Since HIV was first identified in 1980s in patients with AIDS,^1-3 it’s estimated that >42 million people have died from AIDS-related illnesses by 2023 and almost 40 million people were living with HIV worldwide.⁴ Global efforts led by The Joint United Nations Programme on HIV/AIDS (UNAIDS) aim to end HIV/AIDS as a public health threat by 2030. Attributed to the broad availability of effective treatment and prevention tools, countries are making progress toward epidemic control. However, there are still many challenges holding the program off interim UNAIDS milestone.⁵

It is essential to identify where and why new HIV infections are occurring to implement cost-effective targeted measures. As early as nearly 20 years ago, there were already attempts in Europe to establish recent infection surveillance to characterize the HIV epidemics using laboratory-based testing for HIV recent infection,^6-8 and this approach was recently piloted in Kenya and Zimbabwe as well.⁹ Laboratory-based recency testing requires a cold chain during specimen transport to a central laboratory, skilled laboratorians, and well-maintained laboratory devices, which can be challenging for a program-based surveillance in resource-limited settings.¹⁰

Asanté HIV-1 Rapid Recency^® Assay (ARRA) is a rapid test that can verify HIV-positive status [positive verification (PV) line] and detect recent HIV-1 infection [long-term (LT) line] in the same device (Sedia Biosciences Corporation, Oregon). It is based on the concept of limiting antigen avidity assay first described by our group in 2010.¹¹ A prototype in rapid test format was developed in 2013,¹² and it is the first and only commercially available test in this genre. The assay’s LT line contains a previously described HIV-1 rIDR-M recombinant antigen¹¹ coated at a specific but limiting concentration, so it only binds HIV-1 antibodies with high avidity, typically found in blood samples of individuals with LT infections. As a rapid test, the assay is easy to use and requires no additional equipment. With appropriate training, testers/health care personnel can conduct the testing in clinic/community settings.¹³ The innovation and successful commercialization of rapid testing for recent infection make it possible to establish a near real-time surveillance for HIV-1 recent infections at scale in resource-limited countries, in conjunction with routine HIV testing services (HTS).¹² Supported by the U.S. President’s Emergency Plan for AIDS Relief (PEPFAR) program, as of 2024, more than 20 countries are using ARRA for HIV recent infection surveillance under the Tracking with Recency Assays to Control the Epidemic (TRACE) initiative.¹⁴

Mean duration of recent infection (MDRI) and false recency rate (FRR) are two key performance parameters of recency assays. MDRI is the average time that subjects remain classified as “recent infection” by recency assay after seroconversion. FRR is the proportion of individuals with a LT infection (i.e., greater than 12 months) misclassified as recently infected.

In this study, we estimated the MDRI and FRR of ARRA using previously collected longitudinal and cross-sectional specimens, respectively, from various sources. In addition to visual interpretation, we used electronic reader supplied by the manufacturer to get quantitative numbers of line intensity to facilitate analysis. Furthermore, we conducted descriptive analyses to determine the probability of individuals testing recent infection on ARRA over time post-seroconversion. These findings provide laboratory evidence to guide the interpretation of data collected through surveillance using ARRA. Use of the ARRA in this context is not to estimate incidence but to identify factors that drive ongoing transmission of HIV-1 infections, as described.^15-17

ARRA is the only commercially available HIV rapid test for detecting recent infection post-seroconversion, which we have further characterized in this report. Results from this study using seroconversion panels demonstrate that MDRI of the test is close to 160–170 days, similar to what was extrapolated from comparison with previously calibrated LAg-Avidity EIA at ODn cutoff of 2.0. ARRA evaluation by the Consortium for the Evaluation and Performance of HIV Incidence Assays (CEPHIA) reported comparable MDRI estimate (175 days since nucleic acid positive).³³ The small difference could be attributed to acute infection detection versus seroconversion estimates or differences in characteristics of the seroconversion panels. It is to be noted that previously collected panels are depleting and collection of new longitudinal panels from recently infected individuals is not possible now due to ethical reasons. Therefore, in the absence of valuable panels, a comparison with previously calibrated assays and extrapolation from the data, as we have done before, is the most obvious and practical approach to estimate MDRI.^30,31,34

It should be noted that the MDRI does not equate to “seroconversion of all individuals within 6 months” since it is a mean of the population. MDRI is useful at the population level, not for individual cases. However, when used in the context of individual cases, they are considered to be infected within past 12 months due to a bell-shaped distribution of recency period around the mean. MDRI is useful for estimating population incidence, which is not the intent here. Although there is a categorical classification, it is meant to be interpreted as “likely infected within 12 months,” considering individual variability and some level of misclassification that can persist. This is quite appropriate for this intended use since not all individuals transition from recent to LT status at or before 6 months.

Our study estimated an overall FRR of about 3% among individuals with LT infections and close to 2% among those without advanced disease. This is consistent with recommended product profile of a recency assay. However, the FRR increases to about 5% among those with CD4 < 200. Accordingly, although the test will classify 95% of individuals with CD4 < 200 correctly as LT cases, some individuals may misclassify as recent due to waning antibodies during late-stage progression. This misclassification can be mitigated or reduced with an appropriate recent infection testing algorithm (RITA) combined with robust case surveillance that can provide information about prior diagnosis, ART use, and advanced HIV disease with low CD4 and their exclusion from recent infection testing. In this context, baseline CD4 should be measured soon after diagnosis as part of the case surveillance and recency testing. Initial drop in CD4 during acute/early infection can be an issue in some cases when including this biomarker as part of the case surveillance. However, the intent here is to improve overall recency classification but not to classify every case correctly, which may not be possible due to complexity of host-virus interaction.

Determination of MDRI is meant to characterize the recency assays to assist in a better understanding of the recent cases. It is to be noted that MDRI represents the population mean of the duration of recency. Although all individuals will start as being recent (Figs. 1 and 4), there are individuals who transition early to LT status and others who transition later to LT status (Fig. 1). Our analysis estimated that after seroconversion about 40%–50% may transition to LT status by 5–6 months as shown by avidity maturation (Fig. 1) and probability profile (Fig. 4). It is to be noted that the assay classified 278 (64.2%) of the 433 specimens (regions A + B) that were within 167 days of seroconversion, as recent while 36.8% were classified as LT. Similarly, 340 (48.4%) of 702 specimens that were within 1 year of seroconversion were classified as recent. Moreover, this proportion can vary depending on the distribution of specimens, since very early specimens are more likely to be classified as recent (Fig. 4).

Accordingly, although the individuals with seroconversion within 6 months were technically true recent cases, only a subset were classified as recent because of rapid or slow maturation of antibodies. As the time passes, between 6 and 12 months, more individuals will transition to LT status. Recent cases identified by an assay represent average number of cases within the MDRI, where those missed by the recency assay before the MDRI (Fig. 1, upper left region B) are balanced by detection of additional recent cases after the MDRI (lower right region C) in population studies. This observation should not be misinterpreted as “sensitivity” of the recency assay, as reported by a study in Uganda.³⁵ Longitudinal specimens from seroconverting individuals are used to confirm expected maturation or changes in biomarker profile and estimate MDRI of an assay, not to determine sensitivity of an assay. In routine recent infection surveillance, if individuals are known to be recently infected, based on the testing history with last negative test within last 6–12 months, they should be classified as recent without testing by the recency assay. However, if they are tested, only a subset is likely to classify as recent due to varying maturation kinetics, as shown here. In fact, as pointed out in 2022 technical guidance, “sensitivity” is not a relevant term in the context of recency assays.³⁶

PV line of the ARRA, which is meant to verify HIV-positive status while classifying them as recent or LT, had a sensitivity of 99.1%. These results are consistent with previously published studies demonstrating that, if verified and approved by a regulatory agency, ARRA could be used as part of the HIV testing algorithm.^30,31,37

Until recently, laboratory-based tests have been used to identify recently infected individuals.^9,10,16,38 However, this requires specimen collection and robust transportation network to regional or central laboratory to maintain specimen integrity and quality for testing. ARRA is a rapid test that can be performed in non-laboratory settings. The development of a well-characterized rapid recency assay allows recency testing using whole blood specimens at or near the point of collection. However, the focus of the recency testing should be on surveillance. It is not recommended that these results be returned to the individual since it does not have any clinical benefit in the test-and-start era.³⁹ Currently, more than 20 PEPFAR-supported countries are implementing the ARRA to help identify subpopulations with ongoing transmission as part of the program-based surveillance in HTS. The test is well accepted by health care providers demonstrating feasibility of implementation as a surveillance tool.¹³ This implementation is providing useful information about transmission hotspots and risk factors associated with acquiring HIV infections.^17,40-42

We do not recommend use of ARRA for the estimation of incidence since population being tested in HTS settings is biased, may vary over time, and is not representative of the population. Incidence estimate also requires high coverage of the target population. Moreover, the precision of a rapid recency test, with visual interpretation, is not likely to be as high as laboratory-based assay, like LAg-Avidity EIA, which is the preferred method for estimation of incidence in cross-sectional surveys.^43-45 We also do not recommend using the test as part of the HIV diagnostic strategy until the performance has been verified and approved by a regulatory agency.

Although the test is intended to be used among newly diagnosed ART-naïve persons who are identified in HTS settings, results from several countries indicate that a substantial proportion of individuals coming to HTS are prior positives and often on ART. They can be misclassified as recent due to decay of antibodies while on ART. This situation can be mitigated by adding viral load (VL) as part of the RITA. Testing clients coming to hospitals also indicated a significant proportion of individuals with advanced HIV disease/CD4 < 200/mm³. Ideally, both these populations should be excluded with robust case surveillance before the recency testing is performed, thereby reducing the need for VL.

The TRACE initiative, funded by the PEPFAR, is meant to complement cross-sectional surveys, such as population-based HIV impact assessment surveys, to identify recently infected individuals from routine testing services. If and when expanded to cover all newly diagnosed HIV-positives, along with appropriate RITA, it can provide a granular picture of ongoing transmission and help target appropriate prevention strategies.^14,41,46,47

It is reported that the performance of recency assay may vary across HIV subtypes.²⁹ The ARRA-LT line is coated with a multi-subtype recombinant protein rIDR-M to minimize such potential variation. Results didn’t indicate significant differences in MDRI across subtypes with overlapping 95% CI (Table 2). Moreover, these differences are unlikely to significantly affect recent infection surveillance. The primary objective of recent infection surveillance is to identify subpopulations with a high prevalence of recent cases infected within last 12 months rather than to estimate incidence rates.

Galiwango et al. reported a concordance of 72% between the two laboratories using only 45 specimens, all from recently infected individuals collected within 6 months of seroconversion.³⁵ Low concordance may be attributed to variability in training, individual visual acuity, and selection of specimens that are close to the cutoff. Recognizing that inter-laboratory reproducibility is expected to be higher than inter-tester reproducibility, our study demonstrated inter-tester reproducibility of >96% using more than 2,300 specimens with main discordancy observed for specimens close to the threshold cutoffs. Focusing just on specimens collected between 100 and 200 days since seroconversion from our dataset, we observed 177/202 (87.6%) specimens had concordant results, keeping in mind that all these individuals are recent infections (less than 1 year) and flip-flop while transitioning from recent to LT status is normal and expected. Considering inter-tester versus inter-laboratory settings, these results are not very different from Galiwango’s results. A well-planned hands-on training with a well-characterized panel is critical and should be part of a comprehensive quality assurance program to ensure good reproducibility. Trainings that can standardize the practice of testing and result interpretation on the ARRA should minimize variations. Furthermore, in real implementation settings, the proportion of challenging samples should be substantially lower than that of the panel used by Galiwango et al. where all specimens were from incident infections and were within 6 months of seroconversion.

The ARRA is a reliable rapid test for recent infection surveillance in programmatic settings. The determination of MDRI, FRR, and probability of testing recent by ARRA over time after seroconversion provide scientific reference for data interpretation for recency surveillance using Asanté. Overall, the ARRA is a valuable tool in the fight against HIV, helping to identify recent infections and target resources where they are needed most. Nevertheless, ARRA needs to be implemented with appropriate training, quality measures, and surveillance strategy and the surveillance data should be interpreted in the context of the target population.

Supplementary Data S1

Supplementary Data S2