Virological outcomes of various first-line ART regimens in patients harbouring HIV-1 E157Q integrase polymorphism: a multicentre retrospective study

Abstract

Background

Integrase strand transfer inhibitors (INSTIs) are recommended as first-line ART for people living with HIV (PLWH) in most guidelines. The INSTI-resistance-associated mutation E157Q, a highly prevalent (2%–5%) polymorphism of the HIV-1 (human immunodeficiency virus type 1) integrase gene, has limited data on optimal first-line ART regimens. We assessed the virological outcomes of various first-line ART regimens in PLWH with E157Q in real-world settings.

Methods

A multicentre retrospective observational study was conducted on PLWH who underwent integrase genotypic drug-resistance testing before ART initiation between 2008 and 2019 and were found to have E157Q. Viral suppression (<50 copies/mL) rate at 24 and 48 weeks, time to viral suppression and time to viral rebound (≥100 copies/mL) were compared among the first-line ART regimens.

Results

E157Q was detected in 167 (4.1%) of 4043 ART-naïve PLWH. Among them, 144 had available clinical data after ART initiation with a median follow-up of 1888 days. Forty-five started protease inhibitors + 2 NRTIs (PI group), 33 started first-generation INSTI (raltegravir or elvitegravir/cobicistat) + 2 NRTIs (INSTI-1 group), 58 started once-daily second-generation INSTI (dolutegravir or bictegravir) + 2 NRTIs (INSTI-2 group) and eight started other regimens. In the multivariate analysis, the INSTI-2 group showed similar or favourable outcomes compared with the PI group for viral suppression rates, time to viral suppression and time to viral rebound. Two cases in the INSTI-1 group experienced virological failure.

Conclusions

The general guideline recommendation of second-generation INSTI-based first-line ART for most PLWH is also applicable to PLWH harbouring E157Q.

Introduction

Antiretroviral therapy (ART) has dramatically changed the prognosis of people living with human immunodeficiency virus (PLWH). Integrase strand transfer inhibitors (INSTIs), which have high virological potency, good tolerability and favourable pharmacokinetic profiles were introduced in 2007.¹ The second-generation INSTI class, which has a higher barrier to resistance was introduced in 2013 and is now used worldwide.² INSTI-based ARTs are currently recommended as first-line regimens by various guidelines.^3–5

The E157Q polymorphism of the HIV-1 (human immunodeficiency virus type 1) integrase gene had been reported as an accessory INSTI-resistance-associated mutation in previous studies.^6,7 Its prevalence is approximately 2%–5% in ART-naïve individuals and is known to vary by HIV-1 subtypes.^8–16 E157Q alone does not reduce susceptibility to first-generation or second-generation INSTIs in vitro, but it acts as a compensatory for INSTI-resistance mutations.^9,17–20 Several studies evaluated the therapeutic outcomes of a small number of cases with E157Q at baseline and found no association between the baseline mutation and virological outcomes.^21–24 However, there were some reported cases of virological failure and selection of E157Q mutation under INSTI-based ART.^25–27 The E157Q is the most prevalent polymorphism among those listed as ‘mutations associated with resistance’ to first-generation INSTIs in the ANRS algorithm⁷ and with a drug-resistance penalty score for first-generation INSTIs in the Stanford Drug Resistance Database, although the Stanford penalty score of 10 (potential low-level resistance) is translated into ‘S’ in an SIR classification.⁶ Nonetheless, no large-scale studies have compared the virological outcomes of INSTI-based ARTs to other ARTs in PLWH harbouring E157Q polymorphisms in real-world settings where patient background and viral sequences vary.

In this study, we aimed to assess short- and long-term virological outcomes of various first-line ART regimens in PLWH harbouring the E157Q polymorphism at baseline, to elucidate the optimal first-line ART regimen for these individuals.

Methods

Study design

A multicentre retrospective observational study was conducted at HIV clinical centres included in the Japanese Drug Resistance HIV-1 Surveillance Network. This network is made up of 18 clinical centres, five public health institutes and other collaborating medical facilities in Japan. It has collected results of HIV-1 genotypic drug-resistance testing in the integrase region of ART-naïve PLWH from participating facilities since 2008.^28–30 The epidemiological characteristics of HIV-1 in Japan and the representativeness of the Japanese Drug Resistance HIV-1 Surveillance Network were described in our previous study.²⁹ In Japan, raltegravir has been available since 2008, elvitegravir since 2013, dolutegravir since 2014 and bictegravir since 2019. This study was approved by the institutional review boards of each of the participating institutions. All the participants provided written informed consent for data collection and subsequent analyses unless the requirement for written informed consent for study participation was waived by an institutional review board.

Participants

We included PLWH who underwent genotypic drug-resistance testing in the integrase region before ART initiation between 2008 and 2019 at the 18 clinical centres, were found to live with a virus bearing E157Q polymorphism, subsequently initiated ART and had available clinical data after ART initiation. The clinical information was collected from the medical records at each clinical centre. Exclusion criteria included: those who did not start ART by April 2021, those with no available clinical information, loss-to-follow-up before starting ART and loss-to-follow-up after starting ART without a single follow-up visit.

Classification of participants

Participants were classified into four groups according to their first-line ART regimen: (i) individuals who started ART with a protease inhibitor and two nucleoside/nucleotide reverse-transcriptase inhibitors (NRTIs), the PI group; (ii) individuals who started ART with raltegravir or elvitegravir/cobicistat and two NRTIs, the INSTI-1 group; (iii) individuals who started ART with once-daily dolutegravir or bictegravir and two NRTIs, the INSTI-2 group and (iv) individuals who started ART with other regimens, the Others group. Antiretroviral drugs other than NRTIs were defined as anchor drugs.

Virological outcomes

We evaluated viral suppression rate, time to achieving viral suppression after ART initiation and time to viral rebound after viral suppression. Viral suppression rate was assessed by the proportion of individuals whose plasma viral load (VL) was <50 copies/mL at 24 (24 ± 4) weeks or 48 (48 ± 8 weeks) weeks after ART initiation. The status at 24 and 48 weeks was classified into six categories: (a) continued the first-line anchor drug with VL <50 copies/mL at the timepoint; (b) continued the first-line anchor drug with VL ≥50 copies/mL at the timepoint; (c) changed the first-line anchor drug due to poor viral control by the timepoint; (d) changed the first-line anchor drug for reasons other than poor viral control by the timepoint; (e) data not taken at the timepoint with continued follow-up and (f) loss-to-follow-up. Viral suppression rates were calculated in two ways: missing = excluded analysis [a/(a + b + c)] and missing = failure analysis [a/(a + b + c + d + e + f)].

The time to achieving viral suppression was evaluated using Kaplan–Meier plots, where the first achievement of VL < 50 copies/mL after ART initiation was treated as an event, and the end of the first-line anchor drug or the last visit as censoring.

In addition, the time from once virally suppressed to viral rebound was evaluated using Kaplan–Meier plots. Viral rebound was defined as VL ≥100 copies/mL after achieving VL < 50 copies/mL. Viral rebound in this definition includes viral blip, virological failure (VF) and low-level viremia (VL 100–200 copies/mL) after once virally suppressed. The end of the first-line anchor drug, or the last visit was treated as censoring.

VF was defined as two consecutive VL >200 copies/mL after achieving VL < 50 copies/mL. Anchor drug change due to poor viral control before or after achieving VL <50 copies/mL without meeting the definition of VF were defined as ‘poor viral control without VF’.

Baseline drug-resistance mutations

Sequences of the protease and reverse-transcriptase region (HXB2: 2253–3269 nt) and integrase region (HXB2: 4230–5093 nt) were analysed for drug-resistance mutation at baseline. For drug-resistance-associated mutations in the integrase region other than E157Q, all mutations listed to any one of the following three algorithms were counted as resistance-associated mutations: Stanford University HIV Drug Resistance Database (mutations with penalty score),⁶ ANRS HIV-1 genotypic drug-resistance interpretation’s algorithms⁷ and IAS-USA HIV drug-resistance mutations.⁵ For drug-resistance mutations in protease and reverse-transcriptase region, mutations listed to the WHO surveillance were counted as resistance mutations.³¹ Subtyping was performed by jpHMM using protease, reverse-transcriptase and integrase nucleotide sequences.³²

Statistical analysis

Fisher’s exact test was used for categorical variables and the Kruskal–Wallis test was used for continuous variables to compare the baseline characteristics of the participants. For viral suppression rates at 24 and 48 weeks, univariate analysis was performed for age, sex, nationality, baseline VL, baseline CD4, first-line anchor drug and NRTI using a logistic regression analysis. Multivariate analysis was performed including first-line anchor drug and baseline VL in addition to factors with P < 0.05 in the univariate analysis. The time to achieving VL <50 copies/mL and the time from achieving VL <50 copies/mL to viral rebound were analysed using Kaplan–Meier plots and Cox proportional hazards model. The univariate Cox proportional hazards model was analysed for age, sex, nationality, baseline VL, baseline CD4, first-line anchor drug and NRTI. The multivariate Cox proportional hazards model included the first-line anchor drug and baseline VL in addition to the factors for which the P value was <0.05 in the univariate model. For the first-line ART regimen in the multivariate analysis, the PI group was used as a reference since PI-based ART was the most commonly used ART regimen other than INSTI-based ART in this study. All statistical analyses were performed using R v.4.1.2 (The R Foundation for Statistical Computing, Vienna, Austria). A P value <0.05 was considered statistically significant.

Results

Study population

A total of 4043 ART-naïve PLWH underwent genotypic drug-resistance testing in the integrase region between 2008 and 2019 at the 18 participating centres, and 167 PLWH (4.1%) were detected with the E157Q polymorphism. After excluding 23 PLWH based on the exclusion criteria, 144 (140 males and four females) individuals with E157Q were included in the study (Figure 1). The median age at diagnosis was 39 years [interquartile range (IQR): 34–48 years]. The median observation duration was 1888 days (IQR 923–2727 days) and the median duration on the first-line anchor drug was 769 days (IQR: 329–1551 days). Of the 144 included individuals, 138 (95.8%) were infected with HIV-1 subtype B.

Figure 1.

Flowchart for the study participants.

As a first-line ART regimen, 45 (31.3%) started protease inhibitor + 2 NRTIs (PI group), 33 (22.9%) started first-generation INSTI + 2 NRTIs (INSTI-1 group), 58 (40.3%) started once-daily second-generation INSTI + 2 NRTIs (INSTI-2 group) and eight (5.6%) started other regimens (Others group). One patient, where non-boosted fosamprenavir + 2 NRTIs was initiated, was classified in the PI group, quickly suppressed VL and maintained sustained viral suppression. Regarding NRTIs, abacavir/lamivudine, tenofovir disoproxil/emtricitabine or emtricitabine/emtricitabine were used as first-line NRTIs; there were two cases who did not use NRTIs that were classified in the Others group, and no cases used a single NRTI. The baseline characteristics of participants were summarized in Table 1. The PI group had a relatively early calendar year of diagnosis, while the INSTI-2 group had a relatively recent calendar year of diagnosis. Also, depending on the components of the available fixed-dose combination single tablet, the breakdown of the NRTIs among the four groups was different.

Table 1.

Baseline characteristics and follow-up durations of the study participants by first-line ART regimen

First-line ART regimen	PI	INSTI-1	INSTI-2	Others		Total
n	45	33	58	8		144
					P value
Sex, male	44 (97.8%)	32 (97.0%)	56 (96.6%)	8 (100%)	1.00	140 (97.2%)
Median age at diagnosis (IQR)	37 (35–40)	39 (33–50)	41.5 (33.25–49.5)	49 (40.5–56.5)	0.061	39 (34–48)
Nationality					0.71
Japanese	43	33	55	8		139
Other	2	0	3	0		5
The median year of diagnosis (range)	2011 (2008–2016)	2013 (2009–2019)	2016 (2011–2019)	2015 (2012–2019)	<0.001	2014 (2012–2017)
Person-years of follow-up	306.4	176.3	217.6	40.7		740.9
Person-years on the first-line anchor drug	123.9	75.9	167.4	12.9		380.2
Median follow-up duration, days (IQR)	2741 (1106–3606)	1647 (1026–2802)	1513 (679.8–2057.2)	1880 (1136–2236)	<0.001	1888 (923–2727)
Median duration on the first-line anchor drug, days (IQR)	826 (203–1634)	771 (268–1129)	794.5 (595.2–1665.2)	363 (84.75–604.25)	0.13	769 (329–1551)
first-line NRTI					<0.001
ABC/3TC	10	3	23	3		39
TDF/FTC	34	20	14	1		69
TAF/FTC	1	10	21	2		34
None	0	0	0	2		2
Log10 VL at pretreatment, median (IQR)	4.89 (4.30–5.48)	5.08(4.68–5.48)	4.90 (4.56–5.41)	5.15 (4.86–6.01)	0.37	4.95 (4.59–5.48)
VL ≥105 copies/mL at pretreatment	18 (40.0%)	18 (54.5%)	25 (43.1%)	5 (62.5%)	0.44	66 (45.8%)
CD4+ T-cell count (/µL) at pretreatment, median (IQR)	194 (61–281)	194 (51–367)	164 (47–350)	56 (47.8–185.3)	0.59	174 (51–335)
Subtype					0.08
B	41	33	57	7		138
Non-B	4	0	1	1		6
Baseline mutations other than IN_E157Q					—
RT_K219N	0	0	1	0		1
PR_L24I	0	0	1	0		1
PR_V82A	0	1	0	0		1
PR_M46I/L	1	1	4	0		6
PR_M46I/L + IN_L74I	1	0	1	0		2
IN_L74I/M	5	1	3	0		9
IN_E92G + IN_L74I	1	0	0	0		1
IN_T97A	0	1	0	0		1
IN_G140S	0	0	1	0		1

The details of first-line ART regimen were as follows: in the PI group, 36 were boosted darunavir + 2 NRTIs, four were boosted lopinavir + 2 NRTIs, two were boosted atazanavir + 2 NRTIs, two were boosted fosamprenavir + 2 NRTIs and one was non-boosted fosamprenavir + 2 NRTIs; in the INSTI-1 group, 12 were elvitegravir/cobicistat/FTC/[TAF or TDF] and 21 were raltegravir + 2 NRTIs and in the INSTI-2 group, 44 were once-daily dolutegravir + 2 NRTIs and 14 were bictegravir/FTC/TAF. In the Others group, three were twice-daily dolutegravir + 2 NRTIs, one was rilpivirine + 2 NRTIs, one was maraviroc + 2 NRTIs, one was once-daily dolutegravir + maraviroc + 2 NRTIs and two were raltegravir + etravirine.

PI, protease inhibitor + 2 NRTIs; INSTI-1, first-generation integrase strand transfer inhibitor + 2 NRTIs; INSTI-2, once-daily second-generation integrase strand transfer inhibitor + 2 NRTIs; Others, other ART regimens; NRTI, nucleoside/nucleotide reverse-transcriptase inhibitor; ABC/3TC, abacavir/lamivudine; TDF/FTC, tenofovir disoproxil fumarate/emtricitabine; TAF/FTC, emtricitabine/emtricitabine; PR, protease; RT, reverse-transcriptase; IN, integrase.

Virological outcome at 24 and 48 weeks

The VL status and CD4 count at 24 and 48 weeks after ART initiation are summarized in Table 2. In the missing = excluded analysis, the viral suppression rate at 24 weeks was 73.7% for the PI group, 87.5% for the INSTI-1 group, 91.7% for the INSTI-2 group and 80.0% for the Others group. In the multivariate model adjusted for baseline VL, the INSTI-2 group was associated with viral suppression at 24 weeks [adjusted odds ratio (aOR): 5.00, 95% confidence interval (CI): 1.24–20.10, PI group as a reference, Table S1 (available as Supplementary data at JAC Online)]. Viral suppression rate at 48 weeks was 87.1% for the PI group, 87.5% for the INSTI-1 group, 98.1% for the INSTI-2 group, and 75.0% for the Others group. The adjusted odds ratio for viral suppression at 48 weeks was high in the INSTI-2 group (aOR: 9.73, 95% CI: 0.98–97.2, PI group as a reference), but not statistically significant (Table S2). In the missing = failure analysis, viral suppression rate was 62.2%, 63.6%, 75.9%, and 50.0% at 24 weeks and 60.0%, 63.6%, 89.7%, and 37.5% at 48 weeks for the PI, INSTI-1, INSTI-2, and Others group, respectively. The multivariate model adjusted for baseline VL showed that INSTI-2 group was associated with viral suppression at 48 weeks (aOR: 6.16, 95% CI 2.15–17.6, PI group as a reference, Table S4). In both analyses, a baseline VL >10⁵ copies/mL was negatively associated with viral suppression at both time points (Tables S1–S4). One patient in the PI group and one patient in the INSTI-1 group changed first-line anchor drug due to poor viral control by 48 weeks. Details of these two cases are shown as ID 1 and ID 2 in Table 3.

Table 2.

Participants’ status according to VL and CD4 count by first-line ART regimen at 24 and 48 weeks after ART initiation

First-line ART regimen	PI	INSTI-1	INSTI-2	Others
N	45	33	58	8
VL status at 24 weeks
(a) Continued the first-line anchor drug with VL <50 copies/mL at 24 weeks	28	21	44	4
(b) Continued the first-line anchor drug with VL ≥50 copies/mL at 24 weeks	10	3	4	1
(c) Changed the first-line anchor drug due to poor viral control by 24 weeks	0	0	0	0
(d) Changed the first-line anchor drug for reasons other than poor viral control by 24 weeks	7	4	0	3
(e) Data not taken at 24 weeks	0	2	8	0
(f) Loss-to-follow-up by 24 weeks	0	3	2	0
Proportion with VL < 50 copies/mL, missing = excluded analysis, [a/(a + b + c)] (95% CI)	73.7% (56.9%–86.6%)	87.5% (67.6%–97.3%)	91.7% (80.0%–97.7%)a	80.0% (28.4%–99.5%)
Proportion with VL < 50 copies/mL, missing = failure analysis, [a/(a + b + c + d + e + f)] (95% CI)	62.2% (46.5%–76.2%)	63.6% (45.1%–79.6%)	75.9% (62.8%–86.1%)	50.0% (15.7%–84.3%)
CD4+ T-cell count (/µL) at 24 weeks (IQR)	317.5 (210.8–502.2)	383 (188.2–514.2)	350.5 (185.8–486)	140 (119–171)
VL status at 48 weeks
(a) Continued the first-line anchor drug with VL <50 copies/mL at 48 weeks	27	21	52	3
(b) Continued the first-line anchor drug with VL ≥50 copies/mL at 48 weeks	3	2	1	1
(c) Changed the first-line anchor drug due to poor viral control by 48 weeks	1b	1b	0	0
(d) Changed the first-line anchor drug for reasons other than poor viral control by 48 weeks	10	6	0	4
(e) Data not taken at 48 weeks	0	0	0	0
(f) Loss-to-follow-up by 48 weeks	4	3	5	0
Proportion with VL < 50 copies/mL, missing = excluded analysis, [a/(a + b + c)] (95% CI)	87.1% (70.2%–96.4%)	87.5% (67.6%–97.3%)	98.1% (89.9%–100.0%)	75.0% (19.4%–99.4%)
Proportion with VL < 50 copies/mL, missing = failure analysis, [a/(a + b + c + d + e + f)] (95% CI)	60.0% (44.3%–74.3%)	63.6% (45.1%–79.6%)	89.7% (78.8%–96.1%)a	37.5% (8.5%–75.5%)
CD4+ T-cell count (/µL) at 48 weeks (IQR)	367.5 (262.2–619.2)	388 (273–537.5)	405 (281–510)	209.5 (192.8–251.5)

PI, protease inhibitor + 2 NRTIs; INSTI-1, first-generation integrase strand transfer inhibitor + 2NRTIs; INSTI-2, once-daily second-generation integrase strand transfer inhibitor + 2NRTIs; Others, other regimens.

^aThe aOR for viral suppression was statistically significant (P < 0.05, PI group was used as a reference).

^bDetails are shown in ID 1 and ID 2 in Table 3.

Table 3.

Cases of VF or poor viral control without meeting the definition of VF

ID	First-line ART regimen	Group	Age at diagnosis	Sex	VL, baseline (copies/mL)	CD4, baseline (/μL)	HIV-1 subtype	Baseline resistance-associated mutations other than IN_E157Q	Days to VL < 50 after starting ART	Type of event	Days to event	VL at event	Acquired drug-resistance mutation
1	DRV + rtv + ABC/3TC	PI	20s	M	2 300 000	468	B	None	154	Poor viral control without VF	244	210	Not tested
2	RAL + TDF/FTC	INSTI-1	30s	M	300 000	173	B	None	not achieved	Poor viral control without VF	329	140	Not tested
3	RAL + ABC/3TC	INSTI-1	60s	M	50 000	4	B	None	25	VF	742	47 000	IN_E92EA, IN_V151VA, IN_N155NH
4	EVG/COBI/TAF/FTC	INSTI-1	20s	M	137 000	636	B	None	91	VF	1135	4540	No DR mutation

VF, virological failure was defined as two consecutive VL > 200 copies/mL after once achieving VL < 50 copies/mL; poor viral control without VF was defined as change of the anchor drug due to poor viral control before or after achieving VL < 50 copies/mL without meeting the definition of VF; DRV, darunavir; ABC/3TC, abacavir/lamivudine; RAL, raltegravir; TDF/FTC, tenofovir disoproxil fumarate/emtricitabine; EVG/COBI/TAF/FTC, elvitegravir/cobicistat/tenofovir alafenamide/emtricitabine; DR, drug resistance.

Time to viral suppression

The Kaplan–Meier analysis of the time from ART initiation to viral suppression (VL <50 copies/mL) is shown in Figure 2. The multivariate analysis adjusted for baseline VL and baseline CD4 count showed that the INSTI-1 group [adjusted hazard ratio (aHR): 4.14, 95% CI: 2.39–7.19, PI group as a reference] and the INSTI-2 group (aHR: 4.76, 95% CI: 2.99–7.57, PI group as a reference) were associated with shorter time to viral suppression (Table S5). Patients with baseline VL >10⁵ copies/mL were associated with a longer time to viral suppression (aHR: 0.34, 95% CI: 0.22–0.51) (Table S5). One patient in the INSTI-1 group (ID 2 in Table 3) changed the anchor drug due to poor viral control before achieving VL < 50 copies/mL, and this point was not treated as censoring.

Figure 2.

Kaplan–Meier analysis of time to viral suppression (VL <50 copies/mL) after ART initiation by first-line ART regimen. PI, protease inhibitor + 2 NRTIs; INSTI-1, first-generation integrase strand transfer inhibitor + 2 NRTIs; INSTI-2, once-daily second-generation integrase strand transfer inhibitor + 2 NRTIs; Others, other ART regimens.

Time from once virally suppressed to viral rebound

The time from once achieving VL < 50 copies/mL to viral rebound (VL ≥100 copies/mL) was analysed by Kaplan–Meier analysis (Figure 3). There were 19 cases of viral rebound (six in the INSTI-1 group, five in the INSTI-2 group, seven in the PI group and one in the Others group). In both univariate and multivariate models, the hazard ratio for viral rebound was low in the INSTI-2 group (crude HR: 0.45, 95% CI: 0.14–1.42, aHR: 0.43, 95%CI: 0.14–1.36, PI group as a reference), but was not statistically significant (Table S6).

Figure 3.

Kaplan–Meier analysis of time from once virally suppressed to viral rebound. Viral rebound was defined as VL ≥100 copies/mL after once achieving VL <50 copies/mL. ART, antiretroviral therapy; PI, protease inhibitor + 2 NRTIs; INSTI-1, first-generation integrase strand transfer inhibitor + 2 NRTIs; INSTI-2, once-daily second-generation integrase strand transfer inhibitor + 2 NRTIs; Others, other ART regimens.

Virological failure (VF) or poor viral control without VF

Of the 144 cases, 380 person-years of follow-up on first-line anchor drugs, there were two cases of VF defined as two consecutive VL >200 copies/mL, both in the INSTI-1 group (ID 3 and ID 4 in Table 3). There were two cases that did not meet the criteria for VF but changed their anchor drug due to poor viral control, as described before (ID1 and ID2 in Table 3). ID 3 started ART with raltegravir + abacavir/lamivudine and developed VF at 742 days after ART initiation with newly acquired mutations: E92EA, V151VA and N155NH in the integrase region. ID 4 started ART with elvitegravir/cobicistat/tenofovir alafenamide/emtricitabine (EVG/cobi/TAF/FTC), and developed VF at 1135 days after ART initiation, but no resistance mutations were detected, and the VL was suppressed again to <50 copies/mL just before the modification of ART regimen.

Cases with drug-resistance-associated mutations in the integrase region other than E157Q at baseline

Drug-resistance-associated mutations in the integrase region other than E157Q at baseline were found in 14 cases (Table 1). Two of them started first-generation INSTI, five started second-generation INSTI and seven started PI-based ART. None of them demonstrate VF or poor viral control without VF (Table S7).

Discussion

This multicentre, retrospective observational study investigated the virological outcomes of 144 ART-naïve individuals harbouring E157Q who initiated ART over a median follow-up of 1888 days. To the best of our knowledge, this is the largest study evaluating the long-term virological outcomes in PLWH harbouring E157Q at baseline in real-world settings. These results demonstrated that the second-generation INSTI-based regimen showed similar or favourable outcomes compared to the PI-based regimen for viral suppression rates at 24 and 48 weeks after ART initiation, time to viral suppression and time from viral suppression to viral rebound.

The impact of baseline E157Q on the virological outcome of INSTI-based first-line ART has been reported in several sub-analyses of randomized controlled trials.^21,22,24 However, there are not many patients with E157Q. One of the largest reports was the NAMSAL study comparing a dolutegravir-based versus a low-dose efavirenz-based regimen conducted in Cameroon, where CRF02_AG predominantly circulates. The study reported that 8.4% (25 in the dolutegravir group and 26 in the low-dose efavirenz group) of the study participants had E157Q at baseline, and that the presence of the baseline mutations was not associated with an absence of viral suppression at 48 and 96 weeks.^21,22 The largest study in subtype B endemic countries is the GS-US-380-1489 and GS-US-380-1490 study. It showed that baseline INSTI-associated mutations, including E157Q, did not significantly affect viral suppression rates at 48 weeks after initiating second-generation INSTI-based first-line ART.²⁴ Moreover, one of the largest previous observational studies reported that among 646 patients who started an INSTI-based regimen as a first-line ART, 104 had INSTI-associated minor mutations at baseline and that these mutations were not associated with virological outcomes.²³ However, in this previous study L74I/M accounted for most of the INSTI-associated mutations and there were only 14 cases with E157Q.²³ Another study found that E157Q was detected in six ART-naïve patients with acute/recent infection and eight ART-naïve patients with chronic infection, five of them started first-generation INSTI and five started second-generation INSTI as a first-line ART, and all of them were virally suppressed by 1 year.³³ However, another previous study reported that two of the eight ART-naïve patients with E157Q who started INSTI-based first-line ART failed to achieve VL < 50 at 24 weeks, and the first-line ART of these two cases were first-generation INSTI-based⁹. In addition, a previous case report of a patient with E157Q at baseline showed that first-line ART with raltegravir + abacavir/lamivudine and second-line ART with dolutegravir + abacavir/lamivudine resulted in a nonvirological response.²⁵ Additionally, two case reports showed that E157Q emerged following raltegravir-based ART as a salvage therapy.^26,27 The present data together with previous reports indicate that a second-generation INSTI-based regimen is the preferred first-line ART in achieving and maintaining viral suppression in individuals with the E157Q polymorphism, while a small number of failures with first-generation INSTI-based regimen could not dispel concerns.

Regarding the coexistence of E157Q with other drug-resistance-associated mutations in the integrase region, there were 11 cases with L74I/M, one with T97A, one with E92G + L74I and one with G140S, and seven of the 14 cases started INSTI-based first-line ART with good virological response. L74I/M is a polymorphic mutation with higher prevalence than E157Q and is an accessory mutation that does not have a penalty score by itself in the Stanford database.⁶ Our study suggests that INSTI-based first-line ART is effective even when L74I/M coexists with E157Q.

Our study has some limitations. First, the study does not compare the virological outcomes with individuals without E157Q due to the difficulty in obtaining data on a sufficient number of individuals without E157Q in our surveillance. Therefore, our study cannot evaluate an E157Q-specific effect on virological outcomes. The comparable or superior viral suppression rate in the second-generation INSTI-based first-line ART compared with PI-based ART has been reported in the FLAMINGO study not in the context of E157Q.³⁴ The results of our study confirm that viral suppression rates in second-generation INSTI-based ART were comparable or superior to those in PI-based ART that is similar to the results of the FLAMINGO study even in the presence of E157Q. Second, our study retrospectively analysed the outcomes of regimen choices made by physicians, influenced by a variety of clinical and sociodemographic backgrounds. Although we conducted multivariate analyses to correct for background factors, we could not adjust for all potential confounders that might influence virological outcomes. Additionally, we could not assess adherence to ART. Third, VF and other poor viral control occurred in only four cases, three of them were in the INSTI-1 group but were too few to perform a statistical analysis. Furthermore, one of the VF cases started EVG/cobi/TAF/FTC and had viral suppression prior to ART modification. Therefore, we cannot conclude from our study that VF is significantly more frequent in the INSTI-1 group.

Conclusion

The general guideline recommendation of second-generation INSTI-based first-line ART for most PLWH is also applicable to PLWH harbouring E157Q in terms of short- and long-term virological outcomes.

Acknowledgements

We are grateful to all individuals registered in our surveillance network and to all medical staff for the time and effort they expended in surveillance. We would like to thank the members of the Japanese Drug Resistance HIV-1 Surveillance Network: Junko HATTORI (National Hospital Organization Nagoya Medical Center and MSD K.K.); Masakazu MATSUDA, Urara SHIGEMI and Reiko OKAZAKI (National Hospital Organization Nagoya Medical Center); Mami NAGASHIMA, Kenji SADAMASU (Tokyo Metropolitan Institute of Public Health); Makiko KONDO (Kanagawa Prefectural Institute of Public Health); Haruyo MORI (Osaka Institute of Public Health); Asako NAKAMURA (Fukuoka Institute of Health and Environmental Sciences); Kaori SATO (Hokkaido University Hospital), Tamayo WATANABE (Ishikawa Prefectural Central Hospital); Kiyonori TAKADA (Ehime University Hospital); Koji SUDO (Keio University School of Medicine and Hanah MediTech); Hiroshi FUJIWARA, Yoshifumi UWAMINO, Sho UCHIDA, Kei MIKITA, Ayumi YOSHIFUJI and Ho NAMKOONG (Keio University School of Medicine); Yuko TOMAKI, Nanae NITTA, Hiroshi KOTANI, Naomi NISHIMATSU, Haruka KANNARI and Yumi UEDA (Keio University Hospital); Yasuo OTA (National Hospital Organization Higashisaitama National Hospital); Hirotomo NAKATA and Takamasa UENO (Kumamoto University); Junichi MASUDA, Teiichiro SHIINO (National Center for Global Health and Medicine) and Machiko OTANI, Shioko KOJIMA (National Institute of Infectious Diseases)

Funding

This work was supported by the Japan Agency for Medical Research and Development (AMED) (JP21fk0410028, JP22fk0410050, JP23fk0410050).

Transparency declarations

D.W. received honoraria from Gilead Sciences, K.K., ViiV Healthcare K.K., MSD K.K. and Janssen Pharmaceutical K.K. out of this work. T.F. received speaking and advisory fees from ViiV Healthcare and Gilead Science. T. Teshima received grants and personal fees from Chugai, grants and personal fees from Kyowa Kirin, grants from Fuji Pharma, grants and personal fees from Nippon Shinyaku, grants and non-financial support from Asahi Kasei Pharma, grants from Eisai, grants, personal fees and non-financial support from Sumitomo Pharma, grants from Ono, grants, personal fees and non-financial support from Astellas, grants from Shionogi, grants from Priothera SAS, grants from LUCA Science, grants from Otsuka, personal fees from AbbVie, personal fees and non-financial support from Novartis, personal fees from Bristol-Myers Squibb, personal fees from Merck Sharp & Dohme, personal fees and non-financial support from Celgene, personal fees and non-financial support from Janssen, non-financial support from Meiji Seika Pharma, non-financial support from Daiichi Sankyo, non-financial support from AstraZeneca, non-financial support from Roche Diagnostics, non-financial support from Sanofi, non-financial support from Takeda out of this work. Y. Yokomaku received honoraria from Gilead Sciences, K.K. W.S. received honoraria from ViiV Healthcare K.K., GSK., and MSD K.K. All other authors declare no conflict of interest.

Author contributions

S.U., T.K., H.G. and M.I. conceived and designed the analyses. All authors were responsible for data collection and management. S.U. and T.K. performed the statistical analysis. S.U. and T.K. wrote the manuscript and generated tables and figures. All authors critically reviewed and approved the final version of the manuscript.

Supplementary data

Tables S1–S7 are available as Supplementary data at JAC Online.