HIV testing and the HIV epidemic –vitally important to prevent HIV becoming endemic

Editor’s notes: Epidemics refer to situations where the number of infections rises (and subsequently falls) more quickly than might be expected compared to a disease that is endemic.  Endemic implies a stable situation, with natural fluctuations in the number of cases.  Medley and Vassal have written a provocative article in Science that considers how differently individuals, communities and society react to epidemic rather than endemic diseases.  They choose to call HIV in 2017 endemic, which carries a serious risk. As the authors state, “The contained public response, and the concurrent shift of responsibility to individuals to protect themselves from risk, means that endemic disease embeds itself further, as those at risk are often the very same people who do not have the private resources to avoid risk or access treatment.”  There are in fact multiple separate epidemics of HIV in different regions and in different populations.  Some are rising and some are falling. The latest UNAIDS’ report emphasizes the heterogeneity of HIV infections in the world.  New HIV infections have fallen by 29% since 2010 in East and Southern Africa, the region with the highest rates.  On the other hand, new HIV infections have risen by an alarming 60% in Eastern Europe and Central Asia over the same period, albeit from a much lower baseline.  There is widespread political consensus to pursue the UN agenda endorsed at the High Level Meeting on Ending AIDS in New York last year.  Let’s not throw in the towel too soon!

HIV testing services remain central to the HIV strategy and, as usual, this month there are several important papers on aspects of HIV testing, many of which illustrate challenges that need to be overcome.

There are several reasons to encourage people living with HIV to know their status.  First and foremost, we know that the earlier treatment is started in the course of HIV, the better the outlook for the individual.  People who start treatment become much less likely to transmit HIV infection to sexual partners. People who know their HIV status are also able to make informed decisions about their lives and their partnerships.  A study this month by Escudero et al. from New York City used agent-based modelling to understand the testing and care continuum for people who inject drugs. Their results remind us of the key role of HIV testing.  They estimated that 53% of the HIV transmission events from people who inject drugs arose from people who did not know their status, and a further 37% from people who had not been started ART.  In other words, they estimate that only 10-11% of infections from people who inject drugs could be prevented by improving quality of care for people on treatment.  The need to find effective ways to encourage people at risk to know their status and start treatment is stark.

Guanzhou is one of the largest cities in China, with a high population of migrants both national and international.  It is among the most prosperous regions of Guangdong province and has the highest rates of HIV.   Chen et al. added some HIV testing related questions to a wider population based health survey in two districts and showed that approximately a quarter of adults had previously been tested for HIV.  HIV testing was almost all provided through free government facilities or blood transfusion centres.  Despite early steps to make HIV self-testing more available, none of the 666 participants who answered the relevant questions in the survey had used a self-test.  Distance from an HIV testing site was a key determinant of the likelihood of getting tested.  It was not clear that people who might be at higher risk were more likely to be tested, although the numbers and sampling focused on the general population rather than people at special risk.

Wang et al. explored the different HIV test kits used in the first line screening in Xi’an.  In line with Chinese guidelines, but not in line with WHO guidance on HIV testing algorithms for low prevalence settings, they used third- or fourth-generation rapid tests and repeated the positive tests.  WHO’s algorithm for low prevalence settings includes three different rapid tests based on different antigens.   Among 665 people found to be positive on rapid tests, only 559 were confirmed to be HIV-positive by Western blotting.  Subsequent follow up with additional Western blots showed that two of the individuals in whom the first Western blot was indeterminate were seroconverting but the other 104 were HIV-negative and had had false-positive results on the original rapid tests.  False positives were more likely with the fourth-generation test (22% of positive tests) compared to the various third-generation tests used (9-11% of positive tests).  Fourth-generation assays are known to be more sensitive, detecting people with HIV around a week or two earlier in the window period than third-generation assays.  However, the authors point out that in low prevalence settings like Xi’an, the known lack of specificity of fourth-generation assays means that they may not provide sufficient advantages to be used as the first line test.  Overall, the paper emphasizes the importance of using clearly defined algorithms.  The WHO algorithms no longer use Western blots, but do recommend using multiple tests based on different antigens for testing people at low risk of infection, and at least two different tests with different antigens for testing people at high risk of infection.  Everyone should have additional confirmatory tests done prior to starting ART.

Harbertson et al. also focused on the accuracy of rapid diagnostic tests.  They screened samples from 459 military personnel in seven African countries who had reported that they were HIV-positive.  Using the WHO algorithm, they compared the results of quality assured HIV testing to the self-reported HIV status of the participants.  They found that, in different country surveys, between three and 91% of people who said that they were living with HIV were in fact HIV-negative.  The authors point out that several studies have demonstrated the importance of following the WHO guidance, and that the positive predictive value of a test (or algorithm) will always fall as the overall prevalence falls.  They discuss possible limitations such as misunderstanding the question or the terminology used, but discount these possibilities as causing many of the false-positive reports, particularly given the highly variable results across different countries.  There was a strong association between the likelihood of a false positive report and lower education level. People whose understanding of HIV was less good were also more likely to report themselves to be positive falsely.  Overall, the authors assume that quality of testing services needs to be an important priority, while not discounting the challenges of using self-reports to collect information about HIV status.

As more and more people chose to know their HIV status, it may be possible to use routine data from the health service to track the epidemiology of HIV, rather than to use special surveys. Traditionally surveys of antenatal mothers have been used to monitor trends in the HIV epidemic over time.  With the widespread adoption of routine testing for mothers, a large proportion of women have an HIV test.  However, the assays used vary.  For surveillance purposes, samples are often stored and transported as dried blood spots and assays are run in batches using automated ELISA technology.  Routine testing (as discussed above) is often done using an algorithm based on a number of different rapid tests.   Pereira et al. have explored the differences between these approaches among almost 40 000 Brazilian mothers who participated in the antenatal surveillance exercise.  They interviewed mothers and linked their routine ANC results to the surveillance database.  Overall the prevalence of HIV among expectant mothers in Brazil was similar whichever approach was used (0.36% or 0.38%).  However, there were interesting differences.  The performance accuracy in those found positive in the surveillance exercise (which was taken as the gold standard) was only 84% overall and varied between regions from 43% to 100%.  So these 14 false negative results among the 88 individuals who were truly positive were compensated for in the overall prevalence estimates by a similar number (18) of false positive results among around 30 000 individuals who were truly negative. This highlights the challenges of providing accurate results to people in low prevalence settings. The 13% of mothers who slipped through the routine services and were not tested or refused to be tested were significantly more likely to be HIV-positive (0.56%), reinforcing the potential biases involved.  Finding 90% of people living with HIV will require considerable attention to the detail and the quality of HIV testing services.

Adolescents are often a population left behind, and regular reports show that adolescents living with HIV are less likely to know their status or to be on treatment or virally supressed.  Simms et al. used provider initiated testing and counselling (PITC) in primary care clinics in Harare, Zimbabwe.  For two years, the research team supported the routine offer of HIV testing to all six to 15 year olds presenting to seven clinics in a well-defined area of Harare.  The authors then conducted a population-based survey to find out how many eight to 17 year olds (who had had two years of exposure to the intervention) were aware of their status. 141 (2.6%) were living with HIV and more than one-third of these were undiagnosed.  Some had rarely been to the clinic, and others had been taken to the clinic by a guardian who was unable to consent to HIV testing on behalf of the child or the child’s parents.  Others had slipped through the PITC net, possibly because, as Lightfoot et al. in an accompanying comment suggest, providers still find it hard to offer HIV tests to everyone, as they assume that people living with HIV will not appear healthy.  This fits with the researchers’ findings that adolescents living with HIV who were currently healthy, had no skin or other problems and had parents who were alive were less likely to be diagnosed.  Both papers suggest that community based testing is needed to find adolescents. However, this also raises challenges in settings with lower prevalence than the high-density suburbs of Harare chosen for this project.  As prevalence falls lower than the 2.6% observed, a huge testing effort is needed, with attendant costs, but also (as explored above) with the risks of inaccurate results and of the very people that we want to find most, not being around for testing at the right moment. 

 

When an emerging disease becomes endemic.

Medley GF, Vassall A. Science. 2017 Jul 14;357(6347):156-158. doi: 10.1126/science.aam8333.

Epidemics, such as HIV in the early 1980s and Ebola in 2014, inspire decisive government investment and action, and individual and societal concern, sometimes bordering on panic. By contrast, endemic diseases, such as HIV in 2017 and tuberculosis, struggle to maintain the same attention. For many, the paradox is that endemic disease, in its totality, continues to impose a far higher public health burden than epidemic disease. Overall, the swift political response to epidemics has resulted in success. It has proven possible to eradicate epidemic diseases, often without the availability of vaccines and other biomedical technologies. In recent times, only HIV has made the transition from epidemic to endemic, but diseases that have existed for centuries continue to cause most of the infectious disease burden.

Abstract access

 

The risk of HIV transmission at each step of the HIV care continuum among people who inject drugs: a modeling study.

Escudero DJ, Lurie MN, Mayer KH, King M, Galea S, Friedman SR, Marshall BL. BMC Public Health. 2017 Jul 25;17(1):614. doi: 10.1186/s12889-017-4528-9.

Background: People who inject drugs (PWID) are at continued risk for HIV in the U.S., and experience disparities across the HIV care continuum compared to other high-risk groups. Estimates of the risk of HIV transmission at each stage of the care continuum may assist in identifying public health priorities for averting incident infections among PWID, in addition to transmissions to sexual partners of PWID.

Methods: We created an agent-based model simulating HIV transmission and the HIV care continuum for PWID in New York City (NYC) in 2012. To account for sexual transmission arising from PWID to non-PWID, the simulation included the entire adult NYC population. Using surveillance data and estimates from the National HIV Behavioral Surveillance system, we simulated a dynamic sexual and injecting network. We estimated the proportion of HIV transmission events attributable to PWID in the following categories, those: without an HIV diagnosis ('Undiagnosed'); diagnosed but not on antiretroviral therapy (ART) ('Diagnosed - not on ART'); those who initiated ART but were not virally suppressed ('Unsuppressed'); and, those who achieved viral suppression ('Suppressed').

Results: We estimated HIV incidence among PWID to be 113 per 100 000 person-years in 2012, with an overall incidence rate for the entire adult NYC population of 33 per 100 000 person-years. Despite accounting for only 33% of the HIV-infected PWID population, the Undiagnosed were associated with 52.6% (95% simulation interval [95% SI]: 47.1-57.0%) of total transmission events. The Diagnosed - not on ART population contributed the second-largest proportion of HIV transmissions, with 36.6% (95% SI: 32.2-41.5%). The Unsuppressed population contributed 8.7% (95% SI: 5.6-11.8%), and Suppressed 2.1% (95% SI: 1.1-3.9%), relatively little of overall transmission.

Conclusion: Among PWID in NYC, more than half (53%) of transmissions were from those who were unaware of their infection status and more than 36% were due to PWID who knew their status, but were not on treatment. Our results indicate the importance of early diagnosis and interventions to engage diagnosed PWID on treatment to further suppress population-level HIV transmission. Future HIV prevention research should focus on the elimination of identified and potential barriers to the testing, diagnosis, and retention of PWID on HIV treatment.

Abstract  Full-text [free] access

 

Is there a relationship between geographic distance and uptake of HIV testing services? A representative population-based study of Chinese adults in Guangzhou, China.

Chen W, Zhou F, Hall BJ, Tucker JD, Latkin C, Renzaho AMN, Ling L. PLoS One. 2017 Jul 20;12(7):e0180801. doi: 10.1371/journal.pone.0180801. eCollection 2017.

Achieving high coverage of HIV testing services is critical in many health systems, especially where HIV testing services remain centralized and inconvenient for many. As a result, planning the optimal spatial distribution of HIV testing sites is increasingly important. We aimed to assess the relationship between geographic distance and uptake of HIV testing services among the general population in Guangzhou, China. Utilizing spatial epidemiological methods and stratified household random sampling, we studied 666 adults aged 18-59. Computer-assisted interviews assessed self-reported HIV testing history. Spatial scan statistic assessed the clustering of participants who have ever been tested for HIV, and two-level logistic regression models assessed the association between uptake of HIV testing and the mean driving distance from the participant's residence to all HIV testing sites in the research sites. The percentage of participants who have ever been tested for HIV was 25.2% (168/666, 95%CI: 21.9%, 28.5%), and the majority (82.7%) of participants tested for HIV in Centres for Disease Control and Prevention, public hospitals or STIs clinics. None reported using self-testing. Spatial clustering analyses found a hotspot included 48 participants who have ever been tested for HIV and 25.8 expected cases (Rate Ratio = 1.86, P = 0.002). Adjusted two-level logistic regression found an inverse relationship between geographic distance (kilometers) and ever being tested for HIV (aOR = 0.90, 95%CI: 0.84, 0.96). Married or cohabiting participants (aOR = 2.14, 95%CI: 1.09, 4.20) and those with greater social support (aOR = 1.04, 95%CI: 1.01, 1.07) were more likely to be tested for HIV. Our findings underscore the importance of considering the geographical distribution of HIV testing sites to increase testing. In addition, expanding HIV testing coverage by introducing non-facility based HIV testing services and self-testing might be useful to achieve the goal that 90% of people living with HIV knowing their HIV status by the year 2020.

Abstract  Full-text [free] access

 

The characteristics of screening and confirmatory test results for HIV in Xi'an, China.

Wang L, Zhou KH, Zhao HP, Wang JH, Zheng HC, Yu Y, Chen W. PLoS One. 2017 Jul 7;12(7):e0180071. doi: 10.1371/journal.pone.0180071. eCollection 2017.

Objectives: Individuals with recent or acute HIV infection are more infectious than those with established infection. Our objective was to analyze the characteristics of detection among HIV infections in Xi'an.

Methods: A 4th-generation kit (Architect HIV Ag/Ab Combo) and three 3rd-generationEIA kits (WanTai, XinChuang and Livzon) were used for HIV screening. Overall, 665 individuals were identified as positive and were tested by western blotting (WB). The characteristics of the screening and confirmatory tests were analyzed, including the band patterns, the early detection performance and the false-positive rates.

Results: In total, 561 of the 665 patients were confirmed as having HIV-1 infection, and no HIV-2 specific band was observed. Among these 561 WB-positive cases, reactivity to greater than or equal to 9 antigens was the most commonly observed pattern (83.18%), and the absence of reactivity to p17, p31 and gp41 was detected in 6.44%, 5.9% and 2.86% of the cases, respectively. Two cases were positive by the 4th-generation assay but negative by the 3rd-generation assay for HIV screening and had seroconversion. The false-positive rate of the Architect HIV Ag/Ab Combo (22.01%) was significantly higher than those of WanTai (9.88%), XinChuang (10.87%) and Livzon (8.93%), p<0.05

Conclusion: HIV infection in Xi'an is mainly caused by HIV-1, and individuals are rarely identified at the early phase. Although the false-positive rate of the 4th-generation assay was higher than that of the 3rd-generation assay, it is still recommended for use as the initial HIV screening test for high-risk individuals. In Xi'an, a 3rd-generation assay for screening could be considered.

Abstract  Full-text [free] access

 

Self-reported HIV-positive status but subsequent HIV-negative test result using rapid diagnostic testing algorithms among seven sub-Saharan African military populations.

Harbertson J, Hale BR, Tran BR, Thomas AG, Grillo M, Jacobs MB, McAnany J, Shaffer RA. PLoS One. 2017 Jul 7;12(7):e0180796. doi: 10.1371/journal.pone.0180796. eCollection 2017.

HIV rapid diagnostic tests (RDTs) combined in an algorithm are the current standard for HIV diagnosis in many sub-Saharan African countries, and extensive laboratory testing has confirmed HIV RDTs have excellent sensitivity and specificity. However, false-positive RDT algorithm results have been reported due to a variety of factors, such as suboptimal quality assurance procedures and inaccurate interpretation of results. We conducted HIV serosurveys in seven sub-Saharan African military populations and recorded the frequency of personnel self-reporting HIV positivity, but subsequently testing HIV-negative during the serosurvey. The frequency of individuals who reported they were HIV-positive but subsequently tested HIV-negative using RDT algorithms ranged from 3.3 to 91.1%, suggesting significant rates of prior false-positive HIV RDT algorithm results, which should be confirmed using biological testing across time in future studies. Simple measures could substantially reduce false-positive results, such as greater adherence to quality assurance guidelines and prevalence-specific HIV testing algorithms as described in the World Health Organization's HIV testing guidelines. Other measures to improve RDT algorithm specificity include classifying individuals with weakly positive test lines as HIV indeterminate and retesting. While expansion of HIV testing in resource-limited countries is critical to identifying HIV-infected individuals for appropriate care and treatment, careful attention to potential causes of false HIV-positive results are needed to prevent the significant medical, psychological, and fiscal costs resulting from individuals receiving a false-positive HIV diagnosis.

Abstract  Full-text [free] access

 

Transitioning from antenatal surveillance surveys to routine HIV testing: a turning point in the mother-to-child transmission prevention programme for HIV surveillance in Brazil.

Pereira GFM, Sabidó M, Caruso A, Benzaken AS. BMC Infect Dis. 2017 Jul 5;17(1):469. doi: 10.1186/s12879-017-2540-4.

Background: In Brazil, due to the rapid increase in programmes for the prevention of mother-to-child transmission (PMTCT), routine programme data are widely available. The objective of this study was to assess the utility of programmatic data to replace HIV surveillance based on the antenatal care (ANC) surveillance survey (SS).

Methods: We analysed ANC SS data from 219 maternity service clinics. PMTCT variables were extracted from the ANC SS data collection form, which allowed us to capture and compare the ANC SS data and PMTCT HIV test results for each pregnant woman who completed the ANC SS. Both the PMTCT programme and the ANC SS tested for HIV using sequential ELISA and western blot for confirmation. We assessed the completeness (% missing) of the PMTC data included in the ANC SS.

Results: Of the 36 713 pregnant women who had ANC SS HIV tests performed, 30 588 also underwent PMTCT HIV testing. The HIV prevalence rate from routine PMTCT testing was 0.36%, compared to 0.38% from the ANC SS testing (relative difference -0.05%; absolute difference -0.02%). The relative difference in prevalence rates between pregnant women in northern Brazil and pregnant women central-west Brazil was -0.98 and 0.66, respectively. Of the 29 856 women who had HIV test results from both the PMTCT and ANC SS, the positive percent agreement of the PMTCT versus the surveillance test was 84.1% (95% confidence interval [CI]: 74.8-91.0), and the negative percent agreement was 99.9% (95% CI: 99.9-100.0). The PMTCT HIV testing uptake was 86.4%. The ANC SS HIV prevalence was 0.33% among PMTCT non-refusers and 0.59% among refusers, with a percent bias of -10.80% and a differential prevalence ratio of 0.56. Syphilis and HIV testing results were complete in 98% and 97.6% of PMTCT reports, respectively. The reported HIV status for the women at clinic entry was missing.

Conclusion: Although there were consistent HIV prevalence estimates from the PMTCT data and the ANC SS, the overall positive percent agreement of 84.1% falls below the World Health Organization benchmark of 94.7%. Therefore, Brazil must continue to reinforce data collection practices and ensure the quality of recently introduced rapid HIV testing before replacing the PMTCT data with surveillance techniques. However, some regions with better results could be prioritized to pilot the use of PMTCT data for surveillance.

Abstract  Full-text [free] access

 

Community burden of undiagnosed HIV infection among adolescents in Zimbabwe following primary healthcare-based provider-initiated HIV testing and counselling: A cross-sectional survey.

Simms V, Dauya E, Dakshina S, Bandason T, McHugh G, Munyati S, Chonzi P, Kranzer K, Ncube G, Masimirembwa C, Thelingwani R, Apollo T, Hayes R, Weiss HA, Ferrand RA. PLoS Med. 2017 Jul 25;14(7):e1002360. doi: 10.1371/journal.pmed.1002360. eCollection 2017 Jul.

Background: Children living with HIV who are not diagnosed in infancy often remain undiagnosed until they present with advanced disease. Provider-initiated testing and counselling (PITC) in health facilities is recommended for high-HIV-prevalence settings, but it is unclear whether this approach is sufficient to achieve universal coverage of HIV testing. We aimed to investigate the change in community burden of undiagnosed HIV infection among older children and adolescents following implementation of PITC in Harare, Zimbabwe.

Methods and Findings: Over the course of 2 years (January 2013-January 2015), 7 primary health clinics (PHCs) in southwestern Harare implemented optimised, opt-out PITC for all attendees aged 6-15 years. In February 2015-December 2015, we conducted a representative cross-sectional survey of 8-17-year-olds living in the 7 communities served by the study PHCs, who would have had 2 years of exposure to PITC. Knowledge of HIV status was ascertained through a caregiver questionnaire, and anonymised HIV testing was carried out using oral mucosal transudate (OMT) tests. After 1 participant taking antiretroviral therapy was observed to have a false negative OMT result, from July 2015 urine samples were obtained from all participants providing OMTs and tested for antiretroviral drugs to confirm HIV status. Children who tested positive through PITC were identified from among survey participants using gender, birthdate, and location. Of 7146 children in 4251 eligible households, 5486 (76.8%) children in 3397 households agreed to participate in the survey, and 141 were HIV positive. HIV prevalence was 2.6% (95% CI 2.2%-3.1%), and over a third of participants with HIV were undiagnosed (37.7%; 95% CI 29.8%-46.2%). Similarly, among the subsample of 2643 (48.2%) participants with a urine test result, 34.7% of those living with HIV were undiagnosed (95% CI 23.5%-47.9%). Based on extrapolation from the survey sample to the community, we estimated that PITC over 2 years identified between 18% and 42% of previously undiagnosed children in the community. The main limitation is that prevalence of undiagnosed HIV was defined using a combination of 3 measures (OMT, self-report, and urine test), none of which were perfect.

Conclusions: Facility-based approaches are inadequate in achieving universal coverage of HIV testing among older children and adolescents. Alternative, community-based approaches are required to meet the Joint United Nations Programme on HIV/AIDS (UNAIDS) target of diagnosing 90% of those living with HIV by 2020 in this age group.

Abstract  Full-text [free] access

  • share