Evaluating the new "sepsis tests": MeMed BV™
Part 3 of a series
There’s no diagnostic test for sepsis, which leads to high rates of overdiagnosis, delayed care from missed cases, and flawed clinical research.
Several new blood tests claim to solve the problem of diagnostic imprecision in the early management of suspected sepsis. Several have been FDA-cleared for sale and use in the U.S., and you may soon see them in your emergency department or ICU.
When they arrive, the companies selling and the hospitals buying them will celebrate their dedication to innovation in sepsis care. But you’ll be the one responsible for making high-stakes clinical decisions influenced by the tests’ imperfect performance.
This series looks under the hood at these new tests, focusing on their methods, performance, and potential impacts on care for patients with suspected sepsis.
This post focuses on MeMed BV™ and MeMed Severity™.
MeMed™
What Is It?
Based in Israel and Boston, MeMed™ produces a point-of-care blood test (MeMed BV™) that measures three proteins in the blood: TNF-related apoptosis-induced ligand (TRAIL), interferon γ-induced protein-10 (IP-10), and C-reactive protein (CRP).
MeMed BV™ was granted FDA 510(k) clearance in 2025, allowing it to be sold commercially.
How it Works
A small whole blood or serum sample (100-150 μL) is incubated onto a single-use cartridge pre-coated with specific antibodies and an enzyme. The cartridge is inserted into a proprietary tabletop immunoassay analyzer (MeMed Key™). A chemiluminescent reagent is applied to this complex, producing a glow light-emission reaction.
TRAIL and IP-10 are usually elevated in the blood of patients with viral infections, while CRP is usually elevated in bacterial infections. The analyzer measures the amount of luminescence produced by each protein, and an algorithm integrates the combination into a numeric score indexed to previously observed patterns in patients with known bacterial or viral infections. The score is stratified into one of five probability bands for bacterial or viral infection.
MeMed BV™ is considered unreliable in patients with acute gastrointestinal infections, chronic infections (hepatitis, tuberculosis, fungal or parasitic infections), or inflammatory conditions.
MeMed BV™ appears to be in use at some health systems in the U.S. (as suggested by the customer support tab on their website).
The company is also developing MeMed Severity™, a test that analyzes undisclosed biomarkers to produce a risk score for patient deterioration. MeMed Severity™ was granted FDA breakthrough device designation but is not yet commercially available.
Time to Results
About 15 minutes, according to MeMed.
How Was It Validated?
MeMed BV™ has been tested in multiple cohorts totaling thousands of adults and children since at least 2017. The two most recent validation studies are described here.
1. Halabi et al
In a trial published in 2023, 415 adults who presented with suspected pneumonia to three emergency departments in Israel in 2017-2018 were prospectively tested, with treating physicians blinded to MeMed BV™ test results.
The reference standard was adjudication of bacterial or viral infection by a panel of three physicians. They could not agree in 101 of the 415 cases (24%).
Separately, the test produced equivocal results in 10% of the cases.
The investigators removed all these indeterminate cases from the analysis.
This resulted in outstanding performance for the test: a sensitivity for bacterial infection of 98.1%, specificity 88.4% and negative predictive value 98.8%.
They did not force adjudication of the 24% of cases with indeterminate reference standards, so a more complete or accurate picture of the test’s performance could not be calculated.
Multiple MeMed employees are listed as authors, including the last author.
The incidence of sepsis or rates of ICU admission were not described, and only about one-third of patients were admitted to the hospital, suggesting milder disease.
Had clinicians been unblinded, no missed bacterial infections would have been identified, as 100% of adjudicated bacterial cases received antibiotics by the blinded clinicians.
The test misclassified only two out of ~104 bacterial infections as viral (in the selected sample); neither case had microbiological confirmation.
Authors argued that the test could have significantly reduced antibiotic overuse in the patients with adjudicated viral infections, 55% of whom received antibiotics.
2. Apollo Study
MeMed BV™ was also tested in the Apollo study (published 2024) enrolling 476 adults and children at EDs and urgent care centers in the U.S. and Israel, mostly with mild illness.
This time the adjudicator panel agreed more frequently:
372 were adjudicated as viral,
44 as bacterial,
60 indeterminate cases (13%).
With fewer excluded equivocal diagnoses, the test performed less well (reporting 90% sensitivity and 93% specificity for bacterial vs viral infection). There were only 7% equivocal test results.
MeMed BV™ missed 4 of the 44 bacterial infections (calling them viral), a 9% false-negative rate with the equivocal patients excluded. Three of the four had pharyngitis due to group A Streptococcus (strep throat) and one had periorbital cellulitis.
These were mostly urgent care-level patients in severity, with only 13% admitted to the hospital.
A later abstract mined the Apollo data to try to produce a post hoc “sepsis” cohort based on the patients with 2+ SIRS criteria, and reported very high accuracy at identifying “sepsis”. However, SIRS is no longer used for sepsis case identification, and most of their “sepsis” cases appeared to have been discharged home. Clinicians should disregard this data until better-quality prospective data are collected and published on more severely ill patients.
MeMed BV™ is also being tested prospectively in a randomized trial of 260 patients with suspected pneumonia presenting to 11 EDs and urgent care centers in Israel and the U.S., with the treating physicians unblinded to the results. This will yield information on MeMed BV’s™ influence on clinical practice.
That trial is not cluster-randomized; this produces a high risk for bias when patients receiving the intervention are treated alongside those who aren’t, usually by the same physicians. Preliminary results reported at a national emergency medicine meeting in 2024 (ACEP) stated antibiotic overuse was reduced in infections the test predicted were viral.
In an interesting retrospective study published in 2023, authors concluded that the test’s performance was superior to clinician judgment at the time of pediatric patient evaluation.
Overall, when used in lower-risk populations with the indeterminate cases excluded, MeMed BV™ had about 90% sensitivity, or about a 10% false-negative rate (i.e., missed bacterial infections).
Performance characteristics of diagnostic tests are highly dependent on the prevalence of the condition in the population studied and the cut-off values chosen for the test.
Conclusions
MeMed BV™ is an FDA-approved test that measures the levels of three proteins (CRP, IP-10 and TRAIL), quantified by immunoassay and indexed to an algorithm to predict the likelihood of bacterial vs. viral infections.
MeMed BV™ has been tested in thousands of patients, generating the largest body of clinical evidence of the next-generation infection tests reviewed here thus far. However, the cohorts tested were generally low-risk and the test does not seem to yet have been evaluated in cohorts with even a moderate prevalence of sepsis.
MeMed BV™ appears to be a potentially useful adjunct for clinicians treating lower-risk patients in the ED or urgent care settings. The test shows promise in reducing antibiotic overuse in people with viral illness. However, a false negative rate of at least 10-20% should be assumed, posing a risk for undertreatment of bacterial infections.
Its reported data almost certainly overstate real-world performance, because (like other industry players) MeMed’s reported performance is improved by the exclusion of equivocal cases. A randomized trial is currently underway to evaluate the test’s performance in clinical practice but is at significant risk for bias.
MeMed BV™ cannot yet be considered a “sepsis test,” because it has not been tested in cohorts with severe illness. Despite that fact, the company’s website pitch (including a presentation at ACEP by their director of product innovation), seems to encourage the use of the test in suspected sepsis, based on a retrospective analysis of the very much not-septic-seeming patients in its previous trials. That data has not yet been made public. The company did not respond to an email request.
Until prospective validation studies are conducted in sicker patients, it does not seem prudent to use MeMed BV™ to exclude or diagnose potentially severe infections.
MeMed Severity™: Coming Soon
MeMed’s new product, MeMed Severity™, will likely seek to address MeMed BV’s™ potential weaknesses. The severity test analyzes an undisclosed set of biomarkers to predict patient deterioration over subsequent days. Clinicians could conceivably run MeMed BV™ on all patients with suspected infection, and the severity test on those they are more worried about. The severity prediction test is not yet FDA-cleared for sale.