Xu K, Wu DH, Zeng CJ, Guo JY, Xi DY, Wang MJ, Yao ZY, Feng AQ, Ji F, Yan XB, Ye LL, Li CY. Clinical predictors of multidrug-resistant Gram-negative pyogenic liver abscess and nomogram construction: A retrospective analysis. World J Gastroenterol 2025; 31(42): 112478 [DOI: 10.3748/wjg.v31.i42.112478]
Corresponding Author of This Article
Chun-Yang Li, PhD, Chief Physician, Department of Infectious Diseases, The Affiliated Hospital of Xuzhou Medical University, No. 9 Kunpeng North Road, Economic and Technological Development Zone, Xuzhou 221132, Jiangsu Province, China. lichunyangyelili@126.com
Research Domain of This Article
Infectious Diseases
Article-Type of This Article
Retrospective Study
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Nov 14, 2025 (publication date) through Nov 16, 2025
Times Cited of This Article
Times Cited (0)
Journal Information of This Article
Publication Name
World Journal of Gastroenterology
ISSN
1007-9327
Publisher of This Article
Baishideng Publishing Group Inc, 7041 Koll Center Parkway, Suite 160, Pleasanton, CA 94566, USA
Share the Article
Xu K, Wu DH, Zeng CJ, Guo JY, Xi DY, Wang MJ, Yao ZY, Feng AQ, Ji F, Yan XB, Ye LL, Li CY. Clinical predictors of multidrug-resistant Gram-negative pyogenic liver abscess and nomogram construction: A retrospective analysis. World J Gastroenterol 2025; 31(42): 112478 [DOI: 10.3748/wjg.v31.i42.112478]
Ke Xu, Dong-Hui Wu, Chu-Jia Zeng, Jia-Yi Guo, Zhi-Yuan Yao, Graduate School, Xuzhou Medical University, Xuzhou 221006, Jiangsu Province, China
De-Yang Xi, Department of Intensive Care Unit, Xuzhou Central Hospital, Xuzhou 221009, Jiangsu Province, China
Meng-Jiao Wang, Department of Gastroenterology, Beilun District People’s Hospital, Ningbo 315800, Zhejiang Province, China
An-Qiang Feng, Department of Digestive Disease, Xuzhou Central Hospital, Xuzhou 221009, Jiangsu Province, China
Fang Ji, Xue-Bing Yan, Chun-Yang Li, Department of Infectious Diseases, The Affiliated Hospital of Xuzhou Medical University, Xuzhou 221132, Jiangsu Province, China
Li-Li Ye, Department of Pediatrics, The Affiliated Hospital of Xuzhou Medical University, Xuzhou 221006, Jiangsu Province, China
Co-corresponding authors: Li-Li Ye and Chun-Yang Li.
Author contributions: Xu K and Wu DH share co-first authorship based on their equivalent substantive contributions to experimental execution and analytical rigor throughout this collaborative research; Li CY and Xu K contributed to the study design; Xu K conducted literature screening, quality assessment, and manuscript drafting; Xu K, Wu DH, and Zeng CJ performed statistical analyses; Guo JY executed literature mining; Wu DH, Xi DY, Wang MJ, Yao ZY, Feng AQ, and Ye LL conducted data analysis; Li CY, Yan XB, and Ye LL edited manuscript; Ji F provided conceptual guidance and secured funding; Li CY and Ye LL share co-corresponding authorship for their joint leadership in research coordination, intellectual direction, and conceptual innovation, reflecting the project’s collaborative ethos. All authors reviewed and approved the final manuscript.
Institutional review board statement: The study was reviewed and approved by the Medical Ethics Committee of The Affiliated Hospital of Xuzhou Medical University, No. XYFY2025-KL262-01.
Informed consent statement: Due to the retrospective and observational nature of the study without active interventions, the institutional review board waived the necessity of obtaining individual informed consent.
Conflict-of-interest statement: All the authors report no relevant conflicts of interest for this article.
Data sharing statement: Technical appendix, statistical code, and dataset available from the corresponding author. Participants gave informed consent for data sharing.
Open Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Chun-Yang Li, PhD, Chief Physician, Department of Infectious Diseases, The Affiliated Hospital of Xuzhou Medical University, No. 9 Kunpeng North Road, Economic and Technological Development Zone, Xuzhou 221132, Jiangsu Province, China. lichunyangyelili@126.com
Received: July 29, 2025 Revised: August 19, 2025 Accepted: October 14, 2025 Published online: November 14, 2025 Processing time: 108 Days and 1.5 Hours
Abstract
BACKGROUND
In recent years, there has been a significant increase in pyogenic liver abscesses (PLAs) caused by multidrug-resistant (MDR) Gram-negative bacteria (GNB), predominantly Klebsiella pneumoniae and Escherichia coli.
AIM
To clarify the clinical characteristics and risk factors associated with MDR-GNB-related PLAs, develop a predictive nomogram for personalized risk assessment, and enhance the timeliness of empirical antibiotic selection.
METHODS
Based on the antibiotic susceptibility profiles, enrolled patients were divided into two groups: A MDR group comprising 105 individuals and a non-resistant group comprising 163 individuals. A systematic collection of demographic characteristics, laboratory findings, and prognostic indicators was performed. A predictive nomogram was established using multivariate stepwise regression modeling. Model effectiveness was evaluated by examining its discriminative capability, calibration accuracy, and clinical utility through receiver operating characteristic curves with corresponding area under the curve values, calibration graphs, and decision curve analysis. Continuous data were analyzed using the independent-sample t-test if they met normality criteria; otherwise, the Wilcoxon rank-sum test was adopted. For categorical data, Fisher’s exact test was chosen when the expected count in any cell was below five; in all other instances, the χ2 test was applied.
RESULTS
This retrospective study analyzed clinical and laboratory data from 268 patients diagnosed with Gram-negative PLA at a major healthcare facility from January 2019 to February 2025. Among these, 105 cases (39%) were associated with MDR-GNB, primarily Klebsiella pneumoniae (43%) and Escherichia coli (42%). Mixed infections were rare, accounting for only 3% of cases. Multivariate regression revealed five independent predictors of MDR-GNB liver abscesses: Age ≥ 60 years, diabetes, presence of a malignant tumor, lower C-reactive protein levels, and prolonged prothrombin time. These variables were integrated into a nomogram to facilitate individualized risk assessment.
CONCLUSION
The results imply that being aged over 60, diabetes, malignant tumor, lower C-reactive protein levels, and higher prothrombin time levels can accurately forecast MDR-GNB infections in PLAs, highlighting the importance of early screening to enable more targeted antibiotic treatments. However, as this was a single-center study without external validation, the generalizability of our model remains limited. Future multicenter, multi-ethnic prospective studies are needed to validate and extend these findings.
Core Tip: This study identifies age ≥ 60, diabetes, malignant tumor, lower C-reactive protein levels, and prolonged prothrombin time as independent predictors of multidrug-resistant Gram-negative bacteria pyogenic liver abscess, primarily caused by Klebsiella pneumoniae and Escherichia coli. A validated nomogram incorporating these factors enables personalized risk assessment, facilitating early screening and optimized empirical antibiotic therapy for enhanced clinical outcomes. Findings are based on retrospective analysis of 268 Gram-negative pyogenic liver abscess cases.
Citation: Xu K, Wu DH, Zeng CJ, Guo JY, Xi DY, Wang MJ, Yao ZY, Feng AQ, Ji F, Yan XB, Ye LL, Li CY. Clinical predictors of multidrug-resistant Gram-negative pyogenic liver abscess and nomogram construction: A retrospective analysis. World J Gastroenterol 2025; 31(42): 112478
Pyogenic liver abscess (PLA) is a severe infection with substantial regional variation in incidence. In mainland China, annual incidence rates range from 1.1 to 5.4 cases per 100000 population, whereas several other Asian countries report higher rates (12-18 per 100000), significantly exceeding those observed in Western nations (1.1-2.3 per 100000)[1-4]. The epidemiological landscape of PLA has shifted due to rising diabetes prevalence, increased malignancy incidence, greater use of immunosuppressive therapies, frequent invasive hepatobiliary interventions, and the emergence of multidrug-resistant (MDR) and hypervirulent bacterial strains, introducing additional clinical complexities[4,5]. MDR organisms, defined as microbes resistant to at least three antimicrobial classes, commonly include challenging Gram-negative pathogens such as MDR Enterobacteriaceae and Acinetobacter[6]. The increasing prevalence of MDR-Gram-negative bacteria (GNB) infections poses a significant healthcare challenge[7,8], worsening underlying conditions like diabetes and biliary tract diseases, and heightening the risk of severe complications. This study aims to delineate clinical characteristics and risk factors associated with MDR-GNB-related PLA, develop a predictive nomogram for personalized risk assessment, and ultimately enhance the timeliness and accuracy of empirical antibiotic selection.
MATERIALS AND METHODS
Ethical approval for the research protocol
Ethical clearance for this retrospective investigation was obtained from the institutional review board at The Affiliated Hospital of Xuzhou Medical University, No. XYFY2025-KL262-01. Due to the retrospective and observational nature of the study without active interventions, the institutional review board waived the necessity of obtaining individual informed consent.
Patients
Clinical data were retrieved from the database of PLA patients at The Affiliated Hospital of Xuzhou Medical University, spanning from January 2019 to February 2025. Inclusion criteria consisted of the following: (1) Clinical symptoms, such as fever and abdominal pain; (2) Imaging confirmation via abdominal ultrasound or computed tomography; (3) Microbiological verification through blood or pus cultures; and (4) Abscess confirmation via percutaneous biopsy or surgical intervention. Patients with amebic or tuberculous liver abscesses, other non-bacterial abscesses, or incomplete medical records were excluded.
Admission data, including demographic characteristics [gender, age, and body mass index (BMI)], clinical symptoms (fever and abdominal pain), and pre-existing conditions (hypertension, diabetes, and malignancies), were collected from electronic medical records. Laboratory parameters encompassed hematological, biochemical, and microbiological indicators, while imaging data provided details on liver abscess size, location, and number. Additionally, hospital management strategies (abscess drainage, antibiotic therapy) and discharge outcomes (clinical resolution, persistent infection, relapse, or death) were documented.
Assessment of microbiological characteristics
The focus of this research was specifically placed on mixed infections involving at least two pathogens, notably infections caused by carbapenem-resistant or extended-spectrum β-lactamase (ESBL)-producing Enterobacteriales. Enterobacterales producing ESBL primarily exhibit resistance to third-generation cephalosporins and aztreonam, whereas carbapenem-resistant strains display resistance against carbapenems including meropenem, imipenem, and ertapenem[9,10]. Antibiotic susceptibility evaluations were carried out using both the Kirby-Bauer disk diffusion method and automated testing platforms, strictly following procedural guidelines and interpretative criteria provided by the Clinical and Laboratory Standards Institute.
Statistical analysis
We used SPSS 25.0 and R 4.2.2 to for statistical analysis, setting the significance level at P < 0.05. Continuous data were displayed as means and standard deviations, whereas non-normally distributed data were shown as medians and interquartile ranges. For categorical data, we presented frequencies and proportions. We compared groups using t-tests for normally distributed continuous variables and Wilcoxon rank-sum tests for data without a normal distribution. Categorical data were compared using Fisher’s exact test in the case of low expected frequencies and χ2 tests otherwise. Missing data were handled using random forest imputation. To identify predictive variables, we incorporated baseline clinical features and laboratory results into a multivariate stepwise regression model. The model included the optimal mix of variables from the univariate analysis, and backward elimination was used to streamline the final model.
RESULTS
Pathogen distribution in PLAs
Blood and drainage fluid cultures from 268 patients yielded 278 pathogenic isolates. The vast majority of cases (97.0%, n = 260) involved single-pathogen infections, while a small subset exhibited polymicrobial involvement: Six patients (2.2%) with dual pathogens and two (0.7%) with three or more. Among the Gram-negative isolates, Klebsiella pneumoniae (K. pneumoniae) dominated (72.7%, n = 202), followed by Escherichia coli (E. coli, 20.9%, n = 58). Less frequently isolated pathogens included: Serratia spp. (n = 3), Enterobacter cloacae (n = 3), Klebsiella aerogenes (n = 3), Enterobacter asburiae (n = 1), Morganella morganii (n = 1), Klebsiella variicola (n = 1), Klebsiella oxytoca (n = 1), Acinetobacter baumannii (n = 1), and Pseudomonas spp. (n = 4) (Figure 1).
Figure 1 Pathogen distribution of Gram-negative bacteria in bacterial liver abscesses.
Antimicrobial resistance of PLA pathogens
All 268 patients with Gram-negative PLA underwent comprehensive susceptibility testing. MDR pathogens were detected in 39.2% (n = 105), accounting for 115 isolates. Five patients carried two MDR strains, one had three, and another had four (Table 1). Resistance analysis revealed that ESBL-producing bacteria were most prevalent (60.0%, n = 69), including primarily E. coli (28.7%, n = 33) and K. pneumoniae (21.7%, n = 25). Carbapenem-resistant strains constituted 8.7% of all bacteria (n = 10), mostly K. pneumoniae (n = 5). Other resistant pathogens included MDR Pseudomonas (n = 3) and Acinetobacter baumannii (n = 1) (Figure 2).
This cohort included 268 confirmed Gram-negative PLA cases, with key baseline characteristics summarized in Table 2. The MDR group accounted for 39.2% of patients, characterized by a male predominance (67.5%) and half being aged ≥ 60 years. Median BMI was 23.8 kg/m2, and the median Charlson Comorbidity Index (CCI) score was 3.0. Common comorbidities included type 2 diabetes (59.3%), hypertension (26.9%), malignancies (17.9%), and biliary tract disease (34.3%). Complications such as pneumonia and septic shock occurred in 34.3% and 10.4% of cases, respectively. Pleural effusion (28.0%) and ascites (6.7%) were documented, whereas concurrent heart failure and renal insufficiency were uncommon (6.0% and 5.6%, respectively). Clinically, the most frequent symptom was fever (88.4%), followed by abdominal pain (45.2%) and gastrointestinal complaints (59.0%). Imaging findings indicated that abscesses were solitary in 68.3% of patients, primarily involving the right lobe of the liver (77.6%), with a median size of 65.5 mm. Laboratory tests revealed significant increases in inflammatory biomarkers, including median levels of C-reactive protein (CRP; 167.7 mg/L), interleukin-6 (IL-6; 118.3 pg/mL), and procalcitonin (PCT; 6.7 ng/mL), accompanied by moderate elevations in hepatic enzyme levels, such as aspartate aminotransferase (44 U/L) and gamma-glutamyl transferase (119.5 U/L). Treatments included antibiotics alone (16.8%) or combined with drainage (83.2%). Outcomes were favorable in 91.8% of cases, with low rates of treatment failure (4.9%), recurrence (1.5%), and mortality (1.9%).
Table 2 Comparison of clinical characteristics, laboratory parameters, comorbidities, and clinical outcomes in patients with pyogenic liver abscesses caused by multidrug-resistant-Gram-negative bacteria vs non-multidrug-resistant-Gram-negative bacteria, n (%)/median (interquartile ranges).
Comparative analysis of MDR and non-MDR patient profiles
Based on patterns of antimicrobial susceptibility, the 268 included patients were stratified into a MDR group (n = 105) and a non-MDR group (n = 163). According to data summarized in Table 2, MDR-group patients exhibited older age, lower BMI values, higher CCI scores, and an increased prevalence of malignancies. Fever was more common in non-MDR patients, while abdominal pain and other gastrointestinal symptoms showed no significant differences between the groups. Laboratory results revealed that the MDR group exhibited lower levels of CRP, alanine aminotransferase, and albumin, along with smaller abscess diameters (P = 0.038). Conversely, the prothrombin time (PT) and IL-6 Levels were elevated among MDR patients. Abscess locations did not differ significantly between groups.
Comorbidity and outcome differences
The prevalence of septic shock, pleural effusion, ascites, heart failure, and renal insufficiency were comparable between the two groups (Table 2). However, biliary tract conditions, such as gallstones and gallbladder polyps, occurred significantly more frequently in MDR patients (45.0% vs 24.2%, P < 0.001). Pneumonia rates were similar across both groups. Clinically, non-MDR patients had a significantly higher rate of improvement (95.09% vs 86.67%, P = 0.018), while MDR cases exhibited a higher likelihood of persistent or recurrent infections.
Risk factor assessment and predictive model for MDR Gram-negative PLAs
Univariate and multivariate analyses (Table 3) revealed key independent risk factors for MDR Gram-negative PLAs: Age ≥ 60 [odds ratio (OR) = 1.957; 95% confidence interval (CI): 1.099-3.484; P = 0.023], diabetes (OR = 1.998; 95%CI: 1.089-3.663; P = 0.025), malignant tumor (OR = 3.968; 95%CI: 1.810-8.699; P < 0.001), lower CRP levels (OR = 0.993; 95%CI: 0.990-0.997; P = 0.001), and prolonged PT (OR = 1.065; 95%CI: 1.007-1.127; P = 0.028). To streamline clinical decision-making, a nomogram was created (Figure 3), assigning point values to each predictor (Table 4). By totaling these points, healthcare providers can visually estimate the likelihood of an MDR Gram-negative infection (Table 5).
The nomogram demonstrates considerable clinical utility, with an area under the receiver operating characteristic curve of 0.769 (95%CI: 0.675-0.857). The optimal cutoff probability, determined by the Youden index, was 0.654. At the optimal cutoff value of 65.4%, the nomogram demonstrated a sensitivity of 82.1% and specificity of 76.5%. Clinically, when a patient’s predicted MDR risk exceeds this threshold, empirical escalation to broad-spectrum antibiotics (e.g., carbapenems or β-lactamase inhibitor combinations) is recommended (Figure 4A). Goodness-of-fit testing (Hosmer-Lemeshow, P = 0.777) and bootstrap-validated calibration (mean absolute error = 0.017) confirmed the model’s reliability (Figure 4B). Decision curve analysis (Figure 4C) further supported its clinical utility. Collinearity was negligible, with variance inflation factors ranging from 1.08 to 1.48 (Table 6).
The current research analyzed microbiological profiles and antibiotic resistance characteristics of Gram-negative PLA at a major tertiary hospital located in Jiangsu Province, China. K. pneumoniae emerged as the most frequent organism, identified in 72.7% (202/278) of isolates, significantly exceeding E. coli at 20.9% (58/278). Consistent with prior findings from mainland China[11,12], over the past three decades K. pneumoniae has become the predominant pathogen responsible for PLA, likely due to its superior invasive capacity into hepatic tissues compared to other GNB.
The global health crisis of antimicrobial resistance has notably increased both the clinical and economic burdens associated with PLAs[13,14]. Alarmingly, 39.2% (105/268) of the cohort involved MDR-GNB, exceeding rates previously reported in comparable domestic and international studies[12,15]. Resistance profiling identified ESBL-producing strains as the most common resistant phenotype (60.0%, 69/115), with ESBL-positive E. coli (28.7%, 33/115) outnumbering ESBL-positive K. pneumoniae (21.7%, 25/115). Of particular concern were the 10 carbapenem-resistant isolates, half of which (5/10) were carbapenem-resistant K. pneumoniae, a high-risk pathogen notorious for multidrug resistance and significant public health implications[16,17].
Clinically, fever was the predominant symptom (88.43%, 237/268), followed by abdominal pain (45.15%, 121/268) and gastrointestinal symptoms (58.96%). Altered mental status or hypotension frequently indicated septic shock, especially in elderly patients, those with malignancies, and individuals with systolic blood pressure < 100 mmHg and lactate levels ≥ 2 mmol/L[18]. Notably, concurrent biliary disorders (e.g., cholecystitis, gallbladder polyps) were more prevalent in the MDR group. Consistent with the findings of Long et al[12], this observation supports a significantly higher proportion of biliary-origin infections in MDR cases compared with non-MDR cases[12]. Laboratory abnormalities, including deranged hematologic parameters, elevated liver enzymes, and coagulopathy, were pervasive among PLA patients. The MDR group demonstrated significantly depressed hemoglobin (108 g/L vs 118 g/L, P < 0.01), CRP (149.2 mg/L vs 175.6 mg/L, P = 0.016), alanine aminotransferase (48 U/L vs 66 U/L, P = 0.007), and albumin (29.1 g/L vs 30.9 g/L, P = 0.047), alongside prolonged PT (13.7 seconds vs 13.1 seconds, P = 0.007) and elevated IL-6 (153 pg/mL vs 89 pg/mL, P = 0.003). The MDR group exhibited significantly smaller abscess diameters (63.0 mm vs 68.0 mm, P = 0.038). While this finding appears to contradict conventional associations between abscess size and disease severity, it may be explained by distinct clinical characteristics of our cohort. Specifically, the higher prevalence of immunocompromised states (e.g., malignancy: 31.4% vs 9.2%) in the MDR group could lead to attenuated inflammatory responses and atypical clinical presentations, resulting in delayed diagnosis and earlier imaging intervention during abscess development. This hypothesis aligns with Long et al’s observation of smaller abscesses in MDR patients[12], yet contrasts with studies linking larger abscess volumes to K. pneumoniae infections[19]. Notably, the predominance of E. coli (which often associates with biliary pathologies and smaller abscesses) over K. pneumoniae in our MDR cohort may further contribute to this morphological discrepancy. Clinically, these findings underscore that smaller abscess dimensions should not be misinterpreted as mild disease; rather, they may signal occult MDR infections, particularly in immunocompromised hosts, warranting early empiric escalation to broad-spectrum antimicrobial regimens.
In the present study, although the MDR group exhibited significantly elevated IL-6 Levels, IL-6 did not emerge as an independent predictor in the univariate analysis. This suggests that within our cohort, the predictive information conveyed by IL-6 may be overlapping with or superseded by the stronger, combined effect of the final model variables. However, given IL-6’s established role as a pivotal early mediator of acute inflammatory responses and its prognostic value in sepsis and other infectious contexts[20,21], its potential utility in the context of MDR-GNB PLA should not be dismissed. Relevant studies have demonstrated an elevated trend in median IL-6 Levels in MDR tuberculosis compared to drug-susceptible tuberculosis, suggesting that increased IL-6 may indicate more severe systemic inflammatory responses in diverse infectious contexts[21]. Varga et al[22] identified IL-6 as a prognostic biomarker in sepsis. Therefore, we recommend that future investigations explore the integration of IL-6 into multi-biomarker panels or dynamic monitoring protocols alongside established markers like CRP and PCT. For instance, machine learning models capable of handling complex interactions might better capture the nuanced contribution of IL-6. Alternatively, tracking the kinetics of IL-6 Levels early in infection, rather than relying on a single baseline measurement, could prove more informative for risk stratification and early guidance of empiric therapy. Future multicenter studies should validate the combined predictive utility of IL-6 and PCT for early identification of MDR-GNB infections and risk stratification. These disparities likely reflect the MDR group’s higher malignancy rate (31.4% vs 9.2%), chronic malnutrition, and sepsis-driven coagulopathy and cytokine activation, culminating in severe complications like septic shock (11.4% vs 9.8%)[23-25].
Outcomes were markedly worse in MDR cases, with higher rates of treatment failure and recurrence, and diminished clinical improvement; a consequence of both the intrinsic virulence of MDR-GNB and patients’ compromised baseline health. All five carbapenem-resistant K. pneumoniae associated PLAs resulted in poor prognoses, underscoring the therapeutic challenges posed by this highly resistant, transmissible pathogen[26]. Meta-analyses link chronic comorbidities (e.g., hypertension, diabetes), malignancy, and elevated Acute Physiology and Chronic Health Evaluation II scores in association with adverse PLA outcomes[27], while a higher CCI reliably predicts poorer prognosis[28]. Here, the MDR group’s significantly higher CCI values aligned with their unfavorable clinical trajectories, consistent findings have been reported in previous studies[12].
Multivariate analysis pinpointed age ≥ 60, diabetes, malignancy, lower CRP levels, and prolonged PT as independent risk factors for MDR-GNB PLA. In building the predictive model, CRP was analyzed as a continuous variable to preserve its full statistical information and avoid the potential loss of predictive power associated with dichotomization using an arbitrary cutoff. This approach allows the model to capture the graded nature of the inflammatory response across the entire range of CRP values, thereby enhancing the nomogram’s accuracy for risk stratification. The observed association of lower CRP levels with MDR-GNB infection, while counterintuitive, may be elucidated by the immunocompromised status prevalent in the MDR cohort. A relevant study indicates that serum CRP interacts with circulating megakaryocytes via the secreted protein acidic and rich in cysteine pathway, thereby modulating systemic inflammatory responses, a positive correlation exists between CRP levels and the proportion of circulating megakaryocytes[29]. Notably, the MDR cohort exhibited a significantly higher malignancy prevalence than the non-MDR group. In these patients, chemotherapy and radiotherapy regimens may suppress bone marrow hematopoiesis, thereby impairing megakaryocyte proliferation and differentiation, a process that could subsequently dysregulate CRP production. Furthermore, studies have observed dynamic changes in CRP levels among patients receiving immunotherapy, suggesting that immune-related factors may influence CRP expression. This dynamic variation in immunocompromised hosts may also help explain the lower CRP levels observed in our MDR group. Although this study enrolled first-admission patients with newly diagnosed bacterial PLAs, and the majority had not received empirical antibiotic therapy at presentation, effectively reducing the elevated risk of drug-resistant infections associated with prior empirical antimicrobial use, potential selection bias persists. We cannot entirely rule out pre-admission antimicrobial exposure within the preceding 6 months among MDR-group patients, which may have blunted the acute-phase response and contributed to relatively lower CRP levels. Future work will impose stricter controls for potential confounders and further investigate this association.
In this study, the nomogram provides a practical tool for optimizing empirical antibiotic decision-making. However, due to the limitations inherent in its single-center, retrospective design and sample size, the model’s generalizability requires further validation through multicenter prospective studies. For clinical implementation, the optimal risk threshold should be tailored to the specific healthcare setting. In tertiary care hospitals with a high prevalence of resistant pathogens, a lower threshold is recommended to maximize sensitivity and ensure early coverage of MDR organisms. In primary care settings, a higher threshold may be adopted to align with antimicrobial stewardship goals, prioritizing treatment specificity. Integration into workflow could involve embedding the tool as a clinical decision support module within electronic health record systems in advanced settings, enabling automated risk calculation and alerts. In resource-limited community hospitals or clinics, it could be simplified into a rapid bedside scorecard to facilitate initial risk stratification and guide therapeutic decisions.
This study has several limitations. First, beyond the identified predictors (age, diabetes, and malignant tumor), unmeasured host factors such as chronic malnutrition, prior antibiotic exposure, and cellular immune competence may modulate MDR susceptibility. For instance, hypoalbuminemia (lower in the MDR group: 29.1 g/L vs 30.9 g/L, P = 0.047) may reflect nutritional deficits that compromise immune defense. Future predictive models should incorporate such variables to enhance risk stratification. Second, the single-center, retrospective nature of this study may introduce selection bias, particularly given that patients admitted to tertiary referral centers often present with more complex comorbidities. Moreover, the lack of external validation further limits the generalizability of our findings. Finally, the predominance of K. pneumoniae (72.7%) in our cohort reflects the regional epidemiology of East China, however, this pattern may not be generalizable to regions where E. coli is the predominant pathogen. Therefore, future multicenter, prospective studies are warranted to externally validate and expand upon our findings across diverse healthcare settings.
CONCLUSION
PLAs caused by MDR-GNB possess distinctive etiological characteristics. E. coli and K. pneumoniae remain the leading pathogens, and their substantial antimicrobial resistance complicates clinical management. Our findings confirm that individuals aged ≥ 60 years, and those with diabetes, malignancy, lower CRP levels, and prolonged PT, have a higher risk of MDR-GNB infections. Recognizing these risk factors is critical for personalized patient management, facilitating early empirical initiation of broad-spectrum antibiotics, and supporting evidence-based antimicrobial stewardship and infection control strategies. Future research should focus on elucidating resistance gene transmission and host-pathogen interactions in MDR-GNB PLAs to refine prevention and treatment strategies, ultimately improving patient outcomes.
Footnotes
Provenance and peer review: Unsolicited article; Externally peer reviewed.
Peer-review model: Single blind
Specialty type: Gastroenterology and hepatology
Country of origin: China
Peer-review report’s classification
Scientific Quality: Grade B, Grade B, Grade B, Grade C
Novelty: Grade A, Grade B, Grade B, Grade B
Creativity or Innovation: Grade B, Grade B, Grade B, Grade B
Scientific Significance: Grade A, Grade B, Grade B, Grade C
P-Reviewer: Khan S, Research Fellow, Pakistan; Tu ZH, MD, Consultant, Professor, Vice Director, China; Twohig P, MD, Assistant Professor, FRCPC, United States S-Editor: Wu S L-Editor: A P-Editor: Yu HG
Chinese Society of Emergency Medicine. [Expert Consensus on Diagnosis and Treatment of Bacterial Liver Abscess in Emergency Medicine].Zhonghua Jizhen Yixue Zazhi. 2022;31:273-280.
[PubMed] [DOI] [Full Text]
Mohan BP, Meyyur Aravamudan V, Khan SR, Chandan S, Ponnada S, Asokkumar R, Navaneethan U, Adler DG. Prevalence of colorectal cancer in cryptogenic pyogenic liver abscess patients. Do they need screening colonoscopy? A systematic review and meta-analysis.Dig Liver Dis. 2019;51:1641-1645.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 21][Cited by in RCA: 21][Article Influence: 3.5][Reference Citation Analysis (0)]
Schmoch T, Möhnle P, Weigand MA, Briegel J, Bauer M, Bloos F, Meybohm P, Keh D, Löffler M, Elke G, Brenner T, Bogatsch H; SepNet–Critical Care Trials Group. The prevalence of sepsis-induced coagulopathy in patients with sepsis - a secondary analysis of two German multicenter randomized controlled trials.Ann Intensive Care. 2023;13:3.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 32][Reference Citation Analysis (0)]
Rossi G, Nguyen Y, Lafont E, Rossi B, Canouï E, Roux O, Dokmak S, Bert F, Fantin B, Lefort A. Large retrospective study analysing predictive factors of primary treatment failure, recurrence and death in pyogenic liver abscesses.Infection. 2022;50:1205-1215.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 16][Reference Citation Analysis (0)]
