September 2009 Clinical Laboratory News: Get Ready for Novel H1N1 Influenza

CLN Banner Logo

September 2009: Volume 35, Number 9


Get Ready for Novel H1N1 Influenza
What’s the Role of Rapid Tests?
By Genna Rollins

Following initial reports of an influenza-like illness in Veracruz, Mexico in early March 2009, the first confirmed U.S. case of what became known as Swine Flu appeared on March 28. The emergence of what now is formally called novel H1N1 influenza virus quickly grabbed headlines and placed considerable strains on both the public and private healthcare systems. Physicians—particularly pediatricians—were challenged to provide definitive answers to anxious patients and family members, public officials at all levels grappled to craft appropriate responses and guidance based on the changing dynamics of the outbreak, hospitals and clinics contended with a burgeoning influx of patients, and labs struggled to keep up with testing demands.

With the case load rising rapidly, the federal government declared a public health emergency on April 26, and as the virus continued to spread across the globe, WHO declared a pandemic on June 11. CDC sequenced the virus’s genome within days of identifying the novel strain and received emergency use authorization (EUA) from the FDA to distribute a real-time RT-PCR detection panel to public health labs on April 27. The EUA will remain in effect for one year, or until the public health emergency is terminated or the EUA is revoked, whichever occurs first. With only CDC and public health labs initially able to definitively identify the strain, many public health labs were thrust from their traditional reference and surveillance functions to serving as defacto diagnostic labs. On April 29, CDC protocols for the real-time RT-PCR detection panel, including primers and probes, were posted on the WHO website, and some labs subsequently developed their own versions of the test, including Quest Diagnostics' Focus Diagnostics business unit, which received EUA for its assay on July 24 , the first commercial lab to do so.

Now, as the virus spreads through the Southern Hemisphere, public health officials, providers, and labs are gearing up for a possible resurgence in the U.S. Laboratorians involved in this process will need to carefully consider testing algorithms, communication and coordination requirements, and staffing and resource needs, according to experts. “A lot of this is a matter of recognizing the issues, knowing who to talk with, what the decision points are, and determining where to put our efforts based on what's happening. It’ll be really important to work together,” explained Marie Landry, MD, professor and vice chair of the department of laboratory medicine and director of the clinical virology laboratory at the Yale University School of Medicine and Yale-New Haven Hospital.

Responding to the Outbreak

Landry’s experience with the novel H1N1 outbreak has given her a unique perspective on what labs can expect should the virus resurge in the Northern Hemisphere this fall. Her lab at Yale-New Haven Hospital implemented a real-time RT-PCR method based on CDC's published primers and probes, and is probably the only private hospital lab in Connecticut to have done so. By the end of July, the lab had performed more than 600 influenza PCR tests, with more than 300 samples identified as the novel H1N1 strain.

When word got out that the hospital had a test that could not only detect but differentiate the outbreak strain, patients flocked to the emergency department for testing, and the hospital also was inundated with specimens from pediatric offices. As test requests went up, the lab temporarily deviated from its normal policy of reserving PCR influenza testing for inpatients and high-risk ambulatory care patients. “Initially we were doing a lot more PCR than we would normally do in order to track this strain,” Landry recalled. After it became clear that novel H1N1 was the predominant strain, the lab reverted to its standard protocols. “I told the head of our pediatric emergency department that we would not be doing PCR on outpatients except those at high risk of complications. When people came to the ED they would be told when they arrived what testing was or was not going to be available. The volume of testing did go down at that point, but it's hard to know whether we'd hit the peak and it would have gone down anyway,” said Landry.

Detecting and Differentiating the Virus

Determining the most effective testing algorithm and properly interpreting the results has been a particular challenge in this outbreak. Available diagnostic tests include viral culture, serology, rapid antigen testing, PCR, and immunofluorescence assays. At CLN press time, only the CDC real-time RT-PCR detection panel and Focus Diagnostics’ influenza A H1N1 RNA real-time RT-PCR test had EUAs to specifically detect and differentiate the novel H1N1 virus from seasonal strains. Other tests can detect the virus as type A, but are not able to subtype, meaning they can’t differentiate novel H1N1 from seasonal type A strains, such as H1N1 or H3N2. (See “A Word About H1N1,” below, for more information about the virus.)

A Word About H1N1

The novel H1N1 virus is a variation of seasonal influenza, but with a seminal difference. The virus has a unique genetic composition not seen before. Novel H1N1 virus has in common with its seasonal influenza type A cousins a segmented genome that enables it to reassort with another type A strain if two viruses infect the same cell. In the case of novel H1N1, six genes came from triple-reassortant swine influenza viruses that had been identified previously in North American pigs. Two genes—matrix, or M, and neuraminidase, or NA— had lineage from Eurasian swine influenza viruses.

Novel H1N1 also shares with its influenza type A relatives a conserved portion of the matrix gene, and this allowed labs to determine quickly that the new virus was type A strain. However, the hemagglutinin genes of novel H1N1 are distinct from seasonal influenza viruses, so assays available when the virus appeared could detect that it was type A but not identify the subtype, such as H1, H3, or H5. This was an immediate tip-off that the strain had not been seen before, and indeed, nonsubtypeable or novel type A influenza viruses are considered reportable diseases.

At this point, the virus overall does not appear to be particularly virulent, except for recent reports of severe illness in pregnant women. In contrast to seasonal influenza, which disproportionately affects the elderly, novel H1N1 virus appears to be more of a disease of the young, with about two-thirds of cases in young adults and children. At CLN press time, CDC was reporting nearly 44,000 lab-confirmed cases with 5,514 hospitalizations and 353 deaths in the U.S., while WHO was reporting 162,380 lab-confirmed cases and 1,154 deaths in 168 countries.

However, in July, Austin, Texas-based Luminex received FDA clearance to update labeling for the company’s xTAG respiratory viral panel (RVP) multiplex PCR test to include data about the test's performance in patients with novel H1N1. “The RVP can detect novel H1N1 but it can’t identify it specifically as swine flu. The influenza matrix gene is highly preserved so it will tell you you’ve got an influenza A in your sample, but it’s going to be negative for the human-specific H1 and H3 genes, and therefore you will know it's something new that's been detected but not specifically identified as swine flu,” explained James Mahony, PhD, director of the regional virology lab at St. Joseph’s Hospital and professor of pathology and molecular medicine at McMaster University in Hamilton, Ontario. Mahony recently published a report of the RVP detecting the presence of novel H1N1 in 20 of 20 patients infected with the outbreak strain (J Clin Virol 2009; 45: 200–202), and other studies also have reported on the test’s performance (J Clin Virol 2009; 45:191–195; J Clin Microbiol 2009;47:2347–2348). The test has a place in quickly identifying probable H1N1 cases, according to Patrick Balthrop, president and CEO of Luminex. “It can be a critical first line of defense in virus surveillance and in separating patients with more common respiratory infections from novel cases,” he said.

Meanwhile, Boulder, Co.-based InDevR has received supplemental funding from NIH to provide FluChip kits to a limited number of state public health labs. The chips use a non-enzymatic signal amplification method based on the matrix gene version of the test being able to detect novel H1N1 and distinguish the virus from seasonal H1N1 and H3N2. The company also plans to apply for EUA clearance in September, according to Kathy Rowlen, CEO.

In May, WHO issued guidance for lab diagnosis of novel H1N1 with the notation that “molecular diagnostics are currently the method of choice” for the outbreak strain. The purpose of doing so “was to say effectively that only the molecular test can truly identify this novel strain and distinguish it from seasonal H1N1, and that remains the case,” explained Michael Shaw, PhD, associate director for laboratory science in the influenza division at CDC.

Rapid Testing Challenges

Due to technical and cost issues, most hospital labs do not have the capability to offer RT-PCR testing based on CDC’s protocols, and are unlikely to do so by the time seasonal influenza season starts. Rapid influenza diagnostic tests (RIDT), which detect influenza viral nucleoprotein antigen, have become mainstays in seasonal influenza diagnosis and the tests figured prominently in testing algorithms for novel H1N1, but not always with the best results. RIDTs offer the advantage of speedy turnaround times within 30 minutes. But there are a number of limitations, too. The RIDTs cleared by the FDA have different diagnostic capabilities. Some detect and distinguish between influenza A and B viruses; others detect both influenza A and B viruses but do not distinguish between the two, while others detect only influenza A. However, none of the currently available versions distinguish between influenza A subtypes, nor provide antiviral drug susceptibility information.

More importantly, the RIDTs are not particularly accurate, and there is quite a difference in the performance of individual tests. CDC recently evaluated three—BinaxNOW Influenza A&B, Directigen EZ Flu A+B, and QuickVue A+B—and found their sensitivities in detecting novel H1N1 ranged between 40% and 69% (MMWR 2009;58:826–829). Another recent study by the WHO Collaborating Center for Reference and Research on Influenza in Melbourne, Australia, found that in comparison to RT-PCR, the sensitivities of the same three tests ranged from 10% to 70% (Influenza and Other Respiratory Viruses 2009;3:171–176). A third analysis by U.S. researchers found that the overall sensitivity of RIDT was 17.8% when compared to molecular based detection using the RVP assay, with a considerable variation between tests. BinaxNOW Influenza A&B had a reported sensitivity of 9.6%, and 3M Rapid Detection Flu A+B, 40% (J Clin Virol 2009;45:191–195). While the specificity was relatively high (93.6%), the negative predictive value was poor (47.9%). The study also evaluated the performance of direct immunofluorescence (DFA) (46.7% sensitivity), R-Mix culture (88.9% sensitivity), and RVP (97.8%).

“Our results are probably more accurate than a formal clinical trial because they reflect what really goes on in the daily life of the clinical lab with multiple people running the assays and not undergoing the same tightly controlled situation you would have in a clinical trial,” noted lead author Christine Ginocchio, PhD, senior director of the division of microbiology, virology and molecular diagnostics at North Shore-LIJ Health System Laboratories in Lake Success, N.Y. Ginocchio’s study was based on North Shore-LIJ Health System’s experience during the first 5 weeks of the novel H1N1 outbreak. In a period of 9 weeks, the lab performed more than 34,000 influenza tests on about 13,000 patients.

The inaccuracy of the RIDTs in particular was not generally well-understood as the novel outbreak progressed, and may have lead to some sub-optimal clinical and public-health related decisions. For example, Focus Diagnostics was receiving specimens for confirmatory testing from a hospital that was using RIDT in the emergency department. “The hospital had been sending positives to us for confirmation, but they mistakenly sent some negatives. Our test determined that these were actually positives,” recalled Jay M. Lieberman, MD, medical director of Focus Diagnostics. “I talked to the head of the lab and explained that a negative test doesn’t rule out influenza. Initially the director was questioning our test, but when I showed her the published data that looked at the same assay she was using, she understood that they were missing cases." Indeed, both the CDC and Australian studies found that the RIDTs performed better when there were high levels of virus in specimens. According to CDC, “these findings indicate that, although a positive RIDT result can be used in making treatment decisions, a negative result does not rule out infection with novel influenza A (H1N1) virus.”

Relying on a negative test result more than a patient’s clinical presentation would have been more of a problem as the outbreak kicked into high gear, Lieberman added. “If the disease is extremely prevalent, the patient has classic illness, and there’s nothing else circulating, then a negative result is likely a false negative,” he said. At the bookends of the seasonal influenza season, when both seasonal and outbreak strains may be circulating, “it will be important to have more accurate, sensitive tests that can differentiate between the viruses,” he commented.

Flipping a Coin

Given the poor performance of RIDT, some laboratorians like Mahony argue that they have only a very circumscribed role in influenza diagnostics. “If you have an outbreak in an institution such as a nursing home or school, you can collect samples from five patients, for example. Because the sensitivity’s low you’re not going to detect virus in all of them. But if they’re linked and clustered together and one comes up positive and you can subsequently confirm it as a true positive, then the rapid test would have served a purpose,” he contended. “Otherwise, at 50% sensitivity you may as well flip a coin and not even do the test. I don’t think they have a role in clinical medicine.” RIDTs are not used as extensively in Canada as in the U.S., he added.

Others, including Ginocchio, believe the tests do have a place in hospital diagnostics. For instance, the emergency departments at North Shore-LIJ Health System hospitals use RIDTs. “In the 50 to 60 percent sensitivity range for the emergency department and even for admitted patients, if you pick up that many positives early on, it’s very helpful,” she said. However, North Shore-LIJ Health System’s testing algorithm does not rely solely on RIDT, regardless of a positive or negative result. At the onset of the outbreak, “we were getting everybody’s samples. If the rapid tests were positive, we had to differentiate whether it was seasonal or potentially swine flu. If it was negative, our policy always had been to use DFA and viral culture as auto-reflex tests because we knew the sensitivity wasn’t that good.”

A separate study lead by Ginocchio previously had identified that the 3M Rapid Detection Flu A+B had much higher sensitivity than BinaxNOW Influenza A&B in detecting both seasonal influenza A and B, and the North Shore-LIJ Health System was in the process of converting all local labs that serviced the hospital emergency departments to the better performing assay when the novel H1N1 outbreak hit. “We had to stop. There was no way to bring up the new test and train the staff. It was total chaos,” recalled Ginocchio. However, now that the volume of testing has significantly declined, conversion to the 3M Rapid Detection A+B assay is going forward. The testing algorithm will involve reflex testing with RVP both rapid test-negative specimens and positive specimens from patients who are admitted to the hospital. For select cases, influenza A specimens denoted as unsubtypeable by RVP will be tested by Ginocchio’s lab with New York State approval using the CDC protocol for confirmation of novel H1N1.

In recognition of the challenges with the tests, in late July CDC issued interim guidance on their use, which includes a suggested algorithm to assist in the interpretation of RIDT results (See Box, below). “It’s clear people were using a positive result in the way they should, to know whether to prescribe antiviral medications or not. The problem arose in that some were taking a negative result at face value and assuming that meant the flu really wasn’t present,” observed CDC’s Shaw. “What we’re trying to communicate is that isn’t necessarily so because the tests are highly dependent on viral load in the specimen. We wouldn’t want a negative rapid test to be used as the basis for sending a kid back to school if he was still symptomatic.” The same advisory also provided a sample explanatory comment to accompany RIDT results.

Diagnostic Algorithm for H1N1 Virus

In July, CDC issued interim guidance for the detection of novel H1N1 virus using Rapid Influenza Diagnostic Tests. The following algorithm, which was included with the guidance, is intended to assist in the interpretation of RIDT results during periods when influenza viruses are circulating in the community.

Test Result

RIDT (+) for Flu B

RIDT(+) for Flu A

RIDT (-) for Flu A and B

Interpretation

Influenza B virus infection likely

Influenza A virus infection likely. Could be novel H1N1, seasonal H1N1, H3N2, or, rarely, an influenza A virus of animal origin

Cannot rule out Influenza virus infection

Suggested
Response

Treat with antiviral agents, if appropriate.

Consider whether additional diagnostic testing and/or empiric antibiotic therapy for co-infections is indicated.*

Treat with antiviral agents, if appropriate.

Consider whether additional diagnostic testing to determine influenza A subtype and/or if empiric antibiotic therapy for co-infections is indicated.*

Use clinical symptoms, severity, and underlying disease to decide if antiviral treatment is appropriate.

Do NOT use a negative test to send a symptomatic child back to school, to rule out an institutional outbreak, or to dictate infection control measures.

Consider whether further influenza specific testing using viral culture or rRT-PCR is necessary.

Consider whether additional diagnostic testing and/or empiric antibiotic therapy for co-infections is indicated.*

*For guidance on antiviral treatment, go to the CDC website.

The guidance included a sample interpretative comment to accompany RIDT results.

RIDT result: Positive for Influenza Type A
Note: This test cannot distinguish influenza A virus subtypes. For example, this test cannot distinguish influenza infections caused by novel influenza A viruses versus seasonal influenza A viruses.

RIDT result: Negative for Influenza A and B
Note: The sensitivity of this assay has been shown to range between [10–70%**] for the detection of novel influenza A (H1N1) virus and between [20–100%**] for seasonal influenza viruses. A negative result does not exclude influenza virus infection. If influenza is circulating in your community, a diagnosis of influenza should be considered based on a patient’s clinical presentation and empiric antiviral treatment should be considered, if indicated. If more conclusive testing is desired, follow-up confirmatory testing with either [viral culture or RT-PCR**] is warranted.

** Fill in with individual clinical laboratory data and information

The key to interpreting RIDTs or other influenza diagnostics is to “know what’s going on in your community in terms of what’s circulating locally, and to make decisions based on that,” noted Lieberman. For instance, if seasonal influenza is circulating at the same time as the pandemic strain, it would be important to differentiate between them, because they have different antiviral susceptibilities. By and large the novel H1N1 strain has shown susceptibility to both oseltamivir (Tamiflu) and zanamivir (Relenza). However, a small but growing number of cases have been confirmed as oseltamivir resistant. Significantly, none of these cases have been transmitted to contacts of the infected individuals, according to Shaw. The H1N1 outbreak strain first appeared near the end of the seasonal flu season in the Northern Hemisphere, making it hard for epidemiologists to draw conclusions about its behavior in a typical seasonal flu season. So far in the Southern Hemisphere it appears to be edging-out seasonal flu and becoming the predominant strain, Shaw indicated.

The sensitivity of the various influenza diagnostics can be boosted with attention to pre-analytical factors. “The quality of specimen collection, types of swabs and viral transport media used as well as transport conditions will improve how the tests perform,” observed Ginocchio. The timing of sample collection also is crucial, since children in general shed more virus and for longer periods than adults. FDA has advised that the sensitivity of RIDTs are “generally higher in pediatric populations, and when using nasopharyngeal aspirate or nasal wash specimens.” According to Landry, a neighboring hospital uses RIDTs, but has respiratory care practitioners collect the samples to ensure higher quality specimens.

Even though Yale-New Haven Hospital has RT-PCR, DFA still is an important component of the lab’s testing algorithm. “DFA can be the best test or the worst test. It requires a lot of attention to detail,” Landry explained. “When we have DFA negatives that are detected by other methods, we go back and look at the slides and review them with the laboratorian. Was it reasonable to miss this? Why? It keeps people on their toes.” Use of a cytospin to prepare slides and a tight quality control protocol has boosted the test’s performance. During the recent outbreak, DFA detected 81% of novel H1N1 specimens that were positive by PCR. Landry also favors the test because it detects other viruses and provides results within 2 hours.

More information on Novel H1N1 Testing is available at:

  • CDC Interim Guidance for the Detection of Novel Influenza A Virus Using RIDT:
    CDC website
  • FDA Performance and Cautions in Using Rapid Influenza Virus Diagnostic Tests:
    FDA website
  • FDA Cautions in Using Rapid Tests for Detecting Influenza A Viruses:
    FDA website
  • WHO Information for Laboratory Diagnosis of New Influenza A (H1N1) Virus in Humans:
    WHO website
  • CDC Protocol of realtime RT PCR for Influenza A (H1N1):
    WHO website

Preparing for the Unknown

In gearing up for a possible resurgence of novel H1N1, labs will need to stay alert for updates from CDC and FDA, as well as implement communication strategies within their healthcare systems. For example, Landry formed an influenza testing advisory group at Yale-New Haven Hospital to review test performance and algorithms and plan for possible outbreak scenarios. Every year, Ginocchio sends a letter to all physicians affiliated with North Shore-LIJ Health System outlining the various testing options, their advantages and disadvantages.

State public health labs also will be on alert in the coming months. In North Carolina, the state lab held special conference calls with hospitals during the novel H1N1 outbreak and also relied on email alerts and Web updates. As the epidemic abated, state officials asked the hospitals what had gone well for their institutions in terms of staying on top of the fluid situation. “What they told us is it helped to have one person designated to convey guidance and updates from us and CDC. It seemed to make it go better, and helped with communication back to their own labs and institutions, as well as connecting with doctors who were part of their networks,” recalled Leslie Wolf, PhD, public health lab director.

With virology labs pushed to the limit, lab directors will do well to assess the bench skills of the entire lab, something that helped the Texas state lab wade through a veritable avalanche of cases. At the end of April, the lab performed influenza testing for about 12 specimens per day. By May 5, there were more than 1,400 per day, according to Elizabeth Delamater, PhD, manager of the microbiological sciences branch. “As part of continuity of operations planning, look into the background of all the lab staff. You might have someone in clinical chemistry who happens to have a good molecular biology background,” she advised.

Lab workers also should be in line to receive the novel H1N1 vaccination when it becomes available this fall. The CDC’s Advisory Committee on Immunization Practices included health care workers as a priority group to receive the vaccine. CDC also advised in May that diagnostic lab work on suspected H1N1 samples be conducted in BSL2 labs, and that viral culture be done in BSL2 labs with BSL3 practices. However, given the infection profile among lab workers, an update with less stringent requirements was expected in August, according to Shaw.

Regardless of the next stage of the novel H1N1 saga, laboratorians have an essential role in helping clinicians interpret results, according to Mahony. “Communication has to go on between the lab and the users of the lab in terms of the technologies being used and the performance profile of any given test. Everybody needs to be aware of the limitations of each type of test.”

Dr. Ginocchio receives research support from 3M Diagnostics and serves on the scientific advisory board and receives research support from Luminex.

Page Access: