Given that standard testing at the PHE field lab in Port Loko throughout the course of the outbreak in Sierra Leone had involved the use of Altona Diagnostics' RealStar Filovirus Screen RT-PCR kit, the team used that kit as a benchmark to validate the results of the Corgenix test.
They found that the Corgenix test had 100 percent sensitivity and 92 percent specificity in detecting Ebola virus from patient blood samples. However, the team also found that the Altona assay itself performed less than optimally in the PHE lab.
"Surprisingly, the findings also revealed that the standard Altona RT-PCR test, under the conditions deployed in the field, was itself an imperfect reference standard," the authors wrote in a press release. "The Altona RT-PCR assay failed to detect a small number of EVD cases that tested positive by both RDT and by an alternative RT-PCR test, all with relatively low amounts of virus."
The results were unexpected, Boston Children's Hospital's Nira Pollock, the study's senior author, told GenomeWeb. The Altona test was not as sensitive as the team had expected, especially compared to the alternative assay, previously published by a team led by the US Army Medical Research Institute of Infectious Diseases (USAMRIID).
"We didn't expect the Altona [assay] to have this sensitivity limitation in the low viral load sample set. And we only figured that out because we did discordant analysis when we saw that some samples were rapid test-positive and Altona-negative and wanted to figure out whether these were false-positive rapid tests or false-negative Altona," Pollock said.
However, she added, the problematic results can be blamed — at least partially — on the way the test was deployed in the PHE lab, including the platforms that were selected for nucleic acid extraction and amplification, like the Cepheid SmartCycler system. "That may have led the Altona to not perform as well as it could have had it been deployed differently," Pollock said.
Indeed, said Altona's Stephan Oelschlaeger, who develops kits like the RealStar FiloVirus for molecular diagnosis of tropical and emerging diseases , the problem is that the PHE lab used the Altona assay differently than the company recommends. The SmartCycler is not on the list of systems that would produce the most accurate result from the RealStar kit. In fact, Oelschlaeger added, the difference in sensitivity when the kit is used with approved cyclers than when it's used with the SmartCycler is significant. "I am confident that it works very well on the cycler platforms that we specify in our manual," he said.
Pollock seemed to concur. "The Altona package insert specifies certain extraction conditions and certain amplification platforms that it was validated with during the approval process and the PHE lab chose to use some alternative methods," she says. "So it's possible that if the lab had used exactly what was in the package insert, it would have performed better." There's research underway to determine if that's the case, she added.
Currently, Oelschlaeger says, Altona doesn't plan to make the RealStar kit compatible with the SmartCycler system. Labs that only come equipped with SmartCyclers have other options, like the USAMRIID assay. Markus Hess, Altona's head of R&D, said getting the RealStar kit to work on the SmartCycler would not be a matter of making a simple adjustment. "We have assays specifically for the SmartCycler, but it's a different concept," he adds.
But it's important to note that field labs may not all be equipped to use certain assays the way they're recommended, according to Pollock. "Different labs make different choices," to reduce turnaround times or account for sample sizes, she says. "In this case, the PHE lab used smaller volumes of samples for nucleic acid extraction … and that may have decreased the sensitivity. These choices were made for a reason, to try to improve the overall result package in terms of accuracy, speed, and reliability. They were made deliberately to try to optimize testing in an outbreak scenario."
However, she added, the study shouldn't be taken as an indication that there's something generally wrong with PCR-based tests — which are considered the gold standard. What it does is highlight some complications with molecular-based testing in conditions like the middle of an infectious disease outbreak, and shows that there's a need for better point-of-care tests to begin diagnosis.
"It's important to recognize that molecular methods aren't perfect," Pollock says. "They have problems too, particularly in a complicated, dirty laboratory setting in the field. You can have contamination or inhibition. And when you send samples anywhere to another place to have them tested, there are all sorts of issues — not just the transport times, but also the paperwork, samples coming with wrong patient identifiers, results getting reported to the wrong person or the wrong place. And then even if you get results in the lab, you still have to get results out to the clinicians — there could be internet problems or issues with transport of the paper result file." All of these circumstances must be taken into account when considering what kind of test to use to diagnose patients.
"Everyone is always focused on test accuracy, but there's so much more than that," she adds.
For its part, Corgenix is very pleased with the study's results — the company's infectious diseases consultant, Doug Simpson, said he was surprised to see the 100 percent sensitivity. Corgenix will continue to make its rapid diagnostic kit available for any testing being done in West Africa, he adds. The company is also continually improving the test's antibodies so as to be ready if another Ebola outbreak occurs. "We need bigger, better, faster tests so that it doesn't get out of hand like it did last year," Simpson said.
Outbreak preparedness is one goal of the RealStar assay, Oelschlaeger said. In that context, it would be more likely that the kit would be used in a lab that has the recommended cyclers, which would help the assay perform at optimal levels.
The big question now, Pollock said, is how to use all of these tests to get the most accurate diagnostic information. Indeed, she added, what should be done in a large setting is to implement the Corgenix test in conjunction with confirmatory PCR. That way, researchers and doctors can learn the best testing methods to use in different stages of a potential outbreak, and throughout the years that follow. "We can't even speculate how long clinicians in West Africa are going to [need] to test for Ebola if someone presents with compatible symptoms," she said. "Maybe it's forever."