NEW YORK – A multiplex molecular panel to detect pneumonia-causing pathogens has been shown to potentially benefit patient care, according to a handful of recent studies. In one study, the test, from BioMérieux subsidiary BioFire Diagnostics, was shown to have the potential to adjust antimicrobial therapy in approximately 70 percent of cases, sparing patients six days of unnecessary empiric treatment.
Additionally, the panel is finding a place in the COVID-19 diagnostic testing workflow in the UK.
Standard-of-care culture-based testing for pneumonia can have a long turnaround time, as well as challenges in identifying difficult-to-culture organisms, picking up co-infections, and distinguishing infection from contamination or normal flora. Standard culture techniques also don't typically look for viruses, while PCR-based tests for viruses are typically ordered piecemeal by clinicians.
Syndromic panels, meanwhile, have the advantage of more rapid test times and the ability to easily detect co-infections and genes that confer antimicrobial resistance. The tests can be quantitative or semi-quantitative, so thresholding can be employed to tease out infection from colonization. Although phenotypic antimicrobial-susceptibility testing of cultured bacteria can be somewhat more comprehensive than methods that look for resistance markers, molecular tests have the advantage of also being able to just as easily detect nucleic acids from viruses as from bacteria.
The clinical benefits of these advantages include the potential to get patients off broad spectrum empiric therapies and onto the correct drugs sooner, which can improve antimicrobial stewardship and the patient's health, while also lowering costs.
The BioFire FilmArray Pneumonia Panel, which was cleared by the US Food and Drug Administration in 2018, detects 33 common pathogens that can cause pneumonia and is authorized for sputum and bronchoalveolar lavage (BAL) samples, including endotracheal aspirates and mini-BAL samples.
A handful of studies of the panel were published last month in the Journal of Clinical Microbiology.
In one multisite prospective study, the panel was evaluated by researchers from the Medical College of Wisconsin; Ohio State, Loyola, and Indiana universities; the University of Nebraska Medical Center; Froedtert Hospital; and the University of California, Los Angeles, among others.
The study examined BAL samples from 259 adult inpatients comparing the BioFire test to routine bacterial culture for qualitative identification of the 15 bacterial targets. The panel had a combined positive percent agreement of 96 percent and a negative percent agreement of 98 percent with culture.
The panel also picked up viral targets in nearly 18 percent of specimens tested, with 39 percent of these co-infections with a bacterial target.
Importantly, after reviewing the patients' medical records the authors found that 71 percent could have had their antibiotics adjusted if the panel results had been used. This included stopping or de-escalating antibiotics in more than 48 percent of patients, which the authors calculated would spare each patient a little more than six days of unnecessary antibiotic treatment.
Blake Buchan, a clinical pathologist at the MCW and corresponding author on the study said he was "really surprised" by the antibiotic chart review data when he first saw it.
The key was that so many of the patients were receiving multiple antibiotics for presumed bacterial pneumonia that could potentially be discontinued or narrowed based on a sensitive and rapid result, he said in an email. For example, many of the patients were receiving vancomycin to cover potential MRSA infection, but very few actually tested positive for MRSA.
Buchan said that in the study there were also a number of potential pathogens — some of which were reported at a very high number of copies per mL — that were not recovered in culture.
For a majority of these samples, the patients got empiric antimicrobials prior to sample collection, which may have impacted the culturability of the samples. But, "the ability to detect and obtain semi-quantitative information regarding the potential pathogens present in culture-negative specimens provides an opportunity to modify — and typically narrow — antibiotic therapy in cases where traditional culture methods fail and informed antibiotic adjustments would not be possible," Buchan said.
In a second multisite prospective study, also published in JCM, the same team of collaborators validated the assay in analytical and clinical studies, showing 100 percent sensitivity for 15 of the targets with BAL specimens and 10 of the targets with sputum specimens. Other targets had sensitivities greater than 75 percent or were unable to be calculated due to low prevalence in the study population. The study found an overall specificity for all targets of 87 percent compared to culture, with many "false positive" results confirmed by alternative molecular methods.
In a third recent JCM study, a team at Washington University in St. Louis using the BioFire pneumonia panel was able to identify a viral or bacterial target in approximately 59 percent of samples, including Staphylococcus aureus in 22 percent of samples and Haemophilus influenzae in 14 percent, and co-detection of both a viral and bacterial targets in 4 percent of samples.
The WUSTL team also found a 99 percent correlation between the panel and standard methods for detection of viruses, and a 97 percent correlation with standard methods to detect bacteria.
While the absence of certain targets was a strike against the pneumonia panel, the team noted that, overall, it was able to identify bacterial pathogens approximately 44 hours sooner than culture, and determine resistance 56 hours sooner.
Neil Anderson, a clinical microbiologist at WUSTL and Barnes Jewish Hospital said that a particularly interesting finding was the increase in the detection of S. aureus and H. influenzae using the panel, as compared to culture.
"These were detected at low concentrations, based on the BioFire quantitation, suggesting they may have been overlooked or considered as part of normal oral flora for the culture-based approach," Anderson said in an email, noting that the clinical significance of these additional detections is not yet known.
Koby Elias, global product manager at BioFire, said in an interview that the first two JCM publications were part of the prospective clinical study to support FDA submission of the panel. BioFire had been optimistic going into the trials based on early pilot data, Elias said, but the trials demonstrated that the panel can be a "powerful tool for antimicrobial stewardship."
The frequency of viral detection in pneumonia patients, however, was a particularly striking result, Elias said.
In general, respiratory viruses have historically been an underrecognized cause of pneumonia, especially in children and older adults, according to a 2015 surveillance study called Etiology of Pneumonia in the Community, or EPIC.
The EPIC study was among the first to note that viral pneumonia might be more common than presumed, and potentially more common than bacterial pneumonia. Among 2,259 patients in the study with X-ray evidence of pneumonia, culture and standard PCR techniques detected viruses in 23 percent of cases, bacteria in 11 percent, both virus and bacteria in 3 percent, and fungal or mycobacterial pathogen in 1 percent. Importantly, the majority of cases had no detectable pathogen using standard techniques.
MCW's Buchan said that in his group's evaluations, the identification of viral agents in nearly 20 percent of specimens "really opened our eyes to these as potential infectious etiologies," particularly in patients where bacterial pneumonia had been suspected, and so, no tests for respiratory viruses had been ordered.
Specifically, Buchan noted that influenza was detected in a number of specimens that did not have a clinician-ordered test for flu. "These indicate potential missed or delayed diagnosis of viral infections, because they are not often in the differential for hospitalized or ventilated patients," Buchan said.
The INHALE trial and COVID-19 applications
The BioFire panel has been incorporated into a long-term trial in the UK called INHALE that began in 2016 to evaluate molecular diagnostics in hospital-acquired and ventilator-associated pneumonia.
Virve Enne, a clinical microbiologist at University College London and coordinator for the INHALE project, said that initially the trial evaluated pneumonia pathogen detection by comparing the BioFire FilmArray panel to the Curetis Unyvero and metagenomics sequencing using the Oxford Nanopore MinIon.
For a subsequent prospective multisite randomized control trail, however, the five-year study selected the BioFire panel, with testing done at the point of care using the panel on the FilmArray Torch systems.
"FilmArray was chosen due to having both good performance and characteristics making it suitable for point of care testing," such as fast turnaround time, ease of use, and a small footprint, Enne said in an email.
The FilmArray has now been run for about six months in 12 intensive care units across England for patients with ventilator-associated pneumonia. Enne said the most interesting result has been the revelation that, "molecular techniques are far more sensitive than routine bacteriology culture for detection of pneumonia pathogens."
In the current COVID-19 pandemic, detecting and managing bacterial co-infections will likely be even more critical, according to a recent article in The Lancet.
Indeed, BioFire's Elias said that the firm has seen increased utilization of the pneumonia panel during the pandemic.
In COVID-19 infection, secondary bacterial pneumonia appears to be a predictor of mortality, so clinicians want to know what bacterial pathogen is causing the secondary pneumonia, he said, in order to target therapy quickly. This is especially true because broad spectrum antibiotics can be toxic, so having a very sick COVID-19 patient taking them unnecessarily might be dangerous.
Risk of ventilator-associated pneumonia also tends to correlate with days spent on a ventilator, Elias said, adding that the pneumonia panel has a lot of pathogens that are regularly seen in ventilator-associated pneumonias.
In the UK, the INHALE study is now being reoriented toward detection of secondary infections in COVID-19, Enne said.
The panel has been used to date in more than 50 patients at hospitals including University College London Hospital, the Royal Free, Watford General, Chelsea and Westminster Hospital, and Liverpool University Hospital, according to a press statement from UCLH.
In the statement, INHALE's Co-Chief Investigator, David Livermore, said that although COVID-19 disrupted INHALE's original plans, it has also "created a vital medical need," namely to test if real-time information on secondary bacterial pneumonias improves treatment of the sickest, ventilated COVID-19 patients. "We've quickly refocused INHALE to do exactly that," Livermore said.
In the US, neither Buchan at MCW nor Anderson at WUSTL are using the pneumonia panel in the lab currently.
Buchan said his team had begun the process of an in-house verification study and was speaking with clinical colleagues about how best to begin to offer the test, but "Then COVID-19 struck and pretty much took over."
Anderson also said that his group is not currently using the panel, but has been engaging providers in discussions regarding how the panel should be utilized.
The WUSTL group is, however, using the BioFire SARS-CoV-2 Assay in addition to molecular diagnostic tests from Quidel, Diasorin, Danaher's Cepheid, and Roche for the virus. "We have multiple methods in place to make sure we can test adequate volumes of specimen," Anderson said.
Earlier this month BioFire also obtained Emergency Use Authorization for an updated respiratory panel, dubbed RP 2.1, that has added SARS-CoV-2 as a target.
This 2.1 panel is for emergency use, but since the firm tracks the percent positivity of the 26 other targets in the RP panels on its Syndromic Trends website, the SARS-CoV-2 target will now likely need to be incorporated, Elias said. Currently, there are no concrete plans to add SARS-CoV-2 to the pneumonia panel, he said, but the firm is evaluating the technical feasibility.