Section Editor: William J. Greeley.
Studies of preoperative fasting intervals and gastric contents in children suggest that a shortened clear liquid fast of 2-3 h does not significantly increase gastric fluid volume or decrease gastric fluid pH compared with a traditional fasting interval of 6-8 h [1-12]. This implies that there is no increase in the perioperative risk of aspiration when children are allowed to ingest clear liquids closer to the time of anesthesia induction. Many institutions performing pediatric surgery have shortened the preinduction clear liquid fast to 2-3 h based on these data.
With one exception , all of these studies specifically excluded children with gastrointestinal (GI) disorders. Ironically, children with GI disorders often undergo upper endoscopy and receive deep sedation or general anesthesia for this procedure. Because these anesthetic techniques may cause the loss of protective airway reflexes, it is important to know whether there is a potential for large gastric fluid volumes and low gastric fluid pH in this group of patients. How the gastric contents of this presumably at-risk group vary from those of children without GI disorders presenting for routine surgery is not clear. Also, whether there are any discernible differences within this group based on presenting complaint is not known. We hypothesized that there are no differences between the gastric contents of children undergoing endoscopy and those undergoing nonendoscopic surgical procedures. In an attempt to clarify these issues, we analyzed the gastric contents of a series of children presenting for upper endoscopy.
At our institution, all children who present for upper endoscopy receive either deep sedation or general anesthesia. After obtaining institutional review board approval, 248 children (aged 2 mo to 18 yr) presenting for upper endoscopy were prospectively evaluated. The choice of either deep sedation or general anesthesia and the method of airway management were at the discretion of the anesthesiologist. The institutional review board waived the requirement for informed consent because the standard protocol for fasting and endoscopy was not altered. The children were allowed nothing orally (neither solids nor liquids) for at least 6 h (children <6 mo old) or at least 8 h (children >6 mo old). Parents were instructed to continue any GI medications through the morning of the procedure. After the initiation of anesthesia, one of two pediatric gastroenterologists measured gastric fluid volumes under direct endoscopic vision. A Lukens trap was attached to the endoscope suction for accurate volume determinations. To ensure that only stomach contents were collected, the trap was removed when the endoscope was advanced beyond the pylorus. When gastric fluid was collected, pH was measured using pH paper (Baxter Diagnostics Inc., Deerfield, IL).
Patients were categorized into one of five diagnostic/symptom groupings before endoscopy. The groupings were: vomiting, abdominal pain, gastroesophageal reflux, esophageal disease, and others. Demographic data (age, height, weight) were analyzed according to diagnostic grouping by using analysis of variance. Gastric fluid volume and gastric fluid pH were analyzed according to diagnostic grouping by using the Kruskal-Wallis test. Intergroup comparisons were performed by using the Mann-Whitney U-test after Bonferroni's correction had been made to the significance level (to 0.005) for multiple comparisons. Gastric volume and gastric pH were analyzed against age and weight by correlation. The presence of both risk factors (gastric volume >0.4 mL/kg and gastric pH <2.5) was analyzed according to diagnostic grouping by contingency table. Data are presented as mean +/- SD (range). Significance was determined at the P < 0.05 level.
Data were collected from 248 children-121 female and 127 male. Mean age was 8.1 +/- 5.7 yr (range 2 mo to 18 yr), and mean weight was 31.2 +/- 32 kg (range 3.1-115 kg). The mean gastric fluid volume of the entire group was 0.35 +/- 0.45 mL/kg (range 0-3.14 mL/kg) and the mean pH was 1.37 +/- 1.6 (range 1-7). Gastric fluid volume sufficient for pH testing was obtained in 155 patients. Of the entire group, 33% had gastric fluid volumes >0.4 mL/kg, 87% of those from whom gastric fluid was obtained had pH <2.5, and 30% of the entire group met both criteria (Table 1).
(Table 2) summarizes results after separating the children into smaller groupings based on preprocedure diagnosis or symptom. Children in the abdominal pain and esophageal disease groups were older and weighed more than children in other groups. There was a significant difference among the groups with respect to gastric fluid volume (P < 0.0001). The abdominal pain group had greater gastric fluid volume than the vomiting group (P < 0.004), the reflux group (P < 0.0006), and the other symptom group (P < 0.0001). The gastric fluid volume of the abdominal pain group was not significantly higher than that of the esophageal group (P < 0.02; Bonferroni corrected significance level of P < 0.005). There were no differences in pH among the groups. Table 3 lists the percentage of children in each group with gastric fluid volume >0.4 mL/kg, pH <2.5, and those whose gastric contents met both criteria. The incidence of both risk factors was significantly greater in the abdominal pain group compared with the gastroesophageal reflux, esophageal disease, and other symptom groups (P < 0.004). Thirteen children (5%) received GI medications (H2-blockers, prokinetic agents, proton inhibitors) on the day of endoscopy. Their gastric fluid volume and pH were not significantly different from those who did not receive GI medications. Neither age nor weight correlated with gastric fluid volume (r2 = 0.12 for age, r2 = 0.08 for weight) or gastric fluid pH (r2 = 0.002 for both age and weight).
An important question confronting anesthesiologists caring for children with a history of GI symptoms is whether these children pose an increased risk for aspiration. Some view symptoms such as vomiting and reflux or esophageal symptoms as sufficient reason to consider a patient at increased risk of aspiration pneumonitis during anesthesia. Because of this, some anesthesiologists prefer endotracheal intubation by rapid-sequence induction. Others believe that intubation is unnecessary (especially for a brief procedure such as endoscopy) and introduces the low morbidity risk associated with endotracheal intubation. Until recently, there has been no data on the nature of the gastric contents in children with GI symptoms presenting for upper endoscopy, or any other surgical procedure, to help put the question of aspiration risk into better perspective. The study by Ingebo et al. , as well as the current study, provide data with which to make more reasonable clinical decisions when treating this group of children.
One third of the children in our study had gastric fluid volumes greater than 0.4 mL/kg. Most (87%) of the children from whom gastric fluid could be recovered had a pH <2.5. Approximately one third of the entire group met both at-risk criteria. For comparison, Table 1 lists published data of gastric fluid volumes and pH in children without GI symptoms who were fasted for a comparable time (the control subjects from studies in which fasting intervals were shortened). The gastric fluid volumes of the children in our study do not seem to be different from those of healthy children without GI symptoms. Although our mean pH was lower than that in all of the other studies, a smaller percentage of our children had a gastric fluid pH <2.5 than in three of the five other studies in which these data were available. This supports the contention that the gastric contents of children with GI symptoms are not appreciably different from the contents of healthy children when both groups are fasted for at least 6-8 h. Cook-Sather et al.  published some of the same data included in our historical control. The mean gastric fluid volume for 611 healthy, fasted children from previously published studies was found to be 0.40 +/- 0.45 mL/kg (range 0-4.1 mL/kg). All of the data from patients from our study fall within this range.
For the purposes of this study, we used the >0.4 mL/kg and pH <2.5 criteria put forward in 1974 by Roberts and Shirley  to describe risk for acid aspiration syndrome. We share the widespread skepticism of these criteria. Many studies have shown that more than 50% of healthy children presenting for surgery have gastric fluid volumes exceeding this limit, despite a prolonged fast [3,4,7,13]. In addition, these risk criteria assume that 100% of gastric contents will be aspirated, which is unlikely. Because other studies of aspiration risk reported their data with respect to these criteria, we did so as well to facilitate comparison.
Children with abdominal pain had larger gastric fluid volumes than children in three of the other four groups. Almost half had gastric fluid volumes >0.4 mL/kg. This was a surprising finding because we expected that children with a history of chronic vomiting or gastroesophageal reflux would be more likely to have the largest gastric residual volumes. The small number of patients in the esophageal group could have been responsible for the lack of a statistical difference between this group and the abdominal pain group. Compared with the other published studies listed in Table 1, children with abdominal pain, when taken as a distinct group, do not seem to have gastric contents different than those of healthy children, despite being our subgroup at highest risk based on volume.
Ingebo et al.  provide the only other data on gastric contents in children with GI symptoms. They studied gastric contents (obtained under direct vision with an endoscope) at different fasting times (0.5-24 h) and found no correlation between fasting interval and gastric fluid volume or pH. They also found no correlation between gastric contents and presenting symptoms, unlike our study. They concluded that there is no advantage to fasting any child (with or without GI symptoms) >2 h. A problem with this conclusion is that children in their study did not undergo randomization or strict control of fasting interval. Based on the data available, we believe that the safety of shortening the fasting interval in children with GI symptoms is questionable.
Our study was prospective and descriptive. Children were not randomized to any particular group. Fasting intervals of at least 6-8 h were chosen based on anesthesia guidelines in our department at the time of the study. For endoscopy patients, these were more restrictive than for children presenting for elective surgery without gastric symptoms (our departmental guidelines allow clear liquids until 3 h before surgery) because there was uncertainty about the risk of aspiration. Because recorded fasting intervals and medication dosing were dependent on parental reporting, this introduces some degree of uncertainty with our data. Another potential criticism of our study is the lack of a concurrent control group of healthy children without GI symptoms. Because endoscopy has associated risks, we could not justify performing this procedure on a group of healthy children presenting for elective surgery. Instead, we used historical controls from the pediatric anesthesia literature for comparison.
We have shown that gastric contents in our group of children are equivalent to the contents in groups of healthy children. We have not demonstrated whether these findings represent a decreased risk in a child who is more likely to vomit or reflux because of a preexisting GI condition. Gastric fluid volume and pH are but two of several risk factors for aspiration (e.g., intraabdominal pressure, gastroesophageal sphincter competence). However, knowing that the gastric contents of children presenting with GI symptoms are not different from those of healthy children lends support for, but does not prove, the safety of not automatically intubating this group of children before sedation or anesthesia.
In summary, we have shown that a group of children with a variety of GI symptoms presenting for upper endoscopy and fasted for at least 6-8 h did not have increased gastric fluid volumes and acidity compared with historical groups of healthy children fasted for an equivalent period of time. These data do not support the argument that children with GI symptoms are a greater risk of aspiration during sedation or anesthesia.
1. Schreiner MS. Preoperative and postoperative fasting in children. Pediatr Clin N Am 1994;41:11-20.
2. Nicolson SC, Schreiner MS, Watcha MF. Preoperative preparation of the child for anesthesia. Amer J Anesth 1996;23:157-62.
3. Ingebo KR, Rayhorn NJ, Hecht RM, et al. Sedation in children: adequacy of two-hour fasting. J Pediatr 1997;131:155-8.
4. Schreiner MS, Triebwasser A, Keon TP. Ingestion of liquids compared with preoperative fasting in pediatric outpatients. Anesthesiology 1990;72:593-7.
5. Nicolson SC, Dorsey AT, Schreiner MS. Shortened preanesthetic fasting interval in pediatric cardiac surgical patients. Anesth Analg 1992;74:694-7.
6. Splinter WM, Stewart JA, Muir JG. The effect of preoperative apple juice on gastric contents, thirst, and hunger in children. Can J Anaesth 1989;36:55-8.
7. Splinter WM, Stewart JA, Muir JG. Large volumes of apple juice preoperatively do not affect gastric pH and volume in children. Can J Anaesth 1990;37:36-9.
8. Splinter WM, Schaefer JD, Zunder IH. Clear fluids three hours before surgery do not affect the gastric fluid contents of children. Can J Anaesth 1990;37:498-501.
9. Sandhar BK, Goresky GV, Maltby JR, et al. Effect of oral liquids and ranitidine on gastric fluid volume and pH in children undergoing outpatient surgery. Anesthesiology 1989;71:327-30.
10. Splinter EM, Schaeffer JD. Ingestion of clear fluids is safe for adolescents up to three hours before anesthesia. Br J Anaesth 1991;66:48-52.
11. Crawford M. Fasting in children prior to surgery. Anesth Analg 1990;71:400-3.
12. Maekawa N, Mikawa K, Yaku H, et al. Effects of two-, four-, and twelve-hour fasting intervals on preoperative gastric fluid pH and volume, and plasma glucose and lipid homeostasis in children. Acta Anaesthesiol Scand 1993;37:783-7.
13. Manchikanti L, Colliver JA, Marrero TC, et al. Assessment of age-related acid aspiration risk factors in pediatric, adult, and geriatric patients. Anesth Analg 1985;64:11-7.
14. Meakin G, Dingwall AE, Addison GM. Effects of fasting and oral premedication on the pH and volume of gastric aspirate in children. Br J Anaesth 1987;59:678-82.
15. Roberts R, Shirley M. Reducing the risk of acid aspiration during cesarean section. Anesth Analg 1974;53:859-68.
16. Cook-Sather SD, Liacouras CA, Previte JP, et al. Gastric fluid measurement by blind aspiration in paediatric patients: a gastroscopic evaluation. Can J Anaesth 1997;44:168-72.