Introduction

Preoperative diagnosis of small bowel obstruction types is crucial for choosing the optimal surgical strategy. Patients with segmental intestinal ischemia due to strangulation, volvulus, intussusception, gallstone, or bezoar pressure injury require immediate surgery, whereas, in the remaining patients, non-surgical treatment is successful [1, 2].

Due to the variety of etiologies, the grade and level of mesenteric vascular compression, the stage of ischemia, and individual patient characteristics, the clinical diagnosis of intestinal ischemia in bowel obstruction is challenging regardless of surgeon experience [3]. Laboratory methods are routinely used to diagnose ischemia. According to the National Clinical Guidelines (NCG), leukocytosis over 14.0×10⁹, acidosis, and high blood amylase are "highly likely" to support strangulation obstruction [4]. However, some studies have found low specificity of laboratory tests [5, 6]. Higher accuracy in the ischemia diagnosis using instrumental methods has been reported [2], although there is considerable variation in the rates. Thus, ultrasound (US) has 53.3–87% and multislice computed tomography (MSCT) 90–95% informative value for diagnosing obstruction [4, 7–9]. Contrast-enhanced X-ray enterography, while not a diagnostic method for intestinal ischemia, as a last-line diagnosis of complete bowel obstruction, probably allows approach standardization and shortens the time to the decision to operate [10]. However, the accuracy of diagnostic methods and their prognostic value may vary considerably due to variations in the healthcare organization in healthcare facilities, regions, and countries.

Objective: to evaluate the accuracy of clinical, laboratory, and instrumental methods for intestinal ischemia diagnosis in small bowel obstruction in acute care hospitals.

Materials and methods

Results from a previously reported [11] multicenter observational retrospective study of 158 consecutive patients with non-cancer-related small bowel obstruction (SBO) treated in four hospitals of the Moscow Healthcare Department between May 2017 and December 2019 were analyzed.

Exclusion criteria were SBO due to strangulation of hernia, tumor obstruction, colonic obstruction, adynamic ileus, and early adhesive obstruction.

Before starting the study, the study organizers informed physicians at the healthcare facilities about the general principles of the primary examination of patients with SBO NCG and international guidelines [2, 4, 12] recommend laboratory tests, routine MSCT use for intestinal ischemia diagnosis, and, if ischemia is ruled out, the water-soluble contrast agent passage test. Abdominal radiography and ultrasound are recommended as screening methods. In the case of non-surgical treatment, continuous clinical monitoring is necessary; monitoring of laboratory values and instrumental examinations is recommended every 12 hours.

Patient data for 123 parameters were prospectively entered into a database in Microsoft Access. The structure of small bowel obstruction according to the etiology and the presence of intestinal ischemia was established. The group of intestinal ischemia includes patients with confirmed compression of mesenteric vessels and/or intestinal wall by a single adhesion or anatomical structure "isolating" intestinal segment as well as due to compression of the intestinal wall by a large intraluminal mass (bezoar, gallstone). Individual characteristics of patients were established, examination results on admission, after 12 hours of inpatient treatment, and overall outcomes at discharge were registered.

The significance of clinical, laboratory, and instrumental methods for diagnosing intestinal ischemia was analyzed. The role of MSCT and water-soluble contrast agent (WSCA) passage test on patient survival was evaluated.

Limitations on examining patients

Limitations on laboratory testing. Some physicians routinely did not use laboratory tests (acid-base status, lactate levels). In one hospital (in Moscow!), the tests were not available 24 hours a day. In addition, in one hospital, the administration has recommended that physicians reduce the usage of laboratory tests. Due to the above limitations, some laboratory tests were not performed on the first day of inpatient treatment.

Limitations on MSCT, ultrasound, and abdominal radiography. Before the study, three hospitals did not use MSCT as a diagnostic method for SBO. During the study, only in one hospital, MSCT was performed on a mandatory basis; in the remaining hospitals, MSCT was used at the surgeon's discretion and if technically possible. Many radiologists from all healthcare institutions throughout the study were unaware of the practical value of diagnosing intestinal ischemia in bowel obstruction. The diagnosis of intestinal ischemia using MSCT was considered accurate, including the initial misdiagnosis corrected after additional consultations with radiologists and other surgeons prior to surgery.

Due to the frequent unavailability of staff radiologists, abdominal radiographs were analyzed by the surgeon on admission; the impression of the images by the radiologist, in most cases, is available post factum. The ultrasonographers who performed the studies were usually "on-call"; they did not participate in the daily data analysis and had no feedback on the results of the studies they performed. In two hospitals, all abdominal ultrasounds were routinely reviewed by "daytime" ultrasonographers.

Statistical analysis. The study data were subjected to statistical analysis using parametric and nonparametric methods. Accumulation, correction, systematization of the initial data, and visualization of the obtained results were carried out in Microsoft Access spreadsheets. Statistical analysis was performed using STATISTICA 13.3 software.

Quantitative parameters were assessed for normal distribution using the Shapiro–Wilk test (for less than 50 subjects) or the Kolmogorov–Smirnov test (for more than 50 subjects); skewness and kurtosis parameters were also used. The Mann-Whitney test was used to compare between-group distributions of quantitative variables and the chi-square test (χ²) for categorical variables.

Risk assessment was performed by the Kaplan–Meier method with the generation of cumulative mortality probability curves. In all cases, at p-value<0.05, the groups were considered to have differences in the studied parameter at the 5% significance level.

The study was approved by the Local Ethics Committee of the Pirogov Russian National Research Medical University. All study participants provided informed consent to be included in the study and to have their personal data processed.

Results

The structure of small bowel obstruction is presented in the Table 1. Seventy-nine (50.0%) patients had surgery. Signs of intestinal ischemia were revealed in 46 (29.1%) patients. Segmental colonic resection was performed in 12 (7.6%) patients, including two patients with no signs of ischemia. The overall mortality was 7.6%; all deaths occurred in the postoperative period. Postoperative mortality was 15.2%. A comparison of clinical and medical history data in patients depending on the presence of ischemia is shown in Table 2.

Table 1. Distribution of patients by diseases, examination and treatment outcomes

Note. BIO — biliary intestinal obstruction; US — ultrasound; WSCA — water-soluble contrast agent; MSCT — multislice computed tomography.

Table 2. Distribution of patients depending on clinical data

Patients with intestinal ischemia had a history of fewer surgeries (p<0.001) or no surgeries ("virgin abdomen," p<0.001), the pain was typically persistent (p=0.005), and there were differences in stool retention time (p=0.004). No history of abdominal surgery ("virgin abdomen") was common in ischemia with a sensitivity of 0.24, specificity of 0.95 (!), and an accuracy of 0.76 [95% CI: 0,69–0,83, p=0,207]. No differences were found for factors traditionally considered as indications for surgery: the volume and nature ("intestinal") of the discharge via gastric tube.

Table 3 shows the comparative results of laboratory and instrumental examinations of patients.

Table 3. Distribution of patients depending on laboratory and instrumental data at admission and 12 hours later

Note. * — acid-base status, including lactate, was assessed in 80 (50.6%) patients at admission and 40 patients after 12 hours of treatment.

Laboratory tests recommended by NCG as criteria for ischemia in SBO: leukocytosis >14×10⁹ (p=1.0), lactate level >2.0 (p=0.28), HR >90 per minute (p=0.71) and fever (p=0.74), did not differ among patients in the studied groups (see Tables 2, 3). The only parameter for which differences were significant was the Na+ level; however, in both groups, the average level was within the reference range.

The test results of 116 patients 12 hours after admission without surgical treatment were analyzed to assess the significance of leukocytosis change over time for the diagnosis of intestinal ischemia. Of these, leukocytosis was monitored in 76 (65.5%) patients (Table 4). In patients with ischemia 12 hours after the beginning of treatment, there was a rising trend of leukocytosis compared with patients without ischemia: 37.5% vs. 14.5% (p=0.07); a downward trend of leukocytosis was less common: 25% vs. 61.3% (p=0.012). However, the leukocytosis trend, including values within the reference range, did not rule out small intestinal ischemia.

Table 4. White blood cell count in patients with small bowel obstruction throughout 12-hour follow-up

Similar results were obtained when comparing ischemia/non-ischemia groups in adhesive obstruction (Table 5). Laboratory tests (leukocytosis, lactate, acidosis, K⁺) after 12 hours of inpatient treatment did not differ between the groups, except for hyponatremia, which was more common in the ischemia group (p=0.03). For patients with strangulated obstruction, it was more common to have no history of surgery (p<0.001).

Table 5. Comparison of clinical and laboratory parameters in patients with adhesive intestinal obstruction

Note. * — % of patients who were not operated on at the time of the study.

Abdominal radiography and ultrasonography diagnosed SBO in 118 (78.1%) and 102 (68.9%) patients, respectively (in 77.5 and 70.9% of patients with adhesions). Fifty-eight patients were identified based on the "complete–partial obstruction" sign on radiography. Of these, the complete intestinal obstruction was more often consistent with ischemia (p=0.004). The diameter of the intestinal loops and the rate of free fluid in the abdominal cavity detected by ultrasound did not differ in the patients of both groups (p>0.05) (see Table 3).

MSCT was performed in 86 (54.4%) patients and identified bowel obstruction in all (100%) patients. In 22 of them, the ischemia of the small intestine was revealed subsequently; however, the diagnosis was established by MSCT in 17 (77.3%) patients; ischemia was not diagnosed in 5 (22.7%) patients (sensitivity for ischemia was 0.77). Another 10 (15.6%) patients were overdiagnosed with intestinal ischemia; the specificity of MSCT for ischemia was 0.84 and the accuracy was 0.83 [95% CI: 0.73–0.90, p=0.05]. MSCT performed in 8 patients with rare types of SBO established the obstruction etiology in 6 cases: biliary intestinal obstruction (BIO, 1), bezoar (2), volvulus (1), intussusception (2).

A univariate Kaplan-Meier analysis found no effect of MSCT on patient survival (see Figure, a): 7.8% [95% CI: 7.6–8.0%] vs. 6.5% [95% CI: 6.3–6.6%] (p=0.786), as for the probability of discharge on day 5: 57.1% [95% CI: 56.5–57.8%] vs. 64.5% [95% CI: 63.8-65.2%], and on day 10: 88.3% [95% CI: 88.1–88.6%] vs. 83.9% [95% CI: 83.4–84.3%].

Kaplan—Meier cumulative risk of mortality (with 95% confidence interval).

a — log-rank test for diagnosis with/without computed tomography; b — log-rank test for diagnosis with/without water-soluble contrast passage.

Logistic regression estimation revealed a 6.4-fold increase in the odds of an accurate diagnosis (ischemia/non-ischemia) in the case of an "ischemia" report by MSCT (95% CI: 0.025–0.85). However, a diagnosis of ischemia based on a CT scan did not increase the chance of a correct diagnosis compared with cases in which MSCT was not performed (95% CI: 0.233–6.347).

A WSCA passage test was performed in 105 (69.9%) patients (see Table 1). In 82 (78.1%) of them WSCA passed into the colon in 14.5±12.7 hours (Me=12), in less then 12 hours (inclusive) after admission in 52 (49.5%) patients, and after 12 hours in 30 (28.6%) patients.

In 23 (21.9%) patients, WSCA did not pass into the colon; these patients underwent surgery. In addition, despite the WSCA passage into the colon, surgery was performed in 3 (2.9%) more patients due to persisting obstruction; two had adhesive SBO, and one had BIO. A total of 26 (24.8%) patients had surgery after WSCA administration, and 11 (10.5%) of them had intestinal ischemia revealed during surgery.

Thus, the contrast agent passage into the colon predicted the obstruction resolution with a sensitivity of 97% and specificity of 100%.

Of the patients operated on after WSCA administration (n=30), 4 (13.3%) patients died, including 1 (9.1%) with small bowel ischemia. Univariate analysis showed that the use of WSCA in adhesive SBO was associated with lower mortality (see Figure, b): 4.1% [95% CI: 4.0–4.2%] vs. 14.3% [95% CI: 13.7–14.9%] (p=0.032) without assessing group comparability.

Discussion

Accurately identifying patients able to benefit from surgery or non-surgical treatment is one of the main areas of research in bowel obstruction. Preoperative diagnosis of the bowel obstruction etiology in all patients is currently impossible; therefore, the surgical approach is based on the probable diagnosis. The ideal surgical approach involves immediate surgery in all patients with strangulation, early identification of patients with absolute indications for surgery, and a minimum of unnecessary surgery [13]. Currently, due to the risk of missed ischemia, many surgeons operate on almost all patients early [14], even though excessive surgical activity is associated with high mortality [15, 16]. Approximately half of the deceased patients with adhesive obstruction in our study had no signs of ischemia and no absolute indication for surgery. Dozens of publications [1, 17], of which, unfortunately, there are very few domestic ones [18], address the issue. The ischemia group accounted for almost 30% of the overall group. A similar strangulation rate was reported by S.G. Shapovalyants [19]. However, most English-language publications report the incidence of ischemia in SBO at about 10% [5, 20].

Most routine laboratory tests in our study showed no specificity for ischemia, as reported in earlier studies [5, 6]. The meta-analysis showed that the sensitivity and specificity of leukocytosis for mesenteric ischemia were only 0.80 (95% CI: 0.66-0.91) and 0.50 (!) (95% CI: 0.31-0.69) [21], respectively, which is consistent with our data. The statistically significant relationship of some factors (hyponatremia, "virgin abdomen") with intestinal ischemia has also been reported in earlier studies [5, 6, 22]. We found no statistically significant differences in lactate levels traditionally used as a criterion for intestinal ischemia [1, 4, 21] (p=0.076). Monitoring laboratory values at 12 hours also showed no difference between the groups. A decrease in leukocytosis after 12 hours of inpatient treatment was more often observed in the absence of intestinal ischemia; however, no change in leukocytosis ultimately ruled it out. It confirms the relevance of the surgical postulate about the need to suspect strangulation in each patient with intestinal obstruction until its complete (!) resolution [3].

The role of the quality of the tests in our negative assessment of the intestinal ischemia laboratory diagnosis is unclear. The accuracy of laboratory tests requires strict adherence to test technology. Thus, L-lactate was shown to be stable in vitro for no more than 15 minutes after blood sampling [23].

In our study, radiography and ultrasound showed low sensitivity for bowel obstruction, consistent with WSES data [2]. Complete obstruction with small bowel dilatation of more than 3 cm, fluid levels, and the absence of gas in the colon on radiography were more often consistent with intestinal ischemia, as previously reported [24]. In our study, free fluid detected by abdominal ultrasound was irrelevant as a criterion for intestinal ischemia, and the method itself was irrelevant for diagnosing intestinal obstruction types. Ultrasound in intestinal obstruction can be beneficial [25], but the study demonstrated inferior diagnostic capabilities in our setting.

MSCT is currently considered a routine method for diagnosing bowel obstruction with a sensitivity and specificity for ischemia of 90–96% [9]. In our study, these figures were achieved even though the diagnostic quality criteria were not stringent, and the results, to some extent, reflect the prospective accuracy of the method. J.H. Kim et al. reported similar MSCT sensitivity and specificity rates (71% and 83%, respectively) for intestinal ischemia [26]. The effectiveness of MSCT for the diagnosis of rare types of BO should be noted. It was reported that the diagnosis of BIO or bezoar without MSCT often takes several days or even weeks after the initial treatment [27].

Routine use of MSCT is a likely factor in reducing mortality, presumably through a more accurate definition of surgery indications [28]. Authoritative international societies and publications emphasize the dependence of MSCT accuracy in intestinal obstruction on the level of equipment and specialists [2, 29]. Of the methods assessed in our study, only MSCT has the prospect of becoming a driver of improvement through training specialists and informed choice of an individual standardized approach.

Analysis of our results of the WSCA passage test does not allow us to draw definitive conclusions about its benefits. On the one hand, almost 80% of patients recovered without surgery after WSCA administration; on the other hand, 10% of patients in the long-term study initially required immediate surgery because of ischemia, although delayed surgery was not associated with increased mortality. One publication reports 2–8 hours from hospital admission as the optimal time before surgery [14]. According to our data, non-surgical resolution of the obstruction occurred significantly later in most patients. It has been reported that the contrast agent passage into the colon 24 hours after administration predicts resolution of obstruction with a sensitivity of 96–97% and specificity of 96–98% [10, 30]. Our data support these observations and the need for "vigilant monitoring" during the study [3]. It should be noted that the evidence for the benefits of routine WSCA passage tests cannot be considered rigorous. Thus, a recent randomized clinical trial has not shown any benefit of the technique [31].

Study limitations

The study group was small, although it included more than 6.0% of the annual number of patients with adhesive intestinal obstruction revealed in Moscow [32]. The intestinal ischemic injury heterogeneity by the extent, duration of ischemia, and reversibility of morphological changes should be noted, which may have influenced the variability of the results. The study had limitations in the completeness of primary and follow-up instrumental and laboratory examinations and showed mortality rates several times higher than the average for Moscow.

Conclusion

Routine laboratory tests were not specific for intestinal ischemia and, therefore, should not be considered the only criteria for the surgical strategy for bowel obstruction. The routine abdominal ultrasound in the emergency setting was irrelevant for diagnosing small bowel obstruction types and intestinal ischemia. The use of a water-soluble contrast agent passage test is probably beneficial, but the need for surgery may be determined with considerable delay provided the patient is monitored vigilantly. Of the evaluated methods, only the multislice computed tomography provided acceptable diagnostic accuracy and is the only method with prospects of diagnostic quality improvement through technical support and training of surgeons and radiologists.

Authors contribution:

Study concept and design — Tyagunov A.E., Sazhin A.V.

Data collection and processing — Tyagunov A.A., Ermakov I.V.

Statistical analysis — Nechay T.V.

Text writing — Tyagunov A.E., Tyagunov A.A., Ermakov I.V.

Editing — Sazhin A.V.

The authors declare no conflicts of interest.