The study, conducted by Fei Xia and Kecheng Li of the Laboratory Department at Ruian People’s Hospital in Zhejiang province and published in BMC Infectious Diseases, followed 141 consecutive adults with laboratory-confirmed C. difficile infection admitted between January 2020 and December 2023. All patients received standard oral therapy, either metronidazole at 500 milligrams three times daily or vancomycin at 125 milligrams four times daily, and each underwent a follow-up PCR test between seven and twenty-one days after treatment began. From the paired measurements the researchers calculated ΔCt, the change in the tcdB cycle threshold between the initial diagnostic sample and the follow-up specimen. A rising Ct over time, and therefore a positive ΔCt, indicates that the bacterial toxin-gene load is falling as therapy takes hold, while a flat or negative ΔCt suggests that the organism persists at high levels despite treatment.

The cohort was strikingly elderly. The median age was 78 years, and nearly six in ten patients, 59.6 percent, were 75 or older, reflecting the population in which C. difficile infection does its greatest damage. Using the severity criteria of the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America, the investigators classified 101 patients, 71.6 percent, as having non-severe disease, defined by a white blood cell count of 15,000 cells per microlitre or less and a serum creatinine below 1.5 milligrams per decilitre. The remaining 40 patients, 28.4 percent, met the criteria for severe infection because their white cell counts exceeded that threshold or their creatinine reached or surpassed 1.5 milligrams per decilitre, markers of systemic inflammation and kidney injury that have long anchored clinical severity scoring in this disease.

When the researchers compared ΔCt values across these severity categories, the separation was unmistakable. Patients with severe disease had a median ΔCt of just +0.3, with an interquartile range spanning −1.0 to +1.3, meaning that the toxin B gene signal in their follow-up samples had barely budged from baseline. Patients with non-severe disease, by contrast, showed a median ΔCt of +5.3, with an interquartile range of 3.0 to 6.6, indicating a substantial decline in detectable bacterial load. The Spearman correlation coefficient between ΔCt and severity was −0.53, with a p value below 0.001, a moderate and highly significant inverse relationship: the smaller the rise in Ct, the more likely the patient was to have severe infection. Notably, the initial Ct value itself did not differ between severity groups, with a p value of 0.486. The single snapshot of bacterial burden at diagnosis carried no severity signal; it was the trajectory, the change over the first weeks of therapy, that mattered.

Exploratory receiver operating characteristic analysis reinforced this picture. For distinguishing severe from non-severe disease, ΔCt achieved an area under the curve of 0.837, with a 95 percent confidence interval of 0.759 to 0.907, a level of discrimination that approaches the performance of established clinical scoring tools. For a laboratory marker derived from a routine diagnostic assay that is otherwise discarded, this is a remarkable result. It suggests that the quantitative output of the nucleic acid amplification test, interpreted dynamically rather than as a one-off reading, could give clinicians an early, objective readout of microbiological response, complementing white cell counts and creatinine measurements that reflect the host’s inflammatory and organ response rather than the pathogen itself.

But the story turns when the endpoint shifts from severity to outcome. Fifty patients in the cohort, 35.5 percent, experienced what the investigators defined as a poor clinical outcome: either death in hospital or discharge against medical advice, a composite that captures both the ultimate failure of treatment and the pragmatic reality that some gravely ill patients leave care before it concludes. When ΔCt was tested against this composite endpoint, the association vanished. The p value was 0.217, and the area under the curve fell to 0.563, with a confidence interval of 0.458 to 0.663, a range that straddles the 0.5 line of a coin flip. A marker that cleanly separates severe from non-severe infection, in other words, does not reliably separate those who will die or abandon treatment from those who will recover and be discharged.

To understand what actually drove poor outcomes, the team turned to multivariable logistic regression, and the answer was emphatically clinical rather than microbiological. Only two factors emerged as independent predictors. Age of 75 years or older carried an adjusted odds ratio of 8.13, with a 95 percent confidence interval of 2.45 to 27.01, meaning that the oldest patients faced roughly eightfold higher odds of a poor outcome after adjustment for other variables. Admission to the intensive care unit was even more dominant, with an adjusted odds ratio of 23.21 and a confidence interval of 7.84 to 68.66. Two other factors, receipt of three or more non-CDI antibiotics and a diagnosis of diabetes mellitus, reached significance only in univariate analysis and did not survive adjustment. The bacterial load trajectory, the very measurement that tracked severity so well, contributed nothing independent once these host factors were accounted for.

The dissociation between severity and outcome is not as paradoxical as it first appears, and it speaks to a fundamental truth about C. difficile infection in ageing hospital populations. Severity criteria such as leukocytosis and elevated creatinine describe the acute physiological storm of the infection, and a falling toxin-gene burden on repeat PCR plausibly mirrors that storm’s intensity. Death and treatment abandonment, however, are shaped by frailty, comorbidity, the trajectory of the underlying illness that brought the patient to hospital, and the accumulated burden of intensive care. An 80-year-old in the ICU may clear C. difficile toxin gene signal admirably on repeat testing and still die of the confluence of age, organ failure and critical illness. ΔCt measures the microbe’s retreat; the outcome is decided largely by the host’s reserves.

The authors are careful about what their findings do and do not license. As a retrospective, single-centre study of 141 patients treated with a limited formulary at one Chinese tertiary hospital, it cannot establish causation, and the composite outcome of in-hospital death or discharge against medical advice, while pragmatic, is heterogeneous. The follow-up PCR window of seven to twenty-one days is wide, and no consensus protocol exists for serial Ct monitoring during CDI treatment, which is precisely the gap the study set out to probe. The researchers conclude that ΔCt monitoring in this window provides clinically relevant information on microbiological response and disease severity but does not independently predict clinical outcomes, and they call for prospective validation using predefined, outcome-based endpoints before any ΔCt-based monitoring strategy enters routine practice.

Even so, the study gestures toward a quietly practical shift. The Ct value is already generated every time a molecular C. difficile test runs; capturing it, archiving it, and comparing it across serial samples costs nothing beyond a change in laboratory habit and reporting. If larger prospective studies confirm that a sluggish rise in Ct flags severe disease early, clinicians could gain a cheap, objective adjunct for triaging which patients need intensified monitoring, escalation to vancomycin or fidaxomicin, or closer surveillance for complications, while reserving prognostic weight for the factors that truly drive outcomes: advanced age and critical illness. For now, the discarded number has earned a second look, but not the final word.

Subject of Research: Serial PCR cycle threshold dynamics of the tcdB gene as a marker of disease severity and outcome in hospitalized patients with Clostridioides difficile infection

Article Title: Serial ΔCt (tcdB) dynamics as a correlate of disease severity but not an independent predictor of poor outcome in hospitalized patients with clostridioides difficile infection: a retrospective cohort study

Article References: Xia, F., & Li, K. (2026). Serial ΔCt (tcdB) dynamics as a correlate of disease severity but not an independent predictor of poor outcome in hospitalized patients with clostridioides difficile infection: a retrospective cohort study. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14609-4

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14609-4

Keywords: Clostridioides difficile, PCR cycle threshold, tcdB, delta Ct, disease severity, clinical outcomes, retrospective cohort, hospitalized patients, IDSA/SHEA criteria, ICU admission, advanced age, diagnostics