August 2009
Feature Story

Anne Ford

“Hot Topics in POC Microbiology”: That was the official theme of the May 17 talks presented by Sheldon Campbell, MD, PhD, and Roger D. Klein, MD, JD, at the general meeting of the American Society for Microbiology in Philadelphia. But in actuality, the pathologists’ presentations boiled those multiple “hot topics” down to just one fiery theme: how to provide the best care possible to as many people as possible at as low a cost as possible.

“If you’re willing to spend a jillion dollars, you can provide high-quality care to everybody,” dryly noted Dr. Campbell, associate professor of laboratory medicine at Yale University School of Medicine, New Haven, Conn., and director of laboratories for the VA Connecticut Healthcare System. Dr. Klein is medical director of molecular oncology at BloodCenter of Wisconsin, Milwaukee.

To help those of us without jillion-dollar budgets supply optimal care to the maximum number of patients, Drs. Campbell and Klein presented assessments of the current state of molecular point-of-care testing, as well as of HIV antibody testing and rapid influenza diagnosis.

Dr. Campbell introduced his first topic—HIV antibody testing—on a cheerful note. “The fact is that HIV serodiagnosis is quite good,” he said. “It’s highly sensitive and highly specific when you do your confirmatory tests.” Many of the problems with HIV testing, in his view, lie in the preanalytical and postanalytical phases. He cited CDC data gathered from five U.S. cities in 2004–2005 that showed that of 2,261 men who had sex with men, 25 percent were infected with HIV, and 48 percent of those infected were unaware of that fact.

Insufficient access to testing, Dr. Campbell said, was to blame. “Despite the fact that we can test people, we don’t,” he said. What’s more, 31 percent of the people who tested positive and 39 percent of those who tested negative did not return to receive their test results, according to the CDC data, which raises the question: “How good is the test if the patient never gets the result? Not very.”

To provide a snapshot of the state of community rapid HIV testing, Dr. Campbell cited a recent study showing that in 2005–2006, only 17 percent of the 375 sites interviewed—private urban nonprofit community clinics and community-based organizations involved in infectious disease treatment and testing—offered rapid HIV testing. “Policies that encourage greater diffusion of rapid testing are needed, especially in community-based organizations and venues with fewer resources and less access to laboratories,” the study authors concluded (Bogart LM, et al. Am J Pub Health. 2008;98:736–742).

Access to testing aside, Dr. Campbell is also concerned about the performance of rapid HIV tests in early infection, particularly since “it’s likely this is a period in which HIV is transmitted from person to person.” He pointed to a 2008 study that examined the results of a panel of antibody tests—rapid, lab-based enzyme immunoassay, and nucleic-acid amplification—on stored-plasma specimens from recently infected individuals. The 42 specimens were NAAT-positive and antibody-negative by at least one screening test. A lab-based EIA, Bio-Rad Laboratories’ GS HIV-1/HIV-2 Plus O, yielded 14 positive results, while of three rapid tests used, the best performer was Trinity Biotech’s Uni-Gold Recombigen HIV, which yielded 11 positive results. However, in followup testing (available for 30 of the 42 patients), the Uni-Gold test as well as Bio-Rad Laboratories’ Multispot HIV-1/ HIV-2 test detected all 30 positives. OraQuick Advance rapid HIV-1/2 (OraSure Technologies) and Clearview HIV 1/2 Stat-Pak (Inver­ness Medical) performed slightly less well, with 26 and 29 positives respectively (Louie B, et al. J Clin Microbiol. 2008; 46:1494–1497). “Remember that pooled PCR does better than any serological test,” Dr. Campbell noted.

In addition, Dr. Campbell brought up the possibility of procedural concerns related to rapid HIV testing. “When you start moving things out of the central laboratory, it’s harder to monitor them,” he noted. As an example, he pointed to the results of a 2006 post-marketing surveillance study that examined the performance of the OraQuick Advance rapid HIV-1/2 antibody test in 17 city and state health departments between August 2004 and June 2005. The study, which examined 135,724 whole-blood tests and 26,066 oral-fluid tests, found generally very high specificity. However, excess oral-fluid false-positives were found at one site. No expired devices or unique lots were found at the site, no single operator was involved, temperatures were found to be within range, and no change in frequency occurred even after a sampling discrepancy was corrected (Weslowski LG, et al. AIDS. 2006;20:1661–1666).

Then, too, “there are also biological issues that may enter into the question,” Dr. Campbell added. In 2004, the Journal of Perinatology published a study that examined a rapid screening program in a labor and delivery setting. Sixty-nine positive results were found out of 9,781 deliveries. Twenty-six of those rapid-test results were confirmed by Western blot. The prevalence of HIV and positive predictive value among subgroups broke down as follows: Caucasians, 1.53 percent and 75 percent; African Americans, 2.43 percent and 82.6 percent; Hispanics, 0.05 percent and 9.8 percent. The study results raise questions about the role of universal screening in low-risk populations, as well as safety concerns (such as domestic violence and loss of breastfeeding benefits) related to the routine disclosure of unconfirmed HIV rapid test results (Zacharias NM, et al. J Perinatol. 2004;24:743–747).

To those safety concerns, add findings from another study (Hillis SD, et al. Int J STD AIDS. 2007;18:120–122; also reported in: Campbell S, et al. Point of Care: The Journal of Near-Patient Testing and Technology. 2009;8:32–35). The study noted that after rapid HIV screening was implemented in an intrapartum HIV screening program in St. Petersburg, Russia, the intrapartum prophylaxis rate rose from 41 percent to 76 percent. However, the infant abandonment rate rose as well—from 26 percent to 50 percent. Among women with unintended pregnancies, that rate rose to 73 percent. “These tests get used in a complicated social environment,” Dr. Campbell said, “and there may be things happening that we didn’t intend to have happen.”

Given all of these concerns about rapid HIV testing, he concluded, please “don’t count the lab out.” Laboratories should work to reduce their turnaround times for laboratory-based HIV testing and consider implementing the random-access HIV tests that are available now on Siemens’ Advia Centaur and Ortho Clinical Diagnostics’ Vitros ECi systems. “The goal could be to give HIV results in the same time frame as troponin,” Dr. Campbell suggested. Laboratories might also consider preliminary reporting of positives before they’re confirmed. “If it can be done with rapids, why not with lab-based tests?” he asked, though he is quick to note the FDA labeling barrier.

For those who do perform rapid HIV testing, he offered these suggested best practices: Do it with oversight from a clinical laboratory; monitor quality, remembering that while false-positive rates are the easiest to check, sensitivity may be an issue as well; be aware of laboratory-based alternatives to point-of-care testing and be intentional about which best serves patients; and examine outcomes in your own population.

Switching topics, Dr. Campbell asked the audience, “You all want to hear about influenza, right?,” given the recent widespread worry about swine flu. “I’m about sick of it,” he confessed, before reminding his listeners that 3 million to 5 million severe cases of influenza are seen worldwide annually, with 250,000 to 500,000 deaths each year. “A pandemic would greatly increase those numbers—and drive us all berserk with no notice, as in April and May” of this year.

How should laboratories test for influenza? Rapid antigen tests, while fast and easy to perform, yield only 50 to 70 percent sensitivity, “despite occasional publications to the contrary,” Dr. Campbell said. Direct fluorescent antibody testing, while fairly fast and sensitive, requires a fluorescent scope and highly trained staff. And newer shell vial methods for culture, which has been considered the gold standard for sensitivity, take 24 to 48 hours to produce a result. That leaves molecular diagnostic testing as the new gold standard, Dr. Campbell said. Real-time molecular methods can produce results in about an hour, and in some cases, real-time PCR even exceeds culture in sensitivity.

But—to mix a metaphor—while molecular testing in this instance may be the new gold standard, it’s hardly a magic bullet. It brings with it high skill and equipment requirements, meaning that “for most facilities, the decision isn’t between molecular and rapid flu test, but between rapid flu test and nothing at all,” Dr. Campbell said. “Most facilities don’t have the ability to set up real-time PCR testing for influenza and run it throughout the season.” That means they must rely on other methods, including the rapid tests.

But, at least in this case, rapid testing is better than no testing at all, he added. He pointed to a recent VA study that reviewed the use of rapid antigen flu testing in older adults. The study found that performing rapid testing when flu symptoms were present greatly improved the ability to diagnose influenza in the acute care setting. The positive predictive value of fever plus cough increased from 32 percent to 92 percent with a positive rapid flu test, and 91 percent of patients with a positive rapid result who had begun experiencing symptoms within the previous 48 hours received antiviral treatment (D’Heilly SJ, et al. J Clin Virol. 2008;42:124–128).

Just how widespread is the use of rapid antigen assays for influenza testing, anyway? In an informal attempt to find out, Dr. Campbell conducted a survey on ClinMicroNet, an online listserv for microbiologists, earlier this year. Nearly three-quarters of the 41 respondents reported using a rapid flu test in their laboratories in the previous year, 84 percent of whom used rapid flu testing in the main laboratory. “This is a highly prevalent practice,” he said. “Most labs use these tests only as fast and cheap starters for a testing algorithm, but solo use or supplementation only with send-outs is significant as well.”

Best practices for influenza testing include—first and foremost—educating providers on specimen collection. “Specimen collection is the critical step in influenza testing,” Dr. Campbell said. Nasopharyngeal washes are preferable over swabs, he added, since “this is not a virus that’s found in a throat swab normally.” It’s critical to obtain virus-infected ciliated epithelial cells, he says.

Second, he recommends testing during the flu season only, since false-positives have potentially severe consequences. When testing early in the season, it’s best to retain the specimen for confirmatory testing. In addition, “remind providers to test early in illness,” he said. “The best therapeutic results are when drugs are started within 48 hours of the onset of symptoms.” Finally, he urged listeners to guide the interpretation of results by reminding providers that rapid tests are insensitive, that out-of-season false-positives outnumber true positives, and that there are known sources of error, such as bloody samples. For those reasons, it’s prudent to conduct supplementary testing, at least on selected patients and off-season or early-season positives.

Considering the future of molecular point-of-care testing more generally, Dr. Campbell remarked that while POC serology can be quite good, current POC tests for pathogen detection are mostly feeble substitutes for laboratory-based methods. Instrumented antigen testing could improve the quality of these tests to some extent, but “for pathogen detection, molecular—not antigen—is the way to go,” he said.

Molecular POC testing was, in turn, the focus of Dr. Klein’s presentation, which emphasized the need for automated, fully integrated instruments that offer sample preparation, amplification and detection, reproducibility, and reliability. Fortunately, such systems are emerging—such as GeneXpert from Ce­ph­eid, he said. He called it “a PCR lab in a disposable cartridge,” and added, “This is really a remarkable instrument. It’s the first self-contained molecular diagnostics instrument that was available, and already there are FDA-cleared tests available on it,” such as for MRSA and enterovirus. One GeneXpert test not yet available in the United States is Cepheid’s Xpert MTB/RIF assay, which provides TB ID results directly from sputum in less than two hours and simultaneously detects rifampin resistance. In Dr. Klein’s view, the assay is even more impressive because sputum has been considered the “final PCR frontier,” since it is usually highly viscous and often purulent, bloody, or both.

Among companies that have introduced real-time PCR systems are HandyLab, IQuum, Enigma Diagnostics, and Smiths Detection, Dr. Klein continued. HandyLab’s offerings include the Jaguar system, which, though not a true POC instrument, could be used in a satellite laboratory, he noted. It offers random sample access, microfluidics, detection in 45 to 90 minutes, contamination control, and an open menu for laboratory-developed tests. Similarly, Enigma Diagnostics’ Enigma FL and ML fully automated, real-time PCR systems offer rapid testing (with results in 30 to 45 minutes) and microfluid sample preparation, among other features. Other instruments include IQuum’s Liat analyzer and Smiths Detection’s BioSeeq Plus instrument

In the future, Dr. Klein said, these and similar instruments will undergo technological advances in performance, speed, and footprint, and offer expanded test menus, including quantitative assays. Until then, he said, “POC molecular ID testing is most useful for immediate decisionmaking and patients who may be lost to followup.”