Background: Acute infections with HIV account for a disproportionate share of forward transmission in certain populations. We hypothesized that persons with acute HIV infection (AHI) would identify more named partners than those with established HIV infection (EHI).
Methods: We reviewed North Carolina surveillance databases to identify all persons aged ≥18 years in whom HIV was diagnosed during November 1, 2002 to October 31, 2007. We compared the number of named partners identified by persons with AHI versus EHI (based on nucleic acid amplification plus serologic testing) using a multivariable model and also compared information regarding HIV testing among partners identified by these groups.
Results: EHI was diagnosed in 9044 persons and AHI in 120 persons during the study period. Persons with AHI named 2.5 times more partners per index patient [95% confidence interval: 2.1 to 3.0] and 1.9 times more partners newly diagnosed with HIV infection per index patient (95% confidence interval: 1.1 to 3.5) as did persons with EHI.
Discussion: In North Carolina, persons with AHI identified proportionately more named partners and more partners newly diagnosed with HIV infections than persons with EHI. Improved detection of AHI offers critical opportunities to intervene and potentially reduce transmission of HIV.
From the *Epidemic Intelligence Service, CDC, Atlanta, GA; †North Carolina Department of Health and Human Services, Raleigh, NC; ‡Department of Epidemiology, University of North Carolina School of Public Health, Chapel Hill, NC; and §Division of Infectious Diseases, University of North Carolina School of Medicine, Chapel Hill, NC. Zack S. Moore is now with North Carolina Department of Health and Human Services; Sandi McCoy is now with RTI International.
Received for publication January 8, 2009; accepted April 7, 2009.
Preliminary findings presented in a poster at the 57th Annual Epidemic Intelligence Service Conference in Atlanta, GA, April 14-18, 2008.
Sources of funding: None.
Required Office of Management and Budget Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Correspondence to: Zack S. Moore, MD, MPH, Communicable Disease Branch, 1902 Mail Service Center, Raleigh, NC 27699-1902 (e-mail: email@example.com).
Phylogenetic and epidemiologic analyses indicate that acute or early infections with HIV can account for a disproportionate share of forward transmission.1,2 Acute HIV infection (AHI) is characterized by high viral concentrations in plasma and genital secretions3 and a higher likelihood of transmission per event compared with established HIV infections (EHIs).4-6 Many persons with AHI are unaware of their infection status; these persons have a higher prevalence of high-risk behaviors compared with persons who are aware of their infection status, further increasing the likelihood of forward transmission.7-9
On November 1, 2002, the North Carolina Department of Health and Human Services began conducting nucleic acid amplification testing (NAAT) on all antibody-negative or indeterminate serum samples from persons consenting for HIV counseling and testing at publicly funded sites.10 This routine NAAT was added to identify persons with AHI and to allow for intervention through partner counseling and referral services (PCRS). PCRS can be an effective means of identifying persons with undiagnosed HIV infection, providing risk-reduction counseling and facilitating entry into care.11,12 However, previous studies have not established whether the number of partners who can be reached for PCRS is different for persons with AHI versus EHI or whether partners of persons with AHI differ from partners of persons with EHI with regard to HIV status. In the present study, we hypothesized that persons with AHI would identify more named partners per index case than those with EHI and that improved identification of AHI (eg, by routine NAAT of antibody-negative or indeterminate samples) would therefore disproportionately increase the number of partners who could be reached for PCRS. To evaluate this hypothesis, we compared the number of partners identified by persons with AHI to the number identified by persons with EHI during the first 5 years of routine NAAT in North Carolina after controlling for differences in selected demographic features and risk behaviors. We also compared the HIV status of partners identified by persons with AHI versus EHI.
We reviewed the Sexually Transmitted Disease Management Information System (STD-MIS) databases from all regions in North Carolina to identify persons aged ≥18 years in whom HIV was diagnosed during November 1, 2002 to October 31, 2007. These databases included all cases of HIV reported to the state during this period, whether identified through public or private laboratories. Cases were classified in the STD-MIS databases at the time of data entry as acute HIV, chronic HIV, or AIDS. For the purposes of this study, we reviewed laboratory data for those cases classified as acute HIV and included as AHI those meeting the following criteria: (1) a positive NAAT from a specimen with a negative or indeterminate antibody result or (2) a positive NAAT, antigen test, or antibody test from the specimen of a person who had a negative or indeterminate HIV test within the preceding 30 days. We excluded from analysis those cases classified in the STD-MIS databases as acute HIV for which these criteria were not met. To identify additional AHI cases that might have been misclassified, we cross-checked the STD-MIS databases against a centralized database of acute HIV cases maintained by the state HIV Prevention and Control Branch. We included cases identified through this method as AHI if the laboratory criteria listed above were met. We included all other cases as EHI.
Testing procedures for specimens submitted to the North Carolina State Laboratory of Public Health have been reported in detail elsewhere.10 Briefly, serum samples were screened for antibodies against HIV by using Vironostika HIV-1 enzyme immunoassay and Western blot (Bio-Rad Laboratories, Hercules, CA) analysis kits. Serum samples that were antibody negative were pooled and screened for HIV-1 RNA by using the Procleix HIV-1 assay (Gen-Probe, San Diego, CA) before July 2005 or the EasyQ HIV-1 quantitative assay (bioMerieux, Marcy l'Ftoile) thereafter, according to a previously described algorithm.10,13 Antibody-indeterminant samples were tested individually for HIV-1 RNA. Individual specimens with positive NAAT results underwent repeated enzyme immunoassay testing or quantitative HIV-1 RNA testing (Roche Amplicor 1.5; Roche Molecular Systems, Pleasanton, CA). All positive results were confirmed by repeated antibody testing on a subsequent specimen.
All persons tested for HIV were notified of their results if they could be located. Those whose specimens tested positive were provided risk-reduction counseling and PCRS by trained disease intervention specialists (DIS). These DIS conducted the initial patient interviews, facilitated confirmatory testing, and made referrals to care. Persons in whom potential AHI was diagnosed were assigned to a team of specially trained DIS functioning within a distinct organizational structure. All patients were asked to identify sex or needle-sharing partners with whom they had had contact during the period of interest. This period was defined as 1 year before interview for persons with EHI and 6 months before interview for persons with AHI. In this study, we refer to partners for whom sufficient information was available to initiate an investigation as “named partners.” DIS attempted to contact all named partners to provide PCRS. Data regarding HIV status of partners who were contacted but not tested through PCRS were reported by the partner and verified through laboratory results when possible. Patients were also asked to estimate the total number of sex and needle-sharing partners with whom they had contact during the period of interest. We refer to this number as the “estimated period partners.”
We compared demographic and behavioral characteristics of persons with AHI and EHI and their partners by using a 2-sided Pearson χ2 or Fisher exact test, as appropriate, to assess for statistical significance. We compared continuous variables by using the Wilcoxon rank sum test.
To determine the effect of AHI versus EHI on the number of named partners identified, we constructed a negative binomial model with the number of identified partners as the outcome and AHI or EHI as the binary exposure. The negative binomial distribution was selected to account for overdispersed count data. We first constructed an unadjusted model to assess the ratio of partners identified by persons with AHI versus EHI. We then adjusted the model for the following covariates: age group (18-23, 24-30, 31-40, or ≥41 years), race/ethnicity (white, non-Hispanic; black, non-Hispanic; or Hispanic/other), drug use (including any injected or noninjected drug other than alcohol), and sex/male same-sex activity (female; male with reported same-sex activity; or male with no reported same-sex activity). We assessed individual covariates for effect measure modification in bivariate models and then tested them for contribution to the fully adjusted model by manually removing each interaction term and performing the likelihood ratio test (LRT) to assess for significance. Interaction terms with likelihood ratio test P values <0.15 were included in the final model. We did not evaluate further for confounding because the main effects of all covariates were retained in the model. Finally, we repeated the negative binomial model with the number of newly diagnosed HIV-positive partners as the outcome. All analyses were performed using SAS v.9.1 (SAS Institute, Cary, NC).
During November 1, 2002 to October 31, 2007, EHI was diagnosed in 9044 persons and AHI in 120 persons in North Carolina. One hundred six persons with AHI had been classified as acute HIV in the STD-MIS databases; 14 others had been misclassified as chronic HIV (13) or AIDS (1) and were identified from the centralized database of acute HIV cases maintained by the state HIV Prevention and Control Branch. Twenty-two cases that had been classified in the STD-MIS databases as acute HIV did not meet our laboratory criteria for AHI and so were excluded from analysis. Among persons with AHI, 83 (69%) were initially identified by routine NAAT of specimens with negative or indeterminate antibody results; 21 (18%) were identified by positive NAAT and negative or indeterminant antibody test results from the same date (ie, individually ordered rather than routine NAAT); and 16 (13%) were identified by positive antibody or NAAT with a prior negative or indeterminant test from a serum sample obtained during the preceding 1-30 days.
Persons with AHI had a younger median age than those with EHI (30 versus 39 years, P < 0.01). Persons with AHI were also more likely to be male (81% versus 71%, P = 0.01), to be white (44% versus 31%, P = 0.02), and to report male same-sex activity (62% versus 34%, P < 0.01) or noninjection drug use (43% versus 27%, P < 0.01) (Table 1). Proportions reporting Hispanic ethnicity or injection drug use were similar between the 2 groups. The median interval between assignment to a DIS and completion of the index patient interview was 1 day for persons with AHI and 7 days for persons with EHI. During 2002-2007, the annual proportion of persons that could not be located for initial patient interview ranged from 0% to 11% among persons with AHI and 11% to 16% among persons with EHI.
Persons with AHI reported a median of 2 estimated period partners during the 6 months before interview [interquartile range (IQR) 1-4], whereas those with EHI reported a median of 1 estimated period partner during the 12 months before interview (IQR 0-2, P < 0.01). The median number of named partners identified by persons in both groups was 1, with an IQR of 1-2 for persons with AHI and 0-1 for persons with EHI (P < 0.01). Persons with AHI were more likely than those with EHI to have fewer named partners than estimated period partners (48% versus 32%, P < 0.01). All 264 named partners identified by persons with AHI were categorized as sex partners. Twenty-seven of 7899 named partners identified by persons with EHI (0.3%) were categorized as needle-sharing or both needle-sharing and sex partners.
Among named partners for whom HIV testing history data were available, approximately half of those named by persons with either AHI or EHI had been tested previously [121 of 231 (52%) versus 3220 of 6735 (48%), P = 0.17]. However, among those tested previously, fewer partners of persons with AHI reported having tested positive [68 of 121 (56%) versus 2247 of 3220 (70%), P < 0.01]. Overall, more than a quarter of named partners in both groups had been previously identified with HIV infections (26% versus 28%, P = 0.34) (Table 2). Partners of persons with AHI were more likely to be tested through PCRS (50% versus 39%, P < 0.01). Similar proportions of partners in both groups were newly identified with HIV infections [12 of 264 (5%) versus 468 of 7899 (6%), P = 0.35]. However, because persons with AHI tended to name more partners than persons with EHI, the average number of newly identified infections per index case was higher among persons with AHI [12 of 120 (0.10) versus 468 of 9044 (0.05), P = 0.02].
Among partners of persons with AHI who were newly identified with HIV infections, 3 of 12 (25%) were classified in the STD-MIS databases as having AHI. One of these persons met our laboratory criteria for AHI. The remaining 2 had been tested through PCRS as contacts to a single index patient with AHI; both were antibody negative on initial testing but antibody positive and NAAT positive on repeat testing 43 days later. Among partners of persons with EHI who were newly identified with HIV infections, 4 of 468 (<1%) were classified as having AHI. All 4 persons met our laboratory criteria for AHI.
In the unadjusted negative binomial model, persons with AHI identified 2.52 times as many total named partners as did persons with EHI (95% confidence interval: 2.08 to 3.04) (Table 3). In the multivariable analysis, the effect of AHI on partner number differed by race/ethnicity; we therefore stratified all adjusted results by race/ethnicity. In all strata, persons with AHI named more partners than persons with EHI. The magnitude of this difference was greatest among whites.
Overall, 11 of 120 persons with AHI and 445 of 9044 persons with EHI identified partners who were newly diagnosed with HIV infection as a result of PCRS (9% versus 5%, P = 0.03) (Table 4). Among index patients whose partners were newly diagnosed with HIV infection, male same-sex activity was reported by 10 (91%) of those with AHI compared with 178 (40%) of those with EHI. In the unadjusted negative binomial model, persons with AHI identified 1.93 times as many newly diagnosed partners as did persons with EHI (95% confidence interval: 1.06 to 3.53). We were not able to examine this relationship in the fully adjusted model due to limited variation in the numbers of newly diagnosed partners.
This is the first population-based study to demonstrate that persons with AHI identify higher numbers of named partners per index case than persons with established infections. In our population, persons with AHI also have a higher proportion of partners who get tested for HIV as a result of PCRS and more newly identified infections per index case. Although the increased probability of transmission per partner during AHI has been well documented, the higher number of partners named, tested, and newly identified with HIV had not previously been established. The results of this analysis lend additional support for routine identification of AHIs as a means to increase the number of partners reached by PCRS and the number of new infections identified.
Identifying persons with AHI is vital to controlling the spread of HIV. Phylogenetic and epidemiologic analyses among other populations have demonstrated the critical role this group plays in forward transmission.1,2 Earlier diagnoses among index patients provide opportunities for more timely identification of HIV-positive and HIV-negative partners and earlier interventions to prevent transmission. Moreover, improved identification of AHI provides opportunities for index patients to learn their own status more rapidly. Previous research by Marks et al8 indicates that persons who are unaware that they are HIV infected transmit at a rate 3.5 times higher than persons who are aware. Finally, persons with AHI might be more likely to identify the source of their infection than those with EHI because of the shorter interval since infection. Therefore, investigations of AHI cases can allow for improved identification of persons or networks that pose an ongoing risk of transmission.
Our findings are consistent with previous reports demonstrating that such third-party partner notification programs as PCRS are successful in identifying previously undiagnosed HIV infections.11,12 PCRS also serves key functions beyond increasing serostatus awareness, including providing risk-reduction counseling and facilitating access to services for infected and uninfected partners. In our population, each AHI case investigated led to identification of more named partners and more newly identified HIV infections on average than each EHI case investigated. Consistent with our findings, a recent study in a majority of white, predominantly gay male population in San Francisco demonstrated that the number of patient interviews needed to detect 1 new HIV infection was lower for persons with acute infections than for those with nonacute or longstanding infections, although these findings were based on only 1 newly identified HIV infection among partners of acute patients.11
Multiple possible explanations exist for our finding that persons with AHI tended to identify more named partners than those with EHI, even after controlling for differences in demographics and selected risk factors, and despite the shorter period of interest. First, the assumption could be made that risk of infection increases with number of partners. Therefore, persons in whom HIV is diagnosed during acute infection might be identified at a time when they are engaging in contact with more sex or needle-sharing partners. Also, persons with AHI have all had, by definition, recent exposures. Therefore, they might be more likely to identify partners involved compared with persons with EHI whose exposures might be more remote. Finally, persons with AHI might be more likely to perceive partner identification as a means to identify the source of their infection and might therefore provide more complete information.
Interpretation of our findings should take into account differences in the way AHI and EHI cases are handled in North Carolina. Index patients with AHI are interviewed more rapidly, and their interviews are conducted by a specially trained team of DIS functioning within a distinct organizational structure. If persons with AHI perceived the situation differently (eg, with a greater sense of urgency or a stronger sense of the benefits of cooperating with DIS), they might have been more willing than persons with EHI to identify their partners. We were not able to control for the effect of these programmatic factors in our analyses. In addition, the proportion of persons that could not be reached for initial interview or initiation of PCRS was larger among persons with EHI. Because these persons were included in the database with zero named partners, this could bias our results toward an increased number of partners among persons with AHI. Finally, the period of interest for persons with AHI was different than that for persons with EHI (6 months versus 1 year), making a direct comparison of partner numbers difficult. However, this would be expected to bias our results toward identification of a higher number of partners by those with EHI.
We cannot exclude the possibility that some cases were misclassified as AHI or EHI in our analyses. Although laboratory data were not reviewed for all patients classified as EHI, it is unlikely that many of these represent unidentified acute infections. Of greater concern is the high proportion of cases initially classified in the STD-MIS databases as acute HIV that was excluded from analysis [22 of 128 (17%)]. Some of these cases might represent acute infections that were excluded due to missing or incomplete laboratory data. However, our findings were not significantly different when the analyses were repeated with these cases included in the AHI group. Finally, data regarding HIV status of partners not tested through PCRS were obtained through partner self-report and could not be verified in all cases.
In North Carolina, persons with AHI identified more named partners per index case than persons with EHI. This led to a higher number of partners per index case who were tested for and newly identified with HIV infection. Routine detection of AHI offers opportunities to intervene with focused PCRS. States, local jurisdictions, and individual providers should consider methods to improve detection of AHIs so that infected persons and their partners can receive needed services, and additional infections can be prevented.
The authors wish to thank John Barnhart and Todd Vanhoy of the North Carolina Department of Health and Human Services for review of the article and assistance with classification of patients and William Miller of the University of North Carolina Schools of Medicine and Public Health for assistance with statistical methods.
1. Yerly S, Vora S, Rizzardi P. Acute HIV infection: impact on the spread of HIV and transmission of drug resistance. AIDS
2. Brenner BG, Roger M, Routy J-P, et al. High rates of forward transmission events after acute HIV-1 infection. J Infect Dis
3. Pilcher CD, Tien HC, Eron JJ. Brief but efficient: acute infection and the sexual transmission of HIV. J Infect Dis
4. Jacquez JA, Koopman JS, Simon CP, et al. Role of the primary infection in epidemics of HIV infection in gay cohorts. J Acquir Immune Defic Syndr
5. Wawer MJ, Gray RH, Sewankambo NK. Rates of HIV-1 transmission per coital act, by stage of HIV-1 infection, in Rakai, Uganda. J Infect Dis
6. Leynaert B, Downs AM, de Vincenzi I. Heterosexual transmission of human immunodeficiency virus: variability of infectivity throughout the course of infection. Am J Epidemiol
7. Colfax GN, Buchbinder SP, Cornelisse PG, et al. Sexual risk behaviors and implications for secondary HIV transmission during and after HIV seroconversion. AIDS
8. Marks G, Crepaz N, Janssen RS. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS
9. Marks G, Crepaz N, Senterfitt W, et al. Meta-analysis of high-risk sexual behavior in persons aware and unaware they are infected with HIV in the United States. J Acquir Immune Defic Syndr
10. Pilcher CD, Fiscus SA, Nguyen TQ, et al. Detection of acute infections during HIV testing in North Carolina. N Engl J Med
11. Ahrens K, Kent CK, Kohn RP, et al. HIV partner notification outcomes for HIV-infected patients by duration of infection, San Francisco, 2004 to 2006. J Acquir Immune Defic Syndr
12. CDC. Partner counseling and referral services to identify persons with undiagnosed HIV-North Carolina, 2001. MMWR Morb Mortal Wkly Rep
13. Pilcher CD, McPherson JT, Leone PA. Real-time, universal screening for acute HIV infection in a routine HIV counseling and testing population. JAMA