NEW YORK (GenomeWeb) – The first performance comparison of point-of-care influenza A/B tests from Roche and Alere has shown lower-than-expected sensitivity for the Alere flu A target in a recent study.
While both the Roche Cobas Liat and Alere i point-of-care platforms have been previously evaluated relative to other testing platforms, they had never been directly compared. The study also found the Roche Liat test required fewer steps and less operator time.
The study was performed by researchers at the Medical University of South Carolina and led by Frederick Nolte, director of clinical laboratories and a professor of pathology.
Published online last week in the Journal of Clinical Microbiology, it used a set of 129 respiratory samples collected in universal viral transport medium. These samples were from both adult and pediatric patients, and included 80 known positives for influenza A and 16 positives for influenza B.
The study used the BioFire Film Array Respiratory Panel as the comparator test, since it happened to be the test of record used by the lab for all suspected respiratory viral infections, Nolte told GenomeWeb in an email.
The specificities and sensitivities of the Roche Liat Cobas Influenza A/B test were 100 percent for both viruses.
The Alere i Influenza A&B test, however, showed specificities of 100 percent for both viruses but sensitivities of around 71 percent and 93 percent for influenza A and B, respectively.
Nolte's lab undertook the comparison after considering replacing the lateral flow influenza immunoassays used in clinics and emergency departments at MUSC, and finding "there were no head-to head comparisons of the performance characteristics of the two nucleic acid amplification tests to inform the decision of which one we should select," he said.
As part of the investigator-initiated research, the group designed a study protocol and consulted with both Roche and Alere.
"They provided input on the study design and provided the test kits and instruments for the evaluation," Nolte said, but added that neither firm was involved in interpretation of the results or the decision to submit the research for publication.
The discrepancy in the flu A sensitivity and overall ease of use may lie in part with the sample type used for the study.
The group chose to test previously frozen patient swabs in viral transport medium rather than direct nasal swabs.
The Roche Liat test is cleared and CLIA waived only for swabs in VTM. The Alere test is cleared for both sample types but only CLIA waived for direct nasal swabs. It is considered moderate complexity for swabs in viral transport medium.
Specifically, Nolte explained that the Alere test is considered moderately complex by the US Food and Drug Administration for VTM because that sample type requires pipetting using a precision pipette not supplied by the manufacturer.
The Liat is considered CLIA waived for the same sample type since the company provides a transfer pipette with the kit.
Also, the Alere sample reservoir contains 2.5 milliliters of elution buffer and only 100 microliters of the diluted sample is used for analysis, while the Liat has no additional sample dilution, such that the entirety of the pipetted sample is used for analysis, Nolte said.
However, "We were assured by Alere that a single freeze-thaw cycle would have no adverse impact on the performance characteristics of their tests," he said.
In addition, previous published evaluations of the Alere test used frozen samples, Nolte said, and the group made sure that the samples went through only one freeze-thaw cycle and that both the Alere and Roche tests were performed concurrently.
"I don’t think the data would be any different if the samples were fresh, and consequently we do not see the need to repeat the study with fresh samples," Nolte said.
Nolte postulated several factors that could have contributed to the difference in sensitivity between the tests.
Alere has the additional dilution of the sample that Liat does not have, and Liat extracts and purifies the nucleic acid while Alere does not, "perhaps making [the Alere test] less robust," Nolte said. Finally, real-time PCR, as is used in the Liat instrument, may be inherently more sensitive than the isothermal nicking enzyme amplification reaction (NEAR) employed by the Alere test, he said.
For detection of influenza A virus, previous evaluations of the Alere i test showed sensitivities and specificities ranging from 80 to 99 percent and 62.5 to 100 percent, respectively, and from 45 to 98 and 54 to 100 percent, respectively, for influenza B virus, according to the study.
The study also reported a difference in ease of use between the two tests. Namely, the Alere test requires multiple components and steps, as well as a wait time of six minutes before the analysis begins and the operator can walk away. The Liat test has "fewer steps that can be completed by the operator in less than one minute, after which time the instrument can be left unattended," it said.
In response to the study, Norman Moore, director of scientific affairs for Alere's infectious diseases division, also highlighted that the sensitivity reported in the study was lower than that found by a number of other groups.
Furthermore, "The use of frozen samples is not supported in our package insert, and this study does not reflect the point-of-care settings and scenarios for which Alere i is intended to be utilized — where an immediately actionable diagnostic result can facilitate a meaningful and beneficial change in patient management," Moore said in an email to GenomeWeb.
As the study itself also highlights, Moore noted that the sample set was enriched for low positive samples in that it contained a number of late-stage influenza patients, and, due to the freeze-thaw cycle, these may have also undergone nucleic acid degradation.
"Therefore, the sensitivity isn’t what would be expected overall when people are walking in the door during the treatable stage of the disease," Moore said.
On the other hand, the study pointed out that the testing was performed "in a clinical laboratory under controlled conditions by a single experienced medical technologist," rather than at the point of care by non-laboratory personnel, and so may represent a "best-case scenario."
Interestingly, although Ct values are not reported by the Liat platform, Nolte was able to submit the Ct curves to Roche to obtain values. From these, it was determined that the Alere test detected about 96 percent of samples below or equal to a Liat Ct of 27 but none of the samples at or above Ct 28, suggesting the Alere test was less able to detect low viral concentrations.
The Alere test uses a combination of NEAR and molecular beacons, and as such could not provide any cycle threshold values.
Alan Wright, CMO at Roche Diagnostics, told GenomeWeb in an email that the firm hopes the study "helps demonstrate for labs and clinicians the advantages in workflow and patient care that point-of-care PCR technology offers over conventional rapid antigen and culture tests for flu."
Roche also recently received 510(k) clearance and a CLIA waiver for its flu A/B and RSV test, Wright said.
"By providing three results with one patient specimen, this multiplex assay allows the detection of more potential co-infections," he said.
Ultimately, uptake and choice of POC molecular flu testing may depend on the interplay between cost and improvement over POC rapid immunoassays.
Nolte noted that POC antigen tests cost in the range of $10 to $12, while molecular tests are between $40 to $50, depending on the volume and number of instruments needed.
"But the Medicare reimbursement for antigen and nucleic acid amplification tests are $16.33 and $115.62 [respectively], so the reimbursement would potentially offset the increased cost," Nolte said.
And although molecular testing may uncover more flu cases, "The big question is how will that impact patient care and utilization of other resources," Nolte said. "We are currently trying to get a handle on that now," he added.