Clinical characteristics of invasive infection in Klebsiella pneumoniae liver abscess and development of a simplified risk classification model: a single-center retrospective study

Objective To compare the clinical characteristics of patients with Klebsiella pneumoniae liver abscess (KPLA) and non-KPLA patients, to describe the phenotype of invasive infection in patients with KPLA, and to develop and internally validate a simplified risk-classification model for identifying patients with invasive complications. Method A total of 300 patients with bacterial liver abscess were retrospectively included, including 246 cases of KPLA and 54 cases of non-KPLA. Categorical variables were compared using the χ 2 test or Fisher's exact test, and continuous variables were compared using the Mann–Whitney U test. With the occurrence of invasive infection in patients with KPLA as the outcome, a binary logistic regression model was constructed comprising three variables: maximum body temperature, lactate, and procalcitonin (PCT). Because peak values during hospitalization were used as predictors, the model was positioned as a cross-sectional risk-classification tool rather than a prospective prediction instrument. Internal validation was performed using 500 bootstrap resamples and repeated stratified 5-fold cross-validation for 50 iterations, and the model's discrimination, overall classification error, calibration, and decision curve performance were evaluated. Result KPLA accounted for 82.0% of all cases (246/300). Compared with non-KPLA patients, those with KPLA were younger, had a higher prevalence of diabetes, and had lower rates of malignancy, mixed infection, and any extrahepatic metastatic infection. The composite invasive outcome (bacteremia and/or extrahepatic metastatic infection) occurred in 46.7% (115/246) of KPLA patients, including 82 (33.3%) with bacteremia and 35 (14.2%) with extrahepatic metastatic infection (EMI). In the invasive group, the median PCT and lactate levels were 31.0 ng/mL and 2.40 mmol/L, respectively. In the non-invasive group, they were 1.50 ng/mL and 1.20 mmol/L, respectively. The differences between the two groups were statistically significant (both P < 0.001). In the final model, the adjusted ORs for highest body temperature, lactate, and PCT were 2.70 (95% CI 1.56–4.67), 2.77 (95% CI 1.52–5.07), and 1.10 (95% CI 1.06–1.14), respectively. Repeated cross-validation showed that the model achieved an AUC of 0.932 ± 0.003, a Brier score of 0.105 ± 0.002, and a calibration slope of 0.91 ± 0.04. Exploratory decision curve analysis suggested that the model may provide positive net benefit across the threshold range examined. Conclusion In this cohort, KPLA and non-KPLA differed in host background and clinical characteristics. A simplified three-variable risk-classification model incorporating maximum body temperature, lactate, and PCT demonstrated stable performance in internal validation for identifying patients with invasive complications (defined as bacteremia and/or extrahepatic metastatic infection). If validated prospectively, this model may facilitate early identification of high-risk patients who could benefit from intensified surveillance, timely adjustment of antimicrobial therapy, and proactive screening for extrahepatic metastatic foci. However, because peak values during hospitalization were used, a temporal relationship between predictors and outcome onset cannot be established, and the model should be interpreted as a cross-sectional risk-classification tool rather than a prospective prediction instrument. Its clinical applicability requires further investigation, including anchoring of predictor measurement to a standardized time point, harmonization of outcome definitions, and independent external validation in multi-center cohorts.

Authors

Institutions

Publication Details

Journal
Frontiers in Medicine
Published
2026-11-17
DOI
https://doi.org/10.3389/fmed.2026.1944558
Primary Topic
Amoebic Infections and Treatments
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Clinical characteristics of invasive infection in Klebsiella pneumoniae liver abscess and development of a simplified risk classification model: a single-center retrospective study

Xiuping Zhan, Yafei Bu, Xueying Wang, Yannan Xie et al.
4 links
Frontiers in Medicine
Amoebic Infections and Treatments
article

Clinical characteristics of invasive infection in Klebsiella pneumoniae liver abscess and development of a simplified risk classification model: a single-center retrospective study

Xiuping Zhan, Yafei Bu, Xueying Wang, Yannan Xie, Xiaomin Zhang, Xuechao Zhang, Cunji Li
article en

Abstract

Objective To compare the clinical characteristics of patients with Klebsiella pneumoniae liver abscess (KPLA) and non-KPLA patients, to describe the phenotype of invasive infection in patients with KPLA, and to develop and internally validate a simplified risk-classification model for identifying patients with invasive complications. Method A total of 300 patients with bacterial liver abscess were retrospectively included, including 246 cases of KPLA and 54 cases of non-KPLA. Categorical variables were compared using the χ 2 test or Fisher's exact test, and continuous variables were compared using the Mann–Whitney U test. With the occurrence of invasive infection in patients with KPLA as the outcome, a binary logistic regression model was constructed comprising three variables: maximum body temperature, lactate, and procalcitonin (PCT). Because peak values during hospitalization were used as predictors, the model was positioned as a cross-sectional risk-classification tool rather than a prospective prediction instrument. Internal validation was performed using 500 bootstrap resamples and repeated stratified 5-fold cross-validation for 50 iterations, and the model's discrimination, overall classification error, calibration, and decision curve performance were evaluated. Result KPLA accounted for 82.0% of all cases (246/300). Compared with non-KPLA patients, those with KPLA were younger, had a higher prevalence of diabetes, and had lower rates of malignancy, mixed infection, and any extrahepatic metastatic infection. The composite invasive outcome (bacteremia and/or extrahepatic metastatic infection) occurred in 46.7% (115/246) of KPLA patients, including 82 (33.3%) with bacteremia and 35 (14.2%) with extrahepatic metastatic infection (EMI). In the invasive group, the median PCT and lactate levels were 31.0 ng/mL and 2.40 mmol/L, respectively. In the non-invasive group, they were 1.50 ng/mL and 1.20 mmol/L, respectively. The differences between the two groups were statistically significant (both P < 0.001). In the final model, the adjusted ORs for highest body temperature, lactate, and PCT were 2.70 (95% CI 1.56–4.67), 2.77 (95% CI 1.52–5.07), and 1.10 (95% CI 1.06–1.14), respectively. Repeated cross-validation showed that the model achieved an AUC of 0.932 ± 0.003, a Brier score of 0.105 ± 0.002, and a calibration slope of 0.91 ± 0.04. Exploratory decision curve analysis suggested that the model may provide positive net benefit across the threshold range examined. Conclusion In this cohort, KPLA and non-KPLA differed in host background and clinical characteristics. A simplified three-variable risk-classification model incorporating maximum body temperature, lactate, and PCT demonstrated stable performance in internal validation for identifying patients with invasive complications (defined as bacteremia and/or extrahepatic metastatic infection). If validated prospectively, this model may facilitate early identification of high-risk patients who could benefit from intensified surveillance, timely adjustment of antimicrobial therapy, and proactive screening for extrahepatic metastatic foci. However, because peak values during hospitalization were used, a temporal relationship between predictors and outcome onset cannot be established, and the model should be interpreted as a cross-sectional risk-classification tool rather than a prospective prediction instrument. Its clinical applicability requires further investigation, including anchoring of predictor measurement to a standardized time point, harmonization of outcome definitions, and independent external validation in multi-center cohorts.

Frontiers in MedicineVol. 13
Ningxia Medical University General Hospital (CN)
Good health and well-being
Amoebic Infections and Treatments
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