Mean Recency Period for Estimation of HIV-1 Incidence with the BED-Capture EIA and Bio-Rad Avidity in Persons Diagnosed in the United States with Subtype B

Use of HIV Recency Assays for HIV Incidence Estimation and Other Surveillance Use Cases: Systematic Review

BACKGROUND

HIV assays designed to detect recent infection, also known as "recency assays," are often used to estimate HIV incidence in a specific country, region, or subpopulation, alone or as part of recent infection testing algorithms (RITAs). Recently, many countries and organizations have become interested in using recency assays within case surveillance systems and routine HIV testing services to measure other indicators beyond incidence, generally referred to as "non-incidence surveillance use cases."

OBJECTIVE

This review aims to identify published evidence that can be used to validate methodological approaches to recency-based incidence estimation and non-incidence use cases. The evidence identified through this review will be used in the forthcoming technical guidance by the World Health Organization (WHO) and United Nations Programme on HIV/AIDS (UNAIDS) on the use of HIV recency assays for identification of epidemic trends, whether for HIV incidence estimation or non-incidence indicators of recency.

METHODS

To identify the best methodological and field implementation practices for the use of recency assays to estimate HIV incidence and trends in recent infections for specific populations or geographic areas, we conducted a systematic review of the literature to (1) understand the use of recency testing for surveillance in programmatic and laboratory settings, (2) review methodologies for implementing recency testing for both incidence estimation and non-incidence use cases, and (3) assess the field performance characteristics of commercially available recency assays.

RESULTS

Among the 167 documents included in the final review, 91 (54.5%) focused on assay or algorithm performance or methodological descriptions, with high-quality evidence of accurate age- and sex-disaggregated HIV incidence estimation at national or regional levels in general population settings, but not at finer geographic levels for prevention prioritization. The remaining 76 (45.5%) described the field use of incidence assays including field-derived incidence (n=45), non-incidence (n=25), and both incidence and non-incidence use cases (n=6). The field use of incidence assays included integrating RITAs into routine surveillance and assisting with molecular genetic analyses, but evidence was generally weaker or only reported on what was done, without validation data or findings related to effectiveness of using non-incidence indicators calculated through the use of recency assays as a proxy for HIV incidence.

CONCLUSIONS

HIV recency assays have been widely validated for estimating HIV incidence in age- and sex-specific populations at national and subnational regional levels; however, there is a lack of evidence validating the accuracy and effectiveness of using recency assays to identify epidemic trends in non-incidence surveillance use cases. More research is needed to validate the use of recency assays within HIV testing services, to ensure findings can be accurately interpreted to guide prioritization of public health programming.

Regression with interval-censored covariates: Application to cross-sectional incidence estimation

A method for generalized linear regression with interval-censored covariates is described, extending previous approaches. A scenario is considered in which an interval-censored covariate of interest is defined as a function of other variables. Instead of directly modeling the distribution of the interval-censored covariate of interest, the distributions of the variables which determine that covariate are modeled, and the distribution of the covariate of interest is inferred indirectly. This approach leads to an estimation procedure using the Expectation-Maximization (EM) algorithm. The performance of this approach is compared to two alternative approaches, one in which the censoring interval midpoints are used as estimates of the censored covariate values, and another in which the censored values are multiply imputed using uniform distributions over the censoring intervals. A simulation framework is constructed to assess these methods' accuracies across a range of scenarios. The proposed approach is found to have less bias than midpoint analysis and uniform imputation, at the cost of small increases in standard error.

Characteristics of Persons Who Inject Drugs with Recent HIV Infection in the United States: National HIV Behavioral Surveillance, 2012

We evaluated characteristics associated with recent HIV infection among persons who inject drugs (PWID) from 19 U.S. cities who participated in 2012 National HIV Behavioral Surveillance. Recent infection was defined as having a reactive HIV test, a Bio-Rad Avidity index cutoff ≤ 30%, no reported HIV diagnosis ≥ 12 months before interview, and no evidence of viral suppression. Of 8667 PWID, 50 (0.6%) were recently HIV infected. Having a greater number of sex partners (≥ 2 partners vs. 0) [prevalence ratio (PR) 4.7, 95% confidence interval (CI) 1.3-17.8], injecting heroin and other drugs (PR 3.0, 95% CI 1.3-6.6) or exclusively non-heroin drugs (PR 5.9, 95% CI 1.7-20.7) compared to injecting only heroin, and having male-male sex in the past year (PR 7.1, 95% CI 3.0-16.6) were associated with recent infection. Promoting not only safe injection practices but also safe sex practices will be key to preventing new HIV infections.

Exploring HIV-1 Transmission Dynamics by Combining Phylogenetic Analysis and Infection Timing

HIV-1 pol sequences obtained through baseline drug resistance testing of patients newly diagnosed between 2013 and 2017 were analyzed for genetic similarity. For 927 patients the information on genetic similarity was combined with demographic data and with information on the recency of infection. Overall, 48.3% of the patients were genetically linked with 11.4% belonging to a pair and 36.9% involved in a cluster of ≥3 members. The percentage of early diagnosed (≤4 months after infection) was 28.6%. Patients of Belgian origin were more frequently involved in transmission clusters (49.7% compared to 15.3%) and diagnosed earlier (37.4% compared to 12.2%) than patients of Sub-Saharan African origin. Of the infections reported to be locally acquired, 69.5% were linked (14.1% paired and 55.4% in a cluster). Equal parts of early and late diagnosed individuals (59.9% and 52.4%, respectively) were involved in clusters. The identification of a genetically linked individual for the majority of locally infected patients suggests a high rate of diagnosis in this population. Diagnosis however is often delayed for >4 months after infection increasing the opportunities for onward transmission. Prevention of local infection should focus on earlier diagnosis and protection of the still uninfected members of sexual networks with human immunodeficiency virus (HIV)-infected members.

Identifying Clusters of Recent and Rapid HIV Transmission Through Analysis of Molecular Surveillance Data

BACKGROUND

Detecting recent and rapid spread of HIV can help prioritize prevention and early treatment for those at highest risk of transmission. HIV genetic sequence data can identify transmission clusters, but previous approaches have not distinguished clusters of recent, rapid transmission. We assessed an analytic approach to identify such clusters in the United States.

METHODS

We analyzed 156,553 partial HIV-1 polymerase sequences reported to the National HIV Surveillance System and inferred transmission clusters using 2 genetic distance thresholds (0.5% and 1.5%) and 2 periods for diagnoses (all years and 2013-2015, ie, recent diagnoses). For rapidly growing clusters (with ≥5 diagnoses during 2015), molecular clock phylogenetic analysis estimated the time to most recent common ancestor for all divergence events within the cluster. Cluster transmission rates were estimated using these phylogenies.

RESULTS

A distance threshold of 1.5% identified 103 rapidly growing clusters using all diagnoses and 73 using recent diagnoses; at 0.5%, 15 clusters were identified using all diagnoses and 13 using recent diagnoses. Molecular clock analysis estimated that the 13 clusters identified at 0.5% using recent diagnoses had been diversifying for a median of 4.7 years, compared with 6.5-13.2 years using other approaches. The 13 clusters at 0.5% had a transmission rate of 33/100 person-years, compared with previous national estimates of 4/100 person-years.

CONCLUSIONS

Our approach identified clusters with transmission rates 8 times those of previous national estimates. This method can identify groups involved in rapid transmission and help programs effectively direct and prioritize limited public health resources.

Evaluation of the Performance of Three Biomarker Assays for Recent HIV Infection Using a Well-Characterized HIV-1 Subtype C Incidence Cohort

Biomarkers for detecting early HIV infection and estimating HIV incidence should minimize false-recent rates (FRRs) while maximizing mean duration of recent infection (MDRI). We compared HIV subtypes B, E and D (BED) capture enzyme immunoassay (BED), Sedia limiting antigen (LAg) avidity enzyme immunoassay, and Bio-Rad avidity incidence (BRAI) assays using samples from Zimbabwean postpartum women infected with clade C HIV. We calculated MDRIs using 590 samples from 351 seroconverting postpartum women, and FRRs using samples from 2,825 women known to be HIV positive for >12 months. Antibody kinetics were more predictable with LAg and had higher precision compared with BED or BRAI. BRAI also exhibited more variability, and avidity reversal in some cases. For BED, LAg, and BRAI, used alone or with viral load, MDRI values in days were: BED-188 and 170 at normalized optical density (ODn) 0.8; LAg-104 and 100 at ODn cutoff 1.5; BRAI-135 and 134 at avidity index cutoff 30%. Corresponding FRRs were: BRAI 1.1% and 1.0% and LAg 0.57% and 0.35%: these were 3.8-10.9 times lower than BED values of 4.8% and 3.8%. BRAI and LAg have significantly lower FRRs and MDRIs than in published studies, and much lower than BED and could be used to estimate incidence in perinatal women and to measure population-level HIV incidence in HIV control operations in Africa.

The concordance of the limiting antigen and the Bio-Rad avidity assays in persons from Estonia infected mainly with HIV-1 CRF06_cpx

Background

Serological assays to determine HIV incidence have contributed to estimates of HIV incidence, monitoring of HIV spread, and evaluation of prevention strategies. Two frequently used incidence assays are the Sedia HIV-1 LAg-Avidity EIA (LAg) and the Bio-Rad avidity incidence (BRAI) assays with a mean duration of recent infection (MDRI) of 130 and 240 days for subtype B infections, respectively. Little is known about how these assays perform with recombinant HIV-1 strains. We evaluated the concordance of these assays in a population infected mainly with HIV-1 CRF06_cpx.

Material/Methods

Remnant serum samples (n = 288) collected from confirmed, newly-diagnosed HIV-positive persons from Estonia in 2013 were tested. Demographic and clinical data were extracted from clinical databases. LAg was performed according to the manufacturer’s protocol and BRAI testing was done using a validated protocol. Samples with LAg-pending or BRAI-invalid results were reclassified as recent if they were from persons with viral loads <1000 copies/mL or were reclassified as long-term if presenting with AIDS.

Results

In total 325 new HIV infections were diagnosed in 2013 in Estonia. Of those 276 persons were tested with both LAg and BRAI. Using assay results only, the recency rate was 44% and 70% by LAg and BRAI, respectively. The majority of samples (92%) recent by LAg were recent by BRAI. Similarly, 89% of samples long-term by BRAI were long-term by LAg. After clinical information was included in the analysis, the recency rate was 44% and 62% for LAg and BRAI, respectively. The majority of samples (86%) recent by LAg were recent by BRAI and 91% of long-term infections by BRAI were long-term by LAg.

Conclusions

Comparison of LAg and BRAI results in this mostly CRF06_cpx-infected population showed good concordance for incidence classification. Our finding of a higher recency rate with BRAI in this population is likely related to the longer MDRI for this assay.

A Multiplex HIV Incidence Assay for Inferring Recent HIV-1 Transmission and Time of Infection

BACKGROUND

Laboratory assays for determining recent HIV-1 infection are an important public health tool for aiding in the estimation of HIV incidence. Some incidence assay analytes are remarkably predictive of time since seroconversion and may be useful for additional applications, such as predicting recent transmission events during HIV outbreaks and informing prevention strategies.

METHODS

Plasma samples (n = 154) from a recent HIV-1 outbreak in a rural community in Indiana were tested with the customized HIV-1 Multiplex assay, based on the Bio-Rad Bio-Plex platform, which measures antibody response to HIV envelope antigens, gp120, gp160, and gp41. Assay cutoffs for each analyte were established to determine whether an individual seroconverted within 30, 60, or 90 days of the sample collection date. In addition, a novel bioinformatics method was implemented to infer infection dates of persons newly diagnosed with HIV during the outbreak.

RESULTS

Sensitivity/specificity of the HIV-1 Multiplex assay for predicting seroconversion within 30, 60, and 90 days, based on a training data set, was 90.5%/95.4%, 94.1%/90%, and 89.4%/82.9%, respectively. Of 154 new diagnoses in Indiana between December 2014 and August 2016, the majority (71%) of recent infections (≤3 months since seroconversion) were identified between February and May 2016. The epidemiologic curve derived from the bioinformatics analysis indicated HIV transmission began as early as 2010, grew exponentially in 2014, and leveled off in April 2015.

CONCLUSIONS

The HIV-1 Multiplex assay has the potential to identify and monitor trends in recent infection during an epidemic to assess the efficacy of programmatic or treatment interventions.

Evaluation of a BioRad Avidity assay for identification of recent HIV-1 infections using dried serum or plasma spots

Serological methods to differentiate between recently acquired and established HIV-1 infections are a useful tool in the HIV-surveillance to characterize the epidemic, identify groups at risk and assess HIV-preventive interventions. Therefore, an avidity-based, modified BioRad Genscreen™ HIV-1/2 assay (BRA_EUR) was evaluated according to the avidity-based, modified BioRad HIV-1/2 Plus O protocol (BRA_USA). Overall, 692 well defined samples (82.5% B and 17.5% non-B subtypes) from recent (<180 days, n = 239), intermediate (181-364 days, n = 35) or long term infections (≥365 days, n = 419) were used to determine a 'mean duration of recent infection' (MDRI), a 'median DRI' (MdDRI), the false recent rate (FRR), and concordance between the BRAs and the Sedia BED HIV-1 Capture enzyme immunoassay (BED). The optimal avidity index cut-off was determined to be 70% resulting in an MDRI of 233 days (95% IQR: 174-351) and an MdDRI of 171 days (95% IQR: 142-212). Concordance with the BRA_USA was high with 96.4%. The FRR of 6.0% as well as the MdDRI are similar to the BED (8.4%; 170 (139-214) days). Therefore, the BRA_EUR is a suitable alternative to replace the BED and trend analysis will be feasible after minimal adjustments for the MdDRI and the MDRI.

Foreign-Born Individuals with HIV in King County, WA: A Glimpse of the Future of HIV?

To better understand country of birth-related shifts in the demography of people newly diagnosed with HIV infection, we compared demographic and clinical characteristics of foreign-born and U.S.-born residents of King County, WA diagnosed with HIV from 2006 to 2015. The proportion of cases that were foreign-born increased from 23 to 34% during this time. Most foreign-born cases were born in Africa (34%), Latin America (32%), Asia (22%), or Europe (7%). Latin Americans and Asians were similar to U.S.-born individuals by HIV risk factor and gender, while Africans were more likely to be female and less often men who have sex with men. In 2015, approximately 15% of cases presumptively newly diagnosed in King County were foreign-born individuals who self-reported a pre-immigration HIV diagnosis. Increases in foreign-born individuals previously diagnosed out of country may lead to inaccuracy in the count of new diagnoses, including an over-estimate of community-acquired HIV in King County.

Decision tree for accurate infection timing in individuals newly diagnosed with HIV-1 infection

Background

There is today no gold standard method to accurately define the time passed since infection at HIV diagnosis. Infection timing and incidence measurement is however essential to better monitor the dynamics of local epidemics and the effect of prevention initiatives.

Methods

Three methods for infection timing were evaluated using 237 serial samples from documented seroconversions and 566 cross sectional samples from newly diagnosed patients: identification of antibodies against the HIV p31 protein in INNO-LIA, SediaTM BED CEIA and SediaTM LAg-Avidity EIA. A multi-assay decision tree for infection timing was developed.

Results

Clear differences in recency window between BED CEIA, LAg-Avidity EIA and p31 antibody presence were observed with a switch from recent to long term infection a median of 169.5, 108.0 and 64.5 days after collection of the pre-seroconversion sample respectively. BED showed high reliability for identification of long term infections while LAg-Avidity is highly accurate for identification of recent infections. Using BED as initial assay to identify the long term infections and LAg-Avidity as a confirmatory assay for those classified as recent infection by BED, explores the strengths of both while reduces the workload. The short recency window of p31 antibodies allows to discriminate very early from early infections based on this marker. BED recent infection results not confirmed by LAg-Avidity are considered to reflect a period more distant from the infection time. False recency predictions in this group can be minimized by elimination of patients with a CD4 count of less than 100 cells/mm3 or without no p31 antibodies. For 566 cross sectional sample the outcome of the decision tree confirmed the infection timing based on the results of all 3 markers but reduced the overall cost from 13.2 USD to 5.2 USD per sample.

Conclusions

A step-wise multi assay decision tree allows accurate timing of the HIV infection at diagnosis at affordable effort and cost and can be an important new tool in studies analyzing the dynamics of local epidemics or the effects of prevention strategies.

HIV avidity index performance using a modified fourth-generation immunoassay to detect recent HIV infections

BACKGROUND

Detecting recent HIV infections is important to evaluate incidence and monitor epidemic trends. We aimed to evaluate the diagnostic performance and accuracy of the avidity index (AI) for discriminating for recent HIV infections.

METHODS

We collected serum samples from HIV-1 positive individuals: A) with known date of infection (midpoint in time between last HIV-negative and first HIV-positive test); B) infected for >1 year. Samples were divided into two aliquots: one diluted with phosphate buffered saline (PBS) and the other with 1 M guanidine. Both aliquots were assayed by the Architect HIV Ag/Ab Combo 4th generation assay (Abbott). We compared AI found in recent (RI=<6 months from seroconversion) and established (EI) infections. The diagnostic accuracy was evaluated by receiver operating characteristic (ROC) curve analysis. The proportion of samples misclassified as recent (FRR) was calculated.

RESULTS

In total, 647 samples were collected: 455 in group A (51.6% RI and 48.4% EI) and 192 in group B. Among these, sixteen samples were from elite controllers, 294 from treated patients, 328 from patients infected with non-B subtypes. Samples before antiretroviral initiation showed a mean AI significantly lower among RI compared to EI (0.66+0.28 vs. 1.00±0.12; p<0.000). The FRR was 0% using a cut-off of ≤0.70. An extremely low FRR was observed among elite controllers, samples with low VL or CD4. HIV subtype had no impact on AI misclassifications. All individuals in group A reached the AI threshold of 0.80 within 24 months after seroconversion.

CONCLUSIONS

The AI is an accurate serological marker for discriminating recent from established HIV infections and meets WHO requirements for HIV incidence assays.

Detailed Transmission Network Analysis of a Large Opiate-Driven Outbreak of HIV Infection in the United States

In January 2015, an outbreak of undiagnosed human immunodeficiency virus (HIV) infections among persons who inject drugs (PWID) was recognized in rural Indiana. By September 2016, 205 persons in this community of approximately 4400 had received a diagnosis of HIV infection. We report results of new approaches to analyzing epidemiologic and laboratory data to understand transmission during this outbreak. HIV genetic distances were calculated using the polymerase region. Networks were generated using data about reported high-risk contacts, viral genetic similarity, and their most parsimonious combinations. Sample collection dates and recency assay results were used to infer dates of infection. Epidemiologic and laboratory data each generated large and dense networks. Integration of these data revealed subgroups with epidemiologic and genetic commonalities, one of which appeared to contain the earliest infections. Predicted infection dates suggest that transmission began in 2011, underwent explosive growth in mid-2014, and slowed after the declaration of a public health emergency. Results from this phylodynamic analysis suggest that the majority of infections had likely already occurred when the investigation began and that early transmission may have been associated with sexual activity and injection drug use. Early and sustained efforts are needed to detect infections and prevent or interrupt rapid transmission within networks of uninfected PWID.

Short Communication: Heightened HIV Antibody Responses in Postpartum Women as Exemplified by Recent Infection Assays: Implications for Incidence Estimates

Laboratory assays that identify recent HIV infections are important for assessing impacts of interventions aimed at reducing HIV incidence. Kinetics of HIV humoral responses can vary with inherent assay properties, and between HIV subtypes, populations, and physiological states. They are important in determining mean duration of recent infection (MDRI) for antibody-based assays for detecting recent HIV infections. We determined MDRIs for multi-subtype peptide representing subtypes B, E and D (BED)-capture enzyme immunoassay, limiting antigen (LAg), and Bio-Rad Avidity Incidence (BRAI) assays for 101 seroconverting postpartum women, recruited in Harare from 1997 to 2000 during the Zimbabwe Vitamin A for Mothers and Babies trial, comparing them against published MDRIs estimated from seroconverting cases in the general population. We also compared MDRIs for women who seroconverted either during the first 9 months, or at later stages, postpartum. At cutoffs (C) of 0.8 for BED, 1.5 for LAg, and 40% for BRAI, estimated MDRIs for postpartum mothers were 192, 104, and 144 days, 33%, 32%, and 52% lower than published estimates of 287, 152 and 298 days, respectively, for clade C samples from general populations. Point estimates of MDRI values were 7%-19% shorter for women who seroconverted in the first 9 months postpartum than for those seroconverting later. MDRI values for three HIV incidence biomarkers are longer in the general population than among postpartum women, particularly those who recently gave birth, consistent with heightened immunological activation soon after birth. Our results provide a caution that MDRI may vary significantly between subjects in different physiological states.

Performance characteristics of an antibody-based multiplex kit for determining recent HIV-1 infection

The availability of reliable laboratory methods for determining recent HIV infection is vital for accurate estimation of population-based incidence. The mean duration of recent infection (MDRI) and false recent rate (FRR) are critical parameters for HIV incidence assays, as they impact HIV incidence estimates and provide a measure of assay performance. The HIV-1 Multiplex assay is an in-house developed, magnetic bead-based assay that measures virus-specific antibody levels and avidity to multiple analytes. To ensure quality control and to facilitate transfer of the assay to external laboratories or testing facilities, the in-house assay has been adapted and produced in kit form. Here, we describe the performance characteristics of the multiplex kit and demonstrate the stability of the kit components over a one-year period. Two statistical methods were employed to estimate the MDRI of the individual analytes and five different algorithms, combining multiple analyte values. The MDRI estimates for the individual analytes and five algorithms were all between 200 and 300 days post-seroconversion, with no notable difference between the two statistical approaches. All five algorithms exhibited a 0% FRR with specimens from long-term, subtype B HIV-1-infected individuals. The assay parameters described in this study provide the necessary tools to implement the HIV-1 multiplex assay and improves the utility of the assay for field use.

HIV Trends in the United States: Diagnoses and Estimated Incidence

Background

The best indicator of the impact of human immunodeficiency virus (HIV) prevention programs is the incidence of infection; however, HIV is a chronic infection and HIV diagnoses may include infections that occurred years before diagnosis. Alternative methods to estimate incidence use diagnoses, stage of disease, and laboratory assays of infection recency. Using a consistent, accurate method would allow for timely interpretation of HIV trends.

Objective

The objective of our study was to assess the recent progress toward reducing HIV infections in the United States overall and among selected population segments with available incidence estimation methods.

Methods

Data on cases of HIV infection reported to national surveillance for 2008-2013 were used to compare trends in HIV diagnoses, unadjusted and adjusted for reporting delay, and model-based incidence for the US population aged ≥13 years. Incidence was estimated using a biomarker for recency of infection (stratified extrapolation approach) and 2 back-calculation models (CD4 and Bayesian hierarchical models). HIV testing trends were determined from behavioral surveys for persons aged ≥18 years. Analyses were stratified by sex, race or ethnicity (black, Hispanic or Latino, and white), and transmission category (men who have sex with men, MSM).

Results

On average, HIV diagnoses decreased 4.0% per year from 48,309 in 2008 to 39,270 in 2013 (P<.001). Adjusting for reporting delays, diagnoses decreased 3.1% per year (P<.001). The CD4 model estimated an annual decrease in incidence of 4.6% (P<.001) and the Bayesian hierarchical model 2.6% (P<.001); the stratified extrapolation approach estimated a stable incidence. During these years, overall, the percentage of persons who ever had received an HIV test or had had a test within the past year remained stable; among MSM testing increased. For women, all 3 incidence models corroborated the decreasing trend in HIV diagnoses, and HIV diagnoses and 2 incidence models indicated decreases among blacks and whites. The CD4 and Bayesian hierarchical models, but not the stratified extrapolation approach, indicated decreases in incidence among MSM.

Conclusions

HIV diagnoses and CD4 and Bayesian hierarchical model estimates indicated decreases in HIV incidence overall, among both sexes and all race or ethnicity groups. Further progress depends on effectively reducing HIV incidence among MSM, among whom the majority of new infections occur.

Strengthening HIV surveillance: measurements to track the epidemic in real time

Surveillance for HIV as a public health initiative requires timely, detailed and robust data to systematically understand burden of infection, transmission patterns, direct prevention efforts, guide funding, identify new infections and predict future trends in the epidemic. The methods for HIV surveillance have evolved to reliably track the epidemic and identify new infections in real time. Initially HIV surveillance relied primarily on the reporting of AIDS cases followed by measuring antibodies to HIV to determine prevalence in key populations. With the roll-out of antiretroviral therapy (ART) resulting in better survival and the corresponding increase in HIV prevalence, the landscape of surveillance shifted further to track HIV prevalence and incidence within the context of programmes. Recent developments in laboratory assays that potentially measure and differentiate recent versus established HIV infection offer a cost-effective method for the rapid estimation of HIV incidence. These tests continue to be validated and are increasingly useful in informing the status of the epidemic in real time. Surveillance of heterogeneity of infections contributing to sub-epidemics requires methods to identify affected populations, density, key geographical locations and phylogenetically linked or clustered infections. Such methods could provide a nuanced understanding of the epidemic and prioritise prevention efforts to those most vulnerable. This paper brings together recent developments and challenges facing HIV surveillance, together with the application of newer assays and methods to fast-track the HIV prevention and treatment response.