Prevalence of Undiagnosed Acute and Chronic HIV in a Lower-Prevalence Urban Emergency Department

Abstract

Objectives. We estimated the seroprevalence of both acute and chronic HIV infection by using a random sample of emergency department (ED) patients from a region of the United States with low-to-moderate HIV prevalence.

Methods. This cross-sectional seroprevalence study consecutively enrolled patients aged 18 to 64 years within randomly selected sampling blocks in a Midwestern urban ED in a region of lower HIV prevalence in 2008 to 2009. Participants were compensated for providing a blood sample and health information. After de-identification, we assayed samples for HIV antibody and nucleic acid.

Results. There were 926 participants who consented and enrolled. Overall, prevalence of undiagnosed HIV was 0.76% (95% confidence interval [CI] = 0.30%, 1.56%). Three participants (0.32%; 95% CI = 0.09%, 0.86%) were nucleic acid–positive but antibody-negative and 4 (0.43%; 95% CI = 0.15%, 1.02%) were antibody-positive.

Conclusions. Even when the absolute prevalence is low, a considerable proportion of undetected HIV cases in an ED population are acute. Identification of acute HIV in ED settings should receive increased priority.

HIV screening is recommended by the US Centers for Disease Control and Prevention as an essential component of the nation’s HIV prevention effort.^1,2 Emergency departments (EDs) are particularly emphasized as venues for HIV screening.^3–5 Emergency departments serve more than 100 million patients annually, readily accessing vulnerable populations with a high prevalence of undetected HIV.^1,4–8

To date, most attention has been focused on detection of HIV in the chronic phase, after seroconversion, by assay for antibodies. Yet identification of patients during acute HIV infection could have a significant impact on further transmission.^9,10 Testing for acute HIV infection is accomplished by assays that detect viral proteins or viral genetic material before antibody detection is possible. This testing is more expensive, complex, or may delay results compared with antibody testing.^9,11,12 Despite these disadvantages, screening for acute HIV is increasingly suggested by various authors.^9,13–19 Acute HIV infection is thought to contribute disproportionately to HIV incidence because of high viral replication and increased infectiousness during this phase.^15,20–22 Diagnosis prompts many individuals to reduce transmission behaviors,²³ and partner notification efforts may be more successful.²⁴ There is also renewed interest in treatment during acute HIV infection, to lower infectiousness and improve long-term patient health outcomes.^21,25–27 In light of these benefits, screening for acute HIV infection may ultimately be cost-effective and worthy of increased logistical challenges.^9,28

Unfortunately, the controversies and implementation barriers in HIV screening have yet to be fully resolved,^29–35 particularly in ED settings where patient volumes exceed capacity and acute stabilization takes precedence over preventive health.^36–38 Screening in the ED for acute HIV infection will be even more challenging than screening for chronic HIV if it entails additional complexity and expense. Motivation to surmount such barriers is likely to be less in regions of lower HIV prevalence, in which disease incidence would also be presumed lower. Improving our understanding of acute HIV epidemiology in ED settings is fundamental for guiding potential implementation of ED screening interventions targeting acute HIV infection. We estimated the seroprevalence of both acute and chronic HIV infection by using a random sample of ED patients from a low-to-moderate HIV prevalence region of the United States.

METHODS

This was a cross-sectional, observational study in which we used HIV nucleic acid and HIV antibody assays to estimate the prevalence of acute and chronic HIV in an urban ED, in a region of lower HIV prevalence.

Setting and Participants

The ED is located in a Midwestern, urban, 450-bed teaching hospital. There are about 90 000 ED patient encounters annually. Pediatric patients are rarely seen; there is a large pediatric ED located nearby. A publicly funded HIV counseling and testing program has operated in this ED since 1998.^39–41 During this study, about 50% of ED patients were Black, 0.5% were Hispanic, and 40% were uninsured. The cumulative diagnosis rate of HIV/AIDS in the surrounding county was 233 per 100 000 persons.⁴² All patients aged between 18 and 64 years were eligible for this study.

Participants were enrolled during randomly allocated sampling windows between January 2008 and December 2009. We defined sampling windows by time of day and location in the ED.⁴³ Staffing was sufficient to allow consecutive approach of all patients within assigned study windows to offer participation in a compensated study “of diseases of public health importance.” Patients were offered $10 reimbursement for providing a blood sample and $5 for completing a health interview administered by a research assistant. The consent process emphasized that data would be stripped of all identifiers before any analysis. Although HIV was disclosed as 1 of the diseases of interest, it was not emphasized and was included among a list of other conditions.

Data Collection and HIV Assay

The health interview collected information about HIV risk integrated within a broad health history. We determined HIV antibody status by a sequential method involving review of test results from ED records, followed by assay of the collected sample if status was unknown. Thus, if an HIV test was performed on a clinical basis on or after the date the patient was enrolled and the result was negative, we presumed the sample to be HIV antibody–negative. We assayed all other samples for HIV antibody. All samples deemed antibody-negative, including those declared antibody-negative on the basis of clinical test results, we then assayed for nucleic acid.

Antibody assay.

We combined samples requiring antibody assay, (i.e., not known to be negative from clinical records) in pools of 10 samples (100 µL serum per sample).^44–46 We tested pools for HIV antibodies by both standard immunoassay (HIVAB HIV-1/HIV-2 [rDNA] EIA, Abbott, Abbott Park, IL) and rapid immunoassay (OraQuick ADVANCE Rapid HIV-1/2 Antibody Test, OraSure Technologies Inc, Bethlehem, PA). We retested initial positive pools sequentially in pools of 5 participants and then, for pools of 5 that were positive, we tested each sample individually. We tested individual positive samples a second time and then confirmed them by standard Western blot. We validated this method before testing participant samples by combining 9 known negative samples with 10 to 100 microliters of serum from persons with known reactive immunoassay and Western blot.

Nucleic acid assay.

We assayed samples in pools⁴⁷ containing 10 to 100 microliters of serum from 100 participants, that were created from constituent pools of 10 participants. We tested pools for HIV-1 RNA by polymerase chain reaction (AMPLICOR HIV-1 MONITOR UltraSensitive Version 1.5 viral load assay, Roche Molecular Diagnostics, Branchburg, NJ) with a lower limit of detection at less than 50 copies per milliliter. We retested initial positive pools sequentially in pools of 10 participants and then each sample individually. We tested individual positive samples a second time. We validated this method before testing participant samples by spiking known aliquots of 1 milliliter of plasma with 10 microliters of plasma from persons with detectable HIV-1 RNA at a level less than 10 000 copies per milliliter.

Analysis

We excluded duplicate enrollments from further analysis; we only retained the final enrollment. We also excluded patients from analysis if they were previously known to be HIV-positive, either by self-report or review of medical records, and if their HIV status could not be determined (e.g., sample amount insufficient). We excluded 2 samples that were inadvertently assigned the same study identification number.

The primary outcome is reported as the point estimate of prevalence with 95% confidence intervals (CIs). We conducted all statistical analyses with SPSS version 20.0 (IBM Corporation, Somers, NY).

RESULTS

Participant characteristics are shown in Table 1. The majority (98%) reported at least 1 factor potentially placing them at some risk for HIV. Figure 1 illustrates the flow of patients through the study. Staff approached 1488 patients, of whom 1034 (70%) consented. After exclusions, there were 926 samples available for antibody assay. There were 9 participants included for whom antibody testing results were obtained, but RNA results could not be obtained secondary to insufficient sample. There were 7 (0.76%; 95% CI = 0.30%, 1.56%) patients with HIV infection. Of these, 4 (0.43%; 95% CI = 0.15%, 1.02%) were antibody-positive, and 3 (0.32%; 95% CI = 0.09%, 0.86%) were nucleic acid–positive but antibody-negative (HIV RNA levels of 74 818, 50 880, and 1615 copies per milliliter).

TABLE 1—

Characteristics of Participants for Whom Samples Were Tested for Acute and Chronic HIV in a Midwestern Urban Emergency Department: 2008–2009

Characteristics	No. (%)
Age, y
18–29	268 (29.2)
30–39	167 (18.2)
40–49	219 (23.9)
50–64	264 (28.8)
Male gender	457 (49.7)
Race
White	387 (42.1)
Black	503 (54.7)
Other	30 (3.3)
Most recent HIV test
Never	199 (21.8)
Last year	320 (35.1)
> 1 year ago	393 (43.1)
Risk factors ever
Having vaginal intercourse	887 (95.9)
Having any anal intercourse	175 (18.9)
Receiving anal intercourse	108 (11.7)
Giving anal intercourse	83 (9.0)
Having multiple sexual partners, not always using condoms	242 (26.2)
Having a previous STD	315 (34.1)
Having sexual intercourse under the influence of alcohol or drugs	557 (60.2)
Having sexual intercourse with an at-risk partner	281 (30.4)
Man having sexual intercourse with men	19 (4.2)
Having sexual intercourse with a prisoner	200 (21.6)
Having sexual intercourse with an injection drug user	70 (7.6)
Having sexual intercourse with an STD-positive partner	63 (6.8)
Having sexual intercourse with an HIV-positive partner	2 (0.2)
Trading drugs or money for sexual intercourse	76 (8.2)
Using injection drugs	68 (7.4)
Using cocaine	273 (29.5)

Note. STD = sexually transmitted disease.

FIGURE 1—

Flow diagram of participant samples tested for acute and chronic HIV in a Midwestern urban emergency department: 2008–2009.

DISCUSSION

We characterized the prevalence of acute HIV infection relative to chronic HIV among a general and representative sample of patients presenting to an urban academic ED in a region of low-to-moderate HIV prevalence. We found that acute HIV infection constituted nearly half of the undetected HIV cases in this setting. This is a finding of considerable importance, despite the low absolute number of acute HIV infections detected. Low absolute numbers of new HIV diagnoses are a reality when one is screening in settings with lower disease prevalence, particularly when one is using nontargeted approaches,²⁹ yet screening is still deemed cost-effective.^48,49 The 3 instances of acute HIV infection found in this study appear even more significant when crudely extrapolated to the overall number of annual ED patients; for every 1000 patients aged between 18 and 64 years, about 3 are acutely infected and would be missed with traditional antibody testing. Despite progress in implementation of ED screening for chronic HIV, it is therefore likely that thousands of acute HIV infection cases go undetected in US emergency departments every year. These missed opportunities are particularly damaging to individual and public health.^9,13–17

The high proportion of acute HIV infection relative to chronic HIV in our sample is somewhat surprising. The only other ED study similar to this report was from a high-prevalence region and found only 1 of 15 undiagnosed cases to be acute HIV infection.⁵⁰ The proportion of undiagnosed cases that are from acute infection has been as low as 1.4% at first antenatal visit in rural South Africa,⁵¹ 21% in a high-risk population in San Francisco,⁹ 11% to 14% among men who have sex with men and transgender persons,^12,18 and 9.7% in a San Francisco sexually transmitted disease clinic.⁵² In 1 urban urgent care center, fewer than 50% of newly detected cases were from acute infection even when specifically targeting a subset with flu-like illness and identified HIV risk factors.⁵³ Because the incidence of a transmissible disease presumably increases in proportion to its prevalence, we expected acute HIV to be quite rare in the lower-prevalence setting evaluated in this study. Although this was not the case, our results are roughly consistent with 2 recent reports from actual ED screening programs in higher-prevalence areas that employed assays capable of detecting acute HIV infection.^17,54

Overall, the emerging picture suggests that EDs may have an unusually high prevalence of acute HIV infection relative to chronic infection and relative to other venues, both in lower- and higher-prevalence regions. Why the proportion of ED patients with acute HIV infection would be especially high in EDs compared with other settings is unknown, but it is easy to imagine that individuals with acute HIV infection preferentially select the ED to receive care, either because of the acuity of their symptoms^53,55 or because of impaired access^56,57 to other health care resources. If so, EDs will be even more important as sites for HIV screening efforts, and pressure for EDs to incorporate advanced diagnostic algorithms for detection of acute HIV infection will increase. Moreover, the importance of incident case detection by EDs may persist even after most chronic cases are already diagnosed.

Limitations

Our results should be considered in light of several limitations. This single center is not representative of community EDs, rural EDs, and urban EDs with particularly high HIV prevalence. Nonetheless, the lower-prevalence setting used in this study has broad implications: a relatively high prevalence of acute HIV even when overall disease burden is low suggests that screening for acute HIV may be uniquely important in the ED setting and that the need to do so extends beyond higher-prevalence EDs.

The study sample was rigorously obtained and included fresh blood from all ED patients even if they were not otherwise receiving venipuncture as part of their ED care. However, the sample was biased by lack of consent for study participation despite the availability of compensation. In addition, the availability of compensation could have preferentially selected a sample with lower socioeconomic status than the overall ED population. We do not know the true accuracy of self-reported HIV status; it is possible that participants with undetected HIV had actually been diagnosed previously, though this may be unlikely if infected patients intent on nondisclosure declined study participation. Nine of the participants who tested negative for HIV antibody had insufficient sample for nucleic acid assay and we presumed these to be negative, but we cannot exclude the possibility of additional cases of acute infection. Despite the large size of the study, confidence intervals calculated for proportions remain wide, raising the possibility that true proportion of acute HIV in this population is different than that found in this study.

One of the samples counted as an acute infection had a low viral load; recent evidence suggests that low viral loads can be encountered in acute infection.⁵⁸ However, there was no confirmatory antibody testing in this study. If this sample is not counted, the proportion with acute infection would be 0.2% (95% CI = 0.05%, 0.70%), with one third of undetected infections being from acute HIV. Finally, we do not know if the acute cases in this study would have been detectable by newer fourth-generation HIV combination antigen–antibody assays, which have recently shown utility in detecting most cases of acute HIV infection while avoiding the expense and operational challenges of nucleic acid detection.^17,59

Conclusions

We have demonstrated that there are patients with acute HIV infection in unselected, lower-HIV-prevalence ED populations. Although the absolute number is low, acute HIV infection may nonetheless constitute a considerable and unusually high proportion of the undetected HIV cases in an ED population. Large-scale ED screening for acute HIV infection should be an urgent subject of expanded investigation. Meanwhile, emergency physicians should be highly attuned to the possibility of acute HIV infection, particularly if working in an urban ED serving disadvantaged patient populations.