NEW YORK – New rapid multiplex PCR panels for joint infections could soon offer significant improvements over culture or the off-label use of panels intended for respiratory and bloodstream infections.
BioMérieux's BioFire subsidiary recently submitted a 39-target Bone and Joint Infections (BJI) panel to the US Food and Drug Administration, while OpGen's Curetis subsidiary is developing an assay it calls the Invasive Joint Infection panel (IJI). If successfully launched, the tests could provide a quicker and more direct way to diagnose an ailment that, to date, cannot be diagnosed in a straightforward manner.
In a preview of the BJI panel at the annual IDWeek conference last month, BioFire presented for the first time extensive clinical trial data supporting its de novo FDA application, including data from 13 global sites evaluating the BJI panel on more than 1,500 patient samples over approximately two years.
Infections of bones and joints can sometimes be traced to difficult-to-culture anaerobic bacteria or antibiotic-resistant, hospital-acquired pathogens. The imperative to get a patient on the correct antibiotic treatment quickly can be stymied by time-consuming and labor-intensive bacterial culture workflows.
Fluid around a joint, also called synovial fluid, is usually collected to test for infection, or in the case of an infected prosthetic joint, the prosthesis itself is sometimes removed and can be sonicated to collect pathogen.
The standard approach then uses bacterial culture of this fluid, followed by testing for drug resistance. The process takes many days, and is less sensitive to anaerobic organisms or strains such as biofilm-forming bacteria that can thrive on prostheses.
While they wait for the culture results, patients with joint infections usually get best-guess empiric antibiotics, but if it turns out that the antibiotic choice was not ideal, the patient could suffer side effects needlessly. Meanwhile, their infection is not being adequately treated. Inappropriate antibiotic use is also thought to drive global increases in antimicrobial resistance.
With no dedicated joint infection panels available, clinical microbiology labs have pioneered other PCR- and sequencing-based approaches to overcome challenges caused by bacterial culture, including using existing multiplex PCR panels in an off-label fashion.
For example, there have been a number of studies published describing the off-label use of the BioFire Blood Culture Identification (BCID) panel and Curetis Implant and Tissue Infection (ITI) panel to detect the pathogens that cause joint infections. Shotgun sequencing has also been used in a research setting.
Meantime, both BioFire and OpGen have been developing new, dedicated panels for joint infections.
BioFire submitted the BJI panel to the FDA for de novo review in late October. Although the ongoing SARS-CoV-2 pandemic makes forecasting a bit more challenging, Jennifer Baird, global product manager at BioFire Diagnostics, said in an email that commercial launch following FDA clearance is anticipated in the latter half of 2021.
"In about an hour, the BioFire BJI Panel simultaneously targets a broad grouping of 31 causative pathogens and eight antimicrobial resistance markers associated with BJIs," Baird said. The test also uses a mere 200 microliters of synovial fluid, "which confers an advantage over competing send out tests," she said.
BioFire does not intend to seek regulatory clearance for other sample types, such as sonicates of prostheses, or bone and tissue homogenates, due to potential contamination issues. Baird said that preliminary market research suggests that the BJI panel may be used to guide surgical decision-making and streamline antibiotic selection.
Faranak Atrzadeh, chief marketing and scientific affairs officer at OpGen, confirmed that the IJI panel for the Unyvero instrument is in development, but she said the firm has not disclosed anything publicly regarding a timeline for launch or what targets will be on the panel.
That said, "It will be comprehensive and very much in line with some of the other Unyvero panels already commercially available," she said, adding that these are in the range of 30 to 50 targets and cover up to 130 individual species with broad resistance coverage of 10 or more genes with results in less than five hours.
Drilling down on BJI
At IDWeek, which was held virtually, Bart Kensinger, associate director of clinical affairs at BioFire Diagnostics, presented an overview of the firm's new 39-target BJI panel, as well as preliminary findings of a prospective clinical evaluation. The firm also presented two posters describing the analytical and clinical studies.
With 39 targets, the BJI is one of the firm's largest rapid molecular diagnostic panels to date, the largest perhaps being the 43-target BCID2 that the firm submitted to FDA in January and was subsequently cleared in March. For the AMR gene markers, the panel is also designed to report those only if there is also detection of a bacterial strain that can carry the marker, Kensinger said during his IDWeek presentation.
Baird noted in a follow-up that the BJI panel aims to improve on current culture-based diagnostics for joint infection, in part by including detection of anaerobes, such as Finegoldia magna, Kingella kingae, Parvimonas micra, Peptoniphilus species, and others. However, she emphasized that panel results are meant to be used in conjunction with other clinical, epidemiological, and laboratory data, and the BJI Panel is not intended to be used as the sole basis for diagnosis, treatment, or other management decisions.
Although joints can theoretically be infected by bacteria that are common skin flora, Kensinger said BioFire has chosen to leave two such organisms — Staphylococcus epidermidis and Cutibacterim acnes — off the BJI panel because they are also typical contaminants of the testing workflow.
Overall, BioFire claimed a 92 percent sensitivity for the panel compared to routine culture, with nearly 100 percent specificity.
Kensinger said the detection of on-panel organisms that had been missed by culture meant the panel is likely to have a higher diagnostic yield and provide more answers than routine culture.
Furthermore, "For the organisms that are on the panel, we think the benefit is that you're going to get an answer in 60 minutes, and you're also going to get an answer for a lot of the antimicrobial resistance genes ... which could potentially allow a clinician to narrow therapy," he said.
The prospective clinical evaluations of the panel enrolled patients at 10 US sites, as well as three sites in Europe — in The Netherlands, France, and Spain — between May 2018 and March 2020.
The evaluations looked at leftover synovial fluid specimens, half of which were frozen before testing, using the BJI panel at the clinical site on the FilmArray systems, as well as on the high-throughput Torch system. Results were compared to bacterial enrichment and routine culture and followed by MALDI-TOF or microscan pathogen identification. Comparator sequencing-based identification was done at BioFire's labs on frozen aliquots of the samples.
The trial involved 1,544 specimens from unique subjects. Nearly 80 percent of the specimens were collected from knee joint infections, and about half of the samples were native joint infections rather than infections of prostheses.
The BJI detected an organism in nearly 16 percent of samples, or 242 synovial fluid specimens, while culture detected an organism in approximately 13 percent, or 202 specimens. The 40 additional positive detections were of 17 different bacteria.
However, some organisms found by culture were not targeted by the panel. These included the skin flora C. acnes and coagulase-negative Staphylococci, the most prevalent species of which is S. epidermis, as well as Corynebacterium, and 12 others that were each observed once, Kensinger said.
The evaluation found approximately three quarters of BJI positives involved detection of a single gram-positive bacteria type, while nearly 17 percent were a single gram-negative bacterial strain, and nearly 2 percent of detections were a yeast organism. Assorted types of co-detections occurred in about 7 percent of specimens.
Across the board, the specificity of all of the analytes in the test was very high. Sensitivity for gram-positive organisms was "excellent for all analytes," Kensinger said, in the 90 and 100 percent ranges for most organisms excluding four analytes for which there were no positives observed. The same held true for gram-negative bacteria, with the exception of non-detected targets and organisms for which there were a low number of observations in the study.
For the two yeast targets, there were only a few cases and Kensinger attributed a reported sensitivity in the 60 percent range to this circumstance.
The firm also investigated all the discrepant results with additional PCR. It found that 76 of the 79 cases where the panel detected an organism, but the reference method did not, actually had the positive confirmed with PCR, and only three were unresolved. Among the 20 cases where the reference method found a pathogen but the panel did not, 14 were confirmed by additional PCR, two were misidentified by the lab, and four were unable to be confirmed.
"Our observed ratio of false positive to false negative results suggests and increased diagnostic yield of the BioFire BJI panel compared to routine culture," Kensinger said.
Preliminary performance data for the AMR genes also showed 100 percent sensitivity, albeit with few detections and only four resistance genes represented in the sample the vast majority of which were methicillin-resistant S. aureus (MRSA).
Unclear clinical utility
Microbiologists and infectious disease specialists with expertise in diagnosing joint infections were cautious in appraising the new panels, for which there has yet to be independent validation.
Laura Certain, an infectious disease specialist at the University of Utah, said that in general she is glad that companies are working to improve diagnostics for prosthetic joint infections. "They are hugely challenging to treat, and part of successful treatment is using antibiotics appropriately," she said.
But, she added, "It remains to be seen whether having a result faster than culture is beneficial to these patients."
Certain, who is also section chief of joint infections at the Salt Lake City Veterans Affairs Hospital, presented at IDWeek on sequencing-based approaches to prosthetic joint infections specifically, including approaches like 16s PCR-based detection, multiplex PCR panels, and metagenomics.
In her presentation, she noted that there is significant mortality associated with joint infections, and prosthetic joint infections, in particular, are on the rise. By the age of 80, approximately 10 percent of people are expected to have a prosthetic joint of some kind, and 1 to 2 percent of these will become infected, she said. Unfortunately, up to one quarter of infections are culture-negative.
Although sequencing-based methods can potentially identify more pathogens faster, the problem with using existing panels off label is that these panels were not designed for this purpose and they don't seem to work terribly well, Certain said.
OpGen's Atrzadeh concurred. She noted that the BCID and ITI panel do have bacterial targets that can also be present in joint infections, and researchers and labs, particularly outside of the US, "will do anything and everything that they want to do" in their labs. Yet, "Those panels are just not optimized or intended for use with synovial fluid," she said. "Any data or conclusions drawn from those studies really are rather irrelevant."
Indeed, in PCR-based molecular diagnostics, each panel has a very specific specimen type indication, Atrzadeh said. "The ITI panel is a tissue infection panel; our synovial fluid-specific indication will be the joint panel that is currently in development." The IJI panel will have different set of targets, "with an appropriate set of analytes that are intended specifically for joint infections," she said.
Baird at BioFire also commented that the BJI panel is optimized for synovial fluid, much the way that BCID panels are optimized for positive blood culture samples. "We cannot recommend the use of off-label sample types on either panel," she said.
Certain presented some data showing that metagenomic approaches seem to fare somewhat better than off-label panel use. One study showed a low false-positive rate and good match with culture, for example.
Robin Patel, a professor of microbiology and medicine at the Mayo Clinic, is an expert in sequencing-based approaches to joint infections and said her lab routinely uses both the BioFire BCID off-label and next-generation sequencing in its clinical diagnosis.
Previously, the group studied off-label use of BCID for prosthetic joint infections, and found that the overall sensitivity of BCID was 53 percent, compared to 69 percent for culture. The panel specificity, meanwhile, was 99 percent, and the team concluded the BCID could potentially be a useful adjunct for diagnosis.
The lab is now using the BCID panel clinically to diagnose bone and joint infection with an improved workflow, Patel said in an email.
The Mayo team has also validated an approach using metagenomic sequencing performed on Illumina systems, as well as an informatics pipeline from CosmosID for prosthetic joint infection. Last year, a team in the UK also used Oxford Nanopore sequencing to identify pathogens in prosthetic joint infections.
In general, Patel said that sensitivity and specificity need to be considered in conjunction with organism burden in the infection. "Having an assay with a low limit of detection can be especially relevant, but that may bring with it issues of specificity, especially for microorganisms that may be part of the normal skin microbiota," she said. "There are also issues of turnaround time, ease of use and cost to be considered."
Indeed, OpGen's Atrzadeh said that while sequencing and PCR may essentially do the same thing, "It is really comparing apples to oranges."
Access to sequencers in clinical labs is typically limited, they tend to be more expensive, and generating and interpreting sequencing data generally requires specialized expertise.
"For a clinical lab, PCR panels are a better choice because there is more familiarity, they are more accessible, and more easily adoptable, and most if not all of them at some point go through FDA clearance," she said, adding that they are arguably more cost effective as well.
Certain said the true test of any multiplex PCR or metagenomic analysis will be whether tailoring antibiotics to the identified pathogen improves outcomes over empiric antibiotics in cases in which cultures were negative but the sequencing-based detection was positive.
"Until they demonstrate that benefit, the role of such diagnostics is unclear," Certain said.
Studies to assess outcomes associated with the BJI panel as compared to standard of care are currently underway, BioFire's Baird said.