We retrospectively reviewed the records of health examinations at Konkuk University Medical Center (Seoul, Korea), from September 2013 to April 2015. Among these, records of individuals over 18 years of age who underwent a gastroscopy were included in the study. To reduce the effect of variations between endoscopists, only gastroscopies carried out by a single experienced endoscopist (S.P.L.) were included. The study was approved by the Institutional Review Board (IRB) of Konkuk University School of Medicine, which confirmed that the study was in accordance with the ethical guidelines of the Helsinki Declaration (KUH1010664), and it was registered in the Clinical Research Information Service (CRIS) ID: KCT0001512.

All the examinees filled out a self-reporting questionnaire, which included the following items: underlying disease, medication history, smoking, alcohol intake, dyspepsia symptoms, and levels of emotional stress and depression. Subjects’ age, gender, height, weight, and endoscopic and histologic findings were also investigated. Heavy drinking was defined as consuming 15 drinks or more per week for men and eight drinks or more per week for women.

Gastroscopy was performed using a standard upper endoscope (GF-260, Olympus, Tokyo, Japan; or EPK-I, Pentax, Tokyo, Japan). Lidocaine spray was routinely used. All examinees provided written consent to undergo the endoscopic procedures. Midazolam for procedural sedation was selectively administered just before the start of the procedure in accordance with the examinee’s preference. The usual starting dose for producing moderate sedation was 1 to 3 mg based on age, gender, and medical conditions. The dose was increased to achieve an adequate level of sedation [6]. According to the American Society of Anesthesiologists guidelines, moderate sedation is defined as a state in which patients respond purposefully to verbal command and their cardiovascular function is maintained [9]. If a severe paradoxical reaction to midazolam occurred, it was immediately treated with flumazenil. If symptoms of midazolam overdose or midazolam side effects were noted, we tried to stabilize the vital signs of the patient first, and we considered the use of flumazenil. Symptoms of overdose or side effects included hypoxia, hypoventilation, hypotension, apnea, and central nervous system (CNS) depression. Oxygen supplementation was provided via a nasal cannula if the oxygen saturation level remained < 90%, despite verbal and physical stimuli. Hypoxia was recorded only when supplemental oxygen was needed. The incidence of acute complications (abdominal pain, nausea, and vomiting) and late-onset complications (persisting throat pain, bowel perforation, and gastrointestinal [GI] bleeding) were investigated through chart review.

The examinee’s cooperation and the completeness of endoscopic examination were evaluated by the operator’s subjective judgment and recorded directly after the examination. Information on the level of cooperation was collected through a review of the endoscopic reports. “Poor cooperation” meant lack of cooperation of the examinees during the procedure regardless of one’s will. The immediate causes of poor cooperation included severe belching, retching, coughing, and non-compliance of the examinees. “Belching” was defined as the forceful expulsion of gas from stomach through the mouth and “retching” was defined as an involuntary spasm of ineffectual vomiting. “Coughing” was defined as a sudden noisy expulsion of air from the bronchus or lungs, which is triggered by aspiration. “Non-compliance” referred to an action in contrast to bearing and managing impulses to enable the operator to perform the procedure. The study was considered incomplete when some parts of the stomach could not be sufficiently visualized during the procedure. The causes of incomplete study included poor cooperation and a large amount of food materials in the stomach.

Endoscopic and histologic records were reviewed to identify the presence of digestive diseases, such as peptic ulcer disease, reflux esophagitis, GI tract masses, nodular gastritis, chronic atrophic gastritis, and metaplastic gastritis. The endoscopic findings were described based on the criteria of the Sydney classification [10]. A diagnosis of erosive esophagitis was made in cases with hyperemic streaks or mucosal breaks on the lower esophagus. Minimal change lesions of the esophagus were excluded from this diagnosis. Peptic ulcer disease was diagnosed as a deep mucosal defect with evidence of submucosal invasion, excluding ulcer scars. Chronic atrophic gastritis was diagnosed when the mucosal and submucosal vessels were visible without full air distension [11]. Nodular gastritis was defined as chicken-skin like mucosal changes or unusual goose-flesh appearance, and metaplastic gastritis was defined as uneven, irregular whitish patches or plaques [12]. Diagnoses of adenoma and carcinoma were made based on the results of tissue biopsy.

Depression and emotional stress were assessed using the Korean version of the Beck’s Depression Inventory (K-BDI) and the Brief Encounter Psychosocial Instrument-Korean version (BEPSI-K) [13,14]. The BEPSI-K consists of five self-administered items. A total score of 0 to 1.8 indicates a low level of stress; 1.8 to 2.8, a moderate level of stress; and more than 2.8, a high level of stress. The cut off value of 2.4 as a significant stressful status was set based on previous studies [15,16]. The K-BDI is a 21-item self-report questionnaire. A total score of 0 to 9 signifies no depression; 10 to 15, mild depression; 16 to 23, moderate depression; and 24 to 63, severe depression. We considered a total score of ≥ 10 as positive for depression.

Based on the Rome III definition, functional dyspepsia was diagnosed when the subjects had symptoms of dyspepsia without any evidence of reflux esophagitis, peptic ulcer disease, postoperative stomach, upper GI cancers, or any structural disease on upper GI endoscopy. Dyspepsia was defined as one or more of the following symptoms occurring at least once per week in the last 3 months: (1) bothersome postprandial fullness after ordinary-sized meals, (2) early satiety that prevents finishing a regular meal, and (3) epigastric pain or burning. Symptoms with an onset occurring more than 6 months prior to the survey were not asked about in our questionnaire.

Continuous variables and categorical variables were presented as mean ± standard deviation (SD) and number (%), respectively. Subjects were divided into the poor cooperation group and the good cooperation group, according to their level of cooperation during endoscopy. The characteristics of the two groups were compared using the chi-square test, Student t test, and Mann-Whitney U test. Logistic regression analyses were performed to estimate the odds ratio (OR) and 95% confidence interval (CI) for factors that could result in poor cooperation during endoscopy. Statistical analyses were conducted using SPSS software version 19.0 (IBM Co., Armonk, NY, USA), and a p < 0.05 was considered statistically significant in all tests.

A total of 4,422 subjects (56.5% male) were included in the study. The overall mean age and body mass index (BMI) were 45.19 ± 9.87 years (range, 18 to 86) and 23.57 ± 3.27 kg/m² (range, 14.6 to 39.7), respectively (Table 1). The number of subjects who received sedation (midazolam) during endoscopy was 3,164 (71.6%). Cooperation during the endoscopic procedure was poor in 358 subjects (8.1%). The other characteristics of the study population are shown in Table 1.

In the poor cooperation group, the cause of poor cooperation was not recorded in 214 cases. Of the patients with causes recorded, the causes of poor cooperation were belching in 114 subjects (79.2%), retching in 94 (65.3%), non-compliance in 18 (12.5%), and coughing in two (1.4%) (Fig. 1). Often, poor cooperation was caused by multiple reasons (79 subjects, 54.9%).

Poor cooperation was more common among younger subjects and in women (p < 0.001 and p = 0.003, respectively) (Table 2). When broken down by age group, cooperation was poor in 9.6% of those in their 20s or younger (18 of 187), 10.5% of those in their 30s (123 of 1,174), 7.9% of those in their 40s (134 of 1,702), 6.2% of those in their 50s (60 of 968), 5.9% of those in their 60s (17 of 287), and 5.8% of those in their 70s or older (six of 104) (Fig. 2). Subjects in the poor cooperation group had a higher proportion of non-smokers and a lower proportion of past smokers (p = 0.049). Poor cooperation was also related to a high BMI, hiatal hernia, and procedural sedation (p = 0.001, p = 0.017, and p < 0.001, respectively). Emotional stress, depression, functional dyspepsia, alcohol intake, and medication history were not related with poor cooperation.

Among the subjects who received procedural sedation, 10 (3.2%) in the poor cooperation group and one in the good cooperation group were administered the antidote (flumazenil). In the good cooperation group, the reason for using the antidote was hypoxia that could not be alleviated with nasal cannula oxygenation. Three subjects in the poor cooperation group and 12 subjects in the good cooperation group developed complications after the procedure, with no significant difference between the two groups (p = 0.116). The endoscopic examination of 44 subjects was an “incomplete study.” The causes of incomplete study were poor cooperation in 36 cases and large amounts of food materials in the stomach in eight cases.

Logistic regression analysis revealed that that young age (< 40 years) and female sex were independent predisposing factors for an examinee’s poor cooperation during upper endoscopy (OR, 1.423; 95% CI, 1.125 to 1.801; p = 0.003) (OR, 1.670; 95% CI, 1.216 to 2.294; p = 0.002) (Table 3). In addition, high BMI (≥ 25), hiatal hernia, and procedural sedation were also independent predisposing factors for poor cooperation (OR, 1.501; 95% CI, 1.165 to 1.935; p = 0.002) (OR, 2.235; 95% CI, 1.300 to 3.844; p = 0.004) (OR, 2.918; 95% CI, 2.083 to 4.088; p < 0.001). After logistic analysis, smoking history was not related with examinees’ poor cooperation.