The Last Mile In Preventing A Deadly Hospital Infection

By Deborah Borfitz

December 6, 2023 | The “power of the team approach” was on full display for a study designed to understand the contribution of asymptomatic carriers to the spread of Clostridioides difficile in hospitals, according to Vincent Young, M.D., Ph.D., a member of the departments of internal medicine/infectious diseases and microbiology & immunology at University of Michigan Medical School. He was part of a cross-disciplinary team who has found that even the most aggressive infection prevention and control measures have their limitations, since infections aren’t necessarily a consequence of person-to-person transmission. 

They can result from dormant conditions associated with the patients themselves. That’s the inescapable conclusion of a nine-month investigation where researchers analyzed daily fecal samples from nearly 1,200 patients admitted to the intensive care unit (ICU) at RUSH University Medical Center in Chicago. (Nature Medicine, DOI: 10.1038/s41591-023-02549-4).  

Just under half of the 425 isolated C. diff strains were genetically distinct from one another, and only six genomically supported transmissions occurred over the study period. About 9% of patients had C. diff detected in at least one fecal sample and, despite posing little risk to others, they themselves had a 24-times greater risk than noncarriers of developing full-blown infection. 

The surprising findings emerged from an advantaged, high-resource unit equipped with all-private patient rooms routinely cleaned with a disinfectant that kills C. diff spores, which is not the case at all hospitals even in high-risk areas like the ICU, says RUSH infectious disease specialist Mary Hayden, M.D. Reaching the last mile with C. diff will take as-yet unknown steps beyond rigorous hand hygiene and environmental cleaning. 

Practical Issues

The longitudinal genomic surveillance study was an add on to a broader year-long project, funded by the Centers for Disease Control and Prevention (CDC), where Hayden was analyzing microbiome features predictive of the acquisition of multi-drug-resistant organisms. She says it was a “fantastic opportunity” to study patient-to-patient transmission in more detail and to work with Young and his colleague Evan Snitkin, Ph.D., an expert in computational biology and genomics. 

Because the study required the enrollment of most patients in the ICU, and rectal swab sampling was a routine infection prevention practice, researchers were given waiver of written informed consent by the institutional review board at RUSH, says Hayden. But they did need to get verbal consent from participants and provide them with a description of the study and post those details in every patient’s room along with who to contact with questions. Very few patients opted out, she adds. 

For context, it is thought that 3% to 5% of the general population outside of a healthcare setting has C. diff in their gut—either from birth or acquired via accidental ingestion—and about half of those isolates don’t produce toxins damaging to cells lining the intestines, Young says. But patients landing in the ICU tend to have had a lot of prior contact with the medical system with differing infection prevention practices as well as a much higher likelihood of receiving an antibiotic due to a severe or life-threatening infection. 

The bacterium lives in the intestines of humans and animals and spreads through their feces into the environment. Why some individuals get colonized, and others don’t, is unknown, says Young. It is also perplexing why the colonization rate isn’t higher given the volume of antibiotics people take that spawned the antibiotic resistance crisis.  

Preventing Transition

If results of the study are generalizable and reproducible, discussions and policies around infection prevention practices may need to change, says Hayden. At the very least, it raises the question about where efforts should be directed. 

At hospitals like RUSH that are already doing everything they can to prevent most cross-transmission of C. diff, next steps might be “an even greater focus on antimicrobial stewardship,” she says. The Centers for Medicare & Medicaid Services might also want to reexamine its policy of penalizing hospitals for hospital-acquired C diff infections over which they have limited control. 

The fact that nearly one-third of the 425 C. diff strains identified in the study were genetically distinct patient to patient suggests that the process of transitioning from colonization to infection may happen in different ways, says Snitkin. “We picked up one organism that causes one disease, but the genome sequence of the isolates... can vary by as much as half [their] genes.” This is perhaps unsurprising, given that the common ancestor for all the different phenotypic strains has existed for tens of millions of years. 

Prophylactically treating patients colonized with C. diff, before disease development, is a controversial practice limited largely to bone marrow transplant populations, says Hayden. Studies on the practice have yielded mixed results where vancomycin has sometimes caused more disruption to recipients’ gut microbiota. New approaches are clearly needed, including perhaps fecal microbiota-based live biotherapeutics. 

Snitkin’s colleagues at the University of Michigan (Jenna Wiens and Krishna Rao) are currently leading a study using machine learning to predict which patients are at risk of developing a C. diff infection without knowing their colonization status. Tactics that include “reverse isolation,” to protect them from others, are then employed in hopes of preventing them from transitioning to infection. More targeted interventions can’t confidently be adopted until the many unknown surrounding the colonization-to-infection process are learned, he says.  

A series of papers produced by infectious disease and machine learning specialists at the University of Michigan previously demonstrated via a retrospective analysis the possibility of using only the electronic medical record of patients to predict with seven-fold increased accuracy their risk of developing C. diff infection in the next five days, which might be enough time to intervene, adds Young, pointing to a 2018 article that published in Infection Control and Hospital Epidemiology (DOI: 10.1017/ice.2018.16). That has since been done prospectively throughout Michigan Medicine, the University of Michigan's academic medical center, and when patients cross that risk threshold all manner of infection prevention is employed—from reverse isolation involving soap and water handwashing before entering a susceptible patient’s room to a higher level of pharmacy review to ensure the medications they’re given aren’t increasing their susceptibility to C. diff.  

“The spores of C. diff are the form that gets transmitted from patient to patient and are resistant to killing by ethanol-based hand sanitizers,” Young notes. 

Pathways of Transmission

Microbiota restorative therapies are being advanced that seek to interfere with the C. diff colonization-to-infection transition, reports Snitkin. Michigan Medicine has also turned to sequencing-based approaches to look at clinical exposures of patients and changes in their microbiota over time, and at scale, in hopes of discovering causal factors for the colonization-to-infection transition. “It is a highly clinical question that needs a great deal more study in different populations in different healthcare settings to identify appropriate interventions.” 

Artificial intelligence is taking the entire infectious disease field by storm, says Young. Computers are good at taking large-scale data generated from day-to-day care of a multitude of patients and homing in on what is in fact “the right therapy for the right patient at the right time”—the holy grail of personalized medicine. 

It takes a village, he stresses. The latest study pulled from the know-how of top-notch epidemiologists, infection preventionists, pharmacists, microbiologists, computational biologists, genomic scientists, and basic scientists working with animal models. It’s the same scenario with machine learning—“you need a lot of different inputs to be able to come up with a strong output.” 

It’s an exciting time to be in the infection prevention field, says Hayden. A “new era” has arrived where it is possible to start teasing out with much greater precision the pathways of transmission for not only C. diff but also for microbes that cause other healthcare-associated infections such as bloodstream infections and pneumonia.  

C. diff is the culprit of almost half a million infections in the U.S. each year, according to the CDC. A total of about 1.7 million healthcare-associated infections occur annually, resulting in 99,000 deaths.