Jake” was a 17-year-old high school student who came to see me with his supportive but anxious mother. Four months earlier, Jake’s pediatrician, having read the CDC recommendations for routine testing of all patients aged 13 to 64,
In a previous Editorial here, I discussed the need for broader uptake of the 2006 CDC recommendations for routine voluntary HIV screening.1 In response to those recommendations, Maryland law was changed in July 2008, eliminating the need for written consent and formal pre- and post-test counseling. Ever since the passage of that law, I’ve been wondering when I would begin seeing patients who were diagnosed based on routine screening. A few months ago, I saw my first such patient.
“Jake” was a 17-year-old high school student who came to see me with his supportive but anxious mother. Four months earlier, Jake’s pediatrician, having read the CDC recommendations for routine testing of all patients aged 13 to 64, ordered an HIV test, which came back positive. This news hit Jake and his mother hard. The health department got involved; his girlfriend and her family had to be notified (she tested negative); Jake’s sexuality was repeatedly questioned, and his ability to focus on school and post-graduation planning suffered. During the 4 months that it took for Jake and his mother to learn of his diagnosis, find an HIV consultant, and wait for an appointment, he and his mother read about HIV infection. Although what they read was largely reassuring, they were still reeling from the shock of this completely unexpected diagnosis. Jake’s mother worried that he wasn’t mature enough to take medications regularly. She also worried about the financial implications: She was raising him alone on an income that barely covered living expenses. Her group insurance was costly and required high copayments. She worried about what would happen when it was time for him to leave home and was no longer insured.
Jake’s mother brought a copy of the original lab test, which showed a fully positive enzyme immunoassay (EIA) and Western blot. Jake’s medical history was unremarkable. He had no chronic medical conditions and no physical complaints. After his mother left the room, I asked him about risk behaviors. He said he had had unprotected sex with two girls from his high school, his current girlfriend, who tested negative, and another girl, who had moved away a year ago and hadn’t been found or notified. He denied having had sex with men or having used illicit drugs other than marijuana. He had a few piercings and tattoos, but “who doesn’t?” His physical examination was entirely normal.
I ordered laboratory tests, including a CD4 count and an HIV viral load. When the results came back showing a CD4+ cell count of 846/µL and an undetectable viral load, I asked Jake to return for a repeat HIV serology, which was completely negative. In comparing Jake’s original laboratory tests with those I’d ordered, a few interesting discrepancies caught my eye. With only a 4-month interval between tests, the mean corpuscular volume differed by 10 points, and while the current total protein and albumin were normal, the original test showed a high total protein and a mildly low albumin. In addition, urine tests performed on the same day 4 months earlier showed pyuria with a positive Chlamydia probe. Jake said he was treated for Chlamydia based on that test result, but he had never had symptoms.
It was clear that Jake’s laboratory test results belonged to someone else-someone who was presumably walking around with undiagnosed HIV infection and an untreated sexually transmitted disease. I had the happy task of telling Jake and his mother that Jake was HIV-negative. My second and last visit with Jake ended with smiles, an admonition to use condoms, and mutually expressed wishes that our paths would not cross again.
The combination of an EIA and a confirmatory Western blot test is highly accurate. The US Preventive Services Task Force reports that the sensitivity and specificity of the EIA alone are 99.7% and 98.5%, respectively; when confirmed by a reactive Western blot, the chances of a false-positive result in a low-prevalence setting is approximately 1 in 250,000.2 Other studies have confirmed the accuracy of this approach to testing, reporting false-positive rates of 0.0004% to 0.0007% in the United States.3-5 When false-positive results do occur, they are often due to clerical error or error on the part of laboratory technicians or phlebotomists, as was almost certainly the case with Jake.
Unlike EIAs, rapid HIV tests are not automatically confirmed, making them more prone to false-positive results than the standard serology. The positive predictive value of a rapid test can be low in individuals at low risk or in low-prevalence areas. For a test with a sensitivity of 99.9% and a specificity of 99.8%, which is similar to those of many rapid HIV tests and EIAs, the positive predictive value would fall from 98% in an area with a seroprevalence of 10%, to 33% in an area with a seroprevalence of 0.1%. Therefore, all positive rapid test results should be considered preliminary and should be confirmed by further testing, preferably with a Western blot, immunofluorescence assay, or an RNA test.6 In some settings, a second rapid test from a different manufacturer is used for confirmatory purposes. Although this approach increases the positive predictive value of the result, the CDC considers it less definitive than confirmation with serology or RNA tests. However, these considerations apply to emergency rooms, walk-in clinics, and testing centers rather than to office-based practices, where the CDC recommendation is to include a standard HIV serology in the battery of blood tests being ordered by the clinician for routine testing or screening.
Jake’s case points out that it’s not enough to advocate routine HIV screening and appropriate referral of those who test positive. Jake’s pediatrician is to be commended for following those recommendations, but the lives of Jake, his mother, and his girlfriend were still thrown into unnecessary upheaval for several months because of an incorrect diagnosis. Primary care clinicians who follow CDC recommendations for routine voluntary HIV screening must be taught to go one step further before referring their seropositive patients. Confirmation of the diagnosis is essential, usually by obtaining the follow-up lab tests that an HIV expert will want to see at an initial visit: a CD4 count, an HIV viral load, and a genotypic resistance test. In rare cases where the viral load is undetectable, a repeat HIV serology can determine whether the initial test result is incorrect or whether the patient is infected but controlling the virus without therapy (an “elite controller”). And while the CDC has moved away from risk behavior assessment as a prerequisite for routine screening, an assessment of risk in this setting can be helpful in counseling the patient.
If Jake’s story was to be believed, his “pre-test probability” of being infected was extremely low, increasing the probability that his test results were incorrect. As we begin testing more people who are at low risk for infection, we will naturally see more false-positive results, despite the high degree of accuracy of the HIV serology.
Primary care practitioners will also need to know how to respond to equivocal or indeterminate tests, which will be more common than true false positives. In the absence of recent exposure, patients with positive EIAs but negative Western blots are almost invariably uninfected, although this pattern can occur with very recent infection.7 Those who test positive by EIA and who have bands on Western blot that do not meet diagnostic criteria for a positive result must be counseled and retested. Those at low risk can be reassured that infection is unlikely; those at high risk are more likely to be in the short “window period” before seroconversion. In either case, confirmation is necessary, either with immediate RNA testing or follow-up serology. Patients should be counseled about prevention of HIV transmission while awaiting confirmatory results.
The HIV serology is about as accurate as a diagnostic test can be, but there can still be subtleties and uncertainties in the diagnosis of HIV infection. It’s tempting to downplay these issues in an attempt to encourage frontline clinicians to embrace the idea of routine screening for HIV, but there is a price to be paid for oversimplification. Providing algorithms for the interpretation and management of HIV test results may ultimately be more useful than simply recommending testing and referral. When there is uncertainty about test results, physicians can contact the National Clinicians Consultation Center (Web site: http://nccc.ucsf.edu; phone: 800-933-3413) for immediate expert advice on test interpretation.
Postscript: Jake’s mother e-mailed me recently to say that the insurance company was refusing to pay for his office visits or laboratory testing. They argued that since Jake didn’t have HIV infection after all, he didn’t need a CD4 count, or viral load, or a visit with an HIV expert!
Gallant JE. Preaching to the choir: advocating routine HIV testing.
Chou R, Huffman LH, Fu R, et al; US Preventive Services Task Force. Screening for HIV: a review of the evidence for the U.S. Preventive Services Task Force.
Ann Intern Med
Kleinman S, Busch MP, Hall L, et al; Retrovirus Epidemiology Donor Study. False-positive HIV-1 test results in a low-risk screening setting of voluntary blood donation.
Burke DS, Brundage JF, Redfield RR, et al. Measurement of the false positive rate in a screening program for human immunodeficiency virus infections.
N Engl J Med
MacDonald KL, Jackson JB, Bowman RJ, et al. Performance characteristics of serologic tests of human immunodeficiency virus type 1 (HIV-1) antibody among Minnesota blood donors. Public health and clinical implications.
Ann Intern Med
Centers for Disease Control and Prevention. General and laboratory considerations: rapid HIV tests currently available in the United States.
. Last modified November 20, 2007. Accessed April 12, 2009.
Owen SM, Yang C, Spira T, et al. Alternative algorithms for human immunodeficiency virus infection diagnosis using tests that are licensed in the United States.
J Clin Microbiol