Comparative performance among inflammatory biomarkers for pre-operative prediction of gangrenous cholecystitis

Abstract

BACKGROUND

This study aimed to compare the diagnostic performance of inflammatory biomarkers derived from routine blood parameters in patients with acute cholecystitis, based on the hypothesis that these biomarkers can predict gangrenous cholecystitis independently of imaging before surgery.

METHODS

Data from 187 consecutive adult patients who underwent laparoscopic cholecystectomy for acute cholecystitis during the same hospitalization were retrospectively analyzed in an observational cohort. After the exclusion of 7 cases, 180 patients constituted the analytic cohort. Gangrenous cholecystitis was defined as intraoperative or pathological verification. C-reactive protein (CRP) and six prespecified haemogram-derived indices – the neutrophil-, platelet- and monocyte-to-lymphocyte ratios (neutrophil-to-lymphocyte ratio [NLR], platelet-to-lymphocyte ratio [PLR], monocyte-to-lymphocyte ratio [MLR]) and three composite indices (systemic immune-inflammation index, systemic inflammation response index [SIRI], aggregate index of systemic inflammation) – were evaluated; diagnostic accuracy was assessed using receiver operating characteristic-area under the curve (ROC-AUC), cutoff points were determined using the Youden index, and an age- and sex-adjusted logistic regression model was established.

RESULTS

Gangrenous cholecystitis was confirmed in 33 of the 180 analyzed patients (18.3%). CRP was 163.00 mg/L (interquartile range [IQR] 39.00–224.00) in the gangrenous group and 15.00 mg/L (IQR 4.02–55.89) in the non-gangrenous group. Among hemogram-derived indices, SIRI (7.10 [4.09–10.03] vs. 2.48 [1.03–4.89]) and MLR (0.62 [0.40–0.79] vs. 0.31 [0.19–0.50]) were significantly higher in gangrenous cases (both p<0.001); NLR was 8.00 (5.32–12.90) vs. 4.00 (2.35–6.66) (p<0.001). In the ROC analysis, CRP showed the highest discriminatory power (AUC 0.81; 95% confidence interval 0.71–0.89); at a cutoff point of 61.0 mg/L, the sensitivity and specificity were 0.73 and 0.78, respectively. Among the derived indices, the highest AUCs were observed for SIRI and AISI. In the model including CRP and SIRI adjusted for age and sex, only CRP remained an independent predictor; each 10 mg/L increase in CRP was associated with a 14% increase in the odds of gangrenous cholecystitis.

CONCLUSION

CRP provides the highest discrimination for pre-operative prediction of gangrenous cholecystitis and was the only independent predictor in multivariable analysis. The hemogram-derived indices showed moderate univariable discrimination, but their additional value beyond CRP remains uncertain and requires further evaluation in larger, prospective cohorts.

INTRODUCTION

Acute cholecystitis is one of the most common complications of gallstone disease and increases morbidity and mortality if surgery is not performed in a timely manner.^[1] Current guidelines recommend early laparoscopic cholecystectomy, but the difficulty level of acute cholecystitis surgery can vary significantly, and there are no clearly defined criteria in the literature for predicting gangrenous cholecystitis preoperatively.^[2] Gangrenous cholecystitis is one of these severe clinical presentations, with an incidence of 10–40% and morbidity of 15–50% among acute cholecystitis cases.^[3] Deciding on the timing of surgery and the balance between risks and benefits is a challenging process that requires clinical experience. This is particularly true for severe variants such as gangrenous cholecystitis, where objective pre-operative risk stratification is still poorly defined. Various clinical and sonographic scoring systems have been developed to predict difficult cholecystectomy and conversion to open surgery, but these scores alone are often insufficient to provide a generalizable prediction of gangrene.^[4-6]

Radiological imaging is the primary tool for diagnosing acute cholecystitis. However, it is less reliable in predicting gangrenous cholecystitis. The sensitivity of computed tomography (CT) in predicting gangrenous cholecystitis has been reported to be 29%. This necessitates the development of new strategies for predicting gangrenous cholecystitis independent of radiological imaging. There has been a shift toward clinical studies focusing on accessible, imaging-independent, inexpensive, and generalizable biomarkers in this field.^[7,8] Hematological and biochemical markers have been extensively researched for this purpose. Numerous studies have been conducted on inflammation biomarkers, such as the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), monocyte-to-lymphocyte ratio (MLR), and C-reactive protein (CRP). However, fewer studies have addressed newer composite biomarkers, such as the systemic immune-ınflammation ındex (SII), systemic inflammation response index (SIRI), and aggregate index of systemic ınflammation (AISI).^[9-12] The current literature predominantly consists of studies that focus on individual markers and are conducted across different cohorts with heterogeneous outcomes. Studies reporting head-to-head comparisons of multiple inflammatory biomarkers within the same cohort, along with calibration and clinical benefit analyses, are limited.

The primary aim of this study was to compare the performance of seven prespecified inflammatory biomarkers derived from routine pre-operative blood parameters, independent of imaging, in predicting gangrenous cholecystitis. The selection of biomarkers was hypothesis-driven. CRP was included as the canonical hepatic acute-phase reactant of established clinical relevance, whereas the six hemogram-derived indices were prespecified to represent biologically complementary axes of systemic inflammation – neutrophil-, platelet-, and monocyte-based ratios against lymphocyte balance, together with the multi-component composites that combine them. These specific indices have been the most extensively investigated hemogram-derived inflammatory ratios in acute surgical and inflammatory conditions, supporting their selection over less-studied alternatives.^[13]

MATERIALS AND METHODS

Study Design and Ethical Approval

This study was designed as a single-center, retrospective, observational cohort analysis of consecutive patients who underwent surgery at a tertiary center for acute cholecystitis between July 2023 and December 2024. The reporting was structured according to the STROBE guidelines for observational studies. Ethical approval for the study was obtained from the Scientific Research Ethics Committee of Sakarya University (E-43012747-050.04–450863-86). This study was conducted in accordance with the principles of the Declaration of Helsinki.

Patient Selection and Data Collection

All consecutive patients aged >18 years who presented to the hospital due to acute cholecystitis and underwent cholecystectomy during the same admission, with accessible pre-operative laboratory records obtained at the time of emergency presentation, surgical operation videos, and pathology slide records, were included in the study population. Patients were excluded at the screening stage if they had a history of medical treatment or percutaneous cholecystostomy during admission for acute cholecystitis, concurrent choledocholithiasis or pancreatitis, or incomplete study records. The final analytic cohort, therefore, comprised 180 patients (Fig. 1).

Figure 1.

Patient flow through the study. Cohort flow diagram structured according to the STROBE statement.

After identifying suitable patients, demographic characteristics, presenting complaints, vital signs, abdominal ultrasound and computed tomography reports, surgical notes, intraoperative video recordings, and final pathology reports were retrospectively extracted from the hospital information management system and transferred to an anonymized electronic database.

Definition of Gangrenous Cholecystitis

Gangrenous cholecystitis was defined a priori as the presence of histopathological evidence of necrosis or gangrene in the resected specimen and/or intraoperative macroscopic findings indicative of gangrene – such as ischemic, greenish-black discoloration of the gallbladder wall, pericholecystic necrosis, or a necrotic odor (Fig. 2). All resected gallbladders underwent routine histopathological examination, and intraoperative video recordings were available for all patients; surgical video recordings were reviewed by two surgeons blinded to the clinical information, with consensus reached in cases of disagreement.

Figure 2.

Representative intraoperative and cross-sectional imaging findings of gangrenous cholecystitis. (a and b) Intraoperative laparoscopic views. (c) Coronal and (d) axial preoperative contrast-enhanced computed tomography images (arrows indicate the gallbladder).

Laboratory Analyses and Inflammatory Biomarkers

Pre-operative blood samples were collected as part of routine care during the initial emergency department visit. Analyses were performed using the institution’s standard hematology and biochemistry infrastructure. All measurements were based on clinical routine units: CRP: mg/L; complete blood count and differential: ×10⁹/L; red cell distribution width (RDW): %. CRP, white blood cell count (WBC), NEU, LYM, MONO, PLT, and RDW were used as raw data for basic variables. Measurements taken at the same sampling time were used for all indices derived from these variables.

Six prespecified hemogram-derived indices were calculated and analyzed alongside CRP, with each index selected a priori for a distinct mechanistic role in systemic inflammation. To minimize statistical and biological redundancy, reciprocal indices and exploratory derivatives were not included in the final univariable, receiver operating characteristic (ROC), or multivariable analyses (see discussion).

• SIRI = (NEU×MONO)/LYM. Rationale: Integrates the combined neutrophil and monocyte response with lymphopenia, reflecting cellular innate immune activation.

• SII = (PLT×NEU)/LYM. Rationale: Combines neutrophilia and lymphopenia with the platelet component to capture immunothrombotic activity.

• AISI ≃ PIV = (NEU×MONO×PLT)/LYM. Rationale: A four-component composite aggregating neutrophil, monocyte, and platelet contributions against lymphocyte count.

• NLR = NEU/LYM. Rationale: The most extensively validated inflammatory ratio has established prognostic value across surgical and infectious conditions.

• PLR = PLT/LYM. Rationale: Highlights the platelet contribution to systemic inflammation and reflects the inflammation–thrombosis axis.

• MLR = MONO/LYM. Rationale: Emphasizes the monocyte–macrophage axis, which is mechanistically relevant to local tissue ischemia and necrosis.

Statistical Analysis

Statistical analyses of the data obtained from the study were performed using Python 3.11 and the Statistical Package for the Social Sciences version 25.0 software (IBM Corp., Armonk, NY, USA). For continuous variables, the mean±standard deviation (SD) or median (interquartile range, [IQR]) was used, whereas for categorical variables, the number and percentage (n [%]) were used. The assumption of normal distribution was assessed using the Shapiro–Wilk test.

Binary comparisons for variables meeting the assumption of normal distribution, the Welch t-test was used; for those not meeting this assumption, the Mann–Whitney U test was used. For categorical variables, the Chi-square test or, in cases where the expected cell value was <5, Fisher’s exact Chi-square test was used. The effect size was reported as the rank-biserial correlation (r) for non-parametric comparisons. In analyses where multiple biomarkers were tested simultaneously, p-values were adjusted using the Benjamini–Hochberg false-discovery-rate (FDR) procedure across the seven prespecified markers (k=7). All statistical tests were two-sided; the conventional threshold of p<0.05 (or Benjamini–Hochberg FDR-adjusted q<0.05 for multi-marker analyses) was used to define statistical significance, and confidence intervals (CI) are reported at the 95% level.

ROC analyses and Youden-optimal cutoff determination were restricted a priori to the seven prespecified inflammatory biomarkers (CRP, NLR, PLR, MLR, SII, SIRI, and AISI). Raw hematological components (WBC, neutrophil, lymphocyte, monocyte, and platelet counts, and RDW) were reported as baseline descriptors in Table 1 only and were not subjected to ROC, FDR, or multivariable testing. For each prespecified marker, the cohort-specific optimal cutoff was identified using the highest Youden index (J), and the corresponding sensitivity and specificity were reported.

Table 1.

Baseline demographic and pre-operative laboratory characteristics by gangrenous cholecystitis status

Characteristic	Overall (n=180)	Non-gangrenous (n=147)	Gangrenous (n=33)	p
Sex, n (%)				0.156
Female	107 (59.4)	91 (61.9)	16 (48.5)
Male	73 (40.6)	56 (38.1)	17 (51.5)
Age, years (mean±SD)	50.7±14.1	49.5±14.1	56.3±13.0	0.010
CRP, mg/L	21.80 (5.94–81.22)	15.00 (4.02–55.89)	163.00 (39.00–224.00)	<0.001
WBC, ×109/L	11.47 (8.70–14.30)	11.00 (8.38–13.88)	13.71 (11.22–17.26)	0.001
Neutrophils, ×109/L	8.85 (5.67–11.80)	8.10 (5.02–11.04)	11.66 (8.80–13.31)	<0.001
Lymphocytes, ×109/L	1.90 (1.27–2.60)	1.94 (1.45–2.61)	1.35 (0.90–2.10)	0.002
Monocytes, ×109/L	0.63 (0.45–0.90)	0.60 (0.43–0.80)	0.81 (0.60–1.08)	0.001
Platelets, ×109/L	260 (225–305)	266 (222–307)	239 (232–293)	0.782
RDW, %	13.60 (13.00–14.33)	13.60 (13.00–14.35)	13.70 (13.00–14.30)	0.796

Values are expressed as mean±SD for age and median (IQR) for laboratory variables. p-values reflect univariable comparisons between non-gangrenous and gangrenous groups. CRP: C-reactive protein; IQR: Interquartile range; RDW: Red cell distribution width; SD: Standard deviation; WBC: White blood cell count.

The multivariable model was prespecified before any data inspection and deliberately parsimonious. CRP was included as the canonical hepatic acute-phase reactant, biologically distinct from cellular indices. SIRI was selected on biological grounds as the most comprehensive cellular composite, integrating neutrophil, monocyte, and lymphocyte components in a single ratio and structurally encompassing the simpler indices NLR (NEU/LYM) and MLR (MONO/LYM). The other hemogram-derived indices (NLR, PLR, MLR, SII, and AISI) were not entered together with SIRI because they share components and would have introduced severe collinearity with limited additional information. The robustness of the model to this choice was assessed in a sensitivity analysis in which SIRI was substituted by each of the other five indices in turn (Supplementary Table S1). Logistic regression was therefore performed with CRP and SIRI as inflammatory predictors, adjusted for age and sex. Robust (HC3) standard errors were used in coefficient estimates; results are presented as odds ratios (OR) and 95% CI. To improve clinical interpretability, the CRP OR is additionally reported per 10 mg/L increment, in addition to the standard per/mg/L estimate.

The discriminatory power of the logistic model was assessed using the area under the curve (AUC), mean Brier score, and average precision.

RESULTS

Study Population and Baseline Characteristics

A total of 187 consecutive patients underwent laparoscopic cholecystectomy for acute cholecystitis during the study period and were screened for inclusion. Seven patients (3.7%) were excluded. The final analytic cohort comprised 180 patients, of whom 33 (18.3%) had gangrenous cholecystitis confirmed by intraoperative and pathological assessment [non-gangrenous, n=147; gangrenous, n=33; (Table 1)]. The gangrenous group was older (mean age, 56.3±13.0 years vs. 49.5±14.1 years; Welch t-test p=0.010), consistent with the mean±SD presentation in Table 1. Sex distribution did not differ significantly between groups (male: 51.5% vs. 38.1%; p=0.156). At presentation, CRP, WBC, neutrophil, lymphocyte, and monocyte counts differed between groups, whereas platelet count and RDW did not (Table 1). Regarding outcome ascertainment, all 33 patients classified as gangrenous (100%) had histopathological confirmation of necrosis or gangrene in the resected specimen, and intraoperative macroscopic findings reviewed on surgical video were concordant with the pathological diagnosis in every case; no patient was classified solely on the basis of one modality.

Biomarker Distributions and Univariable Comparisons

In univariate analyses, CRP was the strongest discriminator between groups, with markedly higher values in gangrenous cases (median 163.00 mg/L [IQR 39.00–224.00] vs. 15.00 mg/L [IQR 4.02–55.89]; p<0.001; rank-biserial r=+0.611). All six prespecified hemogram-derived indices (NLR, PLR, MLR, SII, SIRI, and AISI) were also significantly elevated in gangrenous cases (all p<0.001; FDR-adjusted q<0.001 with k=7; effect sizes r=+0.40 to +0.56), though with consistently smaller effect sizes than CRP. Full distributional summaries are provided in Table 2; selected boxplots are shown in Figure 3. Platelet count and RDW did not differ between groups [platelets: p=0.782; RDW: p=0.796; Table 1].

Table 2.

Univariable comparisons of inflammatory biomarkers between gangrenous and non-gangrenous groups

Biomarker	Non-gangrenous, median (IQR)	Gangrenous, median (IQR)	p-value	FDR q value	Rank-biserial r
CRP	15.00 (4.02–55.89)	163.00 (39.00–224.00)	<0.001	<0.001	+0.611
SIRI	2.48 (1.03–4.89)	7.10 (4.09–10.03)	<0.001	<0.001	+0.556
AISI	605 (294–1274)	1765 (1022–2741)	<0.001	<0.001	+0.551
MLR	0.31 (0.19–0.50)	0.62 (0.40–0.79)	<0.001	<0.001	+0.548
SII	1030 (567–1788)	2236 (1601–3057)	<0.001	<0.001	+0.518
NLR	4.00 (2.35–6.66)	8.00 (5.32–12.90)	<0.001	<0.001	+0.503
PLR	132.5 (100.0–181.1)	202.5 (127.3–276.5)	<0.001	<0.001	+0.401

Effect size is presented as the rank-biserial correlation (r). Analytic n=180; lymphocyte-dependent indices were computed on n=178. AISI: Aggregate index of systemic inflammation, CRP: C-reactive protein; FDR: False discovery rate; IQR: Interquartile range; MLR: Monocyte-to-lymphocyte ratio; NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; SII: Systemic immune-inflammation index; SIRI: Systemic inflammation response index.

Figure 3.

Distribution of CRP and haemogram-derived inflammatory indices by gangrenous cholecystitis status. Box-and-whisker plots for (a) CRP, (b) NLR, (c) PLR, (d) MLR, (e) SII, (f) SIRI, and (g) AISI. Boxes denote the interquartile range, the central horizontal line represents the median, and whiskers extend to ×1.5 the interquartile range; open circles indicate outlier values. Lymphocyte-dependent indices were computed on n=178. AISI: Aggregate index of systemic inflammation; CRP: C-reactive protein; FDR: False discovery rate; MLR: Monocyte-to-lymphocyte ratio; NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; SII: Systemic immune-inflammation index; SIRI: Systemic inflammation response index.

Diagnostic Accuracy and Biomarker Performance

In ROC analysis, CRP demonstrated the highest discriminatory power for gangrenous cholecystitis (AUC 0.81, 95% CI 0.71–0.89), with a Youden-optimal cutoff of 61.0 mg/L yielding sensitivity 0.73 and specificity 0.78. The six hemogram-derived indices showed moderate, broadly comparable discrimination (AUC range 0.70–0.78), without any single index outperforming CRP. Full ROC metrics are summarized in Table 3, and curves are shown in Figure 4.

Table 3.

Diagnostic accuracy of biomarkers for gangrenous cholecystitis

Marker	AUC (95% CI)	Optimal cut-off (Youden)	Sensitivity	Specificity
CRP	0.81 (0.71–0.89)	>61.0	0.73	0.78
SIRI	0.78 (0.69–0.85)	>3.86	0.79	0.68
AISI	0.78 (0.69–0.85)	>1022	0.76	0.70
MLR	0.77 (0.69–0.85)	>0.356	0.85	0.62
SII	0.76 (0.67–0.84)	>1410	0.82	0.66
NLR	0.75 (0.67–0.83)	>4.59	0.85	0.57
PLR	0.70 (0.60–0.81)	>195	0.58	0.81

AUC 95% CIs were estimated via bootstrap resampling (2000 iterations). AISI: Aggregate index of systemic inflammation, AUC: Area under the curve; CI: Confidence interval; CRP: C-reactive protein; MLR: Monocyte-to-lymphocyte ratio; NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; ROC: Receiver operating characteristic; SII: Systemic immune-inflammation index; SIRI: Systemic inflammation response index.

Figure 4.

Receiver operating characteristic (ROC) curves of inflammatory biomarkers for gangrenous cholecystitis. The dashed diagonal represents the line of no discrimination. AISI: Aggregate index of systemic inflammation; AUC: Area under the curve; CRP: C-reactive protein; MLR: Monocyte-to-lymphocyte ratio; NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; ROC: Receiver operating characteristic; SII: Systemic immune-inflammation index; SIRI: Systemic inflammation response index.

Multivariate Logistic Regression and Model Performance

In the prespecified multivariate logistic regression model (CRP+SIRI+age+sex), CRP was the only independent predictor of gangrenous cholecystitis (Table 4). Each 10 mg/L increase in CRP was associated with a 14% increase in the odds of gangrenous cholecystitis (OR 1.14, 95% CI 1.09 to 1.20).

Table 4.

Multivariable logistic regression for gangrenous cholecystitis

Predictor	OR (95% CI)	p
CRP (per 10 mg/L)	1.14 (1.09–1.20)	<0.001
SIRI (per 1 unit)	1.07 (1.00–1.16)	0.057
Age (per 1 year)	1.01 (0.97–1.04)	0.718
Male sex (vs. female)	1.85 (0.73–4.66)	0.192

Complete-case analysis (n=178). The CRP odds ratio is reported per 10 mg/L increment. SIRI and age are reported per one unit increase. Sex is reported as male versus female. Model performance metrics: AUC=0.83; Brier score=0.103; average precision=0.641. AUC: Area under the curve; CI: Confidence interval; CRP: C-reactive protein; HC3: Heteroscedasticity-consistent (type 3) standard errors; OR: Odds ratio; SIRI: Systemic inflammation response index.

SIRI did not reach the conventional significance threshold (p=0.057) and should therefore not be interpreted as an independent predictor in this cohort. Age and male sex did not reach statistical significance, and no significant multicollinearity was detected. In a sensitivity analysis, SIRI was replaced in turn by each of the other five prespecified hemogram-derived indices (Supplementary Table S1). The CRP estimate was stable in every specification (odds ratio 1.013 to 1.015 per 1 mg/L, all p<0.001), and model performance was essentially unchanged (AUC 0.82 to 0.83; Brier score 0.101 to 0.104). The cellular index reached nominal significance when SIRI was replaced by NLR (p=0.041), PLR (p=0.016), SII (p=0.014) or AISI (p=0.041). These p values were not adjusted for multiplicity and the corresponding effect sizes were very small (odds ratios 1.000 to 1.07 per unit increase). Model discrimination and the CRP estimate were therefore stable across substitutions, although the statistical significance of the cellular index varied. (Supplementary Table S1.)

DISCUSSION

In this single-center retrospective cohort study, we evaluated the comparative performance of multiple inflammatory biomarkers in predicting the presence of gangrenous cholecystitis in patients who underwent surgery for acute cholecystitis. In our study, CRP alone demonstrated the highest discriminatory power. Hemogram-derived composite indices exhibited moderate discriminatory power. Furthermore, in the predefined model, CRP was verified as an independent predictor, and overall model performance in this cohort was acceptable, although the model was not externally validated. The most important contribution of this study to the literature is that it directly compared hemogram-derived composite indices (particularly SIRI/SII/AISI) within the same cohort and under the same endpoint definition, thereby establishing a quantitative basis for determining which indicator may provide more meaningful risk stratification in practice.

The fact that CRP alone demonstrated the strongest discriminatory performance is consistent with the findings of studies on the spectrum of complicated acute cholecystitis. Bouassida et al.^[14] reported that CRP is a strong predictor of clinically significant outcomes, such as advanced acute cholecystitis and conversion to open surgery, in a large prospective cohort study. In our study, focusing on the direct presence of gangrene, necrosis as an outcome and reinforcing the definition through pathological and operative assessment supports a close relationship between CRP, an acute-phase reactant reflecting the systemic inflammatory burden and severe tissue damage.

Our findings regarding inflammatory biomarkers are generally consistent with those of previous studies, indicating an association between NLR and the severity and complicated course of acute cholecystitis. Micić et al.^[12] reported that NLR could predict severe acute cholecystitis; similarly, studies conducted in different subgroups have shown that pre-operative NLR may be parallel to an increased inflammatory burden.^[15,16] However, the heterogeneity of evidence regarding NLR has also been emphasized at the meta-analysis level. Kler et al.^[17] reported in a meta-analysis that NLR may be a diagnostic/prognostic indicator for acute cholecystitis; however, they noted that results should be interpreted with caution due to differences in study designs, endpoint definitions, and cutoff values.^[17] In our study, the moderate discriminatory power of NLR and the prominence of CRP can be considered a possible reflection of this heterogeneity.

Chen et al.^[18] demonstrated that these indices can contribute to clinically meaningful scenarios, such as distinguishing between acute and chronic cholecystitis; however, there is a gap in the literature regarding gangrenous cholecystitis. Our study found that inflammatory biomarkers alone can provide more consistent guidance in predicting gangrenous cholecystitis than cell counts. However, it also provides evidence supporting the notion that, as most of these indices are derived from the same hematological components, the claim of a “best” marker cannot currently be definitively established for the derived inflammatory biomarkers. The biological basis of these findings is consistent with the acute-phase response and natural immune activation. In acute inflammation, hepatic acute-phase proteins rise rapidly; the severity of this response may parallel tissue damage and infection load.^[19,20] Hemogram-derived indices, on the other hand, combine different dimensions of systemic inflammation on the same scale by reflecting cellular components such as neutrophilia, relative lymphopenia, and monocyte response.^[21] Indices that include the platelet component can indirectly capture the inflammation–thrombosis interaction and the concept of immunothrombosis; the ability of severe inflammatory conditions to increase the thrombotic tendency at the microvascular level can be explained by these relationships.^[22] These mechanisms account for the univariable elevation of biomarkers such as SIRI/SII/AISI in gangrenous cases observed in our study, but this elevation did not translate into independent predictive value beyond CRP in the multivariable analysis.

The Tokyo Guidelines provide a fundamental reference for severity classification and treatment strategies in the clinical management of acute cholecystitis.^[1,23] However, in practice, the degree of tissue gangrene and necrosis encountered in the operating theater may vary among patients who may fall within the same clinical severity category, which may create difficulties in surgical timing and strategy selection. In this context, markers that can be generated independently of imaging and routine blood tests may provide additional predictive value to clinical assessment alongside the Tokyo Guidelines. Similarly, the existence of biomarker studies aimed at predicting the need for percutaneous cholecystostomy indicates that laboratory-based risk classification can be integrated into decision-making processes.^[23] Our study may enable earlier detection of risk and individualization of treatment management, particularly in cases with suspected severe inflammation or necrosis, rather than serving as an alternative to existing scoring systems.

From a practical surgical-strategy perspective, the present findings suggest several operational implications that warrant prospective evaluation. In patients presenting with markedly elevated CRP at admission – and therefore a higher prior probability of underlying gangrene – clinicians may consider prioritising the case on the emergency operative list, ensuring senior surgical involvement given the recognised increase in operative difficulty in gangrenous disease, and preparing for technically demanding scenarios including subtotal cholecystectomy or early conversion when a safe dissection plane cannot be established, in line with current emergency-surgery consensus guidelines.^[24] The cohort-specific CRP cutoff identified here (61 mg/L) is not yet validated for clinical decision-making, but it may inform prospective protocols linking biomarker thresholds to perioperative checklists such as broader-spectrum antibiotic coverage, intraoperative drain placement, and a more vigilant post-operative monitoring strategy. The hemogram-derived indices, given their lack of independent predictive value beyond CRP, are not recommended as standalone decision aids; however, when integrated into structured laboratory panels alongside CRP and the Tokyo Guidelines-based assessment, they may help to standardize risk communication with patients and surgical teams. These implications remain hypotheses generated by a single-center cohort and require prospective multicenter validation before clinical adoption.

Study Strengths and Limitations

The strengths of this study include its sequential case series design, definition of outcome by pathology and intraoperative assessment, and head-to-head comparison of multiple biomarkers within the same cohort. Several limitations should be acknowledged. First, the single-center retrospective design limits generalizability and exposes the analysis to residual selection bias inherent to retrospective cohorts. Second, the analytic cohort was restricted by design to patients who underwent same-admission laparoscopic cholecystectomy; patients managed conservatively or with percutaneous cholecystostomy were not eligible, because gangrenous cholecystitis could not be ascertained in those settings by the present intraoperative/pathological outcome definition. This introduces a spectrum bias: the findings apply to surgically managed acute cholecystitis and may not generalize to the highest-risk patients diverted to non-operative or interventional pathways. Third, the cutoff values are cohort-specific, and the absence of external validation means that the diagnostic thresholds reported here require confirmation in independent, multicenter cohorts before clinical adoption. Finally, because the hemogram-derived indices share overlapping cellular components, claims of superiority among the retained markers – even within the prespecified seven-marker design (see Methods) – remain provisional in this cohort, and the borderline performance of SIRI in the multivariable model underscores that the additional value of any cellular index beyond CRP remains uncertain.

CONCLUSION

This study evaluated inflammatory markers obtained from routine blood tests for the pre-operative prediction of gangrenous cholecystitis. CRP emerged as the most powerful single discriminating marker and the only independent predictor in multivariable analysis. The hemogram-derived composite indices showed moderate univariable discrimination, but their additional value beyond CRP remains uncertain in the present cohort. CRP, which is rapidly available during emergency presentation, may therefore offer useful pre-operative information when imaging is inadequate. However, the cohort-specific cutoff values reported here are not yet ready for clinical use and require validation in independent multicenter populations before being adopted as decision thresholds. Any potential complementary role of hemogram-derived indices similarly requires further evaluation in larger, prospective cohorts.

Ethics Committee Approval

Ethical approval for the study was obtained from the Scientific Research Ethics Committee of Sakarya University (Date: 20.02.2025, Decision No: E-43012747-050.04–450863-86).

Peer-review

Externally peer-reviewed.

Authorship Contributions

Concept: A.Ö.C., N.F.; Design: A.Ö.C.; Supervision: F.A.; Resource: E.E:; Materials: E.S.; Data collection and/or processing: E.T., M.Y.; Analysis and/or interpretation: A.Ö.C.; Literature review: A.Ö.C.; Writing: A.Ö.C.; Critical review: F.A., N.F.

Conflict of Interest

None declared.

Financial Disclosure

The author declared that this study has received no financial support.

Use of AI for Writing Assistance

Disclosure on the use of Artificial Intelligence Generated Content (AIGC). During the preparation of this manuscript, the authors used an artificial intelligence based tool to improve the readability and grammar of the text, and to assist with the plotting code used to generate the figures from the raw dataset. All figures were produced directly from the study data, and all output was reviewed and edited by the authors, who take full responsibility for the content of the publication.