Genotyping not required for sustained effectiveness of long-acting cabotegravir plus rilpivirine: evidence from the RELATIVITY cohort

Abstract

Objectives:

Long-acting injectable cabotegravir and rilpivirine (LAI CAB+RPV) provides an alternative to daily therapy for people with HIV (PWH) with virologic suppression. Although genotypic testing is recommended before switching, its real-world clinical value is unclear. We assessed outcomes after switching to LAI CAB+RPV with or without available genotypes in the Spanish RELATIVITY cohort.

Design:

RELATIVITY is a multicenter, ambispective cohort study assessing the effectiveness and safety of LAI CAB+RPV in adults with HIV across 58 centers in Spain.

Methods:

Posthoc analysis of 3146 participants, focusing on the availability of genotypic resistance data before switching.

Results:

Of the 3146 participants, 53.5% (n = 1682) did not have genotypes available. The median follow-up was 13.3 months [interquartile range (IQR) 8.6, 18.9] in the no-genotype group and 14.9 months (IQR 9.0, 19.2) in the genotype group (P = 0.003). Both groups maintained high virological suppression rates (>93%) up to the 23rd month of follow-up, with no significant differences observed in virological or immunological outcomes. Virologic failure rates (0.5% vs. 1.0%; P = 0.476) and permanent discontinuation rates (6.1% vs. 6.6%; P = 0.804) were similar. Of the 20 participants with virologic failure, 12 had genotype data. After resuming oral antiretroviral therapy, 8 of those with and 4 of those without the genotype achieved undetectable viral loads. Adherence to injection schedules and changes in body mass index were comparable.

Conclusions:

In this large real-world cohort, the absence of genotypic data did not affect LAI CAB+RPV effectiveness in virologically suppressed PWH. Limitations, including ambispective design, short follow-up, and low non-B subtype prevalence, may limit generalizability.

Introduction

Long-acting injectable (LAI) cabotegravir (CAB) plus rilpivirine (RPV) has emerged as an innovative novel therapeutic option for people with HIV (PWH) who have achieved virologic suppression [1,2]. Clinical trials [3–7] and real-world studies [8–22] have consistently demonstrated the efficacy and safety of this regimen, with high rates of virologic suppression maintained over extended periods. Reflecting these outcomes, major clinical guidelines [23–25] now recommend LAI CAB+RPV as an alternative to daily oral regimens, offering the flexibility of monthly or bimonthly dosing. However, as with any antiretroviral therapy (ART), understanding the factors influencing treatment outcomes is crucial for optimal patient management and long-term success.

Several baseline characteristics have been identified as potential risk factors for virologic failure (VF) in people receiving LAI CAB+RPV [26]. A posthoc multivariable analysis of pooled data from phase III trials (FLAIR, ATLAS, and ATLAS-2M) highlighted body mass index (BMI) >30 kg/m², HIV subtype A1/A6, and the presence of RPV resistance-associated mutations (RAMs) as predictive factors for VF [26]. While these findings have informed clinical decision-making, the relative importance and applicability of each factor in diverse real-world settings remain subjects of ongoing investigation. Notably, the CARES study, with its more extensive and more diverse patient population, is a randomized prospective trial that used retrospective proviral DNA testing to detect archived resistance and suggested that some initially identified risk factors, such as HIV subtype A1 or genotype, may not significantly impact treatment outcomes in all contexts [7].

In this context, the RELATIVITY cohort, a nationwide Spanish study cohort in real-world settings, provides a unique opportunity to examine these issues further [8,27]. This multicenter, ambispective study includes PWH from 58 institutions across Spain, with planned follow-up extending to at least 2029. Remarkably, the RELATIVITY cohort represents the largest European cohort to explore the role of genotyping using real-world data, offering valuable insights beyond previous studies.

This work presents a posthoc subanalysis of the RELATIVITY cohort that compares efficacy outcomes in virologically suppressed people with PWH who switched to LAI CAB+RPV, with or without prior genotypic information available at the time of switching. We aimed to assess the clinical relevance of genotypic testing by analyzing baseline characteristics and outcomes between these groups. By leveraging data from the RELATIVITY cohort, this research sought to provide evidence that informs clinical guidelines on the necessity of genotyping before initiating LAI CAB+RPV.

Methods

Study design and setting

The RELATIVITY study is a nationwide, multicenter, ambispective cohort designed to evaluate the effectiveness and safety of long-acting injectable cabotegravir plus rilpivirine (LAI CAB+RPV) in real-world clinical settings across Spain. The study started in June 2023, shortly after the public funding and availability of LAI CAB+RPV in Spain in December 2022. This ambispective design enables the inclusion of patients who initiated treatment before the study's launch (retrospective component) as well as those who are followed prospectively after enrollment. Participating centers have joined the cohort progressively over the past 2 years, contributing retrospective data on previously treated patients and conducting ongoing prospective follow-up through 2029.

Participants

Participants included virologically suppressed adults (≥18 years) with HIV who had received at least one dose of LAI CAB+RPV. Exclusion criteria included pregnancy, prior participation in LAI CAB+RPV clinical trials, and known resistance mutations to RPV or second-generation integrase strand transfer inhibitors (INSTIs). In participants without prior genotypic resistance data, the absence of RAMs was inferred through a thorough review of medical history, ensuring no prior VF that could compromise the efficacy of CAB or RPV. This approach, combined with sustained viral suppression, supported the exclusion of resistance in these cases.

For the primary analysis in this study, all participants with an undetectable viral load (<50 copies/ml) at the start of LAI CAB+RPV were included. Results for patients who met these criteria are presented in the main tables and figures.

Additionally, a supplementary analysis of virologic outcomes was conducted on all participants who met the approved label criteria for initiating LAI CAB+RPV, which included an undetectable viral load, the absence of RAMs to INSTIs and nonnucleoside reverse transcriptase inhibitors (NNRTIs), and no prior VF on these drug classes. Results from this subpopulation are shown in the supplementary material, Supplemental Digital Content.

Data collection

Data were collected using a standardized electronic case report form (CRF) named REDCAP [28]. The CRF captures comprehensive information, including demographic data, anthropometric measurements, HIV-related information, laboratory data, treatment details, and oral bridge therapies. Participants were assigned to the genotype-available group if they had undergone genotype testing at any time prior to switching therapy, regardless of subsequent virologic outcomes. A rigorous review was conducted to ensure completeness of all variables relevant to the primary study objectives. Missing data were limited to secondary or less clinically relevant variables, such as height, baseline weight, or occasional laboratory parameters. These absences did not affect key variables or the validity of the primary analyses. Therefore, no imputation methods were applied.

Variables

Key data collected included patient demographics, HIV history, previous treatment regimens, comorbidities, substance use, weight, BMI, and genotypic resistance testing, when available. Outcomes included virologic suppression (<50 copies/ml), VF (defined as two consecutive HIV-1 RNA measurements ≥200 copies/ml or a single measurement >500 copies/ml, leading to treatment discontinuation), immunological markers, adherence to the injection schedule, and reasons for discontinuation. Adherence was assessed based on the percentage of days covered by ART and categorized into three groups: complete adherence (100% coverage), high adherence (90–99.9% coverage), and suboptimal adherence (<90% coverage). All variables and outcomes were assessed according to each center's routine clinical practice and international guidelines, with no standardized follow-up schedule across sites.

In addition to the main analyses, a 1 : 1 propensity score matching was performed to adjust for potential baseline confounders, including demographic characteristics (age, sex, nationality), clinical variables (time on ART, documented prior virologic failure, duration of virologic suppression), treatment adherence, and prior ART regimen characteristics. Matching was conducted to balance these characteristics between participants with known and unknown genotype groups, minimizing confounding bias. The matched cohort characteristics and results are presented in Table S1, Supplemental Digital Content.

Statistical analysis

Qualitative variables are presented as absolute frequencies and percentages, while quantitative variables are expressed as median and interquartile range. Comparative analyses are performed using appropriate statistical tests: the chi-square test or Fisher's exact test for qualitative variables, and the independent samples t-test or the Mann–Whitney U test for quantitative variables, as appropriate. Kaplan–Meier survival curves were generated to estimate the probability of remaining free from VF over time, with comparisons between groups with and without available genotype analyzed using the log-rank test.

Ethical considerations

The study protocol has been approved by the Ethics Committee for Research with Medicines of the Burgos and Soria Health Area with code 23-00144. All patients provide informed consent before participation. Data confidentiality is maintained in compliance with Spanish and European Union data protection regulations.

Results

The study included 3146 participants (84.3% males) from 58 institutions across Spain, divided into two groups: 1464 (46.5%) with available prior genotypes and 1682 (53.5%) without genotypic data.

Baseline characteristics by genotype availability

Baseline demographic and clinical characteristics based on genotype availability are presented in Table 1. Baseline virologic and immunologic characteristics by genotype availability are detailed in Table 2. Additionally, baseline characteristics and treatment outcomes after 1 : 1 propensity score matching are shown in Table S1, Supplemental Digital Content, providing adjusted comparisons between participants with known and unknown genotype status.

Table 1.

Baseline demographic and clinical characteristics of participants switched to LAI CAB+RPV in the Relativity cohort by genotype availability.

	Unknown genotype	Known genotype	OR (95% CI)	P-value
n	1682	1464
Age, median (IQR), years	45.9 (37.0, 55.7)	44.0 (37.0, 52.5)		0.002
Sex, n (%)
Female	265 (15.8)	200 (13.7)	1.18 (0.97–1.45)	0.421
Male	1402 (83.7)	1251 (85.6)	0.86 (0.71–1.05)
Transgender male	0 (0.0)	0 (0.0)	-
Transgender female	8 (0.5)	10 (0.7)	0.7 (0.24–1.97)
Nationality, n (%)
Spanish	1105 (66.0)	1079 (73.9)	0.69 (0.59–0.80)	<0.001
Non-Spanish	569 (34.0)	382 (26.1)	1.45 (1.24–1.70)
Region of origin among non-Spanish participants, n (%)
South/Central/North America	467 (80.0)	313 (80.5)	0.97 (0.69–1.35)	0.016
Africa	29 (5.0)	34 (8.7)	0.55 (0.31–0.94)
Central Europe	17 (2.9)	8 (2.1)	1.43 (0.58–3.86)
Western Europe	40 (6.8)	14 (3.6)	1.97 (1.03–3.97)
Eastern Europe	21 (3.6)	19 (4.9)	0.73 (0.37–1.45)
Other Europe	2 (0.3)	0 (0.0)	3.34 (0.13-Inf)
Asia	8 (1.4)	1 (0.3)	5.39 (0.72–239.48)
Weight, median (IQR), kg	75.2 (67.0, 84.0)	76.3 (68.5, 85.8)		0.008
Baseline BMI, median (IQR), kg/m2	24.6 (21.6, 27.5)	24.9 (22.2, 27.8)		0.006
Baseline BMI category, n (%)
Underweight	15 (1.6)	21 (2.0)	0.79 (0.40–1.52)	0.238
Normal weight	429 (46.0)	428 (41.6)	1.20 (1.01–1.43)
Overweight	364 (39.1)	431 (41.8)	0.91 (0.76–1.09)
Obese	124 (13.3)	150 (14.6)	0.90 (0.70–1.15)
Height, median (IQR), cm	1.8 (1.7, 1.8)	1.8 (1.7, 1.8)		0.690
HIV transmission route, n (%)
Men who have sex with men	1008 (59.9)	964 (65.8)	0.78 (0.67–0.90)	<0.001
Heterosexual	288 (17.1)	301 (20.6)	0.8 (0.66–0.96)
Injecting drug use	140 (8.3)	57 (3.9)	2.24 (1.62–3.13)
Vertical transmission	12 (0.7)	16 (1.1)	0.65 (0.28–1.47)
Other	17 (1.0)	23 (1.6)	0.64 (0.32–1.26)
Not available	217 (12.9)	103 (7.0)	1.96 (1.52–2.53)
Substance use habits, n (%)
Tobacco use	502 (29.8)	429 (29.3)	1.03 (0.88–1.20)	0.769
Alcohol use	133 (7.9)	175 (12.0)	0.63 (0.49–0.81)	<0.001
Illicit drug use	501 (32.4)	456 (32.8)	0.98 (0.84–1.15)	0.819
Other substance use	83 (4.9)	82 (5.6)	0.87 (0.63–1.21)	0.449
Chemsex	66 (3.9)	49 (3.3)	1.18 (0.8–1.76)	0.444
Liver-related comorbidities, n (%)
Alcoholic liver disease	7 (0.4)	2 (0.1)	3.05 (0.58–30.18)	0.259
Nonalcoholic fatty liver disease	33 (2.0)	40 (2.7)	1.38 (0.87–2.22)	0.189
Active Hepatitis C	27 (1.6)	17 (1.2)	3.05 (0.58–30.18)	0.365
Active Hepatitis B	3 (0.2)	2 (0.1)	1.38 (0.87–2.22)	1.000
Hepatitis B serology (%)
Vaccinated anti-HBs positive	847 (58.1)	848 (65.4)	0.73 (0.62–0.86)	<0.001
Vaccinated anti-HBs negative	244 (16.7)	151 (11.7)	1.52 (1.22–1.91)
Past infection	203 (13.9)	150 (11.6)	1.23 (0.98–1.56)
Isolated anti-HBc	35 (2.4)	23 (1.8)	1.36 (0.78–2.43)
Seronegative unvaccinated	128 (8.8)	122 (9.4)	0.93 (0.71–1.21)
Chronic liver disease (%)	52 (3.1)	33 (2.3)		0.182
Fibrosis stage in participants with chronic liver disease
F0	2 (4.7)	2 (7.4)	0.61 (0.04–8.97)	0.413
F1	18 (41.9)	11 (40.7)	1.05 (0.35–3.14)
F2	9 (20.9)	5 (18.5)	1.16 (0.30–5.02)
F3	7 (16.3)	1 (3.7)	5.06 (0.58–236.18)
F4	7 (16.3)	8 (29.6)	0.46 (0.12–1.73)
ART regimens at baseline, n (%)
INSTI-based	1276 (75.9)	1150 (78.6)	0.86 (0.73–1.01)	<0.001
NNRTI-based	197 (11.7)	133 (9.1)	1.33 (1.04–1.70)
PI-based	69 (4.1)	102 (7.0)	0.56 (0.39–0.81)
Others	59 (3.5)	28 (1.9)	1.91 (1.18–3.11)
NK	81 (4.8)	51 (3.5)	1.39 (0.96–1.99)
Reasons for switching to LAI CAB+RPV, n (%)
Toxicity	21 (1.2)	28 (1.9)	0.63 (0.34–1.14)	0.175
Drug-drug interactions	6 (0.4)	6 (0.4)	1.00 (0.30–3.33)	1.000
Treatment simplification	391 (23.2)	354 (24.2)	0.95 (0.81–1.11)	0.567
Comfort/quality of life	896 (53.3)	716 (48.9)	1.21 (1.04–1.39)	0.016
Malabsorption	13 (0.8)	29 (2.0)	0.38 (0.19–0.74)	0.005
Swallowing disorders	8 (0.5)	8 (0.5)	1.00 (0.37–2.70)	0.978
Patient request	583 (34.7)	573 (39.1)	0.82 (0.71–0.96)	0.010
Other reasons	147 (8.7)	137 (9.4)	0.92 (0.72–1.18)	0.588

BMI, body mass index; CAB, cabotegravir; CI, confidence interval; HIV, human immunodeficiency virus; INSTI, integrase strand transfer inhibitor; IQR, interquartile range; LAI, long-acting injectable; NK, not known; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; OR, odds ratio; PI, protease inhibitor; RPV, rilpivirine.

Table 2.

Baseline virologic and immunologic characteristics of participants switched to LAI CAB+RPV in the Relativity cohort, stratified by genotype availability.

	Unknown genotype	Known genotype	OR (95% CI)	P-value
n	1682	1464
CD4+ cell count nadir, median (IQR), cells/μl	333.5 (195.0, 500.0)	345.0 (196.0, 496.0)		0.471
Viral load at HIV diagnosis, median (IQR), copies/ml	38 490 (4930, 14 7971)	63 000 (16 800, 210 000)		<0.001
Time from HIV diagnosis to first ART initiation, median (IQR), months	2.0 (1.0, 22.0)	2.0 (0.0, 12.0)		<0.001
AIDS defining condition at baseline, n (%)
Yes	209 (12.4)	180 (12.3)	1.01 (0.81–1.26)	<0.001
No	1325 (78.8)	1243 (84.9)	0.66 (0.55–0.80)
Not available	148 (8.8)	41 (2.8)	3.35 (2.33–4.89)
Duration of ART prior to LAI CAB+RPV initiation, median (IQR), years	10.0 (6.0, 18.0)	8.8 (5.0, 12.4)		<0.001
Duration of virologic suppression prior to LAI CAB+RPV initiation, median (IQR), months	96.0 (38.8, 160.5)	72.0 (30.0, 120.0)		<0.001
History of virological failure with any ART regimen, n (%)
No	1349 (80.2)	1247 (85.2)	0.7 (0.58–0.85)	<0.001
Yes	41 (2.4)	69 (4.7)	0.51 (0.33–0.76)
Not available	292 (17.4)	148 (10.1)	1.87 (1.5–2.32)
Antiretroviral class involved in prior virologic failure, n (%)
INSTI-based	12 (40.0)	14 (23.0)	2.24 (0.78–6.34)	0.194
NNRTI-based	6 (20.0)	20 (32.8)	0.51 (0.15–1.58)
IP-based	12 (40.0)	27 (44.3)	0.84 (0.31–2.22)
Virologic blips in the 5 years prior to LAI CAB/RPV initiation, n (%)a
0	1268 (84.1)	1069 (79.8)	1.34 (1.1–1.64)	0.021
1	147 (9.8)	168 (12.5)	0.75 (0.59–0.96)
2	49 (3.3)	45 (3.4)	0.97 (0.63–1.49)
3	16 (1.1)	26 (1.9)	0.54 (0.27–1.06)
> 3	27 (1.8)	32 (2.4)	0.75 (0.43–1.29)
Oral lead-in phase completion, n (%)	287 (17.1)	117 (8.0)	2.37 (1.88–3.00)	<0.001
HIV subtype, n (%)
B	–	667 (45.6)		–
A1/A2	–	28 (1.9)
F/CRF	–	45 (3.1)
Other	–	59 (4.0)
Not available	–	665 (45.4)
Wild-type virus, n (%)	–	1249 (85.3)		–
Mutation profile, n (%)
NRTI-associated resistance mutations	–	129 (8.8)		–
184V	–	29 (2.0)
Other NRTI mutations	–	99 (6.8)
NNRTI-associated resistance mutations	–	127 (8.7)
K103N	–	43 (2.9)
E138A	–	8 (0.5)
Other NNRTI mutations	–	80 (5.5)
INSTI-associated resistance mutations	–	18 (1.2)
L74M/I/F	–	1 (0.1)
T97A	–	1 (0.1)
Other INSTI mutations	–	16 (1.1)
Number of resistance mutations, n (%)
0	–	1249 (85.3)		–
1	–	166 (11.3)
2	–	47 (3.2)
3	–	2 (0.1)

AIDS, acquired immunodeficiency syndrome; ART, antiretroviral therapy; CAB, cabotegravir; CI, confidence interval; CRF, circulating recombinant form; HIV, human immunodeficiency virus; INSTI, integrase strand transfer inhibitor; IQR, interquartile range; LAI, long-acting injectable; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; OR, odds ratio; RPV, rilpivirine.

^aA blip was defined as a temporary, low-level increase in HIV viral load (50–200 copies/ml) that returned to undetectable on repeat testing.

Participants without available genotypic data were slightly older and had lower proportions of Spanish individuals and of men who have sex with men (MSM), as well as marginally lower median weight and baseline BMI compared to those with genotypic data. No significant differences were found in sex distribution, BMI categories, height, substance use, or comorbidity profiles between groups. BMI evolution over time was also similar, except for minor, transient differences at months 5 and 9 (see Tables 1 and S2, Supplemental Digital Content).

Compared to participants with genotypic data, those without available genotyping exhibited distinct virological characteristics, including a lower viral load at diagnosis, a longer duration of ART, and a more prolonged period of virological suppression before initiating LAI CAB+RPV. They also exhibited a lower rate of previous VF, a higher proportion of individuals without blips in the preceding five years, and a greater frequency of oral lead-in phase usage. Among those with genotyping, subtype and resistance data are provided in Table 2.

The group without genotypic data showed slight differences in preswitch ART regimens compared to the genotyped group, with lower use of PI-based and INSTI-based regimens, but higher use of NNRTI-based regimens. Both groups shared similar motivations for switching to LAI CAB+RPV, with quality-of-life improvement being the most frequently cited reason (Table 1).

Virological and immunological outcomes

The median follow-up duration was slightly shorter in the group without genotype [13.3 months, interquartile range (IQR) 8.6, 18.9] compared to the group with genotype (14.9 months, IQR 9.0, 19.2; P = 0.003).

The proportion of patients with a viral load of <50 copies/ml remained high over time in both groups, with no clinically meaningful differences observed between those with and without available genotype information. Although some time points showed statistically significant differences, both groups maintained viral suppression rates above 93% up to the 23rd month of follow-up (Fig. 1). There were no clinically relevant changes in CD4⁺ cell count, CD8⁺ cell count, or the CD4/CD8 ratio throughout the study, regardless of whether the participants had a known or unknown genotype available (Table S3, Supplemental Digital Content).

Fig. 1.

Proportion of participants with viral load <50 copies/ml by genotype availability over time.

Treatment adherence and injection timeliness

Adherence rates among patients without genotypic data were 85.5% for full coverage (100%), 12.8% for high adherence (90–99.9% coverage), and 1.7% for suboptimal adherence (<90% coverage). In comparison, patients with known genotype had rates of 83.0%, 15.4%, and 1.6%, respectively (P = 0.180). Notably, in both groups, 98.4% of patients achieved at least 90% coverage of treatment days (P = 0.144).

As shown in Figures S1, Supplemental Digital Content and S2, the proportion of participants receiving on-time injections (within ±7 days of the scheduled date) remained high and comparable between groups throughout the follow-up period.

Treatment discontinuation and virologic failure

Oral bridges were made in 1.0% of patients in both the group without genotype and the group with genotype. Permanent discontinuation was reported in 6.1% of the group without genotype and 6.6% of the group with genotype (P = 0.804).

The rates of VF were similar between groups: 0.5% in the group without prior genotypic testing (n = 8) and 1.0% in the group with prior genotypic testing (n = 12) (P = 0.476). Among the 12 individuals with available genotypic data who experienced VF, two had a documented history of prior VF to protease inhibitors. Among the remaining eight individuals with no available genotypic data, seven were confirmed to have no prior history of VF, and for one individual, this information was not available.

RAMs were detected at the time of VF in five of eight participants (62.5%) without prior genotyping and in 7 of 12 participants (58.3%) with prior genotyping. Among those without prior genotyping, NNRTI RAMs were detected in 1 participant, INSTI RAMs in 2 participants, and dual NNRTI and INSTI RAMs in 2 participants. In the group with prior genotyping, NNRTI RAMs were detected in 4 participants, INSTI RAMs in 3 participants, and no dual NNRTI and INSTI RAMs were detected. Following oral treatment resumption, viral suppression was re-achieved in 12 of the 20 participants with VF: 4/8 (50%) in the group without prior genotypic testing and 8/12 (67%) in the group with prior genotypic testing.

Among the participants with VF and no genotype available prior to failure, one participant had INSTI mutations E138K and N155H detected and achieved an undetectable viral load after oral ART (DRV/cob/ABC/3TC). Another participant had INSTI mutations G140S, L74M, and Q148H and achieved an undetectable viral load after oral ART (DRV/cob/FTC/TAF). A third participant exhibited NNRTI mutations 74M, 138K, and 148KR, with viral load also remaining detectable following oral ART (DRV/cob/FTC/TAF). A fourth participant presented NNRTI mutations K103N and Y188L along with INSTI mutations E138eK, Q148R, and L74LM, with viral load remaining detectable after oral ART (DRV/cob/FTC/TAF). The fifth participant showed NNRTI mutations K103N, E138K, and N348I plus INSTI mutations E138K, G140A, and Q148R, with persistent detectable viral load after oral ART (DRV/cob/FTC/TAF). The remaining three participants showed no detectable new RAMs following VF. Of these, two achieved undetectable viral load after oral ART (DRV/cob/FTC/TAF or BIC/FTC/TAF), while one had persistent detectable viral load despite treatment (DRV/cob/FTC/TAF).

Among the 12 participants with VF and baseline genotypes available, 4 had no baseline RAMs and no new RAMs detected postfailure. All four received oral ART following failure, three with DRV/cob/FTC/TAF and one with BIC/FTC/TAF, and each achieved an undetectable viral load after treatment.

Five other participants with wild-type baseline genotypes developed RAMs postfailure. The first developed INSTI mutations E138K and Q148K and achieved undetectable viral load after oral ART (DRV/cob/FTC/TAF). The second developed multiple NNRTI mutations (L100I, K103N, L74M, T97A, G140S, Q148K, Q148R, E157Q) and also achieved viral suppression with oral ART (DRV/cob/FTC/TAF). The third exhibited NNRTI mutations E138A, V179E, P225H, F227L, and N348I and did not achieve viral suppression after oral ART (BIC/FTC/TAF). The fourth had NNRTI mutation K101E and INSTI mutation Q146L and failed to achieve viral suppression after oral ART (BIC/FTC/TAF). Finally, the fifth developed NNRTI mutations K101Ke, V106Vi, and Y181C and did not achieve viral suppression following oral ART (BIC/FTC/TAF).

Finally, three participants had RAMs at baseline. One had NRTI 184V at baseline, developed an additional INSTI 63P mutation postfailure, and did not achieve an undetectable viral load after oral ART (DTG/3TC/dor). Another had baseline NRTI 184V and NNRTI K103N, did not develop any new RAMs, and achieved undetectable viral load following oral ART (DRV/cob/FTC/TAF). The third had baseline NRTI V90I, developed an INSTI Q148R mutation postfailure, and achieved undetectable viral load after oral ART (BIC/FTC/TDF).

Of the 1464 patients with baseline genotyping, subtyping was performed in 799 individuals (Table 2), with no cases of subtype A6 identified. Subtype A6 was detected in one participant at the time of VF, who did not have baseline genotyping available. This participant was a male Portuguese and had a BMI below 30 kg/m². He achieved viral suppression after resuming oral ART with DRV/cob/TAF/FTC at 4 months postfailure.

Other reasons for discontinuation, including local injection site reactions, systemic adverse effects, and other causes, were comparable between groups, with no statistically significant differences observed for any category (Table 3).

Table 3.

Treatment discontinuations in participants switched to LAI CAB+RPV in the relativity cohort by genotype availability.

	Unknown genotype (N = 1682)	Known genotype (N = 1464)	OR (95% CI)	P-value
Type of treatment discontinuation, n (%)
Temporary discontinuation	17 (1.0)	14 (1.0)	1.00 (0.46–2.20)	0.804
Permanent discontinuation	102 (6.1)	97 (6.6)	0.91 (0.67–1.23)
Time to temporary treatment discontinuation, median (IQR), months	15.7 (11.4, 20.4)	11.1 (7.0, 21.2)		0.895
Time to permanent treatment discontinuation, median (IQR), months	7.0 (3.0, 12.3)	8.9 (4.9, 11.1)		0.286
Reason for discontinuation, n (%)a
Local injection site reactions	21 (1.2)	24 (1.6)	1.36 (0.75–2.47)	0.652
Systemic adverse effects	14 (0.8)	15 (1.0)	1.27 (0.62–2.60)	0.821
Virological failure	8 (0.5)	12 (0.8)	0.63 (0.26–1.49)	0.476
Other reasons	64 (3.8)	49 (3.3)	1.17 (0.80–1.72)	0.444

CAB, cabotegravir; IQR, interquartile range; LAI, long-acting injectable; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; OR, odds ratio; RPV, rilpivirine.

^aTreatment discontinuation reasons refer only to permanent discontinuations. Participants may have more than one reason recorded for their discontinuation, except for virological failure, which is exclusive. Therefore, the sum of reasons may exceed the total number of permanent discontinuations.

Among patients without genotypic data, the median time to discontinuation was 7.0 months (IQR 3.1–11.0), compared to 8.6 months (IQR 5.0–11.0) in those with known genotypes (P = 0.251). Similarly, median time to discontinuation due to VF was 7.0 months (IQR 6.8–10.0) in the group without genotypic data and 9.0 months (IQR 7.0–11.0) in the group with genotypic data (P = 0.562).

The survival analysis in Fig. 2 shows that the probability of remaining free from VF over time was very high and did not differ significantly between patients with and without an available genotype (P = 0.270).

Fig. 2.

Survival curve for virologic failure by genotype availability.

The shaded area represents the 95% confidence interval around the survival curve. Numbers in parentheses below the x-axis indicate the number of failures observed during each follow-up period.

Virologic outcomes in on-label long-acting injectable cabotegravir and rilpivirine

The results of the secondary sub-study analyzing patients who initiated LAI CAB+RP strictly following on-label indications included 1307 participants with a known genotype and 1664 with an unknown genotype. Adherence rates among patients without genotypic data were 85.5% for full coverage (100%), 12.9% for high adherence (90–99.9% coverage), and 1.7% for suboptimal adherence (<90% coverage). In comparison, patients with known genotype showed rates of 83.1%, 15.3%, and 1.7%, respectively (P = 0.118). There were small but significant differences in the proportion of participants with undetectable CV throughout the study (months 3, 9, 15, and 19), with the proportion consistently higher in the unknown genotype group (Figure S3, Supplemental Digital Content).

Among patients with unknown genotype, the median time to discontinuation was 7.0 months (IQR 3.1–11.2), compared to 8.4 months (IQR 5.0–11.0) in those with known genotype (P = 0.254). Median time to discontinuation due to VF was 7.0 months (IQR, 6.8–10.0) for the unknown genotype group and 9.0 months (IQR, 7.2–11.0) for the known genotype group (P = 0.387).

The survival analysis revealed a consistently high probability of remaining free from VF over time, with no significant differences observed between patients with available genotypes and those without (P = 0.270) (Figure S4, Supplemental Digital Content).

Discussion

This large real-world cohort study evaluated the impact of genotypic data availability prior to switch on virological outcomes in people living with HIV who switched to LAI CAB+RPV while virologically suppressed. These findings indicate that, despite demographic and clinical differences, both groups maintained similarly high levels of virological suppression, regardless of prior genotypic testing and low rates of VF throughout the follow-up period. These results indicate that, in the context of sustained virological suppression, the absence of prior genotypic information does not compromise the effectiveness or safety of LAI CAB+RPV in routine clinical practice. However, caution should be exercised when generalizing these results to populations with recent viral suppression, complex treatment histories, poor adherence, or a high prevalence of non-B HIV subtypes, including subtype A6, as these factors may influence treatment effectiveness.

Genotypic resistance testing has been emphasized as a key consideration before switching to LAI CAB+RPV. In most pivotal phase III trials, participants underwent baseline resistance testing [3–6,29], and similar practices have been reported in large observational cohorts [11,13,30–34]. However, several cohorts have included patients without a prior genotype available. For example, the ICONA cohort (n = 506) reported genotypic resistance testing for reverse transcriptase and INSTI in 82% and 45.7% of patients, respectively [10]. Similarly, the COMBINE-2 (n = 472) study showed that genotyping was available for 60% of participants before switching [35]. Despite this variability, the rates of VF in these cohorts have remained low and consistent with the outcomes observed in our study. The rate of VF in the ICONA cohort (2/506; 0.4%) and in the COMBINE-2 cohort (3/374; 0.8%) was low and within the range of approximately 1% observed in other real-world series, comparable to rates reported in clinical trials [36]. These findings support the notion that, among individuals who are virologically suppressed at the time of switching, the lack of genotypic data, in the absence of prior VF, does not appear to negatively impact virological outcomes with LAI CAB+RPV in real-life settings. In this context, the RELATIVITY study provides further real-world evidence on scaling LAI CAB+RPV use in settings with limited access to genotypic testing.

The CARES trial [7], conducted across Uganda, Kenya, and South Africa, was a phase 3b randomized study evaluating LAI CAB+RPV in 512 virologically suppressed adults under public health conditions typical of sub-Saharan Africa. Participants with ≥6 months of viral suppression (<50 copies/ml) and no prior VF were randomized 1 : 1 to switch to bimonthly injections or continue oral ART (97% on dolutegravir-based regimens). The trial employed sparse viral load monitoring (every 24 weeks) and conducted retrospective genotyping of archived proviral DNA in all participants, revealing baseline rilpivirine resistance in 8–12% and cabotegravir resistance in 11–16%. At week 48, both arms demonstrated high efficacy, with 96% and 97% suppression, respectively. Notably, there were only two VFs (0.8%) in the injectable group, both associated with emergent resistance mutations (V108I/E138K in RPV; E92E/V/N155H/L74M and G118R in CAB), despite protocol-adherent dosing. These failures occurred in participants with HIV subtypes A1 and D, neither of whom had baseline high-level resistance mutations [7].

Although the design differs, these results align closely with findings from the RELATIVITY cohort (n = 3146), where 53.5% of participants lacked genotyping yet demonstrated comparable virological outcomes. Despite the missing pretreatment resistance data, the Spanish real-world cohort showed nearly identical failure rates between those with and without prior genotyping (1.0% vs. 0.5%, P = 0.476), maintaining greater than 93% viral suppression in both groups over 13–15 months. Both studies challenge the necessity of routine genotyping in long-suppressed populations. However, CARES’ emergent resistance cases underscore the critical need for ongoing virological monitoring, particularly in resource-limited settings where access to genotypic testing remains constrained. The convergence of trial and real-world evidence supports the viability of LAI CAB+RPV across diverse implementation contexts. CARES demonstrates resilience against archived NNRTI resistance in African populations, while RELATIVITY confirms equivalent outcomes when genotyping availability varies.

HIV-1 subtype A6, common in Eastern Europe and Russia, is linked to VF with LAI CAB+RPV due to the L74I integrase polymorphism. While L74I alone doesn’t reduce CAB susceptibility, it enhances viral replication and accelerates resistance development when combined with mutations like Q148R, resulting in significantly reduced CAB efficacy [36]. Clinical trials, including pooled analyses of FLAIR and ATLAS-2M, identified A6 as an independent predictor of failure, with a 15-fold increased risk when combined with obesity or preexisting RPV RAMs. For instance, participants with two or more risk factors (e.g., A6 + RAMs) experienced a 19% failure rate, compared to less than 1% in low-risk groups [26]. The epidemiologic significance of A6 is further underscored by its role in clustered transmission networks in Eastern Europe, where conflict-driven migration has facilitated its spread [37].

Despite A6's established risk in trials, its impact in real-world cohorts (e.g., CARES, RELATIVITY) is less evident due to three factors: geographic underrepresentation: A6 is concentrated in regions like Russia and Ukraine, which are underrepresented in major studies, while Western European cohorts report low A6 prevalence [10,16,35,37]. Limited routine subtyping: Subtyping is rarely performed at baseline unless resistance is suspected [9,12,30,31,33,34,38]. When a VF occurs, it is not typically related to subtype A6 [13,16,33–35]. Confounding variables: A6's risk in trials is amplified by cofactors like obesity or RAMs, which may not align in real-world populations. Notably, the CARES trial in Africa found no increased risk with subtype A1 (common locally), confirming that A6-specific virologic factors drive the risk [7]. These findings emphasize the need for tailored resistance testing in high-prevalence regions and caution against extrapolating trial-based risk models globally.

Genotypic resistance data at VF were available for 20 participants in our cohort, providing insights consistent with larger observational studies [39]. Mutations associated with NNRTIs and INSTIs were observed in approximately 60% of participants. These findings align with a recent systematic review by Ring et al. of 28 observational cohorts including nearly 8000 participants, where 56% of 80 VF events with available genotypic data exhibited NNRTI resistance mutations, 50% showed INSTI resistance mutations, and 41% demonstrated dual-class resistance [39].

Within our cohort, recurrent NNRTI mutations such as E138A/K, K101E, and K103N reflected resistance patterns previously reported in patients failing RPV-containing regimens. Similarly, INSTI mutations, including E138K, Q148R/K, G140S, and N155H, commonly reported in observational studies, were detected [39]. Regarding genotypic susceptibility to future INSTI-containing regimens, substitutions at integrase positions 148 and 263 are considered to have the most significant impact on susceptibility to bictegravir and dolutegravir [39]; while mutations at position 148 were identified in our cohort, no mutations at position 263 were detected.

Following virologic failure, viral re-suppression was achieved in 60% of participants in our cohort after resuming oral ART. These findings are consistent with the review of observational studies conducted by Ring et al.[39], which reported re-suppression in 87.8% (65/74) of VF events where data were available, underscoring the effectiveness of appropriate postfailure management in achieving viral control.

This study has several limitations that should be considered when interpreting its findings. First, as an observational and ambispective cohort, there is potential for selection bias, which may not reflect the broader HIV population. Furthermore, the retrospective component covers a short and limited time period shortly after the approval of LAI CAB+RPV in Spain, which was not specifically measured or analyzed in this study. This restricts the ability to assess secular trends or changes in clinical practice over time and should be considered when interpreting the results. Second, the median follow-up period is relatively short, limiting the assessment of long-term efficacy, durability, and the emergence of rare adverse events or late VFs. Third, our findings are based on a population with prolonged antiretroviral therapy and sustained virologic suppression (median 8 years suppressed in those without genotype data). As such, caution should be exercised in generalizing these results to individuals with shorter treatment duration or more recent viral suppression, where supporting evidence is limited and risk profiles may differ. Fourth, the low prevalence of non-B subtypes, especially A6, in Spain, and the absence of systematic subtyping, restrict the evaluation of subtype-specific risks in this setting. Fifth, the timing of genotypic data availability relative to treatment switch was not captured in the database. Thus, we could not distinguish between recent and remote genotypes, which may affect the interpretation of resistance status and its clinical relevance. This limitation highlights the importance of incorporating detailed temporal genotyping information in future studies to improve the accuracy of resistance assessments and guide optimized treatment strategies. Additionally, lack of pharmacokinetic data, such as trough drug levels, and the absence of adherence verification through objective methods (e.g., drug level measurements or electronic monitoring) represent further limitations of this study. Despite these limitations, the study's strengths include its large, nationwide, multicenter design, the inclusion of a diverse real-world population, and robust, prospectively collected data on virological outcomes and safety. The findings provide valuable evidence supporting the effectiveness and safety of LAI CAB+RPV in routine clinical practice, contributing important real-world insights to complement those from randomized clinical trials. Building on these strengths, we plan to conduct and publish analyses of longer-term follow-up data in future reports to further evaluate the durability and sustained effectiveness of LAI CAB+RPV treatment.

In conclusion, this large, nationwide real-world study demonstrates that LAI CAB+RPV maintains high rates of virologic suppression and low rates of VF in people living with HIV who switch while already suppressed, regardless of the availability of genotypic resistance data. In this context, for long-term virologically suppressed individuals with no history of VF and who are under appropriate clinical follow-up, genotypic testing may not be essential prior to initiating LAI CAB+RPV, as real-world evidence consistently demonstrates high rates of viral suppression and a low incidence of VF, regardless of the availability of genotypic data. Ongoing monitoring and future analyses will be crucial to confirm these outcomes over more extended follow-up periods and in specific subpopulations, such as recent migrants or those with non-B subtypes.

Acknowledgements

*RELATIVITY Cohort Group. Cristina Díez (Hospital General Universitario Gregorio Marañón, Madrid, Spain); Guillermo Soria Fernández-Llamazares (Hospital Universitario de Fuenlabrada, Madrid, Spain); Desiree Pérez Martínez (Hospital Universitario San Agustín, Asturias, Spain); Mireia Cairó Llobell (Hospital Universitario Mutua de Terrassa, Barcelona, Spain); Alberto Romero Palacios (Hospital Universitario de Puerto Real, Cádiz, Spain); Rebeca Cabo Magadan (Hospital Universitario Central de Asturias, Asturias, Spain); Víctor Arenas García (Hospital Universitario de Cabueñes, Asturias, Spain); María Antonia Sepúlveda (Hospital Universitario de Toledo, Toledo, Spain); Antonio Jesús Sánchez Guirao (Hospital General Universitario Morales Meseguer, Murcia, Spain); Cristina Escrich (Hospital Verge de la Cinta de Tortosa, Tarragona, Spain); Carlos Armiñanzas (Hospital Universitario Marqués de Valdecilla, Cantabria, Spain); Eva María Ferreira Pasos (Complejo Asistencial de Segovia, Segovia, Spain); Alberto Juárez Toquero (Hospital Río Hortega de Valladolid, Valladolid, Spain); Ana Lérida Urteaga (Hospital de Viladecans, Barcelona, Spain); Jara Llenas-García (Hospital Vega Baja, Alicante, Spain); Sara García Torras (Hospital de Santa Caterina de Salt, Girona, Spain); Juan E Losa-García (Hospital Universitario de Alcorcón, Madrid, Spain); José Fernando Lluch Perales (Hospital Universitario Doctor José Molina Orosa, Las Palmas, Spain); José Sanz (Hospital Universitario Príncipe de Asturias, Madrid, Spain); Sergio Padilla (Hospital General Universitario de Elche, Alicante, Spain); Hadrián Pernas Pardavila (Complexo Hospitalario Universitario de Pontevedra, Pontevedra, Spain); (Juan José Corte García Hospital de Jove, Asturias, Spain); María Ángeles Garcinuño Jiménez (Complejo Asistencial de Ávila, Ávila, Spain); Juan Carlos Gainzarain (Hospital Universitario de Álava, Álava, Spain); Miriam Estébanez (Hospital Central de la Defensa Gómez Ulla, Madrid, Spain); María del Mar García Navarro (Hospital Universitario de Vinalopo, Alicante, Spain); Patricia Noemí Barragán Gallo (Hospital Residencia Sant Camil de Sant Pere de Ribes, Barcelona, Spain); Noemí Ramos Vicente (Hospital Obispo Polanco, Teruel, Spain); Marta Clavero Olmos (Hospital Universitario Infanta Elena, Madrid, Spain); Marta Milián Sanz (Hospital de Valls, Tarragona, Spain): Mikel del Álamo (Hospital Universitario Cruces, Vizcaya, Spain); Miguel Vicente Egido Murciano (Hospital Universitario San Jorge, Huesca, Spain); Beatriz de la Calle (Hospital General Nuestra Señora del Prado, Castilla La Mancha, Spain); Oscar Luis Ferrero Benéitez (Hospital Universitario de Basurto, Bilbao, Spain).

The authors gratefully acknowledge the RELATIVITY Cohort Group for their essential contributions to this research: (1) Complejo Asistencial de Ávila: Ana Cristina Antolí Royo, Carmen Grande Sáez, María Ángeles Garcinuño Jiménez; (2) Complejo Asistencial de Segovia: Ana Carrero Gras, Eva María Ferreira Pasos, Pablo Bachiller Luque; (3) Complexo Hospitalario Universitario de Pontevedra: Hadrián Pernas Pardavila, Mercedes González, Nuria Vázquez Temprano; (4) Hospital Central de la Defensa Gómez Ulla (Madrid): Miriam Estébanez; (5) Hospital Clínico San Carlos (Madrid): Ana Muñoz Gómez, Jose Reynaldo Homen Fernandez, Juncal Pérez-Somarriba Moreno, Maravillas Carralón, María José Núñez Orantos, Maria Nieves Sanz Perez, Noemí Cabello Clotet, Susana Olmedo Hernández, Valeria Cabral Sousa, Vicente Estrada, Virginia Víctor Palomares; (6) Hospital Clínico San Cecilio (Granada): Clara Martínez Montes, David Vinuesa García, Maite Laperal Martín; (7) Hospital Clínico Universitario de Valencia: Ana Ferrer Ribera, Andreu Belmonte, Carolina Pinto, María José Galindo Puerto, Rosa Oltra Sempere, Sandra Pérez Gómez; (8) Hospital Clínico Universitario Lozano Blesa (Zaragoza): Álvaro Cecilio, Isabel Sanjoaquin Conde, Maria Jose Crusells Canales, Pilar Jimenez Marcen; (9) Hospital Vega Baja (Alicante): Ana Lucas-Dato, Belén Martínez-López, Inmaculada González Cuello, Jara Llenas-García, María García López, Ana Isabel Torres Penalva; (10) Hospital de Burgos: Carolina Navarro, Luis Buzón, María Fernandez; (11) Hospital de Denia (Alicante): Amparo Exojo Morales, Carlos de Andrés David, Inmaculada Poquet Catalá, Karenina Antelo Cuéllar, Laia Navarro Peiró, María Josefa de la Asunción Villaverde, Silvia Lope Bolumar; (12) Hospital de Jove (Asturias): Alba León Barbosa, Esther Nogales Nieves, Eva García Alcalde, Juan José Corte García, Lucía Alonso Alonso, María Vanessa López Peláez, Sara Jaber Carballo; (13) Hospital de Santa Caterina de Salt (Girona): Albert Gómez Lozano, Sara García Torras; (14) Hospital de Valls (Tarragona): Marta Milián Sanz; (15) Hospital de Viladecans: Ana Isabel Lérida Urteaga, Antonia Pérez Barja, Javier López-Nieto Sempere, Magdalena Muelas Fernández; (16) Hospital General Nuestra Señora del Prado (Castilla La Mancha): Beatriz de la Calle; (17) Hospital General Universitario de Elche (Alicante): Ángela Botella Zaragoza, Félix Gutiérrez Rodero, Javier García Abellán, Mar Masiá, Paula Mascarell Arlandis, Sergio Padilla Urrea; (18) Hospital General Universitario Gregorio Marañón (Madrid): Chiara Fanciulli, Cristina Diez Romero, Francisco Tejerina Picado, Juan Carlos López Bernaldo de Quiros, Leire Pérez Latorre, Mª Teresa Aldamiz-Echevarría Lois; (19) Hospital General Universitario Morales Meseguer (Murcia): Antonio Jesús Sánchez Guirao; (20) Hospital General Universitario Reina Sofía (Murcia): Ángeles Muñoz Perez, Antonia Alcaraz García, Cristina Tomás Jimenez, Elena Guijarro Westermeyer, Enrique Bernal Morell, Eva García Villalba, Joaquín Bravo Urbieta, María Dolores Hernández Lorente, Rodrigo Martínez Rodríguez, Román Gonzalez Hipólito; (21) Hospital Obispo Polanco (Teruel): Noemí Ramos Vicente; (22) Hospital Residencia Sant Camil de Sant Pere de Ribes (Barcelona): Patricia Barragán Gallo; (23) Hospital Río Hortega de Valladolid: Alberto Juárez Toquero, Beatriz Valentín Casado, Jesica Abadía Otero, Julia Gómez Barquero, Pablo Bachiller Luque, Yolanda Arranz García; (24) Hospital Universitario 12 de Octubre (Madrid): Adriana Pinto, David Rial, Federico Pulido, Juan Martín, Laura Bermejo Plaza, María De Lagarde, María Teresa López Caballero, Mireia Santacreu, Otilia Bisbal, Rafael Rubio, Roser Navarro, Victoria Macheño del Real; (25) Hospital Universitario Álvaro Cunqueiro (Vigo): Celia Miralles Álvarez, Guillermo Pousada, Henar de las Heras Miralles, Luis Enrique Morano Amado; (26) Hospital Universitario Central de Asturias: Víctor Asensi; (27) Hospital Universitario Cruces (Vizcaya): Laura Guio Carrión, Mikel del Álamo; (28) Hospital Universitario de Álava: Ester Sáez de Adana Arroniz, Juan Carlos Gainzarain Arana; (29) Hospital Universitario de Alcorcón (Madrid): Ana Vegas Serrano, Juan Emilio Losa García, María Velasco Arribas; (30) Hospital Universitario de Basurto (Bilbao): Maite Ganchegui Aguirre, Miriam López Martínez, Óscar Luis Ferrero Beneitez; (31) Hospital Universitario de Bellvitge (Barcelona): Alicia Sedo Mor, Juan Tiraboschi; (32) Hospital Universitario de Cabueñes (Asturias): Josefa Soler González, Víctor Arenas García; (33) Hospital Universitario de Canarias: Nereyda Tosco García, Remedios Alemán Valls; (34) Hospital Universitario de Fuenlabrada (Madrid): Guillermo Soria Fernández-Llamazares, Juan Víctor San Martín López, Ruth Calderón Hernaiz; (35) Hospital Universitario de Guadalajara: Alberto Delgado Fernández, Lorenzo Sánchez, Miguel Torralba González de Suso; (36) Hospital Universitario de Puerto Real (Cádiz): Alberto Romero Palacios; (37) Hospital Universitario de Toledo: Maria Antonia Sepúlveda; (38) Hospital Universitario de Torrejón (Madrid): Alejandra Gimeno García, Alejandro David Bendala Estrada, Carmen Montero Hernández, Diana Corps Fernández; (39) Hospital Universitario de Vinalopo (Alicante): Jose Carlos Escribano Stablè, María del Mar Garcia Navarro, Marouane Menchi Elanzi; (40) Hospital Universitario Doctor José Molina Orosa (Las Palmas): Ana Cerezales Calviño, Bárbara Alonso Moreno, José Fernando Lluch Perales; (41) Hospital Universitario Fundación Jiménez Díaz (Madrid): Alfonso Cabello, Miguel Górgolas Hernández-Mora; (42) Hospital Universitario Infanta Elena (Madrid): María Rubio Olivera, Marta Clavero Olmos; (43) Hospital Universitario Infanta Leonor (Madrid): Belén Escribano Losada, Guillermo Cuevas, Jesús Troya García, Mariano Matarranz, Pablo Ryan, Roberto Pedrero Tomé; (44) Hospital Universitario La Paz (Madrid): Ana Belén Delgado Hierro, Carmen Bucsa, Jose Ignacio Bernardino, Juan González García, Luis Ramos Ruperto, Luz Martín Carbonero, María del Mar Arcos Rueda, María Eulalia Valencia, María Luisa Montes, Rafael Micán, Rocío Montejano; (45) Hospital Universitario La Princesa (Madrid): Andoni Casen Gil, Ignacio de los Santos, María Aguilera García; (46) Hospital Universitario Marqués de Valdecilla (Cantabria): Aitziber Illaro Uranga, Carlos Armiñanzas Castillo, Claudia González Rico, Francisco Arnaiz de las Revillas Almajano, Manuel Gutiérrez Cuadra, Maria Carmen Fariñas Álvarez, Noelia Ruiz Alonso; (47) Hospital Universitario Miguel Servet (Zaragoza): Ángela Forcén Vicente de Vera, María Aranzazu Caudevilla Martínez, Rosa María Martínez Álvarez, Ruth Caballero Asensio; (48) Hospital Universitario Mutua de Terrassa (Barcelona): David Dalmau Juanola, Laura Gisbert Pérez, Mireia Cairó Llobell, Roser Font Canals, Xavier Martínez Lacsa; (49) Hospital Universitario Príncipe de Asturias (Madrid): Jose Sanz; (50) Hospital Universitario Puerta de Hierro (Madrid): Alberto Díaz de Santiago, Ana Fernández Cruz, Sara de la Fuente Moral; (51) Hospital Universitario Ramón y Cajal: Ana Moreno, Daniel de las Heras Gómez, María Jesús Pérez Elías, María Jesús Vivancos Gallego, Mario Pons Guillén; (52) Hospital Universitario San Agustín (Asturias): Desirée Pérez Martínez, Maria Elisa Pino Diaz, Miguel Alberto de Zárraga Fernández; (53) Hospital Universitario San Jorge (Huesca): Miguel Vicente Egido Murciano, Teresa Omiste Sanvicente; (54) Hospital Universitario Son Espases (Palma de Mallorca): Adrián Ferré, Antoni Campins, Francisco Fanjul, Luisa Martín Pena, Melchor Riera; (55) Hospital Universitario Son Llatzer (Palma de Mallorca): Adrián Rodríguez, Patricia Sorní Moreno; (56) Hospital Universitario Virgen de las Nieves (Granada): Carmen Hidalgo; Sergio Sequera Arquelladas; (57) Hospital Verge de la Cinta de Tortosa (Tarragona): Cristina Escrich; (58) Parc Taulí Hospital Universitari (Sabadell): María del Carmen Navarro.

The authors thank Dr Pablo Rivas for providing medical writing support.

Authors’ contributions: All authors gathered participant data from their respective institutions. LB and JTG organized the cohort, conducted the data analysis, and drafted the manuscript. RP assisted in preparing the manuscript and performed the statistical analysis. In addition, all authors participated in drafting the manuscript or revising it critically for important intellectual content. All authors gave final approval of the version to be published and agreed to be accountable for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved.

Funding statement: This work has received no funding.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of interest

L.B. has received lecture fees from ViiV and Janssen; these activities are unrelated to the submitted work. JIB has no conflict of interest. M.J.G.P. has no conflict of interest. J.M. has no conflict of interest. M.R.A.V. has no conflict of interest. M.T.G.S. has no conflict of interest. A.D.S. has no conflict of interest. F.F. has received grants from Gilead and ViiV Healthcare, and speaker fees from Gilead Sciences, Janssen, and ViiV; these activities are unrelated to the submitted work. A.R. has no conflict of interest. A.C.U. has received grants and personal fees from ViiV Healthcare and Gilead, personal fees from Johnson & Johnson and Merck; these activities are unrelated to the submitted work. R.P. has no conflict of interest. M.J.C.C. has no conflict of interest. S.C.I. has no conflict of interest. MAG has no conflict of interest. C.H.T. has no conflict of interest. L.M. has received grants from Gilead, Abbvie, and Janssen, and speaker fees from Abbvie, Gilead Sciences, Janssen, MSD, Roche, and ViiV. He has also served as a consultant for Janssen, Abbvie, and Gilead; these activities are unrelated to the submitted work. D.V.G. has no conflict of interest. C.A.D. has no conflict of interest. E.B.M. has no conflict of interest. R.M.M.A. has no conflict of interest. N.C.C. has received speaker fees from Gilead, ViiV, Janssen, and MSD; these activities are unrelated to the submitted work. J.T. has received lecture fees from ViiV, Janssen, and Gilead; these activities are unrelated to the submitted work. A.G.G. has no conflict of interest. M.J.V. has received grants from Gilead and ViiV Healthcare, and speaker fees from Gilead Sciences, Janssen, and ViiV; these activities are unrelated to the submitted work. J.T.G. has received lecture fees from ViiV, Janssen, and Gilead; these activities are unrelated to the submitted work.

Supplementary Material

aids-40-414-s001.docx ^{(1,008.6KB, docx)}