Development and Internal Validation of a Predictive Score for Nucleic-Acid Amplification Test (NAAT) Positivity in Suspected Clostridioides difficile Infection (CDI) with Glutamate Dehydrogenase (GDH)-Positive and Toxin A/B Enzyme Immunoassay (EIA)-N

Background: Clostridioides difficile infection (CDI) is a common hospital-associated infection. Patients with discordant glutamate dehydrogenase (GDH)-positive and toxin A/B enzyme immunoassay (EIA)-negative tests represent a clinically distinct population, with mortality rates comparable to their toxin A/B EIA-positive counterparts. NAAT positivity may suggest asymptomatic colonization with toxigenic strains, partially treated disease, or mild CDI with toxin levels below the assay’s detection threshold, in which the diagnosis of CDI must be clinically adjudicated. While NAAT results may complement clinical assessment, their access may be limited in specific resource-constrained settings. We aimed to develop and internally validate the Discordant CDI Score to predict NAAT positivity in GDH-positive, toxin A/B EIA-negative patients. Methods: We retrospectively reviewed the electronic medical records (EMRs) of hospitalized patients suspected of CDI with GDH-positive, toxin A/B EIA-negative results from 2019 to 2025. Twelve candidate predictors were predetermined, but their dichotomization thresholds were defined during exploratory unadjusted univariate analysis. Using a held-out internal validation with a temporal split of 70:30, multivariable logistic regression with stepwise variable selection was conducted to identify significant predictors of NAAT positivity in the training cohort. Factors were then converted into a 5-point scoring system based on their beta coefficients. The model’s calibration and discriminatory metrics were assessed, and the optimal score cut-off was selected based on the Youden index. Finally, a paired retrospective treatment-classification analysis was performed to compare model-guided decisions with physicians’ treatment decisions made prior to availability of NAAT results in the same patients. Results: Among 229 participants (160 training, 69 validation), four predictors were identified through multivariable analysis: elevated stool WBC ≥ 10/HPF (2 points), Bristol Stool Form Scale (BSFS) Type ≥ 6, diarrhea frequency ≥ 5 times/day, and abnormal WBC count (<4000 or >10,000 /µL) (1 point each). Calibration assessment suggested potential overfitting, with an optimism-corrected calibration slope and intercept of 0.791 and −0.208, respectively, in the training cohort. The apparent calibration in the validation group had a wide confidence band, while its slope and intercept were 1.15 (95% CI: 0.746–2.275) and −0.193 (95% CI: −0.938–0.835), respectively. At a selected cut-off of ≥3 points, the sensitivity and specificity were 77.8% (14/18; 95% CI: 54.8–91.0) and 80.4% (41/51; 95% CI 67.5–89.0), respectively, in the validation cohort. Score-assisted decisions were associated with a higher proportion of untreated patients compared to physician care, regardless of subsequent NAAT status (158/229, 68.9% vs. 73/229, 31.9%, respectively). While a greater number of untreated/NAAT-positive pairs were observed in the model group (13/18, 22.4% vs. 5/18, 8.6%), this difference was not statistically significant (matched OR 2.60, 95% CI: 0.927–7.299, p = 0.096). Conclusion: The Discordant CDI Score may help predict NAAT positivity in GDH-positive, toxin A/B EIA-negative patients. Several limitations, including a small number of outcome events and model overfitting, warrant further validation. While the score cannot independently establish a CDI diagnosis, its use may complement clinical decision making when access to NAAT is limited.

Authors

Institutions

Publication Details

Journal
Diagnostics
Published
2026-11-16
DOI
https://doi.org/10.3390/diagnostics16223805
Primary Topic
Clostridium difficile and Clostridium perfringens research
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Development and Internal Validation of a Predictive Score for Nucleic-Acid Amplification Test (NAAT) Positivity in Suspected Clostridioides difficile Infection (CDI) with Glutamate Dehydrogenase (GDH)-Positive and Toxin A/B Enzyme Immunoassay (EIA)-N

Asawin Sudcharoen, Nutthawut Laoarphasuwong, Chawan Kritsanaviparkporn, Piyakorn Poonyam et al.
1 links
Diagnostics
Clostridium difficile and Clostridium perfringens research
article

Development and Internal Validation of a Predictive Score for Nucleic-Acid Amplification Test (NAAT) Positivity in Suspected Clostridioides difficile Infection (CDI) with Glutamate Dehydrogenase (GDH)-Positive and Toxin A/B Enzyme Immunoassay (EIA)-N

Asawin Sudcharoen, Nutthawut Laoarphasuwong, Chawan Kritsanaviparkporn, Piyakorn Poonyam, Piyanant Chonmaitree
article en

Abstract

Background: Clostridioides difficile infection (CDI) is a common hospital-associated infection. Patients with discordant glutamate dehydrogenase (GDH)-positive and toxin A/B enzyme immunoassay (EIA)-negative tests represent a clinically distinct population, with mortality rates comparable to their toxin A/B EIA-positive counterparts. NAAT positivity may suggest asymptomatic colonization with toxigenic strains, partially treated disease, or mild CDI with toxin levels below the assay’s detection threshold, in which the diagnosis of CDI must be clinically adjudicated. While NAAT results may complement clinical assessment, their access may be limited in specific resource-constrained settings. We aimed to develop and internally validate the Discordant CDI Score to predict NAAT positivity in GDH-positive, toxin A/B EIA-negative patients. Methods: We retrospectively reviewed the electronic medical records (EMRs) of hospitalized patients suspected of CDI with GDH-positive, toxin A/B EIA-negative results from 2019 to 2025. Twelve candidate predictors were predetermined, but their dichotomization thresholds were defined during exploratory unadjusted univariate analysis. Using a held-out internal validation with a temporal split of 70:30, multivariable logistic regression with stepwise variable selection was conducted to identify significant predictors of NAAT positivity in the training cohort. Factors were then converted into a 5-point scoring system based on their beta coefficients. The model’s calibration and discriminatory metrics were assessed, and the optimal score cut-off was selected based on the Youden index. Finally, a paired retrospective treatment-classification analysis was performed to compare model-guided decisions with physicians’ treatment decisions made prior to availability of NAAT results in the same patients. Results: Among 229 participants (160 training, 69 validation), four predictors were identified through multivariable analysis: elevated stool WBC ≥ 10/HPF (2 points), Bristol Stool Form Scale (BSFS) Type ≥ 6, diarrhea frequency ≥ 5 times/day, and abnormal WBC count (<4000 or >10,000 /µL) (1 point each). Calibration assessment suggested potential overfitting, with an optimism-corrected calibration slope and intercept of 0.791 and −0.208, respectively, in the training cohort. The apparent calibration in the validation group had a wide confidence band, while its slope and intercept were 1.15 (95% CI: 0.746–2.275) and −0.193 (95% CI: −0.938–0.835), respectively. At a selected cut-off of ≥3 points, the sensitivity and specificity were 77.8% (14/18; 95% CI: 54.8–91.0) and 80.4% (41/51; 95% CI 67.5–89.0), respectively, in the validation cohort. Score-assisted decisions were associated with a higher proportion of untreated patients compared to physician care, regardless of subsequent NAAT status (158/229, 68.9% vs. 73/229, 31.9%, respectively). While a greater number of untreated/NAAT-positive pairs were observed in the model group (13/18, 22.4% vs. 5/18, 8.6%), this difference was not statistically significant (matched OR 2.60, 95% CI: 0.927–7.299, p = 0.096). Conclusion: The Discordant CDI Score may help predict NAAT positivity in GDH-positive, toxin A/B EIA-negative patients. Several limitations, including a small number of outcome events and model overfitting, warrant further validation. While the score cannot independently establish a CDI diagnosis, its use may complement clinical decision making when access to NAAT is limited.

DiagnosticsVol. 16(22)
Srinakharinwirot University (TH)
Good health and well-being
Clostridium difficile and Clostridium perfringens research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon