Surveillance data from Europe [1,2], the USA  and Australia  have shown recent increases in the number of individuals testing positive for HIV-1 infection. This is despite awareness of measures to reduce HIV transmission and the widespread availability of highly active antiretroviral therapy (HAART). As conventional techniques utilizing standard HIV-1 antibody testing are unable to distinguish between long-standing and recent infections, it is not possible directly to determine whether this reflects recent increases in HIV transmission .
Identification of recent infections offers a unique opportunity to consider early treatment, facilitate partner notification and prevent onward transmission, monitor trends in incidence and resistance, and assess the effectiveness of intervention measures. This may be of particular importance as recent infections represent a time of increased infectiousness .
The diagnosis of recent infections poses challenges for the clinician and the laboratory. Although 40–90% of recent infections are symptomatic, the clinical manifestations are usually non-specific  and either do not present or go unrecognized by healthcare providers [8,9]. Whereas conventional techniques for identifying recent infections may represent the gold standard, there are limitations to each: many individuals will not have previously tested for HIV ; the possibility of recent infections may not have been considered at the time of clinical presentation and hence RNA/DNA testing had not been performed ; and presentation may be at a time when antibody responses and western blot profiles are already fully developed. Correct identification of recent infections in such individuals would still, however, represent an opportunity for each of the benefits outlined above.
The serological testing algorithm for recent HIV seroconversion (STARHS) relies on the inability of a less sensitive test to detect the low levels of anti-HIV antibody that are present during a window of several months after seroconversion. Specimens that are reactive in a sensitive assay but are below a standardized threshold of reactivity in the ‘detuned’ (less sensitive) assay, are for subtype B indicative of infection within the last 5–6 months .
We describe the incorporation of the STARHS technique into the routine evaluation of individuals diagnosed with HIV infection at a single HIV testing and treatment centre caring for over 1500 HIV-positive patients and discuss the potential benefits of this for routine clinical care and potential limitations for surveillance.
The participants were those individuals who presented to the HIV treatment centre at Brighton and Sussex University Hospitals over a 10-year period between January 1996 and December 2005. Screening for HIV-1 infection was initially done using Abbott AxSYM HIV 1/2 g0 (Abbott laboratories, Abbott Park, Illinois, USA), switching to Abbott AxSYM HIV COMBO 1/2 gO in July 2003 and finally Abbott ARCHITECT HIV Ag/Ab COMBO in May 2005. All reactive specimens were referred for confirmatory testing to the Virus Reference Department, Central Public Health Laboratory (CPHL), now the Health Protection Agency (HPA), Centre for Infections, London, UK. Demographic details (age, sex, risk factors for HIV acquisition, and ethnicity) were recorded at the time of HIV diagnosis. Where an individual had previously been diagnosed with HIV-1 infection at another treatment centre, this was recorded. Conventional diagnosis of recent infections was by a previous negative HIV test within 18 months, detectable HIV RNA with negative HIV antibody test, evolving HIV antibody pattern on serial samples, and/or a western blot banding pattern that was limited to four bands, including at least p24 and gp160.
STARHS testing was performed at the National Reference Laboratory (HPA, London) with the Vironostika HIV-1 Microelisa System assay (bioMérieux UK Ltd, Basingstoke, UK) as previously described . For this study an optical density of less than 1.0 was used to identify recent infections and this cut-off is associated with seroconversion within 5–6 months. Testing was performed retrospectively for those diagnosed in 1996–2000, and prospectively thereafter.
For subjects for whom the STARHS technique suggested incident HIV infection, further information was collected regarding clinical features of primary HIV infection (PHI), prior HIV testing history, clinical status at diagnosis, sexual history, other laboratory features suggestive of recent infection, CD4 cell count, viral load and HAART history (if known previously to have been HIV positive).
Individuals were considered to have had a recent infection if STARHS gave a result consistent with recent infection and this was confirmed by clinical and/or other laboratory features. The proportions of incident to prevalent diagnoses were determined by year and by risk group, and trends over time were analysed using the Kruskal Wallis Test (SPSS software version 12.0; SPSS Inc., Chicago, Illinois, USA).
Between January 1996 and December 2005, 1526 individuals presented as anti-HIV-1 antibody positive to this treatment centre (Table 1). Of these, 714 had previously been diagnosed with HIV infection at another centre. Thus, for 812 this represented a new HIV diagnosis. Of these, 604 (74%) were men who have sex with men (MSM) and 208 (26%) were heterosexual (83 male and 125 female), of whom 110 (53%) were known to have acquired their infection in countries of high HIV prevalence. The median ages were 36 years for the MSM group (range 19 to 82 years) and 34 for the heterosexual group (range 17 to 73 years).
Conventional diagnosis of recent infection
Of those newly diagnosed with HIV infection, 175 of 812 (22%) were initially diagnosed as having recent infection, based on one or more previous negative HIV antibody tests within 18 months (n = 150), evolving serological response (n = 31), incomplete western blot assay profile (n = 8) or antibody negative/PCR positive discordance (n = 1) (Table 1).
Application of serological testing algorithm for recent HIV seroconversion
Of the 1526 specimens collected at first presentation to this treatment centre, 1112 (73%) were available for analysis by the STARHS method, including 715 (88%) of 812 with newly diagnosed infection (Table 2). Among the 1112 tested, 289 (26%) gave a result consistent with an incident infection using the STARHS assay. Considering only the 715 in whom this was their first HIV diagnosis, 228 (32%) individuals were preliminarily identified as incident infection by the STARHS assay.
Identification of false incident serological testing algorithm for recent HIV seroconversion results
Of the 289 individuals identified as incident by STARHS, 79 (27%) were considered to be incorrect based upon the clinical history (Fig. 1). They comprised 38 individuals known to have long-standing HIV infection with an undetectable viral load (< 200 copies/ml) who had been receiving HAART for a median of 45 months (range 6–145 months). A further five, of whom four had a previous AIDS diagnosis, had been receiving antiretroviral therapy for a median of 15 months (range, 3 to 100 months) and were sub-optimally suppressed with a median viral load of 3.83 log copies/ml (range 2.72 to 5.47). A further 27 individuals were not on HAART but had advanced HIV infection with a median CD4 cell count of 37 cells/μl (range 7 to 195) – 19 of these individuals presented with AIDS at diagnosis. A further five were known to have prevalent HIV infection (> 3 years since diagnosis) and were not receiving HAART but had very low levels of detectable virus (< 50 copies/ml in four individuals and 388 copies/ml in another), and a further two had never tested previously but had undetectable viral loads (< 50 copies/ml). For two, of unknown subtype, there was no apparent clinical explanation for an incident STARHS result.
Considering only those in whom this was their first HIV diagnosis, 23 of 228 (10%) were considered to be incorrectly identified as incident based upon the clinical history, and these have been excluded from the data (Table 1). Of these, 17 had a diagnosis of AIDS at presentation and a further four had a median CD4 cell count of 50 cells/μl (range 5–130). The remaining two had undetectable viral loads (< 50 copies/ml).
Combination of serological testing algorithm for recent HIV seroconversion and conventional methods for diagnosing recent infection
Of the 715 individuals in whom this was their first HIV diagnosis and STARHS was applied, excluding those considered to be incorrectly identified as described above, 237 (33%) were determined as incident by incorporating the STARHS technique (Table 2). Without the application of STARHS, only 149 (21%) would conventionally have been identified as recent infections. Therefore, a further 88 (12%) new diagnoses were determined by STARHS to have been infected recently that would otherwise have gone undetected. Behavioural and clinical data were available in 73 of 88 recent infections diagnosed by STARHS alone: unprotected anal or vaginal intercourse with a partner from a high-risk group within the previous 6 months was reported by 60 of 73 (82%); and seroconversion symptoms by 34 of 73 (47%) individuals, as determined at the time of initial presentation.
Of those known to be infected in the previous 6 months by conventional methods (negative test in the previous 6 months, incomplete western blot or evolving serology) and where STARHS was performed, recent infection was corroborated by STARHS in 71 of 74 (96%). In the three remaining cases a negative HIV screening test had been recorded between 4 and 5 months previously (2 subtype B, one unknown).
Conversely, STARHS confirmed nonincident infection in 64 of 64 (100%) individuals known to have been diagnosed HIV-positive more than 18 months previously with a CD4 cell count of > 200 cells/μl, viral load > 500 copies/ml and not on HAART.
Subtype and serological testing algorithm for recent HIV seroconversion results
HIV-1 subtype was known for 485 of 1526 (32%) of the total population; 374 (25%) were infected with subtype B, 111 (7%) with non-B and it was unknown in 1041 (68%). Of those 812 individuals in whom this was their first HIV diagnosis, 366 (45%) were infected with subtype B, 106 (13%) with non-B and it was unknown in 340 (42%).
Of the 71 individuals known to be infected in the previous 6 months by conventional methods and with an incident STARHS result, 58 of 71 (82%) were infected with subtype B and the remaining 13 had an unknown subtype but were UK-born gay white men.
Eleven of 88 (13%) previously unknown recent infections were non-B subtype – documentation was available in ten of these cases and unprotected anal or vaginal sex within the previous 6 months was reported in nine individuals and seroconversion symptoms in six.
Trends in recent infection
The proportion of individuals with diagnosed recent infections (both by established methods and by STARHS) increased over time (Table 1). Considering only those individuals for whom specimens were available for STARHS testing (Table 2), this pattern was even more marked (P < 0.01). There were clear differences between the proportions identified as recent infections by risk category for HIV infection; in the heterosexual group the proportion identified as recent was stable (P > 0.1, NS) and generally low whilst in the MSM group it was higher and increasing (P < 0.01), accounting for the overall increase over time and reaching a level of over 50% of new diagnoses in 2002, 2004 and 2005. Of the 88 additionally identified as recent by STARHS alone, 11 (13%) harboured significant resistant mutations (data not shown).
Over the study period we observed an increasing proportion of individuals newly diagnosed with HIV to have been recently infected, supporting the assertion that the rate of newly acquired HIV-infection is stable  or has increased over recent years [14,15]. This trend is particularly marked amongst MSM. This demonstrates that ongoing HIV transmission is occurring alongside changing sexual behaviour, despite the awareness of effective HIV prevention strategies and the potential for HAART to reduce the transmission of HIV, and may be an indirect result of the beneficial effects of HAART on HIV-related morbidity and mortality [16,17]. An alternative explanation is that, particularly among MSM, individuals are presenting more frequently for HIV testing  thereby making recognition of recent infection more likely.
These data demonstrate that the use of the STARHS technique enables identification of recent infections over and above that which would have otherwise been possible. Utilizing conventional clinical and laboratory techniques for diagnosing recent infections, 21% of new HIV diagnoses were determined to be recent, but with the addition of the STARHS technique a further 12% gave a result consistent with recent HIV infection. Thus 59% more recent infections were diagnosed, representing an increase in overall recent infections from 21 to 33%. Such improved detection of recent versus chronic infection allows the opportunity for discussion of early intervention studies for the individual, and it is likely that sources of infection and contacts at risk of onward transmission may be more easily identified. Furthermore, the incorporation of STARHS into a surveillance scheme that monitors antiretroviral resistance  should permit a comparison of primary resistance between recent and longstanding infections, enabling more accurate prediction of future resistance patterns. The utility of methods such as STARHS to determine recent infection in non-B subtypes requires validation before it can be routinely introduced .
It has been argued that since a STARHS (or similar methodology) result consistent with recent infections can occur in individuals with known prevalent infection this test should not be utilized in a clinical setting [21,22], and our findings confirm the impact of HAART and advanced disease stage on the accuracy of STARHS . Although STARHS was adjudged to have given a false result of recent infections in 79 of 1112 (71.1%) HIV-infected individuals, it was only in nine of these 79 (11%) that an obvious cause for the false incident result could not be easily recognized with adjunctive use of clinical or other laboratory data. Therefore, we found that, when used together, laboratory and clinical data led to a false recent infection rate of only nine of 1112 (0.8%). Thus our data demonstrate that false incident results can be reliably identified and, conversely, suggest that significant limitations may be associated with surveillance applications [24,25] if relevant clinical data are not available for individual specimens.
Finally, alternative serological techniques have recently been described that identify recent HIV infection such as an antibody avidity assay  and the BED ELISA . These assays employ different biological principles, which may be less affected by antiretroviral therapy, advanced disease and subtype and may, either in isolation or in combination with each other, offer more accurate differentiation of recent from longstanding HIV infection. Both of these assays, however, require further validation before their usefulness can be fully assessed.
2. Dukers NTM, Spaargaren J, Geskus RB, Beijnen J, Coutinho RA, Fennema HAS. HIV incidence on the increase among homosexual men attending an Amsterdam sexually transmitted disease clinic: using a novel approach for detecting recent infections. AIDS 2002; 16:F19–F24.
3. Centers for Disease Control and Prevention. HIV/AIDS surveillance report, 2004
. Vol. 16. Atlanta: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2005.
4. National Centre in HIV Epidemiology and Clinical Research. HIV/AIDS, viral hepatitis and sexually transmissible infections in Australia. Annual surveillance report 2005.
The University of New South Wales, Sydney, NSW. Available at: http://web.med.unsw.edu.au/nchecr/Downloads/05_ansurvrp.pdf
[Accessed 11 January 2007.]
5. Murphy G, Charlett A, Jordan LF, Osner N, Gill ON, Parry JV. HIV incidence appears constant in men who have sex with men despite widespread use of effective antiretroviral therapy. AIDS 2004; 18:265–272.
6. Wawer MJ, Gray RH, Sewankambo NK, Serwadda D, Li X, Laeyendecker O, et al
. Rates of HIV-1 transmission per coital act, by stage of HIV-1 infection, in Rakai, Uganda. J Infect Dis 2005; 191:1403–1409.
7. Kaufmann GR, Duncombe C, Zaunders J, Cunningham P, Cooper D. Primary HIV-1 infection: a review of clinical manifestations, immunologic and virologic changes. AIDS Patient Care STDs 1998; 12:759–767.
8. Weintrob AC, Giner J, Menezes P, Patrick E, Benjamin DK Jr, Lennox J, et al
. Infrequent diagnosis of primary human immunodeficiency virus infection: missed opportunities in acute care settings. Arch Intern Med 2003; 163:2097–2100.
9. Schacker T, Collier AC, Hughes J, Shea T, Corey L. Clinical and epidemiologic features of primary HIV infection. Ann Intern Med 1996; 125:257–264.
10. Jin FY, Prestage G, Law MG, Kippax S, Van de Ven P, Rawsthorne P. Predictors of recent HIV testing in homosexual men in Australia. HIV Med 2002; 3:271–276.
11. Burchell AN, Calzavara L, Ramuscak N, Myers T, Major C, Rachlis A, et al
. Symptomatic primary HIV infection or risk experiences? Circumstances surrounding HIV testing and diagnosis among recent seroconverters. Int J STD AIDS 2003; 14:601–608.
12. Janssen RS, Satten GA, Stramer SL, Rawal BD, O'Brien TR, Weiblen BJ, et al.New testing strategy to detect early HIV-1 infection for use in incidence estimates and for clinical and prevention purposes
:42–48. [Erratum in: JAMA
13. Kothe D, Byers RH, Caudill SP, Satten GA, Janssen RS, Hannon WH, et al
. Performance characteristics of a new less sensitive HIV-1 enzyme immunoassay for use in estimating HIV seroincidence. J Acquir Immune Defic Syndr 2003; 33:625–634.
14. Dukers NH, Spaargaren J, Geskus RB, Beijnen J, Coutinho RA, Fennema HS. HIV incidence on the increase among homosexual men attending an Amsterdam sexually transmitted disease clinic: using a novel approach for detecting recent infections. AIDS 2002; 16:F19–F24 [Erratum in: AIDS
15. Giuliani M, Carlo A, Palamara G, Dorrucci M, Latini A, Prignano G, et al
. Increased HIV incidence among men who have sex with men in Rome. AIDS 2005; 19:1429–1431.
16. Elford J. Changing patterns of sexual behaviour in the era of highly active antiretroviral therapy. Curr Opin Infect Dis 2006; 19:26–32.
17. Clements MS, Prestage G, Grulich A, Van de Ven P, Kippax S, Law MG. Modeling trends in HIV incidence among homosexual men in Australia 1995–2006. J Acquir Immune Defic Syndr 2004; 35:401–406.
18. Norton J, Elford J, Sherr L, Miller R, Johnson MA. Repeat HIV testers at a London same-day testing clinic. AIDS 1997; 11:773–781.
19. UK Collaborative Group on Monitoring the Transmission of HIV Drug Resistance. Time trends in primary resistance to HIV drugs in the United Kingdom: multicentre observational study
20. Sakarovitch C, Rouet F, Murphy G, Minga AK, Alioum A, Dabis F, et al
. Do tests devised to detect recent HIV-1 infection provide reliable estimates of incidence in Africa? J Acquir Immune Defic Syndr 2007; 45:115–122.
21. Reed C, Branson B, Janssen R. Interpreting STARHS results for individuals vs estimation of HIV incidence. Presented at the Eleventh Retroconference
, San Francisco, 2004 [abstract P854b].
22. Schwarcz S, Weinstock H, Louie BT, Kellogg TA, Douglas J, LaLota M, et al
. Characteristics of persons with recently acquired HIV infection: application of the serologic testing algorithm for recent HIV seroconversion in 10 US cities. J Acquir Immune Defic Syndr 2007; 44:112–115.
23. Kellogg TA, Loeb L, Dilley J, Adler B, Louie BT, McFarland W. Comparison of three methods to measure HIV incidence among persons seeking voluntary, anonymous counseling and testing. J Acquir Immune Defic Syndr 2005; 39:112–120.
25. Taylor MM, Hawkins K, Gonzalez A, Buchacz K, Aynalem G, Smith LV, et al
. Use of the serologic testing algorithm for recent HIV seroconversion (STARHS) to identify recently acquired HIV infections in men with early syphilis in Los Angeles County. J Acquir Immune Defic Syndr 2005; 38:505–508.
26. Suligoi B, Galli C, Massi M, Di Sora F, Sciandra M, Pezzotti P, et al
. Precision and accuracy of a procedure for detecting recent human immunodeficiency virus infections by calculating the antibody avidity index by an automated immunoassay-based method. J Clin Microbiol 2002; 40:4015–4020.
27. Parekh BS, Kennedy MS, Dobbs T, Pau CP, Byers R, Green T, et al
. Quantitative detection of increasing HIV type 1 antibodies after seroconversion: a simple assay for detecting recent HIV infection and estimating incidence. AIDS Res Hum Retroviruses 2002; 18:295–307.