Hospital cancer units are among the most dangerous places in medicine to catch an infection. The very treatments that give leukemia and transplant patients a chance at survival also strip away the immune defenses that would normally keep opportunistic pathogens in check, and few of those pathogens are as pervasive or as stubborn as Clostridioides difficile. A new study from researchers at North Carolina State University, working with collaborators at Washington University School of Medicine, Vanderbilt University Medical Center and the University of Tennessee, has now mapped in unusual detail how this bacterium moves through oncology wards, and the results upend a long-standing assumption about who is actually spreading the disease.
The study, published Oct. 1, 2026 in the journal Infection Control and Hospital Epidemiology, combined six months of intensive patient surveillance with a sophisticated computational model of two leukemia and hematopoietic cell transplant units in a large US tertiary care hospital. The central finding is stark: patients who were visibly sick with C. difficile infection were not the main engine of new cases. Instead, patients who carried the bacterium silently, without any clinical signs, were responsible for the overwhelming majority of transmissions that produced new colonizations. According to the model, asymptomatic colonized patients accounted for 92 percent of transmission events that resulted in new colonized individuals on the units.
C. difficile is one of the most common hospital-acquired infections and a leading cause of healthcare-associated diarrhea. It thrives when the gut microbiome has been disrupted, typically by antibiotics, and it spreads through spores that can persist on surfaces and equipment for months. For cancer patients, particularly those undergoing treatment for blood malignancies or recovering from hematopoietic cell transplants, the stakes are exceptionally high. These patients have among the highest incidences of C. difficile infection of any hospital population, and a secondary infection on top of immunosuppressive therapy can be fatal.
“Physicians have singled out antimicrobial resistance as a big issue for improvement in cancer patients,” says Cristina Lanzas, professor of infectious disease at North Carolina State University’s College of Veterinary Medicine and corresponding author of the research. “The cancer treatments are improved, but the secondary infections patients get because they are so immunocompromised can be fatal. We decided to look at how C. diff infections were transmitted in oncology wards.”
The challenge the team faced is one that has long frustrated infection-control researchers: transmission is invisible. Standard clinical practice tests patients for C. difficile only when they show symptoms such as diarrhea, which means that people who carry the bacterium without signs of illness slip through undetected. A positive or negative test result on a given day is a snapshot, not a story, and it says nothing about who infected whom. “Testing data only provides a snapshot – are you positive or negative on a given day and time – but transmission events are invisible,” Lanzas explains. “We developed models that traced potential transmission routes and then looked at each model’s ability to replicate actual infection data. Then we chose the model that aligned best with the data we had.”
To make the invisible visible, the researchers built a stochastic, individual-based network model, a type of simulation that tracks every patient, every room and every healthcare worker connection on the two units. Rather than averaging over a population, the model simulated each individual’s state, whether uncolonized, asymptomatically colonized or actively infected, and updated those states over time based on the contact structure of the ward. Healthcare worker room assignments provided the backbone of the network, defining which patients were potentially linked through shared staff, equipment and physical space. Key parameters of the model were estimated using a statistical technique called particle filtering, which iteratively adjusts the simulation until its outputs match the observed data as closely as possible.
The data used to calibrate the model came from an unusually thorough surveillance program. Collaborators at Washington University School of Medicine collected cultures and nucleic acid amplification testing, or NAAT, from patients both at admission and weekly throughout their stay, regardless of whether symptoms were present. Toxin enzyme immunoassay tests were used to diagnose active C. difficile infection. This active surveillance, testing everyone whether sick or not, is what allowed the researchers to distinguish between patients who were merely colonized and those with clinical disease, a distinction that routine testing cannot make.
The model’s output was revealing in several ways. First, it showed that routine, symptom-driven testing alone captures only 23 percent of patients carrying C. difficile, leaving the vast majority of carriers undetected. Second, it quantified the origins of new colonization events: 51 percent of new colonizations, with a range of 38 to 70 percent across model runs, were due to importation, meaning the patients arrived at the unit already carrying the bacterium. Strikingly, 87 percent of those imported cases were admitted as asymptomatic carriers, not as diagnosed infections. The remaining 49 percent of new colonizations, ranging from 30 to 62 percent, arose from transmission within the unit itself, and asymptomatic colonized patients were the source in 92 percent of those transmissions.
Perhaps the most encouraging implication concerns patients with active, symptomatic infections. The model indicated that infection control measures focused on diagnosed CDI patients, such as isolation and contact precautions, appear to be working: symptomatic patients contributed relatively little to onward transmission. The problem lies elsewhere, in the silent reservoir of carriers who are never isolated, never flagged and never treated, because nobody knows they carry the organism. These findings reinforce the role of asymptomatic colonization, particularly through admission importation, in sustaining C. difficile transmission in high-risk hospital settings.
“The issue of people carrying pathogens without knowing about it is very common with all antimicrobial resistant pathogens – we can really be our own worst enemies,” Lanzas says. “But by modeling these events, we can make the invisible visible, and perhaps that can lead to new protocols for patients coming into these wards.” The researchers hope the modeling framework can be adapted to other pathogens, helping hospitals identify the specific contact points where infection risk is highest and design targeted interventions for the patients who need them most. In wards where the population is uniquely vulnerable, knowing where an outbreak will come from before it starts may prove as valuable as any drug.
Subject of Research: Transmission dynamics of Clostridioides difficile in hospital oncology units
Article Title: Model traces C. diff spread in hospital oncology units
Article References: Model traces C. diff spread in hospital oncology units. (n.d.). Original publication
Image Credits: AI Generated
DOI: Not provided
Keywords: Clostridioides difficile, hospital-acquired infection, oncology, asymptomatic colonization, transmission modeling, infection control, immunocompromised patients, stochastic simulation, healthcare epidemiology, leukemia, stem cell transplant, antimicrobial resistance
News Source: Nathaniel Bowman. (October 10, 2026). Asymptomatic patients drive most C. diff spread in hospital cancer wards, model finds. Scienmag.
