Evaluation of Modified Fasting Protocols to Shorten Fasting Time Before Sedation in Children: A Prospective Randomized Noninferiority Trial

ABSTRACT

Backgrounds

Guidelines for fasting before procedural sedation aim to prevent pulmonary aspiration and are primarily targeted for deep sedation. Our study explored whether a shortened fasting protocol is noninferior to the standard protocol by comparing gastric contents evaluated by ultrasound.

Methods

Pediatric patients aged < 3 years, scheduled for elective transthoracic echocardiography under sedation, were randomly allocated to a standard group (4‐h fasting) or a modified group (4‐h fasting for solid and 1‐h fasting for water). Gastric ultrasound was performed to evaluate cross‐sectional area (CSA) in supine and right lateral decubitus positions (RLDP), with the upper body elevated at 45°. The primary outcome was the CSA‐RLDP (CSA_{RLDP 45}). A noninferiority test was performed applying the delta (Δ) of 2.1.

Results

The noninferiority test showed that the modified fasting protocol was noninferior to the standard fasting protocol in terms of CSA_{RLDP 45}, with a mean difference (95% confidence interval) of 0.16 (−0.55 to 0.87) within the noninferiority range of delta.

Conclusion

The modified fasting protocol was noninferior to the standard in pediatric patients undergoing sedation for transthoracic echocardiography, as assessed by gastric ultrasound.

Trial Registration

ClinicalTrials.gov identifier: NCT05810532

1. Introduction

Fasting before procedural sedation is generally implemented to prevent pulmonary aspiration [1]. Although pulmonary aspiration is rare, it can contribute to pneumonitis, pneumonia, and airway obstruction, as well as death [1]. Although fasting guidelines may differ among hospitals, many are based on the guidelines recommended by the American Society of Anesthesiologists (ASA, 2 h for clear fluid, 4 h for breast milk, infant formula, 6 h for nonhuman milk, light meal, and 8 h heavy meal), or the European Society of Anaesthesiology (ESA, allows clear fluid up to 1 h before sedation) [1, 2, 3]. These guidelines emphasize minimizing unnecessarily prolonged fasting times, because they offer no additional safety and instead increase discomfort and dissatisfaction [1, 2, 3].

Gastric ultrasound is a useful point‐of‐care tool for estimating residual gastric volume (GV) and assessing the risk of aspiration in a more individualized manner [4, 5]. Gastric antral cross‐sectional area (CSA) in the right lateral decubitus position (RLDP) measured at 45° upright position (CSA_RLDP45) is recommended as the most reliable surrogate marker for gastric residual volume [4, 5], incorporating it into mathematical models to estimate GV [6, 7]. Alternatively, CSA_RLDP45 can serve as a cutoff value for identifying an empty stomach [8]. Gastric ultrasound offers the advantages of noninvasiveness and rapidity for aspiration risk assessment; however, applicability may vary depending on the patient population and scanning technique, highlighting the need for further studies [7].

The use of gastric ultrasonography (US) has increased in recent times. We considered its application particularly valuable for pediatric patients, who are more susceptible to hypoglycemia and dehydration [9, 10, 11, 12]. Therefore, we aimed to compare a shortened fasting protocol with a standard protocol in pediatric patients undergoing sedated echocardiography. Specifically, we conducted a noninferiority randomized controlled trial applying a delta value of 2.1 to compare the CSA_RLDP45 in pediatric patients undergoing sedated echocardiography under two different fasting protocols: a shortened fasting protocol (allowing clear fluids up to 1 h before sedation) and a standard fasting protocol in pediatric patients undergoing sedated echocardiography.

2. Methods

2.1. Ethics

Approval was obtained from the Institutional Review Board of Kangbuk Samsung Hospital (IRB no. 2023‐02‐026). When echocardiography was scheduled under sedation, the study was explained to the parents of the patients, and written informed consent was obtained from the outpatient clinic.

2.2. Study Participants

This study included pediatric patients aged < 3 years who were scheduled to undergo elective echocardiography under sedation. Patients meeting any of the following criteria were excluded: presence of concomitant diseases other than heart disease or chromosomal disorders, conditions causing delayed gastric emptying, structural abnormalities of the foregut preventing gastric ultrasound, expected difficult airway, neonates, refusal to participate in the study, or cases where sedation was impossible.

2.3. Randomization and Blinding

Before enrolment, a researcher who was not directly engaged in the study used an internet‐based system (http://www.randomization.com) to perform randomization. After randomization, the patients were allocated with equal probability (50%) to the 4‐h fasting group (standard group) or the 1‐h fasting group (modified group). The researcher performing the gastric ultrasound and outcome assessment was blinded to the group allocation.

2.4. Study Groups

In the standard group, patients were not permitted to consume any food or water for at least 4 h before sedation. Conversely, in the modified group, food intake was limited to 4 h before sedation, with an allowance that water could be consumed up to 1 h before sedation.

2.5. Gastric Ultrasound and Sedation

Upon arrival at the echocardiography room, the study participants were identified and gastric ultrasound was performed. Gastric ultrasound was performed by a highly experienced sonographer with > 5 years of expertise who received specialized training and supervision for gastric ultrasound. The equipment used for gastric ultrasound was a VENUGO (GE Healthcare, Wauwatosa, WI, USA) with a 4 MHz curved transducer. The targeted area for gastric ultrasound was the gastric antrum, which was identified by positioning the ultrasound probe over the epigastric region and moving it in the parasagittal plane towards the right. The antrum was visualized between the left lobe of the liver and the pancreas, with the inferior vena cava situated posteriorly.

Gastric antrum was assessed qualitatively and quantitatively. Qualitative assessment included determining whether the gastric antrum was empty or contained fluid or solid material [4, 5]. For quantitative assessment, the cross‐sectional area (CSA) of the gastric antrum was measured by determining two perpendicular diameters: the anteroposterior (AP) and longitudinal (LD) (Figure 1). These measurements were taken between the peristaltic movements and the serosa layer. The CSA was calculated using the formula for the area of an ellipse, which is (AP × LD × π)/4 [7].

FIGURE 1.

Example image of gastric ultrasound measuring two perpendicular diameters (1: longitudinal diameter, 2: anteroposterior diameter). L, liver.

The patient was placed in a supine position with the upper part of the bed elevated to raise the upper body to 45°. A quantitative assessment was performed, followed by CSA measurement (CSA_{supine 45}). Subsequently, the patient was moved to the right lateral decubitus position (RLDP) for both qualitative and CSA measurements (CSA_{RLDP 45}). All ultrasound images were saved for quality review. Later, the stored images were reviewed by a radiologist to exclude any cases where the measurements were incorrect or where incorrect areas were measured. The CSA of the selected images was used to calculate GV according to the following formula [13]:

$$\mathrm{GV}=-7.8+\left(3.5\times {\mathrm{CSA}}_{\text{RLDP}45}\right)+0.127\times \mathrm{age}\left(\text{in months}\right)$$

Subsequently, the calculated GV was divided by the weight of the patient. Stomach risk was defined as a calculated GV > 1.25 mL/kg or the presence of solid content [6].

After gastric ultrasonography, oral chloral hydrate (50 mg/kg) was administered for sedation. The target level of sedation was set to Observer assessment of alertness/sedation scale (OAA/S) of 4 (lethargic response to name spoken in normal tone) [14]. If the sedation was successful, echocardiography was conducted under sedation. If sedation was unsuccessful, a second dose of oral chloral hydrate (25 mg/kg) was administered 20 min after the first administration. If the second sedation failed, echocardiography was performed in the awake state, and the patient was excluded from the study (Figure 2).

FIGURE 2.

Study flow diagram. RLDP: Right lateral decubitus position.

2.6. Outcomes

The primary outcome of our study was CSA_{RLDP 45}. The secondary outcomes were CSA_{supine 45}, GV, GV/kg, stomach risk, qualitative data of gastric ultrasound in the supine position, and RLDP, both in the 45° upright bed. Sedation depth during echocardiography was assessed using the observer assessment of the alertness/sedation scale [14].

Fasting‐ and sedation‐related complications were assessed. Pneumonia was diagnosed based on chest radiography findings. Aspiration was defined as regurgitation of gastric contents followed by clinical signs, such as coughing, wheezing, or desaturation. Hypoglycemia was identified when the blood glucose levels were < 50 mg/dL. Nausea was recorded based on subjective reports of the patients. Dehydration was evaluated based on the presence of dry oral mucosa, decreased urine output, and delayed capillary refill time. Desaturation was defined as a pulse oxygen saturation of < 95%. Bradycardia was defined as a heart rate of < 60 beats per minute.

2.7. Sample Size Estimation

The primary endpoint of this noninferiority study was to investigate whether the modified group was inferior to the standard group by showing that CSA_{RLDP 45} in the modified group was not greater than that in the standard group. The noninferiority delta (Δ) was determined based on the following backgrounds. In a study conducted on pediatric patients, aged 0–18 years (average age: 11–12 years), which includes the age range of our study participants, a median value of 6.1 cm² for Grade 2 (indicating an increased risk of aspiration) CSA, whereas Grade 1 (where gastric contents are present but not enough to increase the risk of pulmonary aspiration) had a median value of 4.0 cm² [13]. Therefore, adopting the median difference of 2.1 as Δ, we evaluated other studies to determine whether this value could be reasonable. In a previous study involving adults, the Δ for comparing gastric emptying was 2.8 cm² [15]. In another study conducted on infants (mean age: 4.5 months), the median difference in the CSA_{RLDP 45} between a Grade 2 stomach and a Grade 1 stomach was 2.1 cm² [16]. Considering that our study targets children aged < 3 years (which is between the two studies) and that the gastric cross‐sectional area increases with age [17], we determined that Δ = 2.1 was reasonable for our study. Therefore, we considered 2.1 a clinically important difference and set 2.1 as Δ. The null hypothesis (H₀) of our study was that the modified fasting protocol (μ ₁) is inferior to the standard fasting protocol (μ ₀) (Δ ≥ 2.1). The alternative hypothesis (H₁) was that the modified fasting protocol (μ ₁) is not inferior to the standard fasting protocol (μ ₀) (Δ < 2.1).

In our pilot study, conducted on 14 participants (unpublished data), CSA_{RLDP 45} was 2.15 ± 0.74 and 3.30 ± 1.97 cm² in the standard (n = 7) and modified fasting groups (n = 7), respectively. Therefore, a total of 68 patients (34 in each group) were necessary based on 80% power and α = 0.05, considering a drop‐out rate of 10%.

2.8. Statistical Analysis

Data are shown as mean ± standard deviation (SD), median (interquartile range, IQR), or number (%), according to the characteristics of the data. The primary outcome was not normally distributed. Therefore, comparisons between the standard and modified groups were performed using the Wilcoxon rank‐sum test. To estimate the mean and 95% confidence interval (CI), we conducted a nonparametric bootstrap analysis using 10 000 resamples. Noninferiority was concluded if the upper bound of the 95% CI was below the noninferiority margin (Δ = 2.1 cm²). Secondary outcomes were compared using the 2‐sample t test, Wilcoxon rank‐sum test, and chi‐square or Fisher's exact tests for normally distributed continuous variables, non‐normally distributed variables, and numeric data, respectively. All statistical analyses were performed using the R software [version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria].

3. Results

In total, 735 individuals were assessed for eligibility. Among them, 667 were excluded: 554 did not meet the inclusion criteria, and 113 declined to participate. Consequently, 68 participants were enrolled and randomized, with 34 allocated to the standard group and 34 to the modified group. In the standard group, 32 participants received the allocated intervention because two did not show up on the study day. All participants in the modified group received the allocated interventions. In the standard group, four participants were lost to follow‐up because of gastric ultrasound examination failure due to poor patient cooperation. In the modified group, two participants were lost to follow‐up: one due to gastric ultrasound examination failure due to poor patient cooperation and another due to solid intake < 4 h before the intervention. After excluding one patient from the standard group due to incorrect measurements, 27 and 32 participants in the standard and modified groups, respectively, were included in the statistical analysis (Figure 3).

FIGURE 3.

CONSORT diagram of the study.

The baseline patient characteristics are shown in Table 1. No significant differences existed in the baseline characteristics between the groups.

TABLE 1.

Baseline patient characteristics.

	Standard group (n = 27)	Modified group (n = 32)
Age, month	16.8 ± 6.0	15.2 ± 5.0
Age distribution, month
≤ 6	0 (0%)	2 (6.2%)
6 <, ≤ 12	7 (25.9%)	7 (21.9%)
12 <	20 (74.1%)	23 (71.9%)
Sex, male	17 (63.0%)	18 (56.3%)
Height, cm	80.6 ± 6.5	78.5 ± 6.5
Weight, kg	10.6 ± 1.6	10.5 ± 1.8
Main cardiac diagnosis
Congenital heart disease
Atrial septal defect	9 (33.3%)	11 (34.4%)
Patent foramen ovale	3 (11.1%)	4 (12.5%)
Ventricular septal defect	1 (3.7%)	1 (3.1%)
Patent ductus arteriosus	6 (22.2%)	7 (21.9%)
Pulmonary stenosis	0 (0%)	1 (3.1%)
Acquired heart disease
Kawasaki disease	6 (22.2%)	6 (18.8%)
Miscellaneous	2 (7.4%)	2 (6.2%)

Note: Data are presented as mean ± SD, and numbers (%).

The noninferiority test for CSA_{RLDP 45} is shown in Figure 4. The mean difference (95% CI) in the CSA_{RLDP 45} between the two groups was 0.16 (−0.55 to 0.87). The upper confidence limit of the mean difference was below the noninferiority limit of 2.1.

FIGURE 4.

Noninferiority diagram. The estimated mean difference in CSA_{RLDP 45} between the standard and modified groups is shown, along with the bootstrap‐based 95% confidence interval (CI) derived from 10 000 resamples. The noninferiority margin (Δ = 2.1 cm²) is illustrated as a red dashed line. The diagram demonstrates that the modified group is noninferior to the standard group.

The data assessed using gastric ultrasound are described in Table 2. The CSA_{supine 45} (p = 0.637) and CSA_{RLDP 45} (p = 0.221) did not differ between the two groups (Figure S1). GV was 2.63 (0.73–4.08) mL in the standard group, whereas it was 3.40 (1.06–4.52) mL in the modified group (p = 0.445). GV/kg did not differ between the groups (p = 0.360). The incidence of GV/kg > 1.25 mL/kg was 3.7% and 6.2% in the standard and modified groups, respectively. Qualitative assessment showed no difference in the supine 45° (p = 0.238). However, it showed a statistical difference (p = 0.034) in the RLDP 45°, with the standard group showing incidences of 51.9% empty, 22.2% liquid, and 25.9% solid and the modified group showing 46.9% empty, 3.1% liquid, and 50.0% solid.

TABLE 2.

Gastric ultrasound assessments.

	Standard group (n = 27)	Modified group (n = 32)	p	SMD
CSAsupine 45, cm2	2.02 (1.62–2.28)	2.00 (1.53–2.63)	0.637	0.143
CSARLDP 45, cm2	2.38 (1.95–2.72)	2.56 (1.93–2.97)	0.221	0.118
GV, mL	2.63 (0.73–4.08)	3.40 (1.06–4.52)	0.445	0.055
GV/kg, mL/kg	0.25 (0.08–0.37)	0.31 (0.09–0.43)	0.360	0.030
GV/kg > 1.25 mL/kg	1 (3.7%)	2 (6.3%)	0.657	0.118
Qualitative assessment
Supine 45°
Empty	19 (70.4%)	20 (62.5%)	0.238	0.452
Liquid	3 (11.1%)	1 (3.1%)
Solid	5 (18.5%)	11 (34.4%)
RLDP 45°
Empty	14 (51.9%)	15 (46.9%)	0.034*	0.708
Liquid	6 (22.2%)	1 (3.1%)
Solid	7 (25.9%)	16 (50.0%)
Risk stomach	17 (29.6%)	17 (53.1%)	0.120	0.491

Note: Data are presented as median (interquartile range), and numbers (%).

Abbreviations: CSA, cross‐sectional area; GV, gastric volume; RLDP, right lateral decubitus position; SMD, standardized mean difference.

^* p < 0.05.

The actual fasting duration was 4 (4.0–5.5) h and 2 (1.5–5.0) h in the standard and modified groups, respectively (p < 0.001). Sedation depth during echocardiography assessed with the observer assessment of alertness/sedation scale was not different between the standard [3 (3, 4)] and modified groups [4 (3, 4), p = 0.706]. No cases of pneumonia, hypoglycemia, nausea, or dehydration occurred among the study participants. The incidence of desaturation was 11.1% and 6.2% in the standard and modified groups, respectively (p = 0.486). Bradycardia was observed in 6.2% of the patients in the modified group compared to 0% in the standard group (p = 0.549) (Table 3).

TABLE 3.

Preprocedural fasting details, sedation quality, satisfaction, and associated complications.

	Standard group (n = 27)	Modified group (n = 32)
Actual fasting duration, h	4 (4.0–5.5)	2 (1.5–3.0)
OAA/S during echocardiography	3 (3–4)	4 (3–4)
Complications regards to fasting or sedation
Pneumonia	0 (0%)	0 (0%)
Aspiration	0 (0%)	0 (0%)
Hypoglycemia	0 (0%)	0 (0%)
Nausea	0 (0%)	0 (0%)
Dehydration	0 (0%)	0 (0%)
Desaturation	3 (11.1%)	1 (3.1%)
Bradycardia	0 (0%)	2 (6.2%)

Note: Data are presented as mean ± SD, median (interquartile range), and numbers (%).

Abbreviations: N/A, not available; OAA/S, observer assessment of alertness/sedation scale.

4. Discussion

Our study demonstrated that the modified fasting protocol (allowing oral intake up to 4 h before sedation and permitting water intake until 1 h before sedation) was noninferior to the standard protocol (oral intake not allowed for 4 h before sedation, including water) when compared with CSA_{RLDP 45}. The mean difference (95% CI) in CSA_{RLDP 45} between the two groups was 0.16 cm² (−0.55 to 0.87), with the upper limit of the confidence interval remaining below the prespecified noninferiority margin of 2.1 cm², indicating no clinically significant increase in gastric content volume.

Preprocedural fasting in children has several disadvantages, including increased irritability, distress, and anxiety due to hunger and thirst [10], as well as the risk of hypoglycemia, especially in infants and young children who have high metabolic rates and low glycogen storage [10, 11, 18]. Moreover, strict fasting requirements can complicate scheduling, leading to delays, rescheduling, or cancelations, which may extend the fasting period and increase parental complaints [19]. Therefore, we believe that our study provides clinically meaningful evidence supporting the noninferiority of minimized preprocedural fasting times in pediatric patients undergoing transthoracic echocardiography under sedation.

Our study allowed all types of oral intake, including solid food, human milk, breast milk, and formula, for up to 4 h before the procedure for both groups. Consequently, solid content was observed in 25.9% and 50% of the standard and the modified groups, respectively. Thus, the incidence of risk stomach (solid or GV/kg > 1.25 mL/kg) [4, 20] was 29.6% and 53.1% in the standard and modified groups, respectively. Although the median CSA_{RLDP 45} values (2.38 and 2.56 cm² in the standard and modified groups, respectively) were close to the empty stomach cutoff of 2.40 cm² for those aged < 24 months, the similarity in CSA does not necessarily imply similar safety profiles between low‐volume solid and liquid contents [8]. Because, unlike liquids, even small amounts of retained solid material may carry a different risk profile for aspiration and should not be assumed to be benign based on CSA alone. Previous studies have suggested that milk, formula, and light meals are generally emptied within 4 h [21, 22], but further research is warranted to clarify whether minimal solid content, if present, can be safely tolerated when fasting is required for sedation.

Our study used gastric ultrasound to compare gastric residual volume between the two fasting protocols. Although caregiver‐reported fasting times are commonly used to estimate gastric emptying in children, this method is often unreliable due to recall bias and underreporting of noncompliance [20, 23]. Therefore, we believe our study is meaningful in demonstrating the value of gastric ultrasound as an objective tool to quantitatively assess gastric residual volume, highlighting its practical utility in real‐world outpatient settings.

Our study had several limitations. First, we did not observe severe sedation‐related complications such as aspiration pneumonia. Our study targeted minimal sedation (OAA/S 4), and if deepened, moderate sedation (OAA/S 3); therefore, the respiratory reflexes remained unchanged. However, as sedation can unintentionally deepen, our study results suggest no difference in gastric CSA between the two fasting protocols but cannot be extrapolated to show that severe complications will not occur [2]. Second, this study was designed as a noninferiority test as a primary endpoint. Although various secondary outcomes were evaluated, formal adjustments for multiple comparisons were not performed. Therefore, the findings of these additional analyses should be interpreted with caution. Third, although we aimed to use a modified fasting protocol of 1 h, the actual fasting time in the modified group ranged from 1.5 to 5 h. This was because of delays in the procedure, waiting times, and recall bias, which are common in our outpatient setting. Despite these discrepancies, we believe that our data provide valuable insights into the practical feasibility and safety of permitting clear fluids in similar clinical settings. Fourth, our study is limited in its ability to generalize the results to a broader population. Our study included a relatively small number of participants (mean age, 59). In addition, although our study targeted patients aged < 3 years, we included only two patients aged < 6 months because several declined to participate. Furthermore, our study excluded patients with conditions associated with delayed gastric emptying for safety reasons. These factors may limit the generalizability of our findings to a more diverse population with a broader age distribution. Finally, our primary outcome was CSA_RLDP, and not GV. Although a high GV is a well‐known risk factor for pulmonary aspiration [24], gastric ultrasonography estimates the GV using CSA_RLDP‐based calculations rather than measuring it directly. This estimation process can introduce errors, sometimes resulting in negative values [25], raising concerns about the reliability of GV measurements obtained by gastric ultrasound. CSA_RLDP can effectively differentiate between an empty and non‐empty stomach in pediatric patients [23]. Given its practicality and the ease of obtaining rapid, intuitive measurements, further studies are needed to validate the use of CSA_RLDP alone for aspiration risk assessment in pediatric patients.

In conclusion, the gastric ultrasound assessment of the gastric antrum suggested that a modified fasting protocol (allowing water 1 h before sedation) did not result in a larger gastric CSA_{RLDP 45} compared to the standard fasting protocol (4 h of fasting) in pediatric patients undergoing transthoracic echocardiography under sedation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1. Violin plot of cross‐sectional area (CSA) in supine and right lateral decubitus position (RLDP) with upper body 45° elevated status in the standard group and modified groups.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.