Early identification of reversible strangulating ischemia in adhesive small bowel obstruction using non-contrast computed tomography and routine laboratory parameters

Abstract

BACKGROUND

Early identification of reversible strangulating ischemia in patients with adhesive small bowel obstruction (ASBO) remains challenging.

AIM

To develop a preliminary predictive model for strangulating ischemia based on risk factors, non-contrast computed tomography (CT) findings, and routine laboratory parameters.

METHODS

This retrospective study included data from 103 patients with surgically confirmed ASBO. Patients were classified into two groups based on the ischemia type: Non-strangulating (n = 74) and strangulating ischemia (n = 29). Clinical variables, laboratory results, and CT features were analyzed using multivariate logistic regression and receiver operating characteristic (ROC) curve analyses.

RESULTS

Shorter symptom duration before admission [odds ratio (OR) = 0.972, 95% confidence interval (CI): 0.953-0.992], mesenteric haziness (OR = 4.723, 95%CI: 1.099-20.294), closed-loop obstruction (OR = 7.181, 95%CI: 1.494-34.526), and lower base excess (OR = 0.604, 95%CI: 0.420-0.869) were independently associated with strangulating ischemia. The combined model yielded an area under the ROC curve of 0.897 (95%CI: 0.822-0.948), with a sensitivity of 82.76% and a specificity of 93.24%.

CONCLUSION

A preliminary model combining non-contrast CT findings and routine laboratory parameters helps identify strangulating ischemia at a potentially reversible stage of ASBO and supports early surgical decision-making.

Key Words: Adhesive small bowel obstruction; Strangulating ischemia; Non-contrast computed tomography; Base excess; Predictive model

Core Tip: This retrospective study of 103 patients with adhesive small bowel obstruction identified four independent predictors of strangulating ischemia: Shorter symptom duration, mesenteric haziness, closed-loop obstruction, and lower base excess. A combined model incorporating non-contrast computed tomography findings and routine laboratory parameters achieved excellent discrimination (area under the curve 0.897). The model aids early identification of potentially reversible strangulating ischemia and supports timely surgical decision-making.

INTRODUCTION

Adhesive small bowel obstruction (ASBO) is the most common cause of small bowel obstruction[1], accounting for approximately 56% of cases with an identifiable etiology and up to 75% of all cases, most commonly as a consequence of previous abdominal surgery[2,3]. ASBO remains a major cause of morbidity, mortality, and healthcare expenditures, with a mean hospital stay of 8 days and an in-hospital mortality rate of 3% per episode[1].

Clinically, ASBO can be broadly classified into simple and strangulating obstruction. Strangulating obstruction is a dynamic pathological process that may initially present as reversible intestinal ischemia but can subsequently progress to irreversible bowel necrosis if timely intervention is not implemented[1,4]. Determining the optimal timing for surgery in patients with ASBO remains a major clinical challenge. Unnecessary surgical intervention may increase surgical trauma and contribute to recurrent adhesion formation, predisposing patients to future episodes of bowel obstruction. In contrast, delayed surgery in patients with occult strangulating ischemia may result in bowel necrosis, bowel resection, and/or poor clinical outcomes. As such, early identification of the reversible ischemic stage is of particular clinical importance because it may provide a critical window for timely surgical intervention before irreversible intestinal injury develops. This issue is especially challenging in patients without overt peritoneal signs, in whom strangulating ischemia is often difficult to recognize based on routine clinical assessment alone. In routine practice, the assessment of strangulating ischemia in ASBO usually depends on an overall evaluation of clinical presentation, laboratory findings, and imaging features. However, the diagnostic utility of any single parameter remains limited, particularly in the early stages or in patients without overt peritonitis[4-6]. A more integrated approach based on readily available preoperative variables may be helpful for improving early risk stratification and supporting surgical decision making.

As such, the present study aimed to identify factors associated with strangulating ischemia in patients with ASBO, and to construct a preliminary predictive model based on laboratory and imaging parameters for early risk stratification.

MATERIALS AND METHODS

Study population

This study complied with the ethical requirements of the Declaration of Helsinki and was approved by the Medical Ethics Committee of Tianjin Union Medical Center, Tianjin, China; Approval No. (2026) Fast Review No. B050. Clinical data from patients diagnosed with ASBO at the Tianjin Union Medical Center between January 2019 and December 2025 were retrospectively collected. The inclusion criteria were as follows: Age ≥ 18 years; complete laboratory and imaging data available before surgery; and intraoperative confirmation of ASBO. The exclusion criteria were concomitant perforation, strangulation with irreversible bowel necrosis, and severe adhesions, including abdominal cocoons.

Methods

For all included patients, the diagnosis of reversible strangulating ischemia was confirmed intraoperatively by the attending surgeon based on predefined standardized criteria after the release of the causative adhesion or internal hernia: (1) The affected bowel segment changed from dark purple or dark red to pink within 15 minutes; (2) Visible peristalsis returned to the involved bowel; and (3) Palpable arterial pulsations were present at the mesenteric border. None of the patients underwent bowel resection. Surgeries were performed or directly supervised by two senior attending surgeons (Zhang JD and Xu SC) to ensure consistency in intraoperative judgment. However, given the retrospective design, the degree of subjective bias in assessing bowel color and peristalsis cannot be completely excluded, which is a limitation. Finally, the present study included only patients with surgically confirmed ASBO. A total of 178 patients were screened, of whom 55 patients with strangulation-related bowel necrosis were excluded, including 35 who developed bowel necrosis and subsequently required bowel resection because surgery was not performed in a timely manner. In addition, nine patients with perforations, five with incomplete laboratory data, four with incomplete imaging data, and two with abdominal cocoons were excluded (Figure 1).

Figure 1  Flowchart of patient selection.

Variables and data collection

Demographic characteristics (age, sex, height, and weight), symptom duration (defined as the interval between symptom onset and hospital admission), and routine laboratory parameters were recorded for all patients. Laboratory data obtained from the first blood tests after hospital admission were collected, including complete blood count parameters [white blood cell (WBC) count, neutrophil count (Neu), lymphocyte count, platelet count], biochemical indicators [albumin (ALB), aspartate aminotransferase, lactate (LAC) dehydrogenase, creatine kinase (CK), CK-MB, blood urea nitrogen], coagulation-related indices [D-dimer (D-D), fibrinogen (Fib)], arterial blood gas parameters [base excess (BE), and LAC]. Inflammatory markers such as the platelet-to-lymphocyte ratio (PLR) were calculated.

Computed tomography imaging assessment

All computed tomography (CT) examinations in this cohort were non-contrast scans performed as part of a routine emergency evaluation. The CT images were reviewed in routine clinical practice by attending radiologists and surgeons. Imaging assessments focused on findings suggestive of ischemic strangulation in patients with ASBO. The evaluated CT features included small bowel fecal signs, bowel wall thickening, decreased bowel wall attenuation, mesenteric haziness, ascites, closed-loop obstruction, whirl sign(s), and mesenteric clustering. CT findings were recorded according to the initial clinical reports and correlated with the intraoperative findings for confirmation (Figures 2 and 3). To reflect real-world diagnostic practice, the present study relied on routine clinical CT interpretation(s) rather than a retrospective blinded image review. The presence or absence of these radiological signs was subsequently analyzed in relation to ischemic strangulation.

Figure 2 Abdominal computed tomography findings. A: Axial view = U-/C-shaped bowel loop (orange box) and localized mesenteric haziness (asterisk); B: Coronal view: Closed-loop mechanism (thin orange arrow), collapsed distal bowel (short thick orange arrow), and dilated proximal bowel (long thick orange arrow); C: Sagittal view = U-/C-shaped bowel loop (orange box) and mesenteric haziness (asterisk); D: Coronal view = U-/C-shaped bowel loop (orange box) and mesenteric haziness (asterisk).

Figure 3 Intraoperative findings. A: The greater omentum adheres to the mesentery of the ascending colon, forming a defect (forceps). A herniated small bowel segment appears dark purple-red (white star); B: After release of strangulation, bowel color improves (white bracket) and peristalsis returns. No bowel resection was performed.

Statistical analysis

Data analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, United States). All statistical tests were two-sided, and differences with P < 0.05 were considered statistically significant. Baseline characteristics and clinical variables were summarized descriptively. Normally distributed continuous variables are expressed as mean ± SD, whereas non-normally distributed continuous variables are expressed as medians with interquartile ranges (IQR, Q₁-Q₃). Categorical variables are reported as n (%). The patients were categorized into two ischemia groups: Non-strangulating and strangulating. Between-group comparisons were performed using the independent-samples t-test or Mann-Whitney U test for continuous variables, as appropriate, and the Pearson χ² test for categorical variables. Variables that were significantly associated with strangulating ischemia in univariate analysis were entered into a multivariable binary logistic regression model to identify independent predictors. Non-strangulating and strangulating ischemia were coded as 0 and 1, respectively. Model fit was evaluated using the Hosmer-Lemeshow goodness-of-fit test. Receiver operating characteristic curve (ROC) curve analysis was performed for significant predictors, and the area under the curve (AUC), cut-off value, sensitivity, specificity, and 95% confidence interval (CI) were calculated. A combined predictive model was developed, and its predictive value was assessed using ROC curve analysis.

RESULTS

Baseline characteristics of the study population

Data from 103 patients [57 females (55.34%) and 46 males (44.66%)] with surgically confirmed ASBO were included in this study. The mean ± SD age of the overall cohort was 62.77 ± 13.06 years and the mean body mass index (BMI) was 22.29 ± 2.54 kg/m². Seventy-two patients (69.90%) had a history of abdominal surgery. Regarding the clinical presentation, abdominal distension was observed in 94 (91.26%) patients, nausea and vomiting in 81 (78.64%), and cessation of flatus and defecation in 80 (77.67%) patients. CT imaging yielded the following findings among the cohort: Small bowel fecal signs, n = 19 (18.45%); bowel wall thickening, n = 23 (22.33%); mesenteric haziness, n = 33 (32.04%); ascites, n = 53 (51.46%); closed-loop obstruction, n = 29 (28.16%); and whirl signs, n = 13 (12.62%). Regarding laboratory findings, the mean ALB and Fib levels were 41.37 ± 5.14 g/L and 3.32 ± 0.87 g/L, respectively. The median symptom duration before admission was 24.00 hours (IQR 24.0-60.0 hours), and the median D-D level was 0.72 mg/L (IQR 0.45-1.11 mg/L). The detailed baseline characteristics are summarized in Table 1.

Table 1 Patient characteristics, n (%)/mean ± SD.

Variables	Overall population (n = 103)
Age	62.77 ± 13.06
Sex
Female	57 (55.34)
Male	46 (44.66)
BMI	22.29 ± 2.54
Neutrophil percentage	75.28 ± 7.95
ALB	41.37 ± 5.14
Fib	3.32 ± 0.87
Symptom duration (hour), M (Q1, Q3)	24.00 (24.00, 60.00)
WBC, M (Q1, Q3)	6.64 (5.12, 8.41)
PLR, M (Q1, Q3)	188.51 (134.57, 238.26)
LAC, M (Q1, Q3)	1.30 (1.02,1.80)
BE, M (Q1, Q3)	-2.00 (-3.3, -0.55)
LDH, M (Q1, Q3)	203.00 (180.00, 240.50)
CK, M (Q1, Q3)	57.00 (44.00, 90.00)
CK-MB, M (Q1, Q3)	5.00 (2.70, 8.00)
BUN, M (Q1, Q3)	6.04 (4.36, 7.29)
D-D, M (Q1, Q3)	0.72 (0.45, 1.11)
History of abdominal surgery
No	31 (30.10)
Yes	72 (69.90)
Abdominal distension
No	9 (8.74)
Yes	94 (91.26)
Nausea and vomitting
No	22 (21.36)
Yes	81 (78.64)
Cessation of flatus and defecation
No	23 (22.33)
Yes	80 (77.67)
Small bowel feces sign
No	84 (81.55)
Yes	19 (18.45)
Bowel wall thickening
No	80 (77.67)
Yes	23 (22.33)
Decreased bowel wall attenuation
No	87 (84.47)
Yes	16 (15.53)
Mesenteric haziness
No	70 (67.96)
Yes	33 (32.04)
Ascites
No	50 (48.54)
Yes	53 (51.46)
Closed-loop obstruction
No	74 (71.84)
Yes	29 (28.16)
Whirl sign
No	90 (87.38)
Yes	13 (12.62)
Mesenteric clustering
No	71 (68.93)
Yes	32 (31.07)

Data are presented as mean ± SD for normally distributed variables and median (Q₁, Q₃) for non-normally distributed variables. t: T-test; Z: Mann-Whitney test; M: Median; Q₁: 1^st quartile; Q₃: 3^st quartile; BMI: Body mass index; Fib: Fibrinogen; ALB: Albumin; WBC: White blood cell count; PLR: Platelet to lymphocyte ratio; LAC: Lactate; BE: Base excess; LDH: Lactate dehydrogenase; CK: Creatine kinase; CK-MB: Creatine kinase-MB; BUN: Blood urea nitrogen; D-D: D-dimer.

Univariate analysis of the non-strangulating and strangulating ischemia groups

The patients were divided into two ischemia groups: Non-strangulating (n = 74) and strangulating (n = 29). Univariate analysis revealed that, among the laboratory variables, the strangulating ischemia group exhibited a higher mean ALB level and higher median PLR and D-D levels, but a lower median BE level than the non-strangulating group, and the differences were statistically significant. In contrast, age, BMI, Neu percentage, Fib level, WBC count, Lac level, and other laboratory parameters showed no significant intergroup differences. In terms of disease course, the median symptom duration before admission was shorter in the strangulating ischemia group than in the non-strangulating group, and the difference was statistically significant. Regarding clinical symptoms, nausea and vomiting were more prevalent in the strangulating ischemia group, whereas abdominal distension, cessation of flatus, and defecation did not differ significantly between the groups.

Regarding the CT findings, mesenteric haziness, ascites, and closed-loop obstruction were more common in the strangulating ischemia group than in the non-strangulating ischemia group, and all were significantly associated with strangulating ischemia. In contrast, signs of small bowel feces, bowel wall thickening, decreased bowel wall attenuation, whirl signs, and the remaining imaging findings were not significantly different between the groups. In addition, sex and history of abdominal surgery were not significantly associated with strangulating ischemia (Table 2).

Table 2 Univariate comparison between non strangulated and strangulated ischemia, n (%)/mean ± SD.

Variables	Non-strangulated (n = 74)	Strangulated ischemia (n = 29)	Statistics	P value
Age	62.96 ± 13.14	62.28 ± 13.07	t = 0.238	0.812
Sex			χ2 = 0.815	0.367
Female	43 (58.11)	14 (48.28)
Male	31 (41.89)	15 (51.72)
BMI	22.11 ± 2.41	22.74 ± 2.83	t = -1.135	0.259
N (%)	75.74 ± 8.55	74.10 ± 6.15	t = 0.939	0.35
ALB	40.42 ± 5.15	43.78 ± 4.32	t = -3.113	0.002
Fib	3.40 ± 0.86	3.12 ± 0.87	t = 1.496	0.138
Symptom duration (hour), M (Q1, Q3)	48.00 (24.00, 72.00)	24.00 (12.00, 43.00)	Z = -3.755	0.001
WBC, M (Q1, Q3)	6.30 (4.79, 8.47)	6.70 (6.26,8.17)	Z = -0.935	0.35
PLR, M (Q1, Q3)	174.00 (121.95, 227.23)	223.30 (171.15, 263.33)	Z = -2.178	0.029
LAC, M (Q1, Q3)	1.30 (1.00, 1.80)	1.40 (1.20, 1.59)	Z = -0.603	0.547
BE, M (Q1, Q3)	-1.25 (-2.80, -0.30)	-3.30 (-4.60, -2.00)	Z = -3.572	0.001
LDH, M (Q1, Q3)	207.00 (179.50, 240.00)	199.00 (181.00, 256.00)	Z = -0.073	0.942
CK, M (Q1, Q3)	56.00 (37.00, 92.50)	66.00 (56.00, 83.00)	Z = -1.320	0.187
CK-MB, M (Q1, Q3)	5.60 (2.70, 7.83)	4.09 (2.80, 8.10)	Z = -0.279	0.781
BUN, M (Q1, Q3)	5.70 (4.14, 7.27)	6.30 (4.91, 7.65)	Z = -1.030	0.303
D-dimer, M (Q1, Q3)	0.63 (0.43, 0.97)	0.98 (0.79, 1.33)	Z = -2.717	0.007
History of abdominal surgery			χ2 = 1.177	0.278
No	20 (27.03)	11 (37.93)
Yes	54 (72.97)	18 (62.07)
Abdominal distension			χ2 = 0.562	0.454
No	5 (6.76)	4 (13.79)
Yes	69 (93.24)	25 (86.21)
Nausea and vomitting			χ2 = 5.027	0.025
No	20 (27.03)	2 (6.90)
Yes	54 (72.97)	27 (93.10)
Cessation of flatus and defecation			χ2 = 1.763	0.184
No	14 (18.92)	9 (31.03)
Yes	60 (81.08)	20 (68.97)
Small bowel feces sign			χ2 = 2.241	0.134
No	63 (85.14)	21 (72.41)
Yes	11 (14.86)	8 (27.59)
Bowel wall thickening			χ2 = 0.076	0.783
No	58 (78.38)	22 (75.86)
Yes	16 (21.62)	7 (24.14)
Decreased bowel wall attenuation			χ2 = 0.362	0.547
No	64 (86.49)	23 (79.31)
Yes	10 (13.51)	6 (20.69)
Mesenteric haziness			χ2 = 25.278	0.001
No	61 (82.43)	9 (31.03)
Yes	13 (17.57)	20 (68.97)
Ascites			χ2 = 4.954	0.026
No	41 (55.41)	9 (31.03)
Yes	33 (44.59)	20 (68.97)
Closed-loop obstruction			χ2 = 22.951	0.001
No	63 (85.14)	11 (37.93)
Yes	11 (14.86)	18 (62.07)
Whirl sign			χ2 = 0.307	0.58
No	66 (89.19)	24 (82.76)
Yes	8 (10.81)	5 (17.24)
Mesenteric clustering			χ2 = 0.905	0.341
No	49 (66.22)	22 (75.86)
Yes	25 (33.78)	7 (24.14)

t: T-test; Z: Mann-Whitney test; M: Median; Q₁: 1^st quartile; Q₃: 3^st quartile; BMI: Body mass index; N (%): Neutrophil percentage; Fib: Fibrinogen; ALB: Albumin; WBC: White blood cell count; PLR: Platelet to lymphocyte ratio; LAC: Lactate; BE: Base excess; LDH: Lactate dehydrogenase; CK: Creatine kinase; CK-MB: Creatine kinase-MB; BUN: Blood urea nitrogen; D-D: D-dimer.

Multivariable regression analysis

Statistically significant variables in the univariate analysis were entered into a multivariate binary logistic regression model. The Hosmer-Lemeshow goodness-of-fit test yielded a value of 4.978 (P = 0.760), indicating a good model fit. After adjusting for potential confounders, four variables were found to be independently associated with strangulating ischemia. Symptom duration was negatively associated with strangulating ischemia [odds ratio (OR) = 0.972, 95%CI: 0.953-0.992], indicating that a longer symptom duration was associated with a lower likelihood of strangulating ischemia. Mesenteric haziness was independently associated with an increased risk of strangulating ischemia (OR = 4.723, 95%CI: 1.099-20.294). Similarly, closed-loop obstruction was associated with a higher likelihood of strangulating ischemia (OR = 7.181, 95%CI: 1.494-34.526). BE was also negatively associated with strangulating ischemia (OR = 0.604, 95%CI: 0.420-0.869), suggesting that higher BE values were associated with a lower risk of strangulating ischemia. In contrast, nausea, vomiting, ascites, ALB level, PLR, and D-D level were not independently associated with strangulating ischemia after multivariate adjustment (Table 3).

Table 3 Multivariable regression analysis.

Variables	β	SE	Z	P value	OR	95%CI
						Low	High
Constant	-11.322	3.556	10.137	0.001
Nausea and vomitting (yes, no)	0.914	1.092	0.700	0.403	2.494	0.293	21.216
Symptom duration (hour)	-0.028	0.010	7.481	0.006	0.972	0.953	0.992
Mesenteric haziness (yes, no)	1.553	0.744	4.357	0.037	4.723	1.099	20.294
Ascites (yes, no)	1.489	0.858	3.009	0.083	4.433	0.824	23.842
Closed loop obstruction (yes, no)	1.971	0.801	6.055	0.014	7.181	1.494	34.526
ALB	0.108	0.069	2.429	0.119	1.114	0.973	1.277
PLR	0.010	0.006	2.864	0.091	1.010	0.998	1.022
BE	-0.504	0.186	7.381	0.007	0.604	0.420	0.869
D-dimer	0.907	0.617	2.160	0.142	2.476	0.739	8.294

Odds ratios and 95% confidence intervals were obtained from a multivariable regression model. Hosmer-Lemeshow = 4.978, P = 0.760 > 0.05. OR: Odds ratio; CI: Confidence interval; ALB: Albumin; PLR: Platelet to lymphocyte ratio; BE: Base excess.

ROC curve analysis of predictors of small bowel necrosis

ROC curve analysis was performed for the four independent predictors identified in the multivariate analysis: Symptom duration before admission, mesenteric haziness, closed loop obstruction, and BE.

The AUC for symptom duration was 0.735 (95%CI: 0.639-0.817), with an optimal cut-off value of ≤ 20 hours, corresponding to a sensitivity of 66.28% and a specificity of 77.84%. The AUC for mesenteric haziness was 0.757 (95%CI: 0.663-0.836), with a cut-off value of > 0, yielding a sensitivity of 68.97% and a specificity of 82.43%. The AUC for closed-loop obstruction was 0.736 (95%CI: 0.640-0.818), with a cut-off value of > 0, corresponding to a sensitivity of 62.07% and a specificity of 85.14%. The AUC for BE was 0.727 (95%CI: 0.631-0.810), with an optimal cut-off value of ≤ -3, yielding a sensitivity of 62.07% and a specificity of 77.38%.

All four individual predictors exhibited moderate predictive values with AUCs ranging from 0.70 to 0.80. A combined predictive model incorporating symptom duration, mesenteric haziness, closed-loop obstruction, and BE was developed. The ROC curve analysis demonstrated that the combined model yielded an AUC of 0.897 (95%CI: 0.822-0.948), with a sensitivity of 82.76% and a specificity of 93.24%. Compared with the individual predictors, the combined model demonstrated superior overall predictive performance for strangulating ischemia in patients with ASBO (Table 4 and Figure 4).

Figure 4 Receiver operating characteristic curves of combined model. Curves represent: Symptom duration [≤ 20 hours, area under the curve (AUC) = 0.735], mesenteric haziness (AUC = 0.757), closed-loop obstruction (AUC = 0.736), base excess (≤ -3, AUC = 0.727), and the combined model (AUC = 0.897). The combined model shows the highest area under the curve.

Table 4 Receiver operating characteristic curve analysis of predictors.

Predictors	AUC	SE	Z	P value	95%CI	Cut-off	Sensitivity (%)	Specificity (%)
Symptom duration (hour)	0.735	0.053	4.441	0.001	0.639-0.817	≤ 20	66.28	77.84
Mesenteric haziness	0.757	0.049	2.238	0.001	0.663-0.836	> 0	68.97	82.43
Closed loop obstruction	0.736	0.050	4.687	0.001	0.640-0.818	> 0	62.07	85.14
BE	0.727	0.058	3.922	0.001	0.631-0.810	≤ -3	62.07	77.38
Combined prediction	0.897	0.038	10.418	0.001	0.822-0.948	-	82.76	93.24

Cut-off values were determined using Youden’s index. AUC: Area under the curve; CI: Confidence interval; BE: Base excess.

DISCUSSION

ASBO is the most common cause of small bowel obstruction in routine surgical practice, making it highly relevant in daily clinical decision-making[1,7]. The key challenge in the management of ASBO is determining the optimal timing for surgical intervention[5,7]. Although early surgery may prevent progression to irreversible bowel injury, unnecessary surgery may also cause avoidable surgical trauma and contribute to recurrent adhesion formation, thereby increasing the risk of future obstructions. Therefore, the most clinically meaningful question is not simply how to identify bowel necrosis but also how to identify patients at the reversible ischemic stage when timely intervention may still prevent irreversible damage.

A major strength of this study was the specific focus on reversible strangulating ischemia rather than treating all strangulating obstructions as a single entity. In many previous studies, patients were classified as either strangulated or non-strangulated or according to whether bowel resection was ultimately performed[8-10]. However, classification based solely on bowel resection may not adequately reflect the dynamic progression from ischemia to necrosis. Some patients with strangulating ischemia may not require bowel resection at the time of surgery; however, this does not necessarily mean that they are comparable to patients with a simple obstruction. Grouping such patients together may obscure the critical transition phase during which surgical intervention is most beneficial. Our analysis addressed a clinically important but relatively underexplored stage by focusing on ischemic strangulation before the progression to necrosis.

Our findings revealed that a shorter symptom duration before admission, mesenteric haziness, closed-loop obstruction, and lower BE levels were independently associated with strangulating ischemia. Importantly, these variables were derived from non-contrast CT scans and routine laboratory investigations, which are readily available in emergency settings. This provides the model with considerable practical value because it can be easily integrated into real-world preoperative assessments without depending on contrast-enhanced imaging or specialized biomarkers. Mesenteric haziness and closed-loop obstruction likely reflect compromised mesenteric perfusion and mechanical strangulation, whereas lower BE levels indicate an early systemic metabolic response to bowel ischemia. The association between a shorter symptom duration and strangulating ischemia may further suggest a more abrupt and aggressive disease course.

Another notable finding in our study was that LAC showed no significant difference between the strangulating and non-strangulating ischemia groups, which appears to contradict the common clinical intuition that elevated LAC is a reliable marker of bowel ischemia. There are several possible explanations for this. First, during the potentially reversible stage of strangulating ischemia, anaerobic metabolism within the compromised bowel may be insufficient to produce a substantial increase in systemic LAC levels. The ischemic insult may be partial or intermittent, limiting LAC release into the circulation. Second, the body’s compensatory mechanisms, including hepatic LAC clearance and buffering capacity, may maintain LAC levels within the normal range during early ischemia, particularly in patients without overt hemodynamic compromise. From a clinical perspective, this finding has an important implication: Waiting for LAC levels to increase may delay surgical intervention until ischemia has already progressed beyond the reversible stage. In contrast, BE, which reflects a more integrated metabolic status and can detect subtle acid-base disturbances earlier, showed a significant independent association with strangulating ischemia in our study. Therefore, BE may serve as a more sensitive and timely metabolic indicator than LAC level in identifying patients in the reversible ischemic stage, supporting its inclusion in the predictive model.

Another notable finding was that strangulating ischemia occurred in the absence of overt peritonitis. In routine practice, waiting for peritoneal signs to trigger surgery may delay intervention until ischemia progresses beyond a reversible stage. Thus, the absence of classical peritoneal irritation should not delay surgical considerations in the presence of other risk factors. The clinical relevance of these findings is further supported by the timing of the evaluation. In this cohort, once strangulation was suspected based on the CT and laboratory findings, surgery was generally performed within 6 h after the CT examination, which largely corresponds to the routine preoperative preparation interval in emergency practice. This means that the imaging and laboratory parameters analyzed here were obtained during the actual decision-making window before surgery, rather than representing delayed or late-stage findings. Accordingly, the model may better reflect the reality of emergency surgical triage.

Our study specifically focused on reversible strangulating ischemia and excluded patients with established bowel necrosis. While this design aligns with our aim of identifying the critical therapeutic window, it may somewhat weaken clinical continuity along the spectrum from simple obstruction to ischemia and ultimately to necrosis. Nevertheless, the four independent predictors identified in our model-shorter symptom duration, mesenteric haziness, closed-loop obstruction, and lower BE-can theoretically be expected to progressively worsen as ischemia advances toward necrosis. For instance, BE would likely become more severely negative (e.g., from -3 to -8 or lower) as anaerobic metabolism intensifies. Mesenteric haziness may increase in extent and density owing to progressive edema and hemorrhagic exudation. Closed-loop obstruction, as a mechanical trigger, may persist or become more pronounced. Our model captures patients at a potentially reversible stage; however, extreme values of these parameters (e.g., BE ≤ -6 or extensive mesenteric haziness involving a larger territory) may indicate a higher probability of progression to necrosis. Future prospective studies should investigate the continuous relationship between these parameters and the risk of bowel necrosis, including the identification of specific thresholds that distinguish reversible ischemia from irreversible injury.

Another unexpected finding was that serum ALB levels were significantly higher in the strangulating ischemia group than in the non-strangulating ischemia group in univariate analysis, although this difference was not significant after multivariate adjustment. This finding appears to contradict the conventional view that hypoalbuminemia is associated with more severe disease. There are several possible explanations for this. First, the strangulating ischemia group had a significantly shorter symptom duration (median, 24 hours vs 48 hours in the non-strangulating ischemia group). The shorter disease course may have been insufficient to induce a measurable decline in ALB due to fasting, protein loss from the edematous bowel wall, or inflammatory consumption. Second, some patients in the non-strangulating ischemia group with a longer symptom duration might have had underlying differences in nutritional status or chronic inflammation, which could have contributed to lower baseline ALB levels. Third, acute hemoconcentration due to fluid sequestration in the obstructed bowel loop can transiently elevate ALB levels in the early phase of strangulation. Importantly, this finding should not be misinterpreted as higher ALB being a risk factor for strangulating ischemia. Rather, the observed difference was likely explained by the shorter disease duration in the strangulating ischemia group and the absence of chronic nutritional depletion. Therefore, ALB should be interpreted together with symptom duration, rather than as an isolated predictor.

The present study has several limitations. First, its retrospective, single-center design and relatively limited sample size may have restricted the generalizability of our findings. Particularly, the small number of patients with strangulating ischemia may have affected the stability of the multivariate model and increased the risk of overfitting. Second, although routine laboratory variables were included, not all potential confounders affecting coagulation function, ALB levels, or base-excess levels were excluded. Third, all CT examinations were non-contrast scans performed in an emergency setting. While this improves real-world applicability, it also precludes direct comparison with contrast-enhanced CT-based diagnostic approaches, including established scoring systems, such as the Millet score[11]. Fourth, the interpretation of CT data was based on routine clinical reports rather than on blinded retrospective image reevaluation, which may have introduced some degree of interobserver variability. Fifth, although the intraoperative criteria for reversible strangulating ischemia were standardized and applied by senior surgeons, the retrospective nature of this study precludes complete elimination of subjective bias in the assessment of bowel color recovery and peristalsis. Sixth, the proposed predictive model should be considered preliminary and requires external validation in a larger prospective cohort.

It is also important to clarify the scope of the proposed model. This study only included patients who underwent surgery because intraoperative confirmation of reversible strangulating ischemia is essential for defining the outcome. Thus, the model identified ischemic risk in patients already selected for surgery, but not across all patients with ASBO. Whether this applies to risk screening during initial conservative treatment remains uncertain. The absence of surgical selection bias in conservatively managed patients may alter predictor performance. Therefore, the current model is best viewed as a decision support tool for determining whether and when to proceed with surgery in patients with active operative considerations. For patients receiving non-operative management, external validation in prospective cohorts is needed before extrapolating our findings.

CONCLUSION

Early identification of strangulating ischemia before progression to irreversible necrosis remains a key challenge in the management of ASBO. A preliminary predictive model integrating non-contrast CT findings and routine laboratory parameters demonstrated good discriminatory performance, and may be a practical tool for early risk stratification and timely surgical intervention.