This study is a prospective observational research project designed to evaluate the incidence of residual gastric content in patients scheduled for elective surgery and to investigate patient-related factors that influence gastric emptying. The study primarily employs ultrasonography (USG) as a tool for assessing the presence of residual gastric content and compares its efficacy with traditional questionnaire-based assessments. Additionally, the study examines the utility and accuracy of various gastric volume estimation formulas, assessing their correlation with observed gastric content and aspiration risk.
Study Design:
Participants were enrolled based on pre-defined inclusion and exclusion criteria, including adult patients scheduled for elective surgery who adhered to standard preoperative fasting guidelines. Following a detailed informed consent process, patients were asked to complete a preoperative questionnaire, which gathered data on demographic characteristics, medical history, and patient-reported symptoms, such as early satiety, history of cholelithiasis, and other relevant comorbidities. Ultrasound examinations were performed by a single trained investigator to ensure consistency in imaging and interpretation.
USG was utilized to assess residual gastric content, including both solid and fluid content. The antrum cross-sectional area was measured in the supine and right lateral decubitus positions, and gastric volume was estimated using four different formulas: Michiko, Bouvet, two fo Perlas\' formulas developed in differnet investigations. Patients were classified into three groups according to the Perlas risk score, which is based on ultrasound findings of gastric content (Grade 0 = low risk, Grade 1 = moderate risk, Grade 2 = high risk of aspiration). The primary outcome was the identification of a full stomach, defined as the presence of solid content or fluid contentbased on ultrasound findings.
Registry Procedures and Quality Control:
Quality assurance procedures were integrated into the study to ensure the reliability of both questionnaire-based data and ultrasound findings. The following processes were implemented:
1. Data Validation: All questionnaire responses and ultrasound measurements were reviewed for consistency and completeness. In the event of data entry errors or inconsistencies, the original data sources (questionnaires or ultrasound logs) were consulted for clarification and correction.
2. Source Data Verification: Ultrasound findings were periodically cross-referenced with patient charts and medical records to verify the accuracy of reported comorbidities and other patient characteristics.
3. Data Dictionary and Coding: The study utilized a detailed data dictionary for all variables, including categorical coding for risk factors, comorbidities, and ultrasound findings. For example, the World Health Organization's MedDRA coding was used for medical conditions, and normal ranges for fasting times and antrum cross-sectional area were predefined based on existing literature.
4. Standard Operating Procedures (SOPs): SOPs were developed for patient recruitment, ultrasound measurements, data collection, data management, and data analysis. These procedures included regular checks for missing or incomplete data, predefined criteria for image quality in ultrasound readings, and protocols for identifying and addressing any out-of-range results.
Sample Size Assessment:
A sample size of 207 patients was calculated using the G-Power program (version 3.1.9.7), with the degrees of freedom set to 3 and an effect size of 0.3. This sample size was based on examining two patient groups with a categorical variable containing four categories, with a type I error rate of 0.05 and a type II error rate of 0.1. To account for subgroup analyses and potential registration errors, a target of at least 300 patients was set. The final number of patients completing both the questionnaire and ultrasound was 404.
Plan for Missing Data:
To address missing data, the following strategies were implemented:
* Any incomplete questionnaire responses were flagged, and patients were contacted to complete the missing information, where possible.
* Missing ultrasound measurements were rare, but if ultrasound images were deemed inadequate (e.g., poor image quality or patient inability to tolerate positioning), these patients were excluded from the specific analyses.
* For missing or inconsistent values, a complete-case analysis was employed, excluding individuals from specific models if key variables were not available.
* Sensitivity analyses were conducted to assess the impact of missing data on the final results.
Statistical Analysis Plan:
Data analysis was conducted using R statistical software (version 4.1.2). Statistical techniques were selected based on the study objectives:
1. Primary Objective (USG Efficacy):
* Logistic regression models were employed to assess the relationship between patient characteristics (e.g., early satiety, cholelithiasis) and full stomach based on USG findings. The dependent variable was the binary outcome of full stomach vs. empty stomach.
* Qualitative scoring based on the Perlas risk score was incorporated to evaluate aspiration risk, with separate logistic models developed to predict Grade 0 (low risk), Grade 1 (moderate risk), and Grade 2 (high risk).
* Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC), and the optimal cutoff points were determined using Youden's index to balance sensitivity and specificity.
2. Secondary Objective (Questionnaire vs. USG):
* Cohen's Kappa statistic was used to assess agreement between questionnaire-based predictions and USG findings. This analysis aimed to determine how well the questionnaire could predict residual gastric content and aspiration risk compared to USG.
* Time-efficiency was analyzed by comparing the average duration of completing the questionnaire (3-5 minutes) vs. the ultrasound examination (2.5 minutes).
3. Gastric Volume Estimation Formulas:
* Correlation coefficients (Spearman's rank) were calculated to determine the association between estimated gastric volumes from the four formulas (Michiko, Bouvet, two of Perlas\' formulas) and the ultrasound findings (full stomach and aspiration risk).
* Logistic regression models were developed to evaluate whether the estimated gastric volumes from each formula were significant predictors of full stomach or aspiration risk. Negative or non-significant results were flagged as indicators of formula limitations.
4. Adjusted Models:
* Multivariate models adjusting for potential confounders (e.g., age, sex, comorbidities such as diabetes) were developed to ensure that observed associations between patient characteristics and gastric content or risk were not confounded by other variables.
* A complete-case analysis was performed for all multivariate models, excluding patients with missing values for any of the key variables.
Limitations:
The study acknowledges several limitations. First, no direct measurement of gastric volume was obtained through orogastric tube aspiration, limiting the ability to definitively compare USG findings to a gold standard. Second, the analysis was constrained by the relatively small number of patients (n = 57) for whom volume estimation was possible, particularly due to negative or invalid results from the gastric volume formulas. Lastly, the study was performed in a single center, and all ultrasound measurements were conducted by a single investigator, which, while minimizing inter-operator variability, may limit the generalizability of the findings to other settings.
Future Directions:
Future studies should focus on refining the gastric volume estimation formulas and exploring additional patient-reported symptoms and clinical factors that may affect gastric emptying. There is also a need to standardize ultrasound protocols and validate the findings in a multi-center, diverse patient population. In addition, comparisons with other preoperative risk assessment tools, such as orogastric tube suction, should be explored.
