Patients and study approval
This retrospective study used the data of patients older than 18 years with acute small bowel obstruction admitted to the Division of Gastroenterology, Department of Internal Medicine, Nihon University School of Medicine between January 2014 and August 2018. The inclusion criteria were as follows: (1) presence of clinical symptoms and physical signs associated with acute small bowel obstruction such as abdominal pain, distention, nausea, and vomiting without stool production; (2) diagnosis of ASBO on radiography with computed tomography (CT), which was confirmed by no less than two clinicians or radiologists (Figure 1); and (3) admission to the Department of Gastroenterology at our hospital within one day after symptom onset. The exclusion criteria were as follows: (1) emergency surgery for suspected strangulation; (2) large bowel obstruction, cancer peritonitis, obstructed abdominal wall, groin hernia, active inflammatory bowel disease including suspected Crohn’s disease; (3) early postoperative small bowel obstruction within 4 weeks after surgery; (4) successful treatment without any decompression therapy; (5) successful treatment using a nasogastric tube; and (6) successful treatment using an ileus tube without gastrografin administration within 24 h after admission. If patients had two or more admissions during the study period, the first admission was considered. Patients were classified into the following two groups: (1) those treated with gastrografin administration within 48 h after admission (early gastrografin administration group; EGA group) and (2) those treated without gastrografin administration within 48 h after admission (non-early gastrografin administration group; NEGA group). Some patients were treated with gastrografin administration after 48 h following admission and others were treated without gastrografin administration during the entire hospital stay.
Ileus tube insertion and gastrografin administration
All enrolled patients underwent gastrointestinal decompression within 24 h after admission. The CLYNY single or double-balloon type tube (Create Medic, Tokyo, Japan) was used as an ileus tube for decompression of the gastrointestinal tract. The length and outer diameter were 300 cm and 5.3 mm (16 Fr), respectively. The ileus tube was inserted using endoscopy and/or continuous radiography with a guidewire (500 cm in length and 1.32 mm in diameter). In the endoscopic insertion method, the guidewire was inserted into the duodenum through the main channel of a trans-nasal ultrathin endoscope (GIF-XP260N; Olympus, Tokyo, Japan). After removal of the endoscope, the ileus tube was inserted through the guidewire, reaching at least the upper jejunum (Figure 2). A small amount of gastrografin was used to obtain contrast images of the small intestinal tract during ileus tube intubation. Patients who failed to insert ileus tube were switched to the treatment with nasogastric tube. Such patients were excluded from this study. After insertion, gastrointestinal decompression was performed via the ileus tube, without any oral intake. Gastrografin was administered when there was no clinical remission of the small bowel obstruction, which was considered as symptom remission with stool. Administration was preformed through the ileus tube (Figure 3). The amount of gastrografin at one administration was 150 ml. If possible, the ileus tube was advanced to the anal side. It was then clamped for 1–2 h after gastrografin administration. Gastrografin administration was sometimes repeated when improvement was not achieved over 24 h after the previous administration. In the EGA group, the first gastrografin administration was performed within 24–48 h after hospital admission. In the NEGA group, decompression was continued using an ileus tube without gastrografin administration for at least 48 h after admission. If clinical improvement was not achieved for over 48 h, gastrografin administration was considered, as in the EGA group.
Outcome measures
Clinical variables, including age, sex, height, weight, body mass index (BMI), type and number of prior surgeries, and laboratory test indexes on admission, were compared between the EGA and NEGA groups. The aim of decompression therapy or gastrografin administration was clinical improvement and surgery avoidance because of ASBO. Therefore, the study outcomes were the rate of surgery avoidance, the period until the first stool, period of ileus tube intubation, and total period of hospital admission. Clinical improvement was defined as remission of abdominal symptom with stool production. After confirming clinical improvement, the ileus tube was clamped and then drinking or liquid food was started. Thereafter, the ileus tube was removed after confirming the absence of symptom recurrence over 24 h. Discharge criterion was the achievement of remaining clinical improvement with taking soft or normal food over 48 h after tube removal. In contrast, patients eventually underwent surgery if clinical improvement was not achieved with decompression therapy and gastrografin administration during admission. Patients who showed no clinical improvement in the initial 48 h after admission were consulted by surgeons. Surgery was finally performed if the clinical symptoms worsened or strangulation was suspected.
Statistical analysis
There were confounding biases between the two study groups, as this was a non-randomized study. Thus, propensity-score matching analysis, which has been used to compensate for confounding factors, was adopted in this study [9-11]. Logistic regression of factors, including background characteristics, and propensity score were calculated. Age (<70 years vs. ≥70 years), sex (male vs. female), number of surgeries (multiple vs. others), type of surgery (gastrointestinal vs. others), height (<160 cm vs. ≥160 cm), weight (<50 kg vs. ≥50 kg), BMI (<21 kg/m2 vs. ≥21 kg/m2), white blood cell count (<10000/μl vs. ≥10000/μl), C-reactive protein level (<0.5 mg/dl vs. ≥0.5 mg/dl), and method of ileus tube insertion (endoscopy with radiography vs. radiography) were included in covariate analysis. Nearest-neighbor matching in a 1:1 ratio from the EGA and NEGA groups was performed in calipers (0.11) with a width equal to 0.25 of the standard deviation of the logit of the propensity score. Results are expressed as median (interquartile range [IQR]) for continuous data distributed abnormally. Fisher’s exact test and the Mann–Whitney test were used to identify differences in categorical data and continuous data distributed abnormally, respectively. All statistical analyses were performed using JMP Pro 13.0 software (SAS Institute, Cary, NC, USA). A P-value <0.05 was considered statistically significant. The sample size could not be calculated owing to the retrospective nature of this study.