Article Text

Original article
Identifying recently acquired HIV infections among newly diagnosed men who have sex with men attending STI clinics in The Netherlands
  1. Jussi Sane1,2,
  2. Titia Heijman3,
  3. Boris Hogema4,
  4. Maarten Koot4,
  5. Maaike van Veen3,
  6. Hannelore Götz5,
  7. Johan Fennema3,
  8. Eline Op de Coul1
  1. 1Unit of Epidemiology and Surveillance, National Institute for Public Health and Environment (RIVM), Centre for Infectious Disease Control, The Netherlands
  2. 2European Programme for Intervention Epidemiology Training (EPIET), European Centre for Disease Prevention and Control (ECDC), Stockholm, Sweden
  3. 3Public Health Service of Amsterdam, Amsterdam, The Netherlands
  4. 4Virus Diagnostic Services, Sanquin Blood Supply, Amsterdam, The Netherlands
  5. 5Department of Infectious Disease Control, Public Health Service Rotterdam-Rijnmond, Rotterdam, The Netherlands
  1. Correspondence to Dr Eline LM Op de Coul, Epidemiology & Surveillance Unit, Centre for Infectious Disease Control, National Institute of Public Health and the Environment, PO Box 1, Bilthoven 3720 BA, The Netherlands; eline.op.de.coul{at}rivm.nl

Abstract

Objectives The current surveillance system in The Netherlands cannot differentiate recent HIV infections from established infections, which is crucial for estimating the HIV incidence; this information is needed for assessing trends of the HIV epidemic and the impact of prevention interventions. We determined the proportion of recent HIV infections (RI) and estimated HIV incidence using a recent infection testing algorithm (RITA) among men who have sex with men (MSM) newly diagnosed as having HIV attending sexually transmitted infection (STI) clinics.

Methods Plasma samples collected between 2009 and 2011 were tested for RI with the Architect HIV Ag/Ab Combo immunoassay. Data on viral load, CD4 count and previous HIV testing were incorporated into the RITA. HIV incidence and 95% CIs were estimated. Logistic regression was used to identify factors associated with RI.

Results Of the 251 samples tested for RI, 78/251 (31%) infections were determined as recent by the RITA. No significant change over time was observed. The estimated HIV incidence in this high-risk MSM population was 3.3 per 100 person-years (95% CI 2.5 to 4.1). The only factor associated with RI in the multivariable model was being tested for HIV ≥3 times in the past (aOR=7.4; 95% CI 2.0 to 27.8).

Conclusions The proportion of RIs was comparable to studies in similar settings in Europe. Implementation of the RITA for routine surveillance in The Netherlands to assess trends in RIs over time, to study the infections in other groups and to inform public health actions, is being planned.

  • HIV
  • HIV Testing
  • Surveillance
  • Epidemiology (General)
  • Gay Men

Statistics from Altmetric.com

Introduction

The current HIV surveillance system in The Netherlands largely relies on reporting new diagnoses of HIV and AIDS cases.1 In 2012, approximately 1200 newly diagnosed HIV infections were registered in The Netherlands. The number of newly diagnosed infections was highest among men who have sex with men (MSM) and has been gradually increasing in The Netherlands over the past decade.2 However, the existing surveillance cannot distinguish between an increase in HIV transmission and increased testing of previously undiagnosed infections. This system is slow to reflect potential changes in HIV transmission and is limited in showing the possible impact of public health interventions on HIV transmission.

In recent years, several serological assays that use a single laboratory specimen for identifying recent HIV infections have been developed. These assays allow the proportion of recent infections and incidence of HIV to be estimated in studies that are easier to conduct, less expensive, timely and require fewer resources.3 ,4 The testing algorithms, referred to as recent infection testing algorithms (RITAs), can be combined with additional laboratory, clinical or epidemiological data and have become increasingly widespread for monitoring recent HIV infections and incidence.1

Following recommendations from the European Centre for Disease Prevention and Control (ECDC) to strengthen the role of RITAs in surveillance in Europe, and to validate RITAs in The Netherlands in order to inform public health actions, we determined the proportion of recent HIV infections and provided an estimation of HIV incidence using a RITA among MSM newly diagnosed as having HIV who were tested at sexually transmitted infection (STI) clinics.

Methods

Study design and data collection

HIV testing in The Netherlands occurs mainly via general practitioners, hospitals or STI clinics.2 Blood specimens of HIV-positive tests were collected as a part of another study, to which MSM newly diagnosed as having HIV attending the STI clinics in Amsterdam and Rotterdam were invited to join between February 2009 and December 2011.5 All specimens were anonymised and study participants filled out a questionnaire on demographics, sexual risk behaviour and previous HIV tests. Furthermore, data on viral load and CD4 cell count were collected at the STI clinics. Data on concurrent diagnoses of other STIs and overall HIV testing data (including recorded dates of previous tests) were only available from the Amsterdam STI clinic.

Laboratory methods

To test for recent infections, the anti-HIV avidity index (AI) was measured in plasma with a fourth-generation commercial enzyme immunoassay, Architect HIV Ag/Ab Combo (Abbott Diagnostics, Wiesbaden, Germany) using a previously implemented protocol.6 The assay-specific period for recent infections is ≤6 months from seroconversion. Thus, the mean RITA duration (window period) was 180 days.

RITA and statistical methods

We applied a RITA based on the guidelines presented by the ECDC.1 Data on viral load, CD4 count and previous HIV testing were incorporated in the RITA to minimise false recent rate (FRR) (figure 1). Cases with AI ≤0.80 and classified as recent based on the avidity assay were reclassified as non-recent if CD4 count was <200 cells/µL and/or viral load was <400 copies/mL and/or they had a previous HIV-positive test predating the sample date by more than the mean RITA duration.

Figure 1

Recent infection testing algorithm (RITA) to determine the number of recently infected individuals (n=251).

HIV incidence and 95% CIs were estimated using a method based on the Janssen formula.1 ,4 ,7 The number of HIV-negative samples for the calculations was adjusted based on the proportion of HIV-positive samples eligible for testing for recent infection during the study period.4 The χ2 test was used for comparison of proportions and time trends. Logistic regression was used to identify factors associated with RI. These analyses were performed only on cases from Amsterdam. A p value of 0.2 was used as the screening criterion for selection of variables for multivariate analyses. p Values of <0.05 were considered statistically significant. STATA software (V.12) was used for the analyses.

Results

In total, 251 plasma samples were obtained for recent infection testing (figure 2). The median age was 35 (range 17–71) and 113 cases (46% of cases with known ethnicity; n=246) were of Dutch origin. Out of these samples, 200 originated from Amsterdam and 51 from Rotterdam. At the STI clinic in Amsterdam, 7417 MSM tested negative and 393 tested positive during the study period. Thus, 200/393 (51%) MSM from Amsterdam newly diagnosed as having HIV were included in the study (figure 2). The remaining non-responders were not significantly different from the responders in age (p=0.2) or ethnicity distribution (p=0.18). From Rotterdam, 51/89 (57%) MSM diagnosed as having HIV were included, but other denominator data at the individual level with basic demographic information was not available.

Figure 2

A summary of the study population included in the testing of recent HIV infection, February 2009–December 2011, The Netherlands. *Specimens and questionnaires were collected as a part of another study5 to which men who have sex with men (MSM) newly diagnosed as having HIV attending sexually transmitted infection (STI) clinics in Amsterdam and Rotterdam were invited between February 2009 and December 2011. Only the participants from that study, with a plasma sample of high enough quantity, could be used for recent infection testing. **Analyses performed only for data from Amsterdam STI clinic due to availability of appropriate data on denominators and questionnaires. RITA, recent infection testing algorithm.

Of the 251 samples, 83 were classified as recent based on the AI. All cases classified as recent had complete data on CD4 count; information for viral load was missing for four cases. Five cases were reclassified as non-recent based on low CD4 count (n=2), viral load (n=2) and history of HIV infection (n=1); thus, 78/251 (31%: 95% CI 25.6 to 37.0) infections were determined as recent by the RITA.

Proportions of recent infections in 2009, 2010 and 2011 were 32%, 28% and 33%, respectively, and no significant change over time was observed. The estimated combined incidence over the study period in Amsterdam was 3.3 per 100 person-years (95% CI 2.5 to 4.1).

The only variable significantly associated with the recent infections was the previous HIV testing frequency (aOR=7.4; 95% CI 2.0 to 27.8) (table 1). With regard to concurrent STIs, a concurrent STI was reported among 30/62 (48.4%) cases with recent infection, whereas 52/136 (38.2%) cases with established infection had a concurrent STI (p=0.18).

Table 1

Univariate and multivariate analysis of factors associated with recent infections among MSM newly diagnosed as having HIV (n=200) in Amsterdam, The Netherlands, 2009–2011

Discussion

We used the recently developed RITA protocol1 for the first time in The Netherlands to determine the proportion of recent infections and provide incidence estimations among MSM attending STI clinics. We found that approximately one-third of MSM newly diagnosed as having HIV during the study period were recently infected, without a clear trend over time, indicating ongoing transmission in this population. The proportion of recent infections was comparable to or somewhat lower than similar studies in Europe, which were primarily conducted in STI clinics, and the estimated incidence in the specific study population was within the range previously reported for MSM.1 Earlier studies from The Netherlands reported similar incidence estimates, 3.2% in 1991–20018 and 3.7% in 1999–20059 among MSM attending STI centres, but a different laboratory method was used and data on CD4 or viral load was lacking to optimise the algorithm.

The ECDC recently published an epidemiological framework for using RITAs in Europe to estimate proportions of recent infections and HIV incidence given the availability of proper data.1 Our pilot followed the suggested steps targeting MSM attending STI clinics, a group that has been identified as being at high risk of HIV infection in several studies.1 ,4 The availability of supplementary laboratory data allowed us to minimise the FRR through a case-based approach. Additional studies on known long-term infected samples could provide further FRR estimates.

The comparison of our results to other studies must be made with caution. Various laboratory methods, subjected to different biases, study design and varying sample size complicate the comparison. Harmonised protocols are needed for appropriate comparison and comprehensive evaluation of different laboratory assays is crucial. Such a project, the Consortium for the Evaluation of the Performance of HIV Incidence Assays (CEPHIA), is currently ongoing and will provide valuable information allowing for better comparison of the results.1 Previous studies have, however, shown that avidity assays are accurate in differentiating recent from established HIV infections, and are less affected by HIV-1 subtype variation then Detuned/BED immune assays10 supporting the use of avidity assays for RITAs. Subtype variation should be considered in our study as the study population included persons from Africa (n=13, 5.3%) where subtypes other than B are common.

The association between recent infection and previous HIV testing frequency is not surprising as the probability of detecting recent infection increases if testing is performed more regularly. Our study did not find any association between age and recent infections. Other studies11 ,12 have indicated that the risk for recent infection decreased with age, although these studies, with larger sample sizes, also included other groups than MSM. A previous study on incidence among MSM attending the STI clinic in Amsterdam suggested that incidence was increasing in older men during the study period in 1999–2005.9 Our study had a considerably smaller sample size and this may explain the observed lack of association to age. The proportion of concurrent STIs was also similar to findings from MSM newly diagnosed as having HIV attending STI clinics in The Netherlands in 2011.2

The major limitation in our study was the potential test-seeking bias; MSM seeking care at STI clinics are likely to do so because of risk behaviour and STI/HIV-related symptoms and thus, the number of recent infections and HIV-incidence are probably biased upwards. In addition, individuals who took part in the study may have different behavioural characteristics than non-responders, further complicating the representatives of the sample. The incidence estimated from our study was considerably higher than estimates from an ongoing cohort study on HIV infections among MSM (between 0.7 and 2.0 per 100 person-years during 2009–2012).2 Men willing to participate in long-term follow-up are probably more aware of the risks; additionally, their age distribution is different to our study, as only men aged ≤30 years can enter the cohort. The results from our study may, however, be considered to reflect similar testing settings in MSM attending STI clinics and individuals with the same risks for HIV infection.

This study confirmed that HIV transmission is ongoing among MSM attending Dutch STI clinics, however, no change in recent infections during the study period was observed. To further assess trends of recent infections and to study recent infections in other groups, implementation of RITAs for routine surveillance in The Netherlands, designed to take the FRR into account, is being planned. Returning RITA results to individual patients is not being considered in The Netherlands for the time being.

Key messages

  • Recent HIV infections were determined among men who have sex with men (MSM) attending STI clinics in The Netherlands using a recent infection testing algorithm (RITA).

  • Approximately one-third of MSM newly diagnosed as having HIV were recently infected, indicating ongoing HIV transmission.

  • Routine surveillance of recent HIV infection in The Netherlands is being planned.

References

Supplementary materials

  • Abstract in Dutch

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Handling editor Jackie A Cassell

  • Acknowledgements We thank Barbara Suligoi from the Istituto Superiore di Sanità in Rome for sharing the laboratory protocol and assistance in the implementation of the assay. We would like to thank Birgit van Benthem and Marianne van der Sande (RIVM) for critically reading the manuscript. Jannie van der Helm, Martijn van Rooijen, Maria Prins and Arjen Speksnijder from the Public Health Service Amsterdam are thanked for their collaboration in providing Amsterdam data and samples.

  • Contributors JS analysed the data and wrote the manuscript, TH revised the manuscript and coordinated the data collection, BH and MK performed the laboratory experiments and revised the manuscript, MvV and HG contributed with data, JF designed the study and, EOdC designed and supervised the study and wrote the manuscript. All authors have commented on and approved the final paper.

  • Funding Study was funded by the National Institute for Public Health and Environment (RIVM).

  • Competing interests The fellowship of JS was funded by the European Programme for Intervention Epidemiology Training, European Centre for Disease Prevention and Control.

  • Ethics approval The delay study was approved by the Medical Ethics Committee of the Free University Amsterdam. Patients gave consent for anonymous use of left over samples from that study for scientific research purposes.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.