Sedation for gastrointestinal endoscopies provides improved patient tolerance to the procedures by relieving anxiety, discomfort, and pain.1 Sedation may also provide improved conditions to the endoscopist and improve the quality of the examination.2 Hypoxia is the most common sedation-related adverse event during endoscopy, and its causes include respiratory depression, airway obstruction, and decreased compliance of the chest wall. The incidence of hypoxemia during gastrointestinal endoscopy has been reported to be in the range of 1.5% to 70%, depending largely on differences in definitions and the types of practitioners.3-9

The incidence of respiratory complications in the immediate post-anesthesia period in patients undergoing surgery is high and well documented.10,11 The American Society of Anesthesiologists (ASA) recommends that patients in the immediate period after moderate sedation be observed and monitored until they are near their baseline level of consciousness and are no longer at increased risk for cardiorespiratory depression.12 Similar but less detailed guidelines were defined by the American Society of Gastroenterologists.13 However, there are only limited data on the incidence of respiratory complications during the post-procedure period in patients undergoing gastrointestinal endoscopy under sedation. One study found that half of hypoxemia cases occurred during the recovery period after the completion of esophagogastroscopy,14 but no other data are available.

The aim of this preliminary study was to assess the incidence of respiratory complications and their risk factors during the recovery period following gastrointestinal endoscopy under sedation performed by gastroenterologists. We assumed that because endoscopies are less invasive than surgical procedures performed under sedation and not under general anesthesia, the incidence of complications will be lower than that in the postoperative period. The study also explored the incidence of respiratory complications leading to emergency room or ambulatory visits in the three-day period following gastrointestinal endoscopy.

In this retrospective study, data were collected from 657 patients undergoing gastroscopy and/or colonoscopy, endoscopic ultrasonography, or endoscopic retrograde cholangiopancreatography (ERCP). The procedure and sedation were performed by a team of gastroenterologists and nurses, respectively. According to institutional protocol, sedation was initiated with intravenous administration of midazolam (1–2 mg), fentanyl (50 µg), and/or propofol (30–50 mg). Propofol was administered in boluses, as needed, during the procedure. Monitoring during the procedure included blood pressure measurement before the beginning of the procedure, and then at five-minute intervals. Heart rate and pulse oximetry oxygen saturation (SpO₂) were continuously monitored and automatically recorded throughout the procedure. Capnography was not used for monitoring. At the end of the procedure, the patients were admitted to the post-procedure recovery area for one-hour observation. During the post-procedure recovery period, patients were continuously monitored for heart rate, SpO₂, and blood pressure at five-minute intervals for the first 30 minutes. Ambulatory patients were discharged after becoming fully awake and able to drink and eat a light meal, and vital signs were similar to baseline values. Patients were instructed to contact the gastroenterology department if they experienced abdominal pain, chest pain, fever, black stool, persistent vomiting, or difficulty breathing. Hospitalized patients were discharged after being fully awake, and their vital signs were similar to baseline values.

Data on patient demographics, procedures, and sedation were collected from the patients' electronic records. Demographic data included age; sex (male/female); body mass index; physical status classification according to the ASA (score 1, 2, or 3); and background diseases (ischemic heart disease [IHD], hypertension [HTN], diabetes mellitus [DM], chronic heart failure, chronic respiratory disease, and obstructive sleep apnea). Data on the procedure included the procedure type and length, and the type and dose of medications administered.

Data on sedation-related respiratory complications were collected from the patients' electronic records. We used the definition of complications suggested by Mason and colleagues, whereby hypoxemia was defined as SpO₂ <90% for less than 60 seconds, SpO₂ <90% for more than 60 seconds, or SpO₂ <75% for any time period.15 Other complications included apnea, airway obstruction, cardiogenic shock, and cardiac arrest, as indicated in the charts by the gastroenterologist and the nurse.

Using patients' electronic medical records, information on emergency department visits or hospital admissions during the three days following the procedure were explored, and their relevance for sedation was defined. Relevant findings were defined as fever >38°C, respiratory complaints (cough and dyspnea), and SpO₂ <94%.

Statistical differences were tested using Pearson’s chi-square test. The correlation between the occurrence of desaturation during the procedure and desaturation in the post-anesthesia care unit was tested using Fisher's exact test. Statistical analysis was performed using R ver. 3.4.4 (R Foundation for Statistical Computing, Vienna, Austria). Ordinal logistic regression was used to calculate both univariate and multivariate p-values and odds ratios (ORs). The multivariate model included age, sex, and any other variables that were identified as significant in the univariate analysis. SAS software ver. 9.4 (SAS Institute, Cary, NC, USA) was used for all calculations.

Sedation is an important tool for facilitating gastroenterological endoscopic procedures. The definition of procedural sedation, as proposed by the International Committee for the Advancement of Procedural Sedation, is the administration of one or more pharmacological agents to facilitate a diagnostic or therapeutic procedure, while targeting a state during which airway patency, spontaneous respiration, protective airway reflexes, and hemodynamic stability are preserved, while alleviating anxiety and pain.16 There are many definitions of procedural adverse events. We chose to use the adverse event sedation reporting tool, which has standardized the reporting of sedation-related adverse events.15

Previous studies reported hemoglobin desaturation incidences during sedation of 6.7%,17 12.8%,18 and 4.7%.19 In our study, the incidence was 12.5%. However, the definition of hemoglobin oxygen desaturation was not uniform in all reports. In a large international study including 160,000 patients that used the same definitions as in our study, the incidence of desaturation was reported to be only 7.8 per 1,000 procedures.20 The higher incidence in our study may be explained by the fact that in our study, the percentage of ASA 3 patients was high (41%), and the data did not include children. Indeed, an ASA score ≥3 was reported to be a predictor of respiratory complications.15 Notably, ASA score and other parameters such as age, body mass index, background disease, procedure type, or duration, which were previously reported to be risk factors for respiratory adverse events, were not found to be predictors for desaturation in our study, possibly stemming from the higher proportion of non-healthy patients.

In our study, we examined the incidence of respiratory adverse events during the post-procedure recovery period. We found that the incidence of minor respiratory events was low and no major adverse events occurred. The incidence of respiratory adverse events in the post-procedure period in the present study was much lower than that in previous reports in the post-surgical and general anesthesia periods. Although not tested directly in our study, these differences may stem from the particular lower-risk features of upper and lower endoscopies compared with most surgical techniques, namely, the procedures are shorter, use only short-acting fentanyl of the opioid family of drugs, did not use muscle relaxants, and the procedures did not involve surgical incision altering lung mechanics.21,22

In the three-day period after gastrointestinal endoscopy, only two patients visited the emergency room, and both had complaints possibly related to the procedure itself, such as melena, vomiting, and abdominal pain. None of the patients had respiratory complaints or symptoms of pneumonia. Hence, the incidence of long-term complications was low.

Our study has some limitations, including the inclusion of different types of procedures (colonoscopy, gastroscopy, endoscopic ultrasonography, and ERCP). The study was performed in a tertiary hospital with a high percentage of patients with ASA 3 scores, and the number of patients was limited to 657. However, the present data indicate that balanced propofol-based sedation is relatively safe, both in the procedural and post-procedural recovery periods, with only infrequent minor events noted in the post-procedural recovery period.

In summary, we found that pulmonary disease, ASA score, length of procedure, and fentanyl use increased the occurrence of desaturation events during the procedure. The ASA score, the comorbidity of IHD, DM, and HTN, and desaturation during the procedure increased the occurrence of desaturation events during recovery from the procedure.

Notwithstanding, meticulous vital sign tracking should be used, especially in patients with identified predictors of complications, such as coexisting disease and desaturation during the procedure. Further studies are needed to determine whether these events have clinical implications, such as a greater risk of pulmonary aspiration and pneumonia.