Carly Maucione, MD, a resident physician at Washington University in St. Louis

Carly Maucione, MD

Blending Education, Machine Learning to Detect IV Fluid Contaminated CBCs, With Carly Maucione, MD

Improved detection of IV fluid contamination in complete blood counts (CBCs) could prevent a subset of unnecessary downstream blood transfusions, whether through clinician education or machine learning, according to new research.

Findings from a multicenter machine learning study identified a surprising frequency of contaminated blood specimens, underscoring the clinical consequences of acting on erroneous laboratory results.

“I would be happy if more people read about this problem and familiarized themselves with what IV fluid contamination could look like in their patient specimens,” said study investigator Carly Maucione, MD, a resident physician at Washington University in St. Louis, in an interview with 

. “When they’re working on the floor, and they do see something suspicious, they might say, ‘I remember hearing about this. I think this might be a contaminated specimen.’ Then they do a double check before ordering blood products, or take a moment to make sure they’re not acting on a false basic metabolic panel result.”

IV fluid contamination can be difficult to distinguish from true clinical changes, particularly in hospitalized patients with multiple lines and ongoing infusions, increasing the risk of downstream clinical interventions. Since blood transfusions are commonly initiated in response to abnormal laboratory values, even small rates of undetected contamination can have meaningful clinical implications.

Maucione and colleagues identified machine learning as a potential method to standardize CBC evaluation and assist pathologists with greater precision. Within laboratory practice, there are limited techniques for detecting IV fluid contamination, and commonly used approaches, such as delta checks, are not designed to have high sensitivity for this issue.

“There’s a lot of potential, at least in the laboratory, for machine learning in areas like flow cytometry,” said Maucione. “A lot of the preprocessing and gating is really manual. It’s pattern recognition, and that’s something machine learning can do really well. That’s a hang-up in our workflow, and anything that could do it faster is where I think we should be putting our efforts.”

The retrospective, multicenter machine learning diagnostic validation study used real-world inpatient data from 2 institutions, developing and testing 2 machine learning models with areas under the receiver operating characteristic curve of 0.972 and 0.957 and areas under the precision-recall curve of 0.723 and 0.619 for identifying IV fluid contamination in CBC results.

Results from the models revealed IV fluid contamination in approximately 2% of CBCs, with investigators estimating 6-9% of subsequent transfusions may have been unnecessary.

After identifying the unexpectedly high frequency of contaminated specimens and potentially avoidable transfusions, Maucione and colleagues aimed to address the gap. She acknowledged, however, that there is no gold standard for confirming IV fluid contamination and that further validation will be necessary to ensure the findings are not attributable to other confounding factors. Improving the positive predictive value of the models would be required before using them to delay the release of blood products when contamination is suspected, although this remains a longer-term goal.

While real-time implementation remains a longer-term goal, Maucione underscored that education represents a more immediate and achievable intervention. Increasing awareness among pathologists, laboratorians, phlebotomists, nurses, and other clinicians who draw blood could help prevent misinterpretation of contaminated specimens and reduce unnecessary transfusions.

Ultimately, the study highlights both the promise and the limitations of machine learning in laboratory medicine. Although artificial intelligence may help close gaps in detecting IV fluid contamination, Maucione stressed that thoughtful integration, rigorous validation, and clinician education will remain essential to ensuring patient safety and improving care.

Editor’s Note: Maucione reports no relevant disclosures.

Maucione C, McLamb N, Zaydman MA, Pearson LN, Metcalf RA, Spies NC. Identification of IV fluid contamination in complete blood counts and subsequent unnecessary red blood cell transfusions using artificial intelligence. 

. Published online August 2026:10.1111/trf.70072. doi:

Spies NC, Farnsworth CW. Impact and frequency of IV fluid contamination on basic metabolic panel results using quality metrics. 

Videos

Study investigator Carly Maucione, MD, discusses how artificial intelligence could support laboratory workflows while emphasizing education as a near-term solution.

Uncovering Hidden IV Fluid Contamination in CBCs, With Carly Maucione, MD

Uncovering Hidden IV Fluid Contamination Through Machine Learning, With Carly Maucione, MD

Machine learning may help close the gap in identifying intravenous (IV) fluid contamination in complete blood counts (CBCs) and prevent unnecessary transfusions, according to new research.

The findings from the multicenter study suggest IV fluid contamination in CBCs may be more common than previously recognized, with machine learning models identifying possible contamination in approximately 2% of samples.

“Finding that there is this unmet need, I think, should shock some people, and I think it’s okay to be shocked,” said study investigator Carly Maucione, MD, a resident physician at Washington University in St. Louis, in an interview with 

. “The results were higher than we expected. But there’s mounting evidence that IV fluid contamination is a significant issue, and it’s something we could address in the laboratory.”

Specimen contamination with IV fluids has long been recognized as a challenge in clinical laboratories, as it can produce erroneous CBC results, influencing diagnostic interpretation and clinical decision-making.

As Maucione explained, in inpatient settings, clinicians may encounter unexplained drops in hemoglobin that later normalize, sometimes without a meaningful impact on patient management. In more concerning scenarios, however, falsely low hemoglobin values may prompt unnecessary transfusion decisions.

Still, identifying IV fluid contamination after samples arrive in the laboratory remains difficult. Detection requires laboratory technologists' expertise, with a lack of standardized methods to train pathologists. As such, recognition depends largely on a technologist’s ability to identify abnormal result patterns and infer possible causes. This approach is inherently variable and difficult to scale across high-volume laboratory workflows.

Study investigators recognized machine learning as a potential solution to standardize care and identify IV fluid contamination with precision and accuracy.

To address this gap, investigators conducted a retrospective, multicenter machine learning diagnostic validation study using real-world inpatient data from 2 institutions. The team developed and tested 2 machine learning models designed to retrospectively identify IV fluid contamination in CBC results. Since no gold standard exists for confirming contamination, model outputs were validated against expert chart review.

The models were trained using simulated IV fluid contamination scenarios. They incorporated prior, current, and post-hemoglobin concentrations, platelet counts, and white blood cell counts. Performance was then assessed using 1 year of inpatient CBC data from each institution to evaluate real-world applicability. Transfusions were only classified as potentially unnecessary when post-transfusion hemoglobin values were unexpectedly higher than pre-transfusion values and exceeded 8 g/dL, aligning with commonly used transfusion thresholds.

The models demonstrated strong discriminatory performance, with areas under the receiver operating characteristic curve of 0.972 and 0.957, and areas under the precision-recall curve of 0.723 and 0.619, respectively. Across both institutions, approximately 2% of inpatient CBC trios were classified as potentially contaminated, as previously mentioned.

Importantly, the investigators evaluated the clinical implications associated with contaminated results. Among inpatient transfusions for which a CBC trio was available, investigators deemed 6%-9% potentially unnecessary based on a conservative, rule-based definition validated through expert chart review.

“We are seeing that there are clinical consequences, and those clinical consequences are things that we can track using an algorithm such as this one,” said Maucione. “Both of those are things that we think are worth addressing and moving forward in either developing similar algorithms to detect this prospectively, using these as measures in order to track these rates.”

. Published online August 2026:10.1111/trf.70072. doi:https://doi.org/10.1111/trf.70072

Newbigging A, Landry N, Brun M, et al. New solutions to old problems: A practical approach to identify samples with intravenous fluid contamination in clinical laboratories. 

. 2024;127-128:110763. doi:https://doi.org/10.1016/j.clinbiochem.2024.110763

Machine models identified 2% of CBCs as possibly contaminated and flagged 6–9% of inpatient transfusions as potentially unnecessary, explains Maucione. 


Carly Maucione, MD

Articles