Early Virologic Success on Antiretroviral Therapy Among Individuals With Breakthrough HIV Acquisition on Long-Acting Cabotegravir Pre-exposure Prophylaxis

Abstract

HIV-1 breakthrough on long-acting cabotegravir (CAB-LA) for HIV prevention is rare but could impact viral suppression on integrase strand transfer inhibitor (INSTI)–based antiretroviral therapy (ART). We report the first study of ART outcomes following CAB-LA breakthrough during routine clinical care. Three individuals acquired HIV-1 on CAB-LA despite on-time injections; 2 had low-frequency major INSTI resistance mutations. All started darunavir-based ART; 1 subsequently switched to bictegravir-based ART. All maintained plasma HIV-1 RNA <50 copies/mL through ≥5 months post–ART initiation, providing early evidence of virologic success with standard ART regimens and paving the way for longer-term studies on optimal ART after CAB-LA breakthrough.

Long-acting injectable cabotegravir (CAB-LA) for HIV-1 prevention is highly effective and is key to advancing global goals of reducing HIV incidence. HIV incidence has been shown to be low in CAB-LA prevention trials, but infections that occur within 6 months of the last CAB-LA injection have been associated with delayed seroconversion and selection of integrase strand transfer inhibitor (INSTI) resistance-associated mutations [1]. Emergent signature mutations in the integrase gene can decrease susceptibility to other INSTI drugs including dolutegravir (DTG) and bictegravir (BIC), with higher levels of resistance occurring as multiple INSTI mutations accumulate [2].

BIC and DTG are recommended as part of initial antiretroviral therapy (ART) for most people with HIV-1, but the subsequent efficacy and durability of INSTI-based regimens in individuals who acquire HIV-1 on CAB-LA are unknown. For persons with HIV-1 diagnosed after CAB-LA pre-exposure prophylaxis (PrEP) exposure, US guidelines currently recommend protease inhibitor (PI)–based ART with cobicistat or ritonavir-boosted darunavir unless INSTI resistance mutations are not detected by standard genotype analysis [3, 4]. The World Health Organization (WHO) has not yet recommended a preferred ART regimen for persons with HIV-1 diagnosed after CAB-LA PrEP [5]. Darunavir-based ART is a potent alternative but is generally prescribed only when necessary due to tolerability issues, drug interactions, and limited access in some global settings [6]. Overall, data are limited to inform the selection of an ART regimen in persons diagnosed with HIV-1 after CAB-LA exposure. Here, we report the first study of treatment outcomes following HIV-1 acquisition on CAB-LA PrEP in routine clinical care in the United States.

METHODS

Study Design

SeroPrEP (seroprep.ucsf.edu) is an observational study assessing low-frequency ARV resistance, drug concentrations, and subsequent HIV treatment outcomes among persons who acquire HIV-1 while using PrEP modalities [7]. SeroPrEP enrolls individuals aged ≥18 years in routine care settings in the United States with new evidence of HIV infection and long-acting PrEP use in the last 18 months and is approved by the University of California, San Francisco, Committee on Human Research. After participants provide informed consent, clinical test results are obtained from routine clinical care, including results of HIV diagnostic antibody/antigen (Ab/Ag) tests, commercial plasma HIV-1 RNA tests, and commercial HIV genotype analysis for drug resistance mutations.

Sensitive Quantitation of HIV-1 RNA, Integrase Gene Mutation Frequency, and Cabotegravir Pharmacokinetics

To quantify plasma HIV-1 RNA levels below routine commercial thresholds, we use a single-copy assay (SCA) with a detection limit of 0.4 HIV-1 RNA copies/mL using the Aptima HIV-1 Quant Dx on the Hologic Panther (Marlborough, MA, USA) as previously described [8]. Low-frequency INSTI-associated mutations are identified in plasma HIV-1 RNA using single-genome sequencing (SGS) of full-length integrase based on previous methods where 45 sequences provide 90% certainty of detecting variants at a ≥5% frequency [9]. SCA and SGS are performed at the University of Pittsburgh Infectious Disease Laboratories (Pittsburgh, PA, USA).

Plasma CAB concentrations of study participants are quantified via liquid chromatography-tandem mass spectrometry at the University of California San Francisco Hair Analytical Laboratory (San Francisco, CA, USA) using methods approved by the National Institutes of Health's Clinical Pharmacology and Quality Assurance Program. The plasma assay has a linear dynamic range of 0.0100–20.0 µg/mL and a lower limit of quantitation (LLOQ) of 0.0100 µg/mL. Plasma cabotegravir concentrations were compared with published pharmacokinetic (PK) targets for protein-adjusted concentrations required for 90% inhibition of virus (PA-IC₉₀) [10].

Study Outcomes

Viral suppression was defined as HIV-1 RNA <50 c/mL through the longest available follow-up for each participant. INSTI resistance mutations were defined by the Stanford University HIV-1 Drug Resistance Database, version 9.6 (2024-03-09; https://hivdb.stanford.edu/dr-summary/resistance-notes/INSTI/).

RESULTS

Three men from the United States were diagnosed with HIV-1 despite CAB-LA PrEP use and subsequently enrolled in the SeroPrEP study. The 3 individuals received 5, 6, and 13 on-time CAB-LA injections, respectively. All had taken oral PrEP (emtricitabine with tenofovir disoproxil fumarate [F/TDF] or tenofovir alafenamide [F/TAF]) before the switch to cabotegravir. Participant 003A received an oral CAB lead-in, while participants 001A and 002A switched directly from oral tenofovir-based PrEP to CAB-LA injections, as approved by the US Food and Drug Administration. All individuals reported multiple sexual partners, a history of sexually transmitted infections, and no injection drug use. Body mass index (BMI) at the time of HIV-1 diagnosis was in the overweight/obese range for Participants 001A (31.5 kg/m²) and 003A (35.5 kg/m²) (Table 1).

Table 1.

Participant Characteristics and Testing History During Use of Long-Acting Cabotegravir Pre-exposure Prophylaxis

	Participant 001A	Participant 002A	Participant 003A
Participant characteristics
Age, y	30	39	67
Sex	Male	Male	Male
BMI at HIV-1 detection, kg/m2	31.5	25.0	35.5
History of sexually transmitted infections	Syphilis, chlamydia	Pharyngeal and rectal gonorrhea	Syphilis, rectal chlamydia, pharyngeal and rectal gonorrhea
Partner history in past 3–6 mo	Multiple partners	Multiple partners	Multiple partners
Injection drug use	No	No	No
Noninjection drug use	Marijuana, cocaine	Marijuana	Marijuana
PrEP history
Prior oral PrEP, duration in mo	F/TAF (24)	F/TDF (22)	F/TAF (9)
No. of CAB-LA injections  before HIV-1 diagnosis	5	6	13
All injections on time	Yes	Yes	Yes
Cabotegravir levels
Concentration in plasma	0.690 µg/mL	2.60 µg/mL	1.05 µg/mL
Comparison to PK targets based on PA-IC90a	4×	16×	6×
Days after last injection	42	21	55
Sample collection timing	Random	Random	Trough

Abbreviations: BMI, body mass index; CAB-LA, cabotegravir long-acting; F/TAF, emtricitabine/tenofovir alafenamide; F/TDF, emtricitabine/tenofovir disoproxil fumarate; PA-IC₉₀, protein-adjusted 90% inhibitory dose; PK, pharmacokinetic; PrEP, pre-exposure prophylaxis.

^aThe 600-mg dose of CAB-LA (every 8 weeks) was targeted to achieve >4 × the protein-adjusted 90% inhibitory dose of 0.66 µg/mL in 80% of individuals and >8 × the PA-IC₉₀ of 1.33 µg/mL in 50% of individuals [10].

Detection of HIV-1 Infection

Participants 001A and 002A had negative point-of-care Ag/Ab tests before each injection and at the time of first detectable HIV-1 RNA. HIV-1 was diagnosed in both individuals based on detectable HIV-1 RNA on 2 samples collected on separate days: 4880 and 256 copies/mL (Participant 001A) and 3940 and 7000 copies/mL (Participant 002A). Participant 003A was monitored for HIV-1 by HIV RNA testing accompanying most injections. HIV-1 was detected after the 13th CAB-LA injection when laboratory results returned with a reactive Ag/Ab result and 2 430 000 HIV-1 RNA copies/mL (Figure 1).

Figure 1.

Antiretroviral treatment outcomes and HIV-1 drug resistance profiles in 3 individuals who acquired HIV-1 after long-acting cabotegravir pre-exposure prophylaxis. For all 3 participants, antiretroviral regimens are depicted above the HIV-1 test results as follows: all individuals had pre-exposure prophylaxis (PrEP) with long-acting cabotegravir (CAB-LA); Participants 001A and 002A switched to the antiretroviral (ART) regimen cobicistat-boosted darunavir/emtricitabine/tenofovir alafenamide (DRV/c/F/TAF); Participant 003A had no regimen for 12 d post-HIV detection, then started DRV/c/F/TAF, and switched to bictegravir/F/TAF. Point-of-care or commercial laboratory–based HIV-1 antibody/antigen (Ab/Ag) diagnostic test results are indicated as negative or reactive, and commercial HIV-1 RNA results are reported as copies/mL in nonbold text. Research laboratory testing (in bold) was done by quantitative polymerase chain reaction single-copy assay with a detection limit of 1 RNA copy/mL, and INSTI-associated mutations were identified in plasma HIV RNA using single-genome sequencing of full-length integrase. The proportion presented is the number of sequences with the noted mutation over the total number of genomes sequenced.

Pharmacokinetic Testing for Cabotegravir

CAB concentrations were measured in plasma samples collected at random time points from Participant 001A 42 days after the last CAB-LA injection (0.69 µg/mL; 4 × PA-IC₉₀) and Participant 002A 21 days after the last CAB-LA injection (2.6 µg/mL; 16 × PA-IC₉₀). For Participant 003A, a trough plasma concentration was measured 55 days (∼8 weeks) after the last CAB-LA injection, with a CAB concentration of 1.05 µg/mL (6 × PA-IC₉₀). All 3 plasma CAB concentrations were above the published PK target of 4 × PA-IC₉₀ achieved by 80% of participants but below the 8 × PA-IC₉₀ achieved by 50% of participants in phase 2 trials [10]. (Table 1). The concentration threshold conferring protection from HIV-1 acquisition on CAB has not yet been defined.

Low-Frequency Detection of Integrase Strand Transfer Inhibitor–Associated Mutations

INSTI mutations were not detected in samples collected from any participant using population genotyping assays 7 to 13 days after the first detectable HIV-1 RNA. SGS, however, identified low-frequency INSTI mutations in Participants 002A and 003A. For Participant 002A, the major INSTI mutation E138K was observed in 1 of 56 sequences from a sample collected 1 week post–HIV-1 detection, the N155K polymorphism in integrase was observed in 1 of 43 sequences 2 weeks post–HIV-1 detection, and the mutation G140R was detected in 1 of 41 sequences 24 weeks post–HIV-1 detection. For Participant 003A, the major INSTI mutations Q148R (13 of 50 sequences) and R263K (1 of 50 sequences) and accessory mutation G163R (1 of 50 sequences) were observed by SGS in a sample that was collected 12 days post–HIV-1 detection (Figure 1). In Participant 001A, INSTI mutations were not detected by SGS 6 weeks after HIV-1 detection; however, due to very low RNA levels (4 c/mL), only 4 sequences were obtained, limiting sensitivity.

ART Regimen Selection and Virologic Outcomes

Clinicians for all 3 participants selected an initial darunavir (PI)-based ART regimen, in line with US HIV-1 guidelines for ART in individuals with prior CAB-LA experience [3, 4]. Participant 001A started darunavir/cobicistat (DRV/c)/F/TAF 7 days after first detected HIV-1 RNA (when confirmatory HIV-1 RNA was drawn) and maintained HIV-1 viral suppression (<50 c/mL) at 24 weeks post–ART initiation. Participant 002A initiated DRV/c/F/TAF 13 days after first detected HIV-1 RNA and remained virally suppressed up to 42 weeks post–ART start. Participant 003A started ART with DRV/c/F/TAF 12 days after first detected RNA. He switched to an INSTI-based regimen of BIC/F/TAF 14 days later due to drug interactions and intolerance and based on a commercial population genotype that showed no INSTI resistance. HIV-1 RNA was 22 copies/mL 4 weeks after switch to BIC/F/TAF. Participant 003A remained virally suppressed at 24 weeks postdiagnosis (after 20 weeks on BIC/F/TAF) with 5 HIV-1 RNA copies/mL by SCA and no INSTI resistance mutations detected in 2 sequences by SGS.

DISCUSSION

Here, we provide some of the first evidence of HIV-1 responses to ART initiated following breakthrough on CAB-LA PrEP in routine clinical care. We also report on the first case in any setting describing viral suppression on INSTI-based ART (with BIC/F/TAF) following CAB-LA PrEP breakthrough in an individual with low-frequency major INSTI mutations detected by SGS. This study demonstrates the possibility that INSTI-based ART could remain an option for individuals with prior CAB-LA PrEP experience and no evidence of high-frequency INSTI resistance, paving the way for future larger-scale studies to address this question.

The effectiveness of INSTI-based ART in individuals who acquired HIV on CAB-LA PrEP is of significant clinical concern. Two participants in our study had low-frequency major INSTI mutations (E138K, Q148R, and R263K) detected by SGS shortly after HIV-1 diagnosis. In HPTN 083, 6 individuals who acquired HIV-1 <6 months after the last injection with on-time injections also had E138K, Q148R, or R263K detected, plus other major INSTI mutations, including N155H and S230R. All 6 subsequently started non-INSTI-containing ART (boosted darunavir or efavirenz with 2 nucleoside reverse transcriptase inhibitors), and 5 achieved sustained virologic suppression <40 copies/mL between 45 and 140 weeks postdiagnosis (follow-up after ART was unknown for 1 individual) [1, 11]. Our study is the first to our knowledge to report on early ART modification to INSTI-based ART (BIC/F/TAF) in an individual with low-frequency INSTI resistance on SGS following HIV-1 acquisition on CAB-LA PrEP. This individual achieved and maintained HIV-1 viral suppression 6 months after HIV diagnosis.

Delayed seroconversion has been reported with HIV-1 acquisition on CAB-LA PrEP and occurred in 2 of the 3 participants in this study, consistent with findings from HPTN 083. Antibody detection remained equivocal 6 weeks after the first detectable HIV-1 RNA level in Participant 001A, and at least 7 weeks in Participant 002A, using a sensitive Ag/Ab immunoassay [12]. Due to the cessation of diagnostic testing after starting ART in clinical practice, the specific duration of delayed seroconversion could not be quantified. However, HIV-1 was detected by RNA (before reactive Ag/Ab) in 2 individuals, which may have prevented the accumulation of INSTI mutations or the evolution of selected INSTI mutations to higher levels of frequency. These data highlight the need for improved diagnostics for persons using long-acting PrEP, particularly where point-of-care rapid tests are currently the only option for HIV-1 testing.

The cause of breakthrough HIV-1 acquisition after the long duration of on-time CAB-LA PrEP use (5 to 13 bimonthly injections) remains unknown. Although all 3 participants had plasma CAB concentrations >4 × PA-IC₉₀ targets at the time of sampling, drug levels could have been lower at the time of a true trough in Participants 001A and 002A, whose plasma was drawn before a trough level. Prior CAB concentrations in all 3 participants could have been lower or higher than the levels observed from a single time point after the last CAB injection and HIV-1 diagnosis. Participants 001A and 003A had a BMI >30 kg/m², raising the possibility of inadequate drug exposure if the injection was not delivered intramuscularly due to differences in body habitus and body fat distribution. An analysis of the FLAIR, ATLAS, and ATLAS-2M studies found an association between BMI ≥30 kg/m² and confirmed virologic failure in individuals treated with CAB/RPV, though a similar association has not been reported in the HPTN studies of CAB as PrEP [13]. A longer 2-inch needle is recommended for individuals with a BMI >30 kg/m²; however, the needle length used for CAB injections for Participants 001A and 003A was not consistently documented [14]. No major drug interactions between CAB and concomitant medications were identified, and all 3 individuals reported no injection drug use. Further research is needed to define thresholds for drug concentrations conferring protection from HIV-1 acquisition and to identify additional risk factors for HIV breakthrough on CAB-LA and long-acting PrEP agents.

As access to CAB-LA and future long-acting PrEP expands, breakthrough HIV-1 acquisition is expected to remain rare but has significant potential consequences for subsequent ART selection [15]. Our data provide early evidence that individuals who acquire HIV-1 on CAB-LA in routine clinical care can achieve virologic success using commonly available ART regimens in the United States, with the limitation that the participant who switched to INSTI-based ART only had 24 weeks of follow-up and 20 weeks on INSTI-based ART. Longer-term follow up is needed to evaluate the durability of virologic success in individuals with low-frequency major INSTI mutations.

Continued investigation to evaluate a larger number of individuals who acquire HIV-1 on long-acting products including CAB-LA will be important for addressing key knowledge gaps: (i) understanding the clinical significance of INSTI mutations selected by CAB that are detected at low frequency; (ii) determining which CAB-selected mutations have the greatest impact on outcomes on INSTI-based ART; and (iii) assessing the long-term durability of viral suppression on first-line dolutegravir-based ART used globally after HIV-1 acquisition on CAB-LA PrEP. SeroPrEP and other studies are addressing these questions to inform optimal ART regimen selection after HIV-1 acquisition on long-acting PrEP [16].