A prediction model for hypoxemia during routine sedation for gastrointestinal endoscopy

Development and validation of a predictive model for PACU hypotension in elderly patients undergoing sedated gastrointestinal endoscopy

BACKGROUND

Hypotension, characterized by abnormally low blood pressure, is a frequently observed adverse event in sedated gastrointestinal endoscopy procedures. Although the examination time is typically short, hypotension during and after gastroscopy procedures is frequently overlooked or remains undetected. This study aimed to construct a risk nomogram for post-anesthesia care unit (PACU) hypotension in elderly patients undergoing sedated gastrointestinal endoscopy.

METHODS

This study involved 2919 elderly patients who underwent sedated gastrointestinal endoscopy. A preoperative questionnaire was used to collect data on patient characteristics; intraoperative medication use and adverse events were also recorded. The primary objective of the study was to evaluate the risk of PACU hypotension in these patients. To achieve this, the least absolute shrinkage and selection operator (LASSO) regression analysis method was used to optimize variable selection, involving cyclic coordinate descent with tenfold cross-validation. Subsequently, multivariable logistic regression analysis was applied to build a predictive model using the selected predictors from the LASSO regression. A nomogram was visually developed based on these variables. To validate the model, a calibration plot, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA) were used. Additionally, external validation was conducted to further assess the model's performance.

RESULTS

The LASSO regression analysis identified predictors associated with an increased risk of adverse events during surgery: age, duration of preoperative water abstinence, intraoperative mean arterial pressure (MAP) <65 mmHg, decreased systolic blood pressure (SBP), and use of norepinephrine (NE). The constructed model based on these predictors demonstrated moderate predictive ability, with an area under the ROC curve of 0.710 in the training set and 0.778 in the validation set. The DCA indicated that the nomogram had clinical applicability when the risk threshold ranged between 20 and 82%, which was subsequently confirmed in the external validation with a range of 18-92%.

CONCLUSION

Incorporating factors such as age, duration of preoperative water abstinence, intraoperative MAP <65 mmHg, decreased SBP, and use of NE in the risk nomogram increased its usefulness for predicting PACU hypotension risk in elderly patient undergoing sedated gastrointestinal endoscopy.

Remimazolam versus propofol for sedation in gastrointestinal endoscopic procedures: a systematic review and meta-analysis

BACKGROUND

Propofol has a favourable efficacy profile in gastrointestinal endoscopic procedures, however adverse events remain frequent. Emerging evidence supports remimazolam use in gastrointestinal endoscopy. This systematic review and meta-analysis compares remimazolam and propofol, both combined with a short-acting opioid, for sedation of adults in gastrointestinal endoscopy.

METHODS

We searched MEDLINE, Embase, and Cochrane databases for randomised controlled trials comparing efficacy-, safety-, and satisfaction-related outcomes between remimazolam and propofol, both combined with short-acting opioids, for sedation of adults undergoing gastrointestinal endoscopy. We performed sensitivity analyses, subgroup assessments by type of short-acting opioid used and age range, and meta-regression analysis using mean patient age as a covariate. We used R statistical software for statistical analyses.

RESULTS

We included 15 trials (4516 subjects). Remimazolam was associated with a significantly lower sedation success rate (risk ratio [RR] 0.991; 95% confidence interval [CI] 0.984-0.998; high-quality evidence) and a slightly longer induction time (mean difference [MD] 9 s; 95% CI 4-13; moderate-quality evidence), whereas there was no significant difference between the sedatives in other time-related outcomes. Remimazolam was associated with significantly lower rates of respiratory depression (RR 0.41; 95% CI 0.30-0.56; high-quality evidence), hypotension (RR 0.43; 95% CI 0.35-0.51; moderate-quality evidence), hypotension requiring treatment (RR 0.25; 95% CI 0.12-0.52; high-quality evidence), and bradycardia (RR 0.42; 95% CI 0.30-0.58; high-quality evidence). There was no difference in patient (MD 0.41; 95% CI -0.07 to 0.89; moderate-quality evidence) and endoscopist satisfaction (MD -0.31; 95% CI -0.65 to 0.04; high-quality evidence) between both drugs.

CONCLUSIONS

Remimazolam has clinically similar efficacy and greater safety when compared with propofol for sedation in gastrointestinal endoscopies.

Comparison of efficacy and safety of equivalent doses of remimazolam versus propofol for gastroscopy anesthesia in elderly patients

Remimazolam, a novel intravenous anesthetic, has been proven to be safe and efficacious in the gastroscopy setting among the elderly. However, reports comparing the effectiveness and safety of using equivalent doses of remimazolam with propofol have not been seen. The aim of this study was to compare the sedation efficacy and safety of the 95% effective doses (ED95) of remimazolam versus propofol combined with sufentanil in the gastroscopy setting among the elderly. In the first step of this two-step study, a modified up-and-down method was used to calculate the ED95 of remimazolam and propofol when combined with 0.1 µg/kg sufentanil in inhibiting body movement of elderly patients undergoing gastroscopy. In the second step, ED95 of both agents calculated in the first step were administered, endpoints of efficacy, safety, and incidence of adverse events were compared. A total of 46 individuals completed the first step. The ED95 of remimazolam was 0.163 mg/kg (95% CI 0.160-0.170 mg/kg), and that of propofol was 1.042 mg/kg (95% CI 1.007-1.112 mg/kg). In the second step, 240 patients completed the trial. The anesthetic effective rates of the remimazolam group and the propofol group were 78% and 83%, respectively, with no statistical difference (P = 0.312). Patients in the remimazolam group had more stable circulatory functions (P < 0.0001) and a lower incidence of pain on injection (3.3% vs. 19.5%, P < 0.0001). The incidence of hypotension was low in the remimazolam versus propofol group (15.6% vs. 39.0%, P < 0.0001). Overall adverse event was low in the remimazolam versus propofol group (21.3% vs. 62.7%, P < 0.0001).In this study, we found that when anesthesia was administered to elderly gastroscopy patients based on 95% effective doses of remimazolam and propofol, remimazolam was as effective as propofol, but was safer with a lower incidence of adverse events.Study registration: Chinese Clinical Trial Registry, ChiCTR2000034234. Registered 29/06/2020, https://www.chictr.org.cn .

Effect of intravenous lidocaine on the ED50 of propofol induction dose in elderly patients undergoing painless gastroscopy: a prospective, randomized, double-blinded, controlled study

BACKGROUND

Intravenous lidocaine could be a potential alternative adjuvant to propofol-based sedation for gastroscopy in elderly patients. This study aimed to evaluate the effect of intravenous lidocaine on the median effective dose (ED50) of propofol induction dose in elderly patients undergoing painless gastroscopy.

METHODS

The study included 70 patients aged ≥ 60 years undergoing painless gastroscopy with 64 randomly assigned to either group L (2% lidocaine 1.5 mg/kg, n = 31) or group N (equal volume normal saline, n = 33). All patients received propofol induction following 0.1 μg/kg intravenous sufentanil. The Dixon "up-and-down" sequential method was used, with a 1.5 mg/kg initial induction dose of propofol followed by a 0.1 mg/kg sequential variable dose. The primary endpoint was the ED50 of the propofol induction dose. The total propofol dose, recovery time, adverse events, and local anesthetic intoxication reactions were also recorded.

RESULTS

The ED50 of propofol induction dose was 0.670 (95% confidence interval [CI] 0.216-0.827) mg/kg in group L and 1.118 (95% CI 0.803-1.232) mg/kg in group N. There was a statistically significant difference between the two groups (p < 0.001). The incidence of hypotension and propofol injection pain were lower in group L than in group N (p < 0.05). Furthermore, the orientation recovery time in group L was shorter compared to group N (p < 0.05). None of the participants in group L observed local anesthetic intoxication reactions after receiving lidocaine.

CONCLUSIONS

The administration of intravenous lidocaine to elderly patients undergoing painless gastroscopy resulted in a significant 40% reduction in the ED50 of propofol induction dose, which may be related to the decreased incidence of hypotension and injection pain, as well as the improved post-gastroscopy orientation recovery.

TRIAL REGISTRATION

ChiCTR, ChiCTR2200065530. Registered on 08 November 2022.

Predicting Hypoxia Using Machine Learning: Systematic Review

Background

Hypoxia is an important risk factor and indicator for the declining health of inpatients. Predicting future hypoxic events using machine learning is a prospective area of study to facilitate time-critical interventions to counter patient health deterioration.

Objective

This systematic review aims to summarize and compare previous efforts to predict hypoxic events in the hospital setting using machine learning with respect to their methodology, predictive performance, and assessed population.

Methods

A systematic literature search was performed using Web of Science, Ovid with Embase and MEDLINE, and Google Scholar. Studies that investigated hypoxia or hypoxemia of hospitalized patients using machine learning models were considered. Risk of bias was assessed using the Prediction Model Risk of Bias Assessment Tool.

Results

After screening, a total of 12 papers were eligible for analysis, from which 32 models were extracted. The included studies showed a variety of population, methodology, and outcome definition. Comparability was further limited due to unclear or high risk of bias for most studies (10/12, 83%). The overall predictive performance ranged from moderate to high. Based on classification metrics, deep learning models performed similar to or outperformed conventional machine learning models within the same studies. Models using only prior peripheral oxygen saturation as a clinical variable showed better performance than models based on multiple variables, with most of these studies (2/3, 67%) using a long short-term memory algorithm.

Conclusions

Machine learning models provide the potential to accurately predict the occurrence of hypoxic events based on retrospective data. The heterogeneity of the studies and limited generalizability of their results highlight the need for further validation studies to assess their predictive performance.

Machine learning-based risk prediction of hypoxemia for outpatients undergoing sedation colonoscopy: a practical clinical tool

OBJECTIVES

Hypoxemia as a common complication in colonoscopy under sedation and may result in serious consequences. Unfortunately, a hypoxemia prediction model for outpatient colonoscopy has not been developed. Consequently, the objective of our study was to develop a practical and accurate model to predict the risk of hypoxemia in outpatient colonoscopy under sedation.

METHODS

In this study, we included patients who received colonoscopy with anesthesia in Nanjing First Hospital from July to September 2021. Risk factors were selected through the least absolute shrinkage and selection operator (LASSO). Prediction models based on logistic regression (LR), random forest classifier (RFC), extreme gradient boosting (XGBoost), support vector machine (SVM), and stacking classifier (SCLF) model were implemented and assessed by standard metrics such as the area under the receiver operating characteristic curve (AUROC), sensitivity and specificity. Then choose the best model to develop an online tool for clinical use.

RESULTS

We ultimately included 839 patients. After LASSO, body mass index (BMI) (coefficient = 0.36), obstructive sleep apnea-hypopnea syndrome (OSAHS) (coefficient = 1.32), basal oxygen saturation (coefficient = -0.14), and remifentanil dosage (coefficient = 0.04) were independent risk factors for hypoxemia. The XGBoost model with an AUROC of 0.913 showed the best performance among the five models.

CONCLUSION

Our study selected the XGBoost as the first model especially for colonoscopy, with over 95% accuracy and excellent specificity. The XGBoost includes four variables that can be quickly obtained. Moreover, an online prediction practical tool has been provided, which helps screen high-risk outpatients with hypoxemia swiftly and conveniently.

To establish a risk prediction model for the occurrence of hypoxemia during painless bronchoscopy

The present study was focused on evaluating the clinical predictors of hypoxemia and establishing a multivariable, predictive model for hypoxemia in painless bronchoscopy. A total of 244 patients were enrolled in the study, and data were collected using a self-designed data collection. The retrospective data collected in this study included the relevant data of patients undergoing the painless bronchoscopy, and we used univariate analysis to deal with these influencing factors. Multivariate logistic regression analysis was used to establish the prediction equation, and receiver operating characteristic curve analysis was carried out. Receiver operating characteristic curves and the Hosmer-Lemeshow test were used to evaluate the model performance. P < .05 was considered to indicate statistical significance. Multivariate logistic regression indicated that body mass index (BMI) (odds ratio [OR]: 1.169; 95% confidence interval [CI]: 1.070-1.277), arterial partial pressure of oxygen (PaO2) (OR: 4.279; 95% CI: 2.378-7.699), alcohol consumption (OR: 2.021; 95% CI: 1.063-3.840), and whether the bronchoscope operation time exceeds 30 minutes (OR: 2.486; 95% CI: 1.174-5.267) were closely related to the occurrence of hypoxemia. The prediction model developed by the logistic regression equation was -4.911 + 1.454 (PaO2) + 0.156 (BMI) + 0.703 (Alcohol consumption) + 0.911 (time > 30th minutes). The prediction model showed that the area under the receiver operating characteristic curve was 0.687. The predictive model was well calibrated with a Hosmer-Lemeshow x2 statistic of 4.869 (P = .772), indicating that our prediction model fit well. The accuracy (number of correct predictions divided by the number of total predictions) was 75%. The prediction model, consisting of BMI, PaO2, alcohol consumption, and whether the bronchoscope operation time exceeds 30 minutes. It is an effective predictor of hypoxemia during sedation for painless bronchoscopy.

An interpretable artificial neural network model for predicting hypoxemia via an online tool in adult (18-64) patients during esophagogastroduodenoscopy

Background

The hypoxemia risk in adult (18-64) patients treated with esophagogastroduodenoscopy (EGD) under sedation often poses a dilemma for anesthesiologists. We aimed to establish an artificial neural network (ANN) model to solve this problem, and introduce the Shapley additive explanations (SHAP) algorithm to further improve the interpretability.

Methods

The relevant data of patients underwent routine anesthesia-assisted EGD were collected. Elastic network was used to filter the optimal features. Airway-ANN and Basic-ANN models were established based on all collected indicators and remaining variables excluding airway assessment indicators, respectively. The performance of Basic-ANN, Airway-ANN and STOP-BANG was evaluated by the area under the precision-recall curve (AUPRC) on temporal validation set. The SHAP was used for revealing the predictive behavior of our best model.

Results

999 patients were eventually included. The AUPRC value of Airway-ANN model was significantly higher than Basic-ANN model in the temporal validation set (0.532 vs 0.429, P < 0.05). And the performance of both two ANN models was significantly better than that of STOP-BANG score (both P < 0.05). The Airway-ANN model was deployed to the cloud (http://njfh-yxb.com.cn:2022/airway_ann).

Conclusion

Our online interpretable Airway-ANN model achieved satisfying ability in identifying the hypoxemia risk in adult (18-64) patients undergoing EGD.

Res-SE-ConvNet: A Deep Neural Network for Hypoxemia Severity Prediction for Hospital In-Patients Using Photoplethysmograph Signal

Determining the severity level of hypoxemia, the scarcity of saturated oxygen (SpO2) in the human body, is very important for the patients, a matter which has become even more significant during the outbreak of Covid-19 variants. Although the widespread usage of Pulse Oximeter has helped the doctors aware of the current level of SpO2 and thereby determine the hypoxemia severity of a particular patient, the high sensitivity of the device can lead to the desensitization of the care-givers, resulting in slower response to actual hypoxemia event. There has been research conducted for the detection of severity level using various parameters and bio-signals and feeding them in a machine learning algorithm. However, in this paper, we have proposed a new residual-squeeze-excitation-attention based convolutional network (Res-SE-ConvNet) using only Photoplethysmography (PPG) signal for the comfortability of the patient. Unlike the other methods, the proposed method has outperformed the standard state-of-art methods as the result shows 96.5% accuracy in determining 3 class severity problems with 0.79 Cohen Kappa score. This method has the potential to aid the patients in receiving the benefit of an automatic and faster clinical decision support system, thus handling the severity of hypoxemia.

Identification of Patient-Related and Procedure-Related Factors Contributing to Hypoxemia in Adult Outpatients Undergoing Esophagogastroduodenoscopy (EGD)

PURPOSE

To assess which patient factors and procedure-related factors contribute to hypoxemia during esophagogastroduodenoscopy (EGD) and determine whether prophylactic oropharyngeal suctioning reduces the rate of hypoxemia when compared to oropharyngeal suctioning when clinically indicated by patient's coughing or secretions.

DESIGN

This was a single-site study taking place at a private practice, outpatient facility with no anesthesia trainees present. Patients were randomized to one of two groups based on birth month. After the administration of sedating medications but before the insertion of the endoscope, Group A was oropharyngeal suctioned by either the anesthesia provider or the proceduralist. Group B was oropharyngeal suctioned only when clinically indicated by coughing or visible copious secretions.

METHODS

Data were collected on a variety of patient and procedure-related factors. Associations between these factors and hypoxemia during esophagogastroduodenoscopy were analyzed using the statistical analysis system application JMP. After analysis and literature review, a protocol for prevention and treatment of hypoxemia during EGD was proposed.

FINDINGS

This study found that chronic obstructive pulmonary disease increases the risk for hypoxemia during esophagogastroduodenoscopy. There were no other statistically significant associations between other factors and hypoxemia.

CONCLUSIONS

This study highlights factors that should be evaluated in the future when considering the risk of hypoxemia during EGD. Although not statistically significant, this study's results indicated that prophylactic oropharyngeal suctioning may reduce rates of hypoxemia, as only 1 of 4 cases of hypoxemia occurred in Group A. Additionally, future studies on hypoxemia during monitored anesthesia care for EGD should include an evaluation of the impact of American Society of Anesthesiologists class, history of chronic obstructive pulmonary disease or asthma, body mass index, obstructive sleep apnea, and opioid administration on hypoxemia risk.

Comparison of three sedation models for same-day painless bidirectional endoscopy: A multicenter randomized controlled trial

BACKGROUND AND AIM

We investigated the most beneficial propofol sedation model for same-day painless bidirectional endoscopy (BDE).

METHODS

Asymptomatic participants scheduled for same-day painless BDE examination from October 2020 to September 2021 were randomized to three groups: sedated esophagogastroduodenoscopy followed by unsedated colonoscopy (Group A); sedated esophagogastroduodenoscopy followed by sedated colonoscopy (Group B); and sedated esophagogastroduodenoscopy followed by sedated insertion colonoscopy (Group C). Patient discomfort, colonoscopy performance, doses of propofol, cardiovascular stress, anesthesia resuscitation, and sedation-related adverse events were evaluated.

RESULTS

A total of 3200 participants were analyzed. Baseline demographics, patient discomfort, cecal intubation rate, adenoma detection rate and sedation-related adverse events were similar in the three groups. Propofol dose was the lowest in Group A (137.65 ± 36.865 mg) compared with Group B (177.71 ± 40.112 mg, P < 0.05) and Group C (161.63 ± 31.789 mg, P < 0.05). Decline in vital signs was most obvious in Group B during the procedure (P < 0.05). Recovery time was the shortest in Group A (5.01 ± 1.404 min) compared with Group B (9.51 ± 2.870 min, P < 0.05) and Group C (5.83 ± 2.594 min, P < 0.05); discharge time was the shortest in Group A (3.53 ± 1.685 min) compared with Group B (11.29 ± 5.172 min, P < 0.05) and Group C (6.47 ± 2.338 min, P < 0.05). Adenomas per positive patient of Group A (2.29 ± 1.055) and Group C (2.28 ± 0.931) were more than that in Group B (2.11 ± 0.946, P < 0.05).

CONCLUSIONS

Sedated esophagogastroduodenoscopy followed by unsedated colonoscopy is the superior model for same-day painless BDE with the benefits of satisfactory patient comfort, reduced sedation dose, less cardiovascular stress, faster recovery, shorter discharge time and high colonoscopy quality.

Incidence and Risk Factors for Hypoxia in Deep Sedation of Propofol for Artificial Abortion Patients

Background

Respiratory depression is a life-threatening adverse effect of deep sedation. This study aimed to investigate the factors related to hypoxia caused by propofol during intravenous anesthesia.

Methods

Three hundred and eight patients who underwent painless artificial abortion in the outpatient department of Shanghai Tenth People’s Hospital between November 1, 2019 and June 30, 2020 were divided into two groups according to whether the patients experienced hypoxia (SpO₂ < 95%). Preoperative anxiety assessments, anesthesia process, and operation-related information of the two groups were analyzed. The univariate analysis results were further incorporated into logistic regression analysis for multivariate analysis to determine the independent risk factors affecting hypoxia.

Results

Univariate analysis revealed that body mass index (BMI) (21.80 ± 2.94 vs. 21.01 ± 2.39; P = 0.038, 95% confidence interval (CI) = [−1.54, −0.04]), propofol dose (15.83 ± 3.21 vs. 14.39 ± 3.01; P = 0.002, CI = [−2.34, −0.53]), menopausal days (49.64 ± 6.03 vs. 52.14 ± 5.73; P = 0.004, CI = [0.79, 4.21]), State Anxiety Inventory score (51.19 ± 7.55 vs. 44.49 ± 8.96; P < 0.001, CI = [−9.26, −4.15]), and Self-rating Anxiety Scale score (45.86 ± 9.48 vs. 42.45 ± 9.88; P = 0.021, CI = [−6.30, −0.53]) were statistically significant risk factors for hypoxia during the operation. Logistic regression analysis showed that propofol dosage, menopausal days, and State Anxiety Inventory score were independent risk factors for hypoxia.

Conclusion

Patient anxiety affects the incidence of hypoxia when undergoing deep intravenous anesthesia with propofol. We can further speculate that alleviating patient anxiety can reduce the incidence of hypoxia.

Clinical Trial Registration

[http://www.chictr.org.cn], identifier [ChiCTR2000032167].

Effect of intravenous lidocaine on propofol consumption in elderly patients undergoing colonoscopy: a double-blinded, randomized, controlled trial

BACKGROUND

Elderly patients undergoing colonoscopy with propofol as sedation are prone to respiratory or cardiovascular complications. Intravenous lidocaine has analgesic efficacy and reduces propofol consumption during surgery. Here, the effect of intravenous lidocaine on propofol consumption was evaluated in elderly patients undergoing colonoscopy.

METHODS

Patients were randomly allocated to receive intravenous lidocaine (1.5 mg/kg bolus dose, followed by a 2 mg/kg/h continuous infusion during the procedure; Group L) or a placebo (saline; Group N). During the procedure, sedation was achieved by propofol. The following outcomes were recorded: total propofol consumption; time to loss of consciousness; number of airway modifications; time to the first airway intervention; incidence of sedation-related events; pain score after awakening; endoscopists' and patients' satisfaction scores; memory level of the procedure; and adverse events within 24 h postoperatively.

RESULTS

Compared with Group N, propofol consumption was reduced by 13.2% in Group L (100.30 ± 25.29 mg vs. 115.58 ± 27.52 mg, respectively, p = 0.008). Kaplan-Meier curves showed that the median time to the loss of consciousness episode was shorter in Group L than in Group N (40 s vs. 55 s, respectively, log rank p < 0.0001). The number of airway modifications, time to the first airway intervention, incidence of sedation-related events, time to awakening, pain score after awakening, endoscopists' and patients' satisfaction scores, memory level of the procedure and adverse events within 24 h postoperatively did not differ between the two groups (p > 0.05).

CONCLUSIONS

Intravenous lidocaine can reduce propofol consumption in elderly patients undergoing colonoscopy, with quicker time to loss of consciousness.

TRIAL REGISTRATION

The clinical trial was registered at (12/01/2021, ChiCTR2100042001 ).

The correlation between the STOP-Bang score and oxygen saturation during spinal anesthesia with dexmedetomidine sedation

BACKGROUND

The STOP-Bang questionnaire is a simple screening tool with high sensitivity for the detection of severe obstructive sleep apnea. Predicting airway obstruction would allow the safe management of sedative patients to prevent intraoperative hypoxia. This study was designed to check the correlation between the STOP-Bang score and oxygen saturation (SpO2) during sedation and confirm the availability of the STOP-Bang questionnaire as a preoperative exam for predicting the incidence of hypoxia in sedative patient management.

METHODS

This study included 56 patients who received spinal anesthesia. The pre-anesthesia evaluation was conducted using the STOP-Bang questionnaire. The patients were under spinal anesthesia with an average block level of T10. Dexmedetomidine was infused with a loading dose of 1 μg/kg over 10 min and a maintenance dose of 0.5 μg/kg/h until the end of the procedure. The SpO2 of the patients was recorded every 5 min.

RESULTS

The STOP-Bang score was negatively correlated with the lowest SpO2 (coefficient = -0.774, 95% confidence interval [CI]: -0.855 to -0.649, standard error [SE] = 0.054, P < 0.001). The item of "observed apnea" was the most correlated one with hypoxic events (odds ratio = 6.00, 95% CI: 1.086 to 33.145).

CONCLUSIONS

The STOP-BANG score was significantly correlated with the lowest SpO2 during spinal anesthesia, which enabled the prediction of meaningful hypoxia before it occurred in the sedated patients.

High-Flow Nasal Cannula Oxygen in Patients Having Anesthesia for Advanced Esophagogastroduodenoscopy: HIFLOW-ENDO, a Randomized Clinical Trial

BACKGROUND

Over 6 million esophagogastroduodenoscopy (EGD) procedures are performed in the United States each year. Patients having anesthesia for advanced EGD procedures, such as interventional procedures, are at high risk for hypoxemia.

METHODS

Our primary study aim was to evaluate whether high-flow nasal cannula (HFNC) oxygen reduces the incidence of hypoxemia during anesthesia for advanced EGD. Secondarily, we studied whether HFNC oxygen reduces hypercarbia or hypotension. After obtaining written informed consent, adults having anesthesia for advanced EGD, expected to last longer than 15 minutes, were randomly assigned to receive HFNC oxygen or standard nasal cannula (SNC) oxygen. The primary outcome was occurrence of one or more hypoxemia events during anesthesia, defined by arterial oxygen saturation <92% for at least 15 consecutive seconds. Secondary outcomes were occurrence of one or more hypercarbia or hypotension events. A hypercarbia event was defined by a transcutaneous CO2 measurement 20 mm Hg or more above baseline, and a hypotension event was defined by a mean arterial blood pressure measurement 25% or more below baseline.

RESULTS

Two hundred seventy-one adult patients were enrolled and randomized, and 262 patients completed study procedures. Eight randomized patients did not complete study procedures due to changes in their anesthesia or endoscopy plan. One patient was excluded from analysis because their procedure was aborted after 1 minute. Patients who received HFNC oxygen (N = 132) had a significantly lower incidence of hypoxemia than those who received SNC oxygen (N = 130; 21.2% vs 33.1%; hazard ratio [HR] = 0.59 [95% confidence interval {CI}, 0.36-0.95]; P = .03). There was no difference in the incidence of hypercarbia or hypotension between the groups. The HR for hypercarbia with HFNC oxygen was 1.29 (95% CI, 0.89-1.88; P = .17), and the HR for hypotension was 1.25 (95% CI, 0.86-1.82; P = .25).

CONCLUSIONS

HFNC oxygen reduces the incidence of hypoxemia during anesthesia for advanced EGD and may offer an opportunity to enhance patient safety during these procedures.

An artificial neural network model for prediction of hypoxemia during sedation for gastrointestinal endoscopy

OBJECTIVE

This study was designed to assess clinical predictors of hypoxemia and develop an artificial neural network (ANN) model for prediction of hypoxemia during sedation for gastrointestinal endoscopy examination.

METHODS

A total of 220 patients were enrolled in this prospective observational study. Data on demographics, chronic concomitant disease information, neck circumference, thyromental distance and anaesthetic dose were collected and statistically analysed.

RESULTS

Univariate analysis indicated that body mass index (BMI), habitual snoring and neck circumference were associated with hypoxemia. An ANN model was developed with three variables (BMI, habitual snoring and neck circumference). The area under the receiver operating characteristic curve for the ANN model was 0.80.

CONCLUSIONS

The ANN model developed here, comprising BMI, habitual snoring and neck circumference, was useful for prediction of hypoxemia during sedation for gastrointestinal endoscopy.