Diagnostic Stewardship Debate Over BioFire Meningitis-Encephalitis Panel Heats Up

NEW YORK (GenomeWeb) – It can be tricky business to determine when and how a new diagnostic technology ought to be used in order to best impact patients and costs. In a current debate over diagnostic stewardship for BioFire Diagnostics' meningitis and encephalitis molecular panel, the company and some laboratorians support unrestricted use of the test, while other lab experts have suggested that restrictions are needed, at least until more studies can be done.

BioFire published a modeling study last year that shows a potential economic benefit to unrestricted use of the ME panel assay in pediatric cases, but there have been some critiques of that study's methodology and conclusions. Meanwhile, two respected microbiologists — Jennifer Dien Bard, director of the clinical microbiology and virology laboratory at Children's Hospital Los Angeles, and Kevin Alby, the director of the clinical microbiology laboratory at the Hospital of the University of Pennsylvania — fired up attendees at the Association of Molecular Pathology meeting in November with a debate on stewardship for the ME panel.

The FilmArray ME panel detects 15 targets – six bacteria, seven viruses, and one type of yeast – from a half a milliliter of cerebrospinal fluid. It was cleared by the US Food and Drug Administration in October 2015 and it remains the only molecular panel assay for meningitis and encephalitis currently available.

Importantly, it is cleared for use on all individuals with signs or symptoms of meningitis or encephalitis, with the exception of specimens collected from indwelling CNS devices, like shunts or catheters. "There, we're talking about a different set of pathogens," Wade Stevenson, senior vice-president of global marketing at BioFire Diagnostics, a BioMérieux company, said in an interview. Otherwise, the FDA clearance and the indicated use is pretty broad, and is essentially for every patient suspected of meningitis and encephalitis. "We feel like that is justified and warranted," Stevenson said.

The model

BioFire attempted to discover the benefits of the ME panel in company-sponsored modeling study published in Future Microbiology. The study modeled the cost of the standard of care compared to using the ME panel, with and without restrictions, in a simulated pediatric population.

The study concluded that syndromic testing of all cases of suspected community-acquired meningitis and encephalitis, rather than targeted testing, would achieve the greatest economic benefits. Specifically, testing all cases led to a savings of $3,481 per case, compared to a savings of $2,157 per case if only patients with more than five white blood cells in their CSF, or plieocytosis, were tested with the ME panel.

However, in a letter to the editor of the journal, Kevin Messacar and his colleagues at the University of Colorado Children's Hospital pointed out that the modeling study used diagnostic codes to gather up the sample data used to make the model, and by doing so the authors automatically excluded patients who had CSF samples taken by lumbar puncture due to suspicion of meningitis or encephalitis, but who ultimately did not have those diagnoses.

"The rate of positivity for spinal fluid testing is exceedingly low if you don't use a diagnostic stewardship approach, meaning you have to test a lot of spinal fluids to get a few positives that may impact care," Messacar explained in an interview. Thus, if a lab were to use the ME panel on every spinal fluid sample collected for suspected ME, it could lead to a lot of unnecessary testing, he said, and the cost would be higher than the modeling study suggests.

Representatives at BioFire insisted in an interview, and in a rebuttal letter, that the study accounts for these negative tests by modeling 25 percent of patients who are given LP but not diagnosed as having ME.

But, Messacar said in a follow-up interview that, in practice as well as in a number of published studies, the proportion of positive tests is much different.

Namely, a Journal of Clinical Microbiology evaluation of a pediatric population with suspected ME co-authored by Dien Bard recorded that roughly 13 percent of tests were positive using the ME panel, while a poster co-authored by Thomas Grys at the Mayo Clinic's Phoenix campus and presented at the most recent American Society for Microbiology Microbe conference showed 6 percent positives in an adult population. A mixed pediatric and adult population studied for the firm's clinical validation of the ME panel and published in JCM revealed about 9 percent of tests were positive for a pathogen.

The model relied on administrative coding for ME, Messacar said, and admission diagnosis, or initial diagnosis, is notoriously inaccurate. Discharge diagnosis relies on confirmation of the diagnosis of meningitis or encephalitis, so would not include patients who underwent LP but had normal CSF, he said.

Basically, "The study has internal validity … amongst patients with ME diagnosis codes, but does not have the external validity to suggest that universal ME panel testing should be used on all patients with suspected ME, as the authors suggest, because this was not the group studied," Messacar said.

In an interview, Christine Ginocchio, vice president of global medical affairs at BioMérieux, highlighted that the study is limited in that it was an exploration of "a very difficult topic to address in a real clinical situation, due to the lack of comprehensive diagnostic data." She noted that the model also estimated the effect if half of the non-ME cases were incorrect diagnoses, and imposed large financial penalties for each missed diagnosis, and still showed a financial benefit.

Unrestricted use?

As Ginocchio suggested, however, the multitude of pathogens that can cause ME are quite rare. This might be why it seems to be challenging to pull together a well-powered study to prove that it should be used on all comers. 

Still, supporters of the unrestricted use of the ME panel emphasize that it covers most pathogens of concern, and it requires a small sample volume, sparing doctors and patients the need to perform multiple lumbar punctures and making it particularly advantageous for use on babies and kids, who have relatively little CSF to begin with.

A stewardship approach that prescribes testing only in cases where there is evidence of an immune response in CFS might miss cases, supporters say.

"With pleocytosis, we've seen a number of cases of viral meningitis and encephalitis where the white count was at zero," Dien Bard said in an interview. "In theory, if absence of pleocytosis were a restriction, we wouldn't have identified these pathogens, at least not as quickly as we did using the FilmArray panel," she said.

Supporters also note that the test is appropriate for labs that have never done molecular. Thomas Grys, director of microbiology at the Mayo Clinic in Arizona said that prior to adopting the panel his lab had done gram staining and then had to send out all of its PCRs to the Mayo labs in Rochester.

"To be able to get 14 answers from CSF in two hours and save the volume of specimen … that's an incredible improvement in our situation," he said. This is particularly so considering the panel is a fixed cost, compared to the cost of six or seven send-out PCRs, which also require additional specimen volume and time spent waiting on results, he said.

Thus, "It is probably the syndromic panel where it matters most what you've been doing prior to the panel," Grys said.

The panel has had a positive impact at Dien Bard's large institution, she said, but she also has colleagues in a tertiary care hospital that had no prior molecular testing in-house for CSF, including HSV PCR, "so this has been a groundbreaking change for them … they can now offer testing of multiple viral targets, but are also able to detect additional bacterial pathogens," she said.

Yet, although labs with no prior molecular testing in-house may have the most to gain, they might also be the ones at the biggest risk of contamination and potential false-positive results.

Dien Bard further noted that the FilmArray workflow is, in general, "very easy," and that this factor also may lull users into losing track of the fact that it is actually a highly sensitive molecular test. But, "As long as you're following molecular protocol and aseptic techniques, you should be able to maintain a low contamination rate," said Dien Bard. 

Her lab uses dedicated lab coats specifically for the meningitis panel, sterilizes and bleaches the workspace prior to setting up, and people performing the test are instructed to change gloves as often as possible. The tests are also set up separately, one at a time, to reduce the risk of contamination, and they are run in a biological safety cabinet. The company also recommends using a hood or dead air box for the test, as well as segregating it by using a separate instrument for respiratory panel testing, for example, particularly for pediatric cases which tend to have high titers of respiratory pathogen.

UPenn's Alby noted, for example, that BioFire specifically mentions in the test's instruction booklet that there is a concern for potential erroneous detections of S. pneumoniae, H. influenzae, or HSV-1related to contamination during the collection or handling of the sample. 

BioFire's Stevenson said this phenomenon highlights a general issue the firm has had to contend with in developing the ME panel. "One of the real big challenges about meningitis is that the concentration of actual pathogen in cerebral spinal fluid in real live infectious cases can be really low," he said. 

Accordingly, the appropriate diagnostic is one that is exquisitely sensitive, but this sensitivity "is also the Achilles heel of the meningitis test … [because] it does make it a little prone to environmental contamination," Stevenson added.

Dien Bard's lab does not restrict testing, but rather allows clinicians to order the ME panel when they deem it appropriate. However, that means in practice that it is not being ordered on every single lumbar puncture, she said.

The economic benefits of the ME panel, especially in experienced microbiology labs, remain to be determined. As a first pass, some have advocated restricting the test to patients with plieocytosis. Since the CSF is supposed to be a sterile body fluid, any evidence of blood or immune cells would theoretically point to infection.

"If there is a high index of suspicion for any one particular disease, a targeted test is potentially better," said Alby. His lab does not offer the ME panel at all. "As it is currently constructed, it does not fit the needs of our patient population," he said, considering his lab serves adult patients with more hospital-acquired rather than community-acquired ME, and his lab is skilled in PCR testing.

Furthermore, bacterial culture may be more beneficial, he said, because it is not restricted to the six species on the BioFire panel, and this may be particularly important for patients who have had neurosurgery or other invasive brain procedure and who are susceptible to a whole range of rare and unusual pathogens.

The utility of the ME panel can also depend on the pathogen, Alby said. He noted a recent case study suggesting that, for Cryptococcus – a fungal pathogen particularly inclined to infect immunocompromised people – molecular testing with the BioFire test that detects Cryptococcus neoformans/gattii may be prone to false-negative results, and a lateral-flow antigen test called CrAg-LF may be more sensitive. Indeed, "compared with culture and CrAg, the sensitivity of the FilmArray ME Panel for Cryptococcus in the [US] appears to be in the mid-60 percent range at best," the authors surmised from published data.

That case study also highlighted the caveat that performing a panel test in this particular instance — a kidney transplant patient with a severe headache — may have negatively impacted the patient. In her case, the test never picked up her low-level Cryptococcus infection. But, unfortunately, it did detect a herpesvirus 6 (HHV-6) infection that may not have been the cause of her symptoms, and she was sent home with treatment for that infection until her condition became severe.

Ginocchio, who herself had extensive lab experience and served as chief of infectious disease diagnostics at North Shore-LIJ Health System before moving to BioMérieux, also pointed out that the HHV-6, HSV-1 and -2, and varicella viruses, are examples of situations where clinicians and labs may need to brush up on the peculiarities of molecular testing of CSF.

"When someone has meningitis, they are immune-compromised to a certain degree, and these viruses can latently reactivate at extremely low levels — if you have, for instance, a high white count in the spinal fluid, you may detect a latently infected cell that may not relate to the actual clinical disease," she said, due instead to viral shedding into the CSF. "These are the things that people have to learn and understand."

More studies needed

Alby suggested more outcomes studies are needed to guide the use of the ME panel, in part because the stakes are higher with meningitis and encephalitis, and false-positive or false-negative results can have much more dire consequences than, say, with a respiratory panel.

"As your pre-test probability [of a positive result] decreases, the likelihood that any positive is a true positive also decreases," he explained, but labs might mitigate false positives, and unnecessary costs, by only testing patients who are immunocompetent, or have white blood cells in their spinal fluid, for example. But, of course, "It depends on your patient population, and that sort of approach does not work well in pediatrics" or among immunocompromised patients, he said.

On the other hand, in an immunocompetent adult population, there are some studies supporting unrestricted use, since there is a low likelihood of detecting the cause of ME in an immunocompetent patient with no plieocytosis, Alby said. This could be due to the fact that the concentration of pathogen is likely to be very low, and for viral targets in particular, the clinician essentially has to be lucky in guessing which PCRs to order.

There are no guidelines for ME panel testing currently, Alby said, but he suggested that there seems to be a general consensus that there needs to be some type of thought around the implementation of this test. "You shouldn't just take the wrapper off and go," he said, but rather labs need to think about who is being tested, who the results will be reported to, and how reports are reviewed and worded, to have the most impact. 

Dien Bard's lab is currently undertaking outcomes studies, and Messacar has recently been funded by the National Institutes of Allergy and Infectious Diseases develop an implementation strategy for the ME panel.

"What we have found in general with syndromic panel testing is that unless they are carefully implemented to direct testing towards high-yield patient populations, to ensure that the test results are appropriately responded to, they potentially could lead to more problems than good," he said. His lab's stewardship for the ME panel currently directs providers to use it for suspected ME in immune-compromised children, children with evidence of encephalitis, and all babies under two months. The latter is in part due to the fact that highest impact population in pediatric ME are infants with enterovirus or parechovirus, he said, and neither of these viruses tend to cause plieocytosis.

"If a patient doesn't meet those three criteria, then we only will run that test if they have plieocytosis greater than five cells in the spinal fluid," Messacar said. His group has shown that this will cut down testing by half without influencing the yield of case detection, and they will now study whether or not this strategy improves care and is cost effective. 

Messacar noted that his lab has found indiscriminate use of syndromic panels "leads to lots of unnecessary testing, lots of false-positive test results, and test results that the clinicians don't know how to respond to," he said. So, his lab also sends results with an interpretation and recommendation for response. 

"We think that is particularly important for this panel because there is such a low positivity rate — you would hate to test 100 spinal fluids to get one positive, and then have that one positive, say for a parechovirus, go to a provider who may not be familiar with what a parechovirus is. Then you've lost your impact of testing 100 patients because you didn't do the one extra step of making sure the provider knows what that test result means, and how to respond to it."

But for BioFire, "It’s not a question of if it should be used, it's how it should be used," Stevenson said. Ginocchio added that, "Every laboratory tests needs diagnostic stewardship," but, she said, "If the test is being used for the indications that are in the package insert, then it is being used appropriately."

Messacar, meanwhile, continues to advocate caution. "I think a carefully-thought-out diagnostic stewardship approach can really decrease the amount of unnecessary testing we do, yet still get the impact of the test — this is going to become exceedingly more important as more syndromic panels are rolled out for clinical use," he said.