Skip to main content
Premium Trial:

Request an Annual Quote

MeMed Study Indicates ImmunoXpert Outperforms Existing Infectious Disease Tests


NEW YORK (360Dx) – Israeli biotech firm MeMed has published a study indicating that its ImmunoXpert infectious disease diagnostic outperforms existing commonly used tests for distinguishing between bacterial and viral infections.

In a paper published last month in the European Journal of Clinical Microbiology & Infectious Disease, the company and its collaborators presented data showing its test could determine whether a patient had a viral or bacterial infection with a specificity of 94.3 percent and a sensitivity of 93.5 percent. The test measures the levels of three proteins to determine whether a patient’s infection is viral or bacterial.

That bested the performance of commonly used measures, including C-reactive protein, procalcitonin, interleukin-6, human neutrophil lipocalin, white blood cell count, absolute neutrophil count, and clinical prediction tools.

For the prospective study, the researchers analyzed samples from 493 patients at two Israeli medical centers who presented with respiratory infections or fever of an unknown source. Patient samples were tested for the measures cited above, and the patient cases were referred, after their resolution, to a panel of three experts who decided whether the infections had been viral or bacterial, based on the patients’ records during the course of their treatment. Patients were only included in the study if the panel was unanimous in declaring their case to be either viral or bacterial. This process left the researchers with 314 patients, 174 of whom had viral infections and 139 or whom had bacterial infections.

While certain measures outperformed the ImmunoXpert test in terms of sensitivity, they had much lower specificity. Overall, the ImmunoXpert test had an accuracy of 94 percent. The next best performing test, a combination of CRP and procalcitonin, had an accuracy of 88.5 percent, while a lab score combining CRP, procalcitonin, and urinalysis had an accuracy of 83.1 percent.

The test also correctly identified 50 of 57 patients who were ultimately determined to have viral infections but who were initially given antibiotics, which suggests it could significantly reduce antibiotic overuse, said Kfir Oved, MeMed’s co-founder and CTO. The study did not look at the extent to which the other tests used could have reduced unnecessary antibiotic use, though.

MeMed intended the study to look at “what is actually the added value we can bring compared to what is out there,” Oved said. ImmunoXpert has the CE-IVD mark and MeMed is marketing it on a limited basis in the EU, Switzerland, and Israel. It has sold more than 10,000 tests to medical centers through an early-access program.

ImmunoXpert measures the levels of three proteins – TRAIL, IP-10, and CRP – in blood using immunoassays. One part of the test design that Oved said likely contributed to its higher performance compared to existing measures is that the three proteins include both analytes that reflect the host response to bacterial infection and analytes that reflect the host response to viral infection.

Many markers for distinguishing between the two infection types have focused on the bacterial side of the equation, Oved said. “As such, they had reasonable specificity usually, but sensitivity could be a problem. If the level [of the marker] is high, then, great, it is bacterial. But if the level is low, you don’t actually know.”

Including markers of both viral and bacterial infection in the same panel “allows us to complete the picture from two opposite directions, and that adds information and accuracy,” he said.

Nathan Ledeboer, medical director of clinical microbiology at the Medical College of Wisconsin, said that tests capable of distinguishing between viral and bacterial infection are of great interest to the field.

“There’s a tremendous need,” he said. “We need very high sensitivity, high specificity tests in order to make a decision that is a bit more informed about whether to start a patient on antibiotics or not.”

Ledeboer, who is familiar with MeMed’s efforts but has no relationship with the company, said the comparisons made in the study were useful ones, noting that, in particular, a test like ImmunoXpert needs to demonstrate advantages compared to CRP and procalcitonin.

“One of the things that all of these tests are judged upon is [comparison to] simple CRP, simple procalcitonin,” he said. “And so if you want to justify doing one of these additional tests, you need to show incremental and usable improvement over a simple CRP or simple procalcitonin.”

Ledeboer also noted, as did the EJCMID paper, that several tests measuring host nucleic acids to distinguish between viral and bacterial infection are also under development, though the study did not compare ImmunoXpert to those assays.

Key to further demonstrating the test’s usefulness will be showing its efficacy in larger, more diverse cohorts, Ledeboer said, noting that this has been a problem for markers like procalcitonin.

“Procalcitonin worldwide has had fairly variable roll-out,” he said. “Initially, there was a lot of real promise associated with [using it for detecting] sepsis. Since then, it seems to have found its home a bit more in the respiratory space. But whether people use procalcitonin or not really comes down to the question of what study you believe.”

Varying performance in different populations and across different pathogens is part of “what has tripped up procalcitonin,” Ledeboer said. “So one of the key components to all of the studies is really understanding how diverse of a population you can validate against. What diversity of different pathogens can we validate against?”

He said that the EJCMID study did a nice job of looking at a variety of different infection types but that the cohort was relatively small and not very geographically diverse. Looking at MeMed’s overall validation efforts, he said they had done well including a diversity of infection types and a patient populations.

In all, MeMed has tested the assay in clinical studies comprising around 3,000 patients and is currently involved in around ten additional trials. The company is also in discussion with the US Food and Drug Administration around setting up clinical trials in the US to support an FDA submission.

In addition to validating ImmunoXpert, MeMed is currently working on a point-of-care device and version of the test that could be run in as little as 15 minutes, with the goal of targeting hospital emergency departments and, ultimately, getting it into outpatient settings like doctors’ offices. Oved said the company hopes to begin using the device in Europe around the start of next year.

While noting that “everybody is trying to build point-of-care,” Ledeboer said he saw some downsides to the approach.

“The promise to it is certainly more timely results,” he said. “But the problem with it is, you often get a lot more variability.”

More generally, he said, clinicians are likely to be more resistant to a test that requires purchasing a new piece of hardware than one that runs on their existing equipment.

“I tell manufacturers this all the time,” he said. “If you can avoid building a new box and requiring us to buy a new box, that’s the way to go.”

That said, there are advantages to having your test on a proprietary platform, Ledeboer acknowledged. “You get the revenue from the box, though that usually isn’t much,” he said. “But you control your destiny. [Y]ou're not dependent upon somebody else selling for you. You're not dependent upon licensing agreements. It's a better way to protect your IP. All those kinds of things.”

Additionally, some tests need more sensitive and less variable measurement of their target analytes than standard laboratory analyzers can provide, Ledeboer said. “That may certainly be the case here.”

Indeed, in a previous interview, MeMed CEO Eran Eden said that the low abundance of one of the ImmunoXpert protein markers presented difficulties for existing platforms they looked at.

According to MeMed, its POC platform is able measure proteins reliably at the picogram per milliliter level. The company has also said it plans to use it for POC assays beyond the ImmunoXpert test, but has not specified yet what those might be.