Application of the BED capture enzyme immunoassay for HIV incidence estimation among female sex workers in Kaiyuan City, China, 2006–2007

Article Outline

• Summary
• 1. Introduction
• 2. Methods
  • 2.1. Study design
  • 2.2. HIV antibody and BED-CEIA testing
  • 2.3. Sequence analysis
  • 2.4. Data management and statistical analysis
• 3. Results
  • 3.1. Baseline characteristics of FSW participants
  • 3.2. Incidence of HIV by longitudinal cohort study
  • 3.3. HIV incidence as measured by BED-CEIA assay
  • 3.4. Detection of seroconverted specimens by BED-CEIA testing
  • 3.5. Distribution of HIV subtypes among FSW participants
• 4. Discussion
• Acknowledgements
• References
• Copyright

Summary 

Objective

To estimate HIV incidence among female sex workers (FSWs) by serial cross-sectional surveys and IgG-capture BED-enzyme immunoassay (BED-CEIA).

Methods

We conducted three cross-sectional surveys, 6 months apart, among all consenting FSWs in Kaiyuan City, China. HIV antibody-positive samples were also tested by BED-CEIA.

Results

Among 1412 unique participants, 475 tested HIV-negative and attended >1 survey (longitudinal cohort). Compared to 786 HIV-negative FSWs who only participated once, the longitudinal cohort reported more illicit drug use (10.9% vs. 7.4%, p=0.03), injected drugs more often in the previous 3 months (8.8% vs. 5.3%, p=0.02), and had more positive urine opiate tests (13.7% vs. 8.9%, p=0.008). Four participants in the longitudinal cohort seroconverted over the year, with an overall incidence of 1.1/100 person-years (95% confidence interval (CI) 0.3–2.8). Crude BED-CEIA incidence was 3.4/100 person-years (95% CI 2.3–4.4) with adjusted rates similar to the cohort incidence: McDougal, 1.5/100 person-years (95% CI 1.0–2.0); Hargrove, 1.6/100 person-years (95% CI 1.1–2.1). The BED-CEIA false-positive rate was 4.4% (10/229) among samples from FSWs known to be infected ≥365 days.

Conclusions

Although limited by power, this study provides additional data towards validating BED-CEIA in China. If confirmed by other studies, BED-CEIA will be a useful tool to estimate HIV incidence rates and trends.

Keywords: HIV, Incidence, IgG capture BED-enzyme immunoassay (BED-CEIA), Prospective cohort

Back to Article Outline

1. Introduction 

Understanding the incidence of HIV infection in a population is critically important to understand the dynamics of HIV transmission. Prospective cohort studies are often used to determine disease incidence but are difficult to conduct well because of the associated time, effort, and expense required. Biases can also be introduced if the cohort recruited is not representative of the larger population. An alternative method to estimate HIV incidence is the IgG-capture BED-enzyme immunoassay (BED-CEIA),¹ which can distinguish recent infection from longer-term infection by measuring the proportion of HIV-IgG to total IgG after HIV-1 seroconversion. With this assay, HIV incidence can be estimated using cross-sectional surveys.^[2], ^[3] The performance of BED-CEIA for recent HIV infection has been validated in North America and the Netherlands (subtype B),^[4], ^[5] Thailand (subtypes B and E),^[1], ^[6], ^[7] Zimbabwe (subtype C),⁸ and Kenya (subtypes A, D, and C).¹ Other validation studies in African countries, however, have found that BED-CEIA overestimated HIV incidence by 2–4-fold compared to prospective studies or models.^[9], ^[10], ^[11], ^[12] Accordingly, the Joint United Nations Programme on HIV/AIDS (UNAIDS) has warned that BED-CEIA-related misclassification may be particularly severe in areas with high HIV prevalence and that estimates may vary by HIV-1 subtype.¹¹ The US Centers for Disease Control and Prevention (CDC) recommend the two formulas of McDougal⁴ and Hargrove⁸ to correct the misclassification. However, the two adjusted formulas have only had limited validation using HIV subtypes A1, B, and C.^[12], ^[13]

In China, where the predominant HIV subtypes are circulating recombinant forms of subtypes B and C (CRF_BC),¹⁴ previous studies using BED-CEIA to estimate HIV incidence were cross-sectional and did not account for subtype.^[15], ^[16], ^[17] The objectives of this study were to determine HIV incidence among female sex workers (FSW) in Kaiyuan City, China, through a longitudinal cohort derived from serial cross-sectional surveys, and to compare this cohort study-derived incidence with the BED-CEIA estimated incidence, while accounting for HIV subtype.

Back to Article Outline

2. Methods 

2.1. Study design 

This study was conducted in Kaiyuan City, Yunnan Province, China as three serial cross-sectional surveys, 6 months apart, between March 2006 and April 2007. A general census design was adopted in each of the three cross-sectional surveys. All available community-based FSWs from local entertainment venues were recruited to the study site by outreach workers, with those meeting the following inclusion criteria enrolled: (1) worked in Kaiyuan entertainment venues; (2) women ≥16 years old; and (3) self-reporting receiving money for sex within the previous 3 months. Participants failing to meet any one of these inclusion criteria were excluded. For this analysis, the baseline characteristics of all initially HIV-negative participants attending more than one survey were compared against HIV-negative participants who attended only one survey.

After providing informed consent and HIV pre-testing counseling, participants were administered an anonymous questionnaire by trained interviewers. The questionnaire included questions about demography, sexual behavior, drug use experience, condom using behavior with male clients, and vaginal douching behavior.

2.2. HIV antibody and BED-CEIA testing 

Blood samples were collected and tested for HIV antibody by ELISA (Organon Teknika, Boxtel, Co., Ltd, the Netherlands). All positive samples were confirmed by Western blot (HIV Blot 2.2 WB; Genelabs Diagnostics, Singapore) and tested for HIV subtype. All Western blot-positive samples, including those previously testing positive, were also tested by BED-CEIA (Calypte Biomedical Corporation, Rockville, MD, USA). Specimens with initial ODn ≤1.2 were tested in triplicate to confirm their ODn values. If the median ODn value from all three tests was <0.8, the specimen was considered recently infected (≤155 days), otherwise, the specimen was classified as chronic infection.

2.3. Sequence analysis 

HIV RNA was extracted by QIAamp Viral RNA Mini Kit (Qiagen Inc., Hilden, Germany) according to the manufacturer's instructions, and resuspended with RNA diluent. The nucleotide sequences of 2.6kb gag-RT and C2V3 of env were determined from viral RNA.¹⁸ The nested-PCR product was purified with the QIAquick Gel Extraction Kit and sequenced with the ABI3100 DNA Sequencer (Applied Biosystems Inc.). Sequence fragments were linked by Contig Express program of Vector NTI Advance 10 (Invitrogen, USA). The sequences were aligned with previously reported HIV-1 strains of various subtypes from the Los Alamos database. Multiple alignments were performed by CLUSTAL W with minor manual adjustments. The Kimura 2-parameter method was used for the determination of the evolutionary distance. The reliability of the branching patterns was assessed by bootstrap analysis with 500 replicates. Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 3.1 (Molecular Evolutionary Genetics Analysis, Tempe, AZ, USA). Simplot version 3.2 was used to identify recombination strains. The bootstrap values were plotted for a window of 200bp moving in increments of 50bp along the alignment.

2.4. Data management and statistical analysis 

The crude BED-CEIA estimated HIV incidence was calculated using the US CDC recommended formula: I=100×[(365/w)×N^inc]/[N^neg+(365/w)×(N^inc/2)],¹³ in which w is the window period (155 days), N^inc is the number of recent HIV infections as determined by BED-CEIA, and N^neg is the total number of HIV-seronegative subjects. The 95% confidence intervals (CI) for estimated BED-CEIA incidence were calculated by: 95% CI=I±1.96(I/√N^inc).¹³ Subjects who tested BED-CEIA-positive but were documented by local CDC records (before the baseline survey) or previous study records (from the baseline survey) already to be HIV-positive at least 365 days before (about twice the BED-CEIA window period) were defined as false-positives (also referred to as misclassifications) and were excluded from the calculation of HIV incidence.¹⁹ The crude HIV incidence as determined by BED-CEIA was then adjusted using the US CDC recommended formulas of McDougal and Hargrove.¹³ Exact 95% CI were calculated for HIV incidence based on the Poisson distribution. Comparisons of high-risk behaviors between groups were performed by Chi-square test, independent t-test or Wilcoxon rank sum test. SAS 9.1 (SAS Institute Inc., Cary, NC, USA) was used for data analysis. The questionnaires and study protocol were approved by the institutional review boards of the China CDC and Yunnan CDC.

Back to Article Outline

3. Results 

3.1. Baseline characteristics of FSW participants 

The three cross-sectional surveys conducted included 737, 747, and 705 FSWs, respectively, and of these, there were 1412 unique participants. Among the participants, 151 initially tested HIV-positive, with 68 (45.0%) returning for a second or third survey. Among the remaining 1261 initially HIV-negative FSWs, 475 participated in more than one survey and were included in the longitudinal cohort (236 FSWs attended all three surveys). These were compared with the 786 HIV-negative FSWs who only participated in one survey (Table 1). FSWs who returned for more than one visit were significantly older at baseline (26.2 vs. 24.4 years, p<0.001), older when they first engaged in commercial sex (23.5 vs. 22.5 years, p=0.002), self-reported more use of illicit drugs (10.9% vs. 7.4%, p=0.03), injected drugs more often in the previous 3 months (8.8% vs. 5.3%, p=0.02), and had more positive urine opiate tests in the study (13.7% vs. 8.9%, p=0.008).

Table 1. Baseline high-risk behaviors comparison between female sex workers initially testing HIV-negative who were included or not included in the longitudinal cohort in Kaiyuan City, China, 2006–2007.

Variable	Included in longitudinal cohort- (n=475)	Not included in longitudinal cohorta (n=786)	p-Value
Mean age (years)	26.2±6.8	24.4±6.1	<0.001
Han nationality (%)	319 (67.2)	537 (68.3)	0.67
Mean schooling years	7.2±3.2	7.1±3.3	0.49
Mean age at first intercourse (years)	18.4±2.3	18.2±2.2	0.062
Mean age at first commercial sex (years)	23.5±6.0	22.5±5.4	0.002
Self-reported drug use history (%)	52 (10.9)	58 (7.4)	0.030
Injected drugs in previous 3 months (%)	42 (8.8)	42 (5.3)	0.016
Positive urine opiate test (%)	65 (13.7)	70 (8.9)	0.008
Self-reports vaginal douching (%)	385 (81.1)	596 (75.8)	0.031
Median number of clients in previous week (IQR)	3 (2–5)	3 (2–7)	0.002
Consistent condoms with clients in previous week (%)	407 (85.7)	659 (83.8)	0.38
Condom using rate with the last client (%)	439 (92.4)	713 (90.7)	0.30

IQR, interquartile range. Variables are described as mean±standard deviation, median (IQR), or rate. Statistical comparisons between groups were carried out by t-test or Wilcoxon rank test for continuous variables and Chi-square test for categorical variables.

3.2. Incidence of HIV by longitudinal cohort study 

At the 6-month survey, two FSWs were found to have seroconverted in the longitudinal cohort. At the 12-month survey, two more FSWs seroconverted, one of whom seroconverted between the 6- and 12-month surveys and the other between the initial and 12-month surveys (she did not participate in the 6-month survey). The overall HIV incidence rate for the year was 1.1/100 person-years (95% CI 0.3–2.8; Table 2). The four HIV incident FSWs had a median age of 31.7 years. All had received less than 9 years of formal education. By self-report, only one of the four was an injection drug user (IDU). The four HIV incident FSWs reported a daily average of 2.5 clients (ranging from two to four clients) and a weekly average of 14.5 clients. One of the four FSWs had not used a condom with her most recent client and three of the four had failed to use a condom with at least one client in the previous week.

Table 2. BED-CEIA and longitudinal cohort study derived incidence of HIV among female sex workers in Kaiyuan City, China, 2006–2007

BED-CEIA, IgG-capture BED-enzyme immunoassay; FSW, female sex workers; I_HIV+, HIV incidence; PY, person-years; CI, confidence interval.

3.3. HIV incidence as measured by BED-CEIA assay 

In the three cross-sectional surveys, all HIV antibody-positive specimens were tested by BED-CEIA, with 7/76, 15/89, and 16/92 of the HIV-positive cases testing BED-CEIA-positive (ODn ≤0.8). Of these, 1/7 (14.3%), 3/15 (20.0%), and 6/16 (37.5%) had a documented HIV-positive result at least 365 days before (either through records from the local CDC or in this study), giving an overall false-positive misclassification ratio of 26.3% (10/38; Table 2) among those testing positive for recent infection. Among the total 257 HIV-positive samples, 229 were known to have been infected for longer than 365 days. Of these, 16 tested BED-CEIA-positive for a false-positive misclassification rate of 4.4% (10/229). Of the nine FSWs taking antiretrovirals during the surveys, none tested BED-CEIA-positive.

The overall annualized crude incidence of HIV by BED-CEIA was 3.4/100 person-years (95% CI 2.3–4.4), which was more than three times the total incidence by cohort study (1.1/100 person-years, 95% CI 0.3–2.8; Table 2), but with overlapping 95% confidence intervals. The overall adjusted incidence of HIV (McDougal, 1.5/100 person-years (95% CI 1.0–2.0); Hargrove, 1.6/100 person-years (95% CI 1.1–2.1)) was similar to the total incidence by longitudinal cohort study (1.1/100 person-years (95% CI 0.3–2.8)), also with overlapping 95% confidence intervals (Table 2).

3.4. Detection of seroconverted specimens by BED-CEIA testing 

Among the four seroconverters in the study, the two who converted between the baseline and 6-month surveys (within 180 days) tested BED-CEIA-positive for recent infection at the 6-month survey but negative for recent infection at the 12-month survey. The participant who converted between the 6- and 12-month surveys tested BED-CEIA-positive for recent infection at the 12-month survey. The participant who converted between the baseline and 12-month survey tested BED-CEIA-negative for recent infection at the 12-month visit.

3.5. Distribution of HIV subtypes among FSW participants 

We were able to subtype 113 (72.9%) of the total 155 HIV-positive samples collected from unique FSWs. Of these, 74 (65.5%) were CRF_08BC; 16 (14.2%) were CRF_07BC; 12 (10.6%) were subtype C; 10 (8.9%) were unclassifiable recombinant strains, and one (0.9%) was CRF_01AE. None of the 10 unclassifiable recombinant strains were classified as false-positive recent infections by BED-CEIA. Of the samples from the four seroconverters, two were successfully subtyped (CRF_08BC and C). Among the 10 FSWs who had false-positive BED-CEIA results, we were able to identify the virus subtype for eight: five were CRF_08BC, two were CRF_07BC, and one was subtype C. This distribution was not significantly different from the overall distribution (p=0.84).

Back to Article Outline

4. Discussion 

The primary purpose of this study was to validate the BED-CEIA method to estimate HIV incidence in a cohort of FSWs in Kaiyuan City, Yunnan Province, China. We adopted a general census design among FSWs at the three cross-sectional surveys, recruiting about 90% of the local FSWs at each survey time point (according to the local CDC, there were about 800 FSWs working in Kaiyuan City), in an effort to decrease selection bias. By using 475 initially HIV-negative FSWs who participated in at least two of three cross-sectional surveys 6 months apart, we found the cohort-derived HIV incidence was 1.1/100 person-years (95% CI 0.3–2.8), which was similar to the McDougal-adjusted (1.5/100 person-years (95% CI 1.0–2.0)) and Hargrove-adjusted (1.6/100 person-years (95% CI 1.1–2.1)) HIV-incidences. Of note, although the crude BED-CEIA calculated annualized incidence (3.4/100 person-years (95% CI 2.3–4.4)) was more than three-fold higher than the cohort incidence, this difference was not statistically significant. HIV subtypes were also determined, but our numbers were too small to make any definitive correlations between the subtypes identified and BED-CEIA misclassification.

Many studies have already validated the relationship between BED-CEIA estimated HIV incidence and cohort study derived HIV incidence.^[1], ^[7], ^[8] In China, studies have used BED-CEIA to estimate crude HIV incidence among IDUs in retrospective, cross-sectional analyses.^[15], ^[16] Another cross-sectional BED-CEIA study found similar HIV incidence rates as the cohort study conducted concurrently in the same population.¹⁷ Our study, the first in China to evaluate the BED-CEIA method by direct comparison with a cohort-identified incidence rate in the same subjects, confirms that the adjusted BED-CEIA rates are comparable to our cohort-identified rate. Although the crude BED-CEIA rate was also not statistically different from our cohort incidence rate despite being over three-fold higher, this was likely due to the low number of HIV seroconversions identified in our cohort and the subsequent lack of sufficient power to discriminate between the two results. Our previous study in Kaiyuan identified an HIV prevalence of 10% overall among the FSWs and 30% among the drug-using FSWs.²⁰ We thus expected to find more seroconversions during this one-year study. Furthermore, based on the demographics of the FSWs included in our longitudinal cohort, it appears that these FSWs had greater risk for acquiring HIV infection due to being older and having significantly more drug use compared to those not included in our cohort (Table 1). Despite this, it is possible that our longitudinal cohort was biased and missed the highest risk FSWs in the area.

Among subjects known to be infected for more than 365 days, our study identified a false-positive misclassification rate of 4.4% (10/229), identical to the previously reported proportion in China.¹⁷ However, among the smaller group of subjects testing positive for recent HIV infection, the false-positive misclassification ratio was 26.3% (10/38). We did not find a correlation between HIV subtype and misclassification, but our results are limited by the small sample size. Additional validation studies in China with larger sample sizes are needed to determine the misclassification rate more accurately and its potential correlation with HIV subtype. These data are needed to know how to interpret correctly future cross-sectional BED-CEIA results at the population level, where prior testing results are not available or have not been obtained.

All three subjects who seroconverted within 180 days had BED-CEIA results demonstrating recent infection. Two of these subjects subsequently participated in the third survey and were correctly identified by BED-CEIA as subjects with long-term infection. The fourth subject, who seroconverted within 365 days, had a BED-CEIA result showing long-term infection. On the other hand, the BED-CEIA test misclassified 10/38 (26.3%) of samples, classifying subjects with long-term infections as having recent infections. This rate of misclassification is less than that reported in Zimbabwe (37.8%, 142/376).⁸

Antiretroviral therapy has also been associated with BED-CEIA misclassification,²¹ but did not play a role in our study, with all nine FSWs who were on treatment in our study correctly testing negative by BED-CEIA. Moreover, the storage, shipment, and testing procedures were conducted by US CDC recommended standards, so these factors should not affect the results significantly.¹³

The primary limitation of our study is the relatively small sample size. Because the numbers of FSWs who seroconverted from HIV-negative to -positive were small, our 95% CI for incidence were relatively large, preventing us from identifying statistically significant differences between the crude and adjusted BED-CEIA incidence rates and our cohort-derived incidence rate. Second, our low retention rate may have biased our cohort study-derived HIV incidence, as the highest risk subjects may not have returned for subsequent testing. Our data indicate, however, that the subjects included in our longitudinal cohort actually had higher rates of drug use than those not included (Table 1) making this selection bias less likely. Third, this study provided voluntary HIV counseling and testing (VCT) to each FSW participant during the survey, which may have partly resulted in decreased HIV-related behaviors, resulting in lower numbers of new HIV infections. From this perspective, our cohort-derived incidence may underestimate the true incidence. Finally, all BED-CEIA related parameters, such as the window period (155 days) and confirmatory ODn value (0.8), were adapted from the US CDC's recommendations. These values may or may not be the same for Chinese HIV subtypes and may need to be adjusted. Further studies in China are needed to understand these issues.

In conclusion, this study provides additional data towards validating the use of the BED-CEIA assay to estimate HIV incidence rates in China, with adjusted BED-CEIA rates similar to the cohort-derived incidence. Differentiation between the cohort-derived and the crude BED-CEIA incidence rates, as well as the correlation between HIV subtypes and BED-CEIA misclassification, was limited by insufficient power. Additional studies with larger sample sizes are needed to evaluate more precisely how well BED-CEIA estimates HIV-incidence and incidence trends in Chinese populations. Should these studies confirm our outcomes, BED-CEIA will be a useful and inexpensive tool for HIV surveillance and intervention evaluations in China.

Conflict of interest: None of the authors have any conflicts of interest to declare.

Back to Article Outline

Acknowledgements 

The authors Junjie Xu, Guowei Ding, Haibo Wang, Guixiang Wang, Xia Jin, Ruiling Dong designed the study and provided the primary study tasks. Authors Junjie Xu and Ray Y. Chen undertook the statistical analysis and wrote the manuscript. Dr Ning Wang and Hong Shang wrote and revised the manuscript. Yan Jiang, Manhong Jia, Jennifer Chu, Kumi Smith, Gerald B. Sharp, Xiaoxu Han provided critical revisions of the manuscript. All authors contributed to the surveys and have approved the final manuscript. The authors wish to thank the staff at the Kaiyuan CDC and the outreach workers for providing their support in subject recruitment and survey interviews. The authors also thank all female sex worker participants in the study for their time and for sharing their information.

This study was supported by the Comprehensive International Program of Research on AIDS (CIPRA) grant from the National Institute of Allergy and Infectious Diseases, US National Institutes of Health (U19 AI51915-05) and PRC Ministry of Science and “The eleventh Five-Year Planning Programs prediction model of HIV/AIDS” (2008ZX10001–003).

Back to Article Outline

References 

1. Parekh BS, Kennedy MS, Dobbs 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
   • View In Article
   • MEDLINE
   • CrossRef
2. Rehle T, Shisana O, Pillay V, et al. National HIV incidence measures: new insights into the South African epidemic. S Afr Med J. 2007;97:194–199
   • View In Article
   • MEDLINE
3. Nesheim S, Parekh B, Sullivan K, et al. Temporal trends in HIV type 1 incidence among inner-city childbearing women in Atlanta: use of the IgG-capture BED-enzyme immunoassay. AIDS Res Hum Retroviruses. 2005;21:537–544
   • View In Article
   • MEDLINE
   • CrossRef
4. McDougal JS, Parekh BS, Peterson ML, et al. Comparison of HIV type 1 incidence observed during longitudinal follow-up with incidence estimated by cross-sectional analysis using the BED capture enzyme immunoassay. AIDS Res Hum Retroviruses. 2006;22:945–952
   • View In Article
   • MEDLINE
   • CrossRef
5. Hall HI, Song R, Rhodes P, et al. Estimation of HIV incidence in the United States. JAMA. 2008;300:520–529
   • View In Article
   • CrossRef
6. Hu DJ, Vanichseni S, Mock PA, et al. HIV type 1 incidence estimates by detection of recent infection from a cross-sectional sampling of injection drug users in Bangkok: use of the IgG capture BED enzyme immunoassay. AIDS Res Hum Retroviruses. 2003;19:727–730
   • View In Article
   • MEDLINE
   • CrossRef
7. Sinthuwattaniwibool C, Wasinrapee P, Chaowanachan T, et al. Application of the BED-CEIA to a cohort of military conscripts in northern Thailand to estimate HIV-1 incidence: a validation study. Abstract No. 731. 12th Conference on Retroviruses and Opportunistic Infections, Boston, Massachusetts, USA, February 22-25, 2005.
   • View In Article
8. Hargrove JW, Humphrey JH, Mutasa K, et al. Improved HIV-1 incidence estimates using the BED capture enzyme immunoassay. AIDS. 2008;22:511–518
   • View In Article
   • CrossRef
9. Bulterys M, Parekh BS, Chao A, et al. Use of an IgG capture BED enzyme immunoassay to investigate recent HIV-1 seroconversions among young women in Rwanda: comparison with 24-month prospective cohort data. Abstract No. MoPeC3458. XV International AIDS Conference, Bangkok, Thailand, July 11-16, 2004.
   • View In Article
10. Kim AA, Parekh B, Marum L, et al. Can HIV incidence be estimated from cross-sectional serologic surveys in Africa: novel application of a new HIV incidence laboratory assay. 55^th Annual Epidemic Intelligence Service Conference, Georgia, USA. Available at: http://www.cdc.gov/EIS/conference/archives/EIS%20Booklet_11.pdf (accessed December 2009).
    • View In Article
11. The United Nations Joint Programme on HIV/AIDS (UNAIDS). Statement on the use of the BED-assay for the estimation of HIV-1 incidence for surveillance or epidemic monitoring. Report of a meeting of the UNAIDS Reference Group for Estimates, Modelling and Projections, Athens, Greece, December 13-15, 2005. Geneva: UNAIDS; 2005.
    • View In Article
12. Karita E, Price M, Hunter E, et al. Investigating the utility of the HIV-1 BED capture enzyme immunoassay using cross-sectional and longitudinal seroconverter specimens from Africa. AIDS. 2007;21:403–408
    • View In Article
    • MEDLINE
13. Centers for Disease Control and Prevention. Interim recommendations for the use of the BED capture enzyme immunoassay for incidence estimation and surveillance. Atlanta, GA: CDC, Office of the Global AIDS Coordinator; 2006;
    • View In Article
14. Wang N. HIV epidemic in China and the World—current situation and challenges. Science Technology and Review. 2005;23:4–8
    • View In Article
15. Jiang Y, Wang MJ, Ni MJ, et al. HIV-1 incidence estimates using IgG-capture BED-enzyme immunoassay from surveillance sites of injection drug users in three cities of China. AIDS. 2007;21(Suppl. 8):S47–S51
    • View In Article
    • CrossRef
16. Han M, Wang YL, Wang MJ, et al. Application of BED-CEIA to estimate the human immunodeficiency virus prevalence among injecting drug users in Chongqing municipality. Chinese J Epidemiol. 2007;28:1077–1080
    • View In Article
17. Xiao Y, Jiang Y, Feng JG, et al. Measuring recent human immunodeficiency virus type 1 infections in China. Clin Vaccine Immunol. 2007;14:1384–1386
    • View In Article
    • CrossRef
18. Yang R, Xia X, Kusagawa S, et al. On-going generation of multiple forms of HIV-1 intersubtype recombinants in the Yunnan Province of China. AIDS. 2002;16:1401–1407
    • View In Article
    • MEDLINE
    • CrossRef
19. Buchacz K, Klausner JD, Kerndt PR, et al. HIV incidence among men diagnosed with early syphilis in Atlanta, San Francisco, and Los Angeles, 2004 to 2005. J Acquir Immune Defic Syndr. 2008;47:234–240
    • View In Article
    • CrossRef
20. Wang HB, Chen YR, Ding GW, et al. Prevalence and predictors of HIV infection among female sex workers in Kaiyuan City, Yunnan Province, China. Int J Infect Dis. 2008;19:365–369
    • View In Article
21. Hayashida T, Gatanaga H, Tanuma J, et al. Effects of low HIV type 1 load and antiretroviral treatment on IgG-capture BED-enzyme immunoassay. AIDS Res Hum Retroviruses. 2008;24:495–498
    • View In Article
    • CrossRef