 95% CI, 1.31–1.83), and death or loss to program (aHR, 3.77; 95% CI, 3.19–4.46) compared with those who switched with a VL <200 copies/mL.
In sensitivity analyses (Supplementary Table 3), substantial power was lost when the sample was limited to patients outside of Kenya, but similar patterns in associations were observed despite wide confidence intervals. When stratifying preswitch VL by the median time until initiating dolutegravir (ie, 56 days), associations between a preswitch VL ≥1000 copies/mL and outcomes were generally similar regardless of timing.
DISCUSSION
Patients who switched to dolutegravir with a VL ≥1000 copies/mL or without a recent VL test had worse HIV treatment outcomes, which emerged shortly after switching, compared with patients who switched with a suppressed VL. Although they represented only 5% of patients in the sample, globally, in LMICs, there were several million people living with HIV with an unknown or unsuppressed VL on NNRTI-containing ART before the widespread implementation of dolutegravir [4]. Therefore, the public health implications of interventions to improve outcomes among these patients could be substantial.
Patients with a preswitch VL ≥1000 copies/mL generally had the most adverse HIV treatment outcomes, but we do not know how their outcomes compared with what their outcomes would have been had they not switched to dolutegravir, that is, the counterfactual. The cohorts included in this analysis do not have ready access to genotypic resistance testing for patients with viral nonsuppression on a first-line regimen, so we were unable to determine if resistance mutations explained the associations with more adverse outcomes. Previous research in this setting indicates that mutations relevant to NRTIs (eg, K65R, M184V/I) are common among patients unsuppressed on nevirapine- or efavirenz-containing regimens [29, 30]. Nevertheless, most patients, even with partial or extensive NRTI resistance, can achieve viral suppression on a dolutegravir-containing regimen [31–33]. Most patients with a preswitch VL ≥1000 copies/mL did not have tenofovir substituted with zidovudine; however, recent data support that this would unlikely compromise at least 48-week viral suppression [32–34]. We think it is plausible that patients with preswitch VL ≥1000 copies/mL had adherence issues that went unresolved after switching to dolutegravir; however, we did not have data on whether additional adherence support was provided. It may be possible that clinicians anticipated that switching to a more potent dolutegravir-containing regimen with a higher genetic barrier to resistance would reverse the risk for adverse treatment outcomes, and adherence issues went unaddressed.
For patients without a preswitch VL test, we would hypothesize that the underlying distribution of actual VL levels included a sizable proportion with an elevated VL based on the associations with HIV treatment outcomes. This group of patients appeared to be at especially high risk of poor adherence because of the high prevalence of prior disengagement from care and increased hazards of death or loss to program. In settings with access to routine VL testing, HIV treatment programs should train providers to use preswitch VL to inform treatment decisions before a planned regimen change to dolutegravir and consider additional clinical monitoring and/or adherence support for patients with an unsuppressed VL. At a population level, ongoing surveillance of patients who switched to dolutegravir without a recent VL test or with an unsuppressed VL may help to identify ways to prevent differentially adverse outcomes among these 2 patient groups. Although this study provides some insight from the initial rollout, examination of the long-term effectiveness of dolutegravir use at scale in real-world settings is essential.
This study provides additional evidence into the use of a threshold of <1000 copies/mL on a VL test in the previous 12 months to recommend switching. Ongoing low-level viral replication is associated with increased risk of HIV drug resistance and virologic failure [35–37]; therefore, this threshold has been questioned as possibly being too high [5]. Our findings support this threshold, as switching above it was associated with the highest hazard ratios for outcomes. However, switching with any unsuppressed VL (≥200 copies/mL) was associated with significantly increased hazards of some outcomes, albeit at a lower magnitude.
Some limitations should be considered in the interpretation of this study. The proportion of patients with a preswitch VL that was ≥1000 copies/mL or unknown is likely larger outside of our sample. Our sample included patients established on ART, most for several years, before switching to dolutegravir who might have had better adherence and viral suppression than patients more recently initiating ART. Furthermore, sites in rural locations and those with a primary level of care, with potentially less access to routine VL testing, were underrepresented [8]. Nevertheless, we would not expect smaller proportions of these exposures to bias associations with HIV treatment outcomes. Females were also underrepresented, and most patients who switched back to an NNRTI-containing regimen were female, both of which were due to now resolved concerns about infant neural tube defects [38]. We anticipate more female patients initiating and continuing dolutegravir with consequentially lower overall discontinuation rates. Although we had a large overall sample size, some exposure categories had small sample sizes; with short-term follow-up, the incidence of outcomes was low. Consequentially, confidence intervals were wide for some associations. Longer-term follow-up would allow for more precise estimates and, importantly, would allow for determination of associations with long-term outcomes. In addition to patients with a persistently elevated VL ≥1000 copies/mL, the VL outcome included patients with a single VL ≥1000 copies/mL and no subsequent testing. As the lack of confirmatory testing does not necessarily mean that resuppression on dolutegravir did not occur, this variable may not be fully prognostic of future virologic failure. Confirmatory testing may not always be done, especially in resource-constrained settings [39]; however, because resuppression with dolutegravir is more common than with NNRTIs [40], this approach may need to be reconsidered.
In conclusion, our study supports that a preswitch VL may be useful in identifying patients at elevated risk for adverse HIV treatment outcomes after switching to dolutegravir who may benefit from additional clinical monitoring and/or adherence support. Surveillance of patients who switched without a recent VL test or with an unsuppressed VL may help identify ways to preserve dolutegravir-containing ART as a long-term treatment option.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Financial support. This work was supported by the National Institutes of Health’s (NIH’s) National Institute of Allergy and Infectious Diseases (NIAID), the Eunice Kennedy Shriver National Institute of Child Health & Human Development, the National Cancer Institute, the National Institute on Drug Abuse, the National Heart, Lung, and Blood Institute, the National Institute on Alcohol Abuse and Alcoholism, the National Institute of Diabetes and Digestive and Kidney Diseases, the Fogarty International Center, the National Library of Medicine, and the Office of the Director under awards U01AI096299 (Central Africa) and U01AI069911 (East Africa). Informatics resources were supported by the Harmonist project (R24AI124872).
Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Contributor Information
Matthew L Romo, Department of Epidemiology and Biostatistics & Institute for Implementation Science in Population Health, CUNY Graduate School of Public Health and Health Policy, City University of New York, New York, New York, USA.
Jessie K Edwards, Department of Epidemiology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Aggrey S Semeere, Infectious Diseases Institute, Makerere University, Kampala, Uganda.
Beverly S Musick, Department of Biostatistics and Health Data Science, Indiana University School of Medicine, Indianapolis, Indiana, USA.
Mark Urassa, National Institute for Medical Research, Mwanza, Tanzania.
Francesca Odhiambo, Centre for Microbiology Research, Kenya Medical Research Institute, Nairobi, Kenya.
Lameck Diero, School of Medicine, College of Health Sciences, Moi University, Eldoret, Kenya.
Charles Kasozi, Masaka Regional Referral Hospital, Masaka, Uganda.
Gad Murenzi, Rwanda Military Hospital, Kigali, Rwanda.
Patricia Lelo, Kalembelembe Pediatric Hospital, Kinshasa, Democratic Republic of the Congo.
Katarzyna Wyka, Department of Epidemiology and Biostatistics & Institute for Implementation Science in Population Health, CUNY Graduate School of Public Health and Health Policy, City University of New York, New York, New York, USA.
Elizabeth A Kelvin, Department of Epidemiology and Biostatistics & Institute for Implementation Science in Population Health, CUNY Graduate School of Public Health and Health Policy, City University of New York, New York, New York, USA.
Annette H Sohn, TREAT Asia, amfAR–The Foundation for AIDS Research, Bangkok, Thailandand.
Kara K Wools-Kaloustian, Department of Medicine, Indiana University School of Medicine, Indianapolis, Indiana, USA.
Denis Nash, Department of Epidemiology and Biostatistics & Institute for Implementation Science in Population Health, CUNY Graduate School of Public Health and Health Policy, City University of New York, New York, New York, USA.
References
- 1. World Health Organization. Update of recommendations on first- and second-line antiretroviral regimens: policy brief. July 2019. Available at: https://apps.who.int/iris/handle/10665/325892. Accessed 16 August 2021.
- 2. World Health Organization. Consolidated guidelines on HIV prevention, testing, treatment, service delivery, and monitoring: recommendations for a public health approach. July 2021. Accessed at: https://www.who.int/publications/i/item/9789240031593. Accessed 16 August 2021.
- 3. Clinton Health Access Initiative. HIV market report 2019. Issue 10, September 2019. Available at: https://3cdmh310dov3470e6x160esb-wpengine.netdna-ssl.com/wp-content/uploads/2019/12/2019-HIV-Market-Report.pdf. Accessed 16 August 2021.
- 4. Vitoria M, Hill A, Ford N, et al. The transition to dolutegravir and other new antiretrovirals in low-income and middle-income countries: what are the issues? AIDS 2018; 32:1551–61. [DOI] [PubMed] [Google Scholar]
- 5. Inzaule SC, Hamers RL, Doherty M, Shafer RW, Bertagnolio S, Rinke de Wit TF.. Curbing the rise of HIV drug resistance in low-income and middle-income countries: the role of dolutegravir-containing regimens. Lancet Infect Dis 2019; 19:e246–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Kanters S, Vitoria M, Zoratti M, et al. Comparative efficacy, tolerability and safety of dolutegravir and efavirenz 400mg among antiretroviral therapies for first-line HIV treatment: a systematic literature review and network meta-analysis. EClinicalMedicine 2020; 28:100573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Gupta A, Juneja S, Vitoria M, et al. Projected uptake of new antiretroviral (ARV) medicines in adults in low- and middle-income countries: a forecast analysis 2015-2025. PLoS One 2016; 11:e0164619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Chammartin F, Dao Ostinelli CH, Anastos K, et al. International epidemiology databases to evaluate AIDS (IeDEA) in sub-Saharan Africa, 2012-2019. BMJ Open 2020; 10:e035246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rutstein SE, Hosseinipour MC, Weinberger M, et al. Predicting resistance as indicator for need to switch from first-line antiretroviral therapy among patients with elevated viral loads: development of a risk score algorithm. BMC Infect Dis 2016; 16:280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. TenoRes Study Group. Global epidemiology of drug resistance after failure of WHO recommended first-line regimens for adult HIV-1 infection: a multicentre retrospective cohort study. Lancet Infect Dis 2016; 16:565–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. World Health Organization. HIV drug resistance report. 2019. Available at: https://www.who.int/publications/i/item/WHO-CDS-HIV-19.21. Accessed 16 August 2021.
- 12. Salou M, Butel C, Comlan AS, et al. Challenges of scale-up to dolutegravir-based regimens in sub-Saharan Africa. AIDS 2020; 34:783–7. [DOI] [PubMed] [Google Scholar]
- 13. Carnimeo V, Pulido Tarquino I, Fuentes S, et al. High level of HIV drug resistance informs dolutegravir roll-out and optimized NRTI backbone strategy in Mozambique. JAC Antimicrob Resist 2021; 3:dlab050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hocqueloux L, Raffi F, Prazuck T, et al. Dolutegravir monotherapy versus dolutegravir/abacavir/lamivudine for virologically suppressed people living with chronic HIV infection: the randomized non-inferiority MONCAY trial. Clin Infect Dis 2019; 69:1498–505. [DOI] [PubMed] [Google Scholar]
- 15. Wijting I, Rokx C, Boucher C, et al. Dolutegravir as maintenance monotherapy for HIV (DOMONO): a phase 2, randomised non-inferiority trial. Lancet HIV 2017; 4:e547–54. [DOI] [PubMed] [Google Scholar]
- 16. Rhee SY, Grant PM, Tzou PL, et al. A systematic review of the genetic mechanisms of dolutegravir resistance. J Antimicrob Chemother 2019; 74:3135–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Roberts T, Cohn J, Bonner K, Hargreaves S.. Scale-up of routine viral load testing in resource-poor settings: current and future implementation challenges. Clin Infect Dis 2016; 62:1043–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Ministry of Health, Community Development, Gender, Elderly, and Children (Tanzania). National AIDS Control Programme. National guidelines for the management of HIV and AIDS. 7th ed. April 2019. Available at: https://differentiatedservicedelivery.org/Portals/0/adam/Content/NqQGryocrU2RTj58iR37uA/File/NATIONAL_GUIDELINES_FOR_THE_MANAGEMENT_OF_HIV_AND_AIDS_2019.pdf. Accessed 16 August 2021. [Google Scholar]
- 19. National AIDS and STI Control Program (Kenya). Guidelines on use of antiretroviral drugs for treating and preventing HIV infection in Kenya. 2018. edition. Available at: https://www.nascop.or.ke/new-guidelines/. Accessed 16 August 2021.
- 20. Jewell BL, Mudimu E, Stover J, et al. Potential effects of disruption to HIV programmes in sub-Saharan Africa caused by COVID-19: results from multiple mathematical models. Lancet HIV 2020; 7:e629–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Duda SN, Farr AM, Lindegren ML, et al. Characteristics and comprehensiveness of adult HIV care and treatment programmes in Asia-Pacific, sub-Saharan Africa and the Americas: results of a site assessment conducted by the International epidemiologic Databases to Evaluate AIDS (IeDEA) Collaboration. J Int AIDS Soc 2014; 17:19045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Chi BH, Yiannoutsos CT, Westfall AO, et al. Universal definition of loss to follow-up in HIV treatment programs: a statistical analysis of 111 facilities in Africa, Asia, and Latin America. PLoS Med 2011; 8:e1001111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Fox MP, Rosen S.. Retention of adult patients on antiretroviral therapy in low- and middle-income countries: systematic review and meta-analysis 2008-2013. J Acquir Immune Defic Syndr 2015; 69:98–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Geng EH, Odeny TA, Lyamuya R, et al. Retention in care and patient-reported reasons for undocumented transfer or stopping care among HIV-infected patients on antiretroviral therapy in eastern Africa: application of a sampling-based approach. Clin Infect Dis 2016; 62:935–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Yonga P, Kalya S, Lynen L, Decroo T.. Temporary disengagement and re-engagement in human immunodeficiency virus care in a rural county serving pastoralist communities in Kenya: a retrospective cohort study. Int Health 2020; 12:95–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Osler M, Hilderbrand K, Goemaere E, et al. The continuing burden of advanced HIV disease over 10 years of increasing antiretroviral therapy coverage in South Africa. Clin Infect Dis 2018; 66:S118–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Egger M, May M, Chêne G, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet 2002; 360:119–29. [DOI] [PubMed] [Google Scholar]
- 28. Ford N, Meintjes G, Calmy A, et al. Managing advanced HIV disease in a public health approach. Clin Infect Dis 2018; 66(suppl_2):S106–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Brooks K, Diero L, DeLong A, et al. Treatment failure and drug resistance in HIV-positive patients on tenofovir-based first-line antiretroviral therapy in western Kenya. J Int AIDS Soc 2016; 19:20798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Ndahimana J, Riedel DJ, Mwumvaneza M, et al. Drug resistance mutations after the first 12 months on antiretroviral therapy and determinants of virological failure in Rwanda. Trop Med Int Health 2016; 21:928–35. [DOI] [PubMed] [Google Scholar]
- 31. Aboud M, Kaplan R, Lombaard J, et al. Dolutegravir versus ritonavir-boosted lopinavir both with dual nucleoside reverse transcriptase inhibitor therapy in adults with HIV-1 infection in whom first-line therapy has failed (DAWNING): an open-label, non-inferiority, phase 3b trial. Lancet Infect Dis 2019; 19:253–64. [DOI] [PubMed] [Google Scholar]
- 32. Paton NI, Musaazi J, Kityo C, et al. Dolutegravir or darunavir in combination with zidovudine or tenofovir to treat HIV. N Engl J Med 2021; 385:330–41. [DOI] [PubMed] [Google Scholar]
- 33. Keene CM, Griesel R, Zhao Y, et al. Virologic efficacy of tenofovir, lamivudine and dolutegravir as second-line in adults failing a tenofovir-based first-line regimen: a prospective cohort study. AIDS 2021; 35:1423–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Temfack E, Jahn A, Kalua T, et al. Prospective enhanced monitoring of dolutegravir-based first line in Malawi (oral abstract 35, session O-03). In: Conference on Retroviruses and Opportunistic Infections, Boston, Massachusetts, 8-11 March 2020. [Google Scholar]
- 35. Hermans LE, Moorhouse M, Carmona S, et al. Effect of HIV-1 low-level viraemia during antiretroviral therapy on treatment outcomes in WHO-guided South African treatment programmes: a multicentre cohort study. Lancet Infect Dis 2018; 18:188–97. [DOI] [PubMed] [Google Scholar]
- 36. Swenson LC, Min JE, Woods CK, et al. HIV drug resistance detected during low-level viraemia is associated with subsequent virologic failure. AIDS 2014; 28:1125–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Taiwo B, Gallien S, Aga E, et al. Antiretroviral drug resistance in HIV-1-infected patients experiencing persistent low-level viremia during first-line therapy. J Infect Dis 2011; 204:515–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Romo ML, Patel RC, Edwards JK, et al. ; for the International epidemiology Databases to Evaluate AIDS (IeDEA). . Disparities in dolutegravir uptake affecting females of reproductive age with HIV in low- and middle-income countries following initial concerns about teratogenicity: an observational study. Ann Intern Med 2021; In Press. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Shroufi A, Van Cutsem G, Cambiano V, et al. Simplifying switch to second-line antiretroviral therapy in sub-Saharan Africa: predicted effect of using a single viral load to define efavirenz-based first-line failure. AIDS 2019; 33:1635–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Pepperrell T, Venter WDF, McCann K, et al. Participants on dolutegravir resuppress human immunodeficiency virus RNA after virologic failure: updated data from the ADVANCE trial. Clin Infect Dis 2021; 73:e1008–10. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.