Variations in blood glucose levels following gastrostomy tube insertion in a paediatric population
- First Online:
- Cite this article as:
- Bernhard, N., McAlpine, K., Moineddin, R. et al. Pediatr Radiol (2014) 44: 863. doi:10.1007/s00247-014-2891-6
- 121 Views
Radiologic insertion of a gastrostomy or gastrojejunostomy tube is a common procedure in children. Glucagon is used to create gastric hypotonia, permitting gastric distension and facilitating percutaneous puncture. Glucagon can cause hyperglycaemia and potentially rebound hypoglycaemia. The safety of glucagon and incidence of hypoglycaemia has not been studied following gastrostomy or gastrojejunostomy tube insertion.
To determine variations in blood glucose in children post gastrostomy or gastrojejunostomy tube insertion. Secondarily, to determine the frequency of hypoglycaemia and hyperglycaemia in children who did or did not receive glucagon.
Materials and methods
This is a retrospective observational study of 210 children undergoing percutaneous gastrostomy or gastrojejunostomy tube insertion over a 2-year period. We studied the children’s clinical and laboratory parameters. Abnormal blood glucose levels were defined according to age-established norms. We used descriptive statistics and ANOVA.
We analysed 210 children with recorded blood glucose levels. More than 50% of the children were less than the third percentile for weight. In the glucagon group (n = 187) hyperglycaemia occurred in 82.3% and hypoglycaemia in 2.7% (n = 5). In the no glucagon group (n = 23), hyperglycaemia occurred in 43.5% and there were no cases of hypoglycaemia. The peak blood glucose occurred within 2 h, with normalization by 6 h post-procedure. Five children became hypoglycaemic, all received glucagon; 4/5 had weights <3rd percentile. Logistic regression analysis revealed no factors significantly associated with hypoglycaemia.
Greatest blood glucose variability occurs between 1 h and 3 h post-procedure. Hyperglycaemia is common and more severe with glucagon, and hypoglycaemia rarely occurs. These findings have assisted in developing clinical guidelines for post-percutaneous gastrostomy/gastrojejunostomy tube insertion.
Glucagon is primarily known for its role in maintaining blood glucose homeostasis. During hypoglycaemic states, endogenous glucagon is naturally released from pancreatic islet cells and quickly initiates the mobilization and breakdown of glycogen stores in the liver. The release of glucose from glycogen stores raises the level of glucose in the blood. Following a sharp increase in blood glucose, a phenomenon known as rebound hypoglycaemia has been observed [1, 2]. This can occur with over-compensation of counter-regulatory hormones including insulin, aimed at lowering blood glucose. The complex relationship between hormones regulating blood glucose can cause a child to enter a hypoglycaemic state [1–3].
In interventional radiology, however, glucagon is usually used for its inhibitory effect on the smooth muscles in the gastrointestinal tract [2, 4]. Its use prevents image degradation from peristalsis and improves image quality during abdominal angiography . It is used to inhibit peristalsis of the cecum during air inflation for image-guided cecostomy tube placement. Most commonly in interventional radiology, glucagon is used during gastrostomy and gastrojejunostomy insertions to constrict the pyloric sphincter and to cause hypotonia of the body of the stomach . These gastric effects permit the stomach to accommodate large amounts of air, and enable gastric distension to be achieved. While the use of glucagon in interventional radiology is generally considered safe, it is important for the paediatric interventionalist to be aware of the potential effects of glucagon on the paediatric patient’s blood sugar.
The insertion of a gastrostomy or gastrojejunostomy tube is a common procedure in children who may be unable or unsafe to feed by mouth, or incapable of obtaining sufficient nutrients through their diet [7, 8]. Children receiving gastrostomy or gastrojejunostomy tube insertions are often undernourished, with failure to thrive and associated depletion of glycogen stores in their liver . At our institution, glucagon has been used in interventional radiology during gastrostomy/gastrojejunostomy tube placement for many years, without routine blood glucose monitoring post-procedure.
A 15-month-old girl with Russell–Silver syndrome received a single dose of glucagon during a gastrostomy tube insertion. In the post-anaesthetic recovery area she developed apnea and seizures. Initially, these symptoms were thought to be narcotic-related; however a bedside blood glucose check done in the course of resuscitation revealed profound hypoglycaemia. The girl recovered following administration of dextrose but required transfer to the critical care unit for close observation.
This event made us aware of the potential for hypoglycaemia when glucagon is used in interventional radiology. It prompted two changes in practice: (1) first, monitoring of blood glucose post-procedure in children who received glucagon and (2) later the addition of dextrose to IV infusions peri-procedure in children receiving glucagon. Additionally it was the stimulus to undertake this retrospective observational study of blood sugar levels in children following the introduction of these two changes. Awareness of the potential for hypoglycaemia has also prompted some interventionalists to attempt gastrostomy tube insertions without glucagon.
The primary purpose of the study was to examine the variation in blood glucose levels in paediatric patients following procedural glucagon, and secondarily to determine the frequency of hypoglycaemia and hyperglycaemia in patients who received or did not receive glucagon with their gastrostomy/gastrojejunostomy tube insertion. This paper serves to increase awareness amongst paediatric interventionalists as to the potential risks of glucagon.
Materials and methods
We obtained institutional research ethics board approval for this single-centre retrospective observational study. Inclusion criteria were all consecutive patients 0–18 years of age who received gastrostomy or gastrojejunostomy tube insertion at our institution between July 2009 and June 2011. Exclusion criteria were patients 18 years or older, those in whom blood glucose measurements were not checked or available, those who were known to be diabetic or previously had known hyperglycaemia, and those receiving glucagon for procedures other than gastrostomy/gastrojejunostomy insertion (e.g., cecostomy tube insertions, abdominal angiography). Data sources included the Picture Archiving Communication System (PACS), a dedicated interventional database (www.esh.ca), and the electronic patient chart.
Demographics recorded included age and weight at time of tube insertion, gender and diagnoses. Pre-procedural data recorded for each child included fasting status, medications received, and intravenous fluids (volume, type and sugar content) [10, 11]. Procedural data included the use of dextrose-containing IV solutions during the procedure, as well as the use, dose and time of glucagon administration. Traditionally the doses of glucagon used have not been calculated on a per-kilogram scale—rather they have been given empirically, broadly based on the overall size/weight of child (approximate dosing: 0.2 mg for neonates, 0.3 mg for children 5–10 kg, 0.4 mg for children 10–20 kg, 0.5 mg for children >20 kg), often erring on a smaller initial dose and giving a second dose if air escapes into the duodenum on inflation of the stomach . Those who received glucagon were called the glucagon group; those who did not receive glucagon were called the no glucagon group. Post-procedural data included records of time and level of any available blood sugars analysed (either as a point-of-care sample or through the laboratory), and any adverse events noted. Blood glucose levels were not routinely checked pre-procedure, so they were not available for review.
The World Health Organization growth percentile charts were used to assess each child’s growth percentile according to his or her age. For children younger than 10 years a weight-to-age chart was referenced . For children ages 10–18 years old, height at the time of the procedure was noted and a body-mass-index-to-age chart was used to assign growth percentile .
This was measured as the length of time prior to the procedure that a child had been without a glucose-containing fluid or solid. Children who were receiving total parenteral nutrition containing glucose were assigned a fasting status of 0 h.
(1) Pre-procedure IV fluids. Fluids running or initiated on arrival to the procedure area (approximately 30 min prior to gastrostomy tube insertion). (2) Procedural fluids. IV fluids run during the enterostomy tube placement. (3) Post-procedure fluids. IV fluids running after transfer to the post-anaesthetic recovery area. The type of IV fluids was categorized as containing dextrose (e.g., total parenteral nutrition, dextrose ≥5% in any form), or not containing dextrose.
Medications administered to each child 24 h prior to the procedure were classified into four categories based on the medication’s potential effects on blood glucose levels: (1) risk of hypoglycaemia, (2) risk of hyperglycaemia, (3) risk of glucose intolerance or (4) no known effect on glucose homeostasis. The classification of each drug was determined by the potential risks listed on Lexicomp .
Abnormal blood glucose levels were defined according to laboratory-established age-appropriate norms (children <1 year = 2.5–5.5 mmol/L; 1–2 years = 2.5–5.0 mmol/L; 3–11 years = 2.8–6.1 mmol/L; ≥12 years = 3.3–6.1) .
Any child with a blood glucose less than the age-appropriate range was classified as hypoglycaemic (e.g., <2.5 mmol/L in a child younger than 3 years; <2.8 mmol/L in a child 3–11 years; <3.3 mmol/L in a child ≥12 years).
Any child with blood glucose greater than the age-appropriate range was classified as hyperglycaemic. Significant hyperglycaemia was assigned as any value above 11 mmol/L as used in previous literature . Children were grouped into these categories accordingly (normal blood glucose, hypoglycaemic or hyperglycaemic).
Evolution of practice and technique
All children had their gastrostomy/gastrojejunostomy tube inserted using local anaesthesia with either general anaesthesia or intravenous sedation monitored by an anaesthesiologist or the intensive care transport team. A percutaneous retrograde image-guided technique was used . Following the index case, the practice was gradually introduced that a blood glucose check was to be requested at the bedside with a point-of-care blood glucose monitor on arrival in the post-anaesthetic recovery area and on arrival to the ward, i.e. minimum of two readings per child. Children with marked aberrations in their blood glucose levels often had more than two readings, either in the laboratory or at the bedside, to verify the accuracy or confirm an initial result.
Data were entered into an Excel database (Excel 2007; Microsoft Corp., Redmond, WA) and analysed using descriptive statistics. Logistic regression analysis was undertaken to assess for potential factors correlated with an increased risk of post-procedure hypoglycaemia. Repeated measure ANOVA was used to compare the glucose levels in children who received glucagon (glucagon group), those who did not receive glucagon (no glucagon group) and those with hypoglycaemia (hypoglycaemia). All analyses were done using SAS software version 9.3 (SAS Institute Inc., Cary, NC).
Patient demographics (n = 210)
1 year, 11 months
2 weeks–17 years 10 months
Below the 3rd
Above the 50th
Procedural information (n = 210)
Length of fast (hours)
Pre-procedural IV (n = 271)
Receiving an any IV fluids
IV containing 5% dextrose
Procedural IV (n = 262**)
Containing 5% dextrose
Not Containing 5% dextrose
Post-procedural IV (n = 271)
Containing 5% dextrose
Not containing 5% dextrose
Glucagon (n = 271)
223 (82.3% 95% CI 77.3 - 86.4%)
Median dose (mg)
Range of doses (mg)
Blood glucose levels
Hypoglycemia and hyperglycemia in Glucagon and No Glucagon Groups
Glucagon (n = 187)
No glucagon (n = 23)
By 4 hours post-procedure
5 /2.7% (95% CI 1.2–6.1%)
Median time of lowest recorded blood glucose
By 4 hours post-procedure
154 /82.3% (95% CI 76.2–87.2%)
10 /43.5% (95% CI 25.6–63.2%)
Median time of highest recorded blood glucose
Summary of Hypoglycaemia cases
Medication Risk Factors
Use of glucagon in interventional radiology procedures is common and is generally considered safe. Glucagon in paediatric interventional radiology is most commonly used for enteric access (gastrostomy, gastrojejunostomy and cecostomy tube insertions). It is a valuable agent but requires awareness amongst interventionalists as to its potential effects on the patient’s blood sugar. Several years prior to the index case described, there was an adverse event surrounding a child on beta blockers who became profoundly hypoglycaemic following multiple doses of glucagon during a difficult renal angioplasty at our institution. At the time it was considered to be an atypical response in a unique set of circumstances. However more recently the index case presented here emphasised that hypoglycaemia can occur following a single dose of glucagon. This latter case prompted this study and the subsequent change in clinical practice.
This study found that a wide range of blood glucose levels are present in children who have radiologically inserted gastrostomy and gastrojejunostomy tubes, most marked amongst those who receive glucagon. Although the index case initially raised concern for hypoglycaemia, the large number of children with hyperglycaemia is quite striking in these results. The time of greatest blood glucose variability was 1–3 h post tube insertion. More than 91% (192 of 210) of children with blood glucose levels recorded post-procedure had at least one reading within the first 3 h. Of these readings, 78% fell outside the normal laboratory ranges. Following this period, the blood glucose levels steadily decreased to within the normal range. Cases of hypoglycaemia (n = 5) occurred predominantly within these 3 h and only in the glucagon group. Thus, the 1- to 3-h post-procedure time frame is critical for blood glucose monitoring.
Glucagon is well described for its role in the management of hypoglycaemia and for its use as a diagnostic aid in reducing gastrointestinal motility [14, 17]. The pharmacokinetics indicate a time of onset that varies from 1 min when given intravenously to 10 min given intramuscularly. The half-life is said to be short (8–18 min), and it has a duration of hyperglycaemic action up to 32 min. In this study, however, the hyperglycaemia effects took a long time to return to normal. In the treatment for hypoglycaemia in adults or children over 20 kg, the recommended dose is 1 mg; in children <20 kg the recommended dose is 0.5 mg (or 0.02–0.03 mg/kg when calculated on per kilogram) [14, 17]. As a diagnostic aid, the stomach is considered to be less sensitive to the effects of glucagon than other parts of the intestinal tract. The recommended dose as a diagnostic aid in adults ranges from 0.25 mg to 2 mg . There is no definite dosing guideline for its gastrointestinal effects in children. So it is our practice to use broad dose ranges roughly based on patient size and to give a small initial dose that can be followed by a second dose if air is seen to escape into the duodenum. We acknowledge that for some children the resultant dose may be less than or more than a per-kilogram calculated dose .
The underlying causes of the variations in blood glucose levels of children included in this study may be multifactorial . With 54% of the children weighing less than the third percentile for their age, undernourishment and failure to thrive were prevalent. Compared to normal healthy preschool children, who are able to maintain glucose homeostasis after 8 h of fasting, the malnourished population is at risk for depleted liver glycogen stores. We postulate that, in combination with the effect of the fasting requirements for the gastrostomy/gastrojejunostomy tube insertion, these underweight children likely have a decreased source of glycogen for gluconeogenesis . Based on this hypothesis, we would predict that without the stored glycogen, impaired glucose homeostasis and hypoglycaemia would be more likely to occur. Supporting this prediction, four out of five children with significant hypoglycaemia had weights <3rd percentile. In addition, 4/5 had prolonged fasting (>9 h), and 1/5 fasted for 7 h with a small-volume oral electrolyte solution containing only 3% dextrose 3 h prior to the procedure. The guideline to run dextrose containing IV infusions in the peri-procedural phase was introduced shortly after the practice to monitor blood glucose. Such changes in practice take time to be adopted. This is reflected in the lack of dextrose-containing pre- and per-procedural fluids in several children in Table 4. All five instances of hypoglycaemia occurred in the first 13 months of initiating this practice change. Since more universal adoption of both changes there have been no hypoglycaemic events.
There are a number of reasons why hyperglycaemia occurs, even in the no glucagon group (Fig. 4). Children fighting critical illness have average blood glucose levels in the hyperglycaemic range . Because gastrostomy and gastrojejunostomy tubes are only inserted after alternative options to sustain adequate nutrition have been exhausted, many of these children are in catabolic states, which are associated with higher than normal levels of circulating counter-regulatory hormones, which in turn might pre-dispose the children to wide fluctuations in blood glucose levels including hyperglycaemia. Sedation or general anaesthesia is also known to induce a stress response in the body, further raising blood sugar levels . Finally, and most obviously, glucagon used during a gastrostomy/gastrojejunostomy tube insertion is known to raise blood glucose levels. When exogenous IV glucagon is administered, its effect on raising blood glucose can occur within 30 s to 1 min . We observed a persistence of these elevated blood glucose levels well after the initial spike expected with glucagon. We postulate that this prolonged elevation is caused by the combination of factors listed above, in addition to the fact that glucagon was administered while concomitantly running IV dextrose.
Although hypoglycaemia did not occur in the no glucagon group, this was not statistically significant because of the small number of children with hypoglycaemia (n = 5) and the relatively small sample size of the no glucagon group (n = 23) compared to the glucagon group (n = 187). However the trend was toward less fluctuation in blood glucose levels when no glucagon was used. Analysis of the five cases of hypoglycaemia found that none of the demographic or procedural factors recorded was correlated with an increased risk of developing hypoglycaemia. Lack of statistical significance is not surprising given the relatively low frequency of hypoglycaemia events compared to the total number of children studied. Given the high numbers of procedures and lack of identifiable significant risk factors for becoming hypoglycaemic, precautionary measures for all children receiving a radiologically inserted gastrostomy or gastrojejunostomy tube appear to be the best method to control the issue at hand.
The findings of this study have implications for clinical practice and have prompted the establishment of clinical practice guidelines and computerized post-procedural order entry sets at our institution. These include discussion at the time-out regarding the need for IV glucagon, and the need for procedural as well as post-procedural IV fluids containing 5% dextrose. Routine monitoring of blood glucose levels at 1 h and 4 h post-procedure is now part of the computerized order entry set. Since the institution of these guidelines, we have had no further instances of hypoglycaemia (unpublished observation). Likely the most significant factor has been the addition of dextrose to the IV fluids in these children receiving glucagon. The results from this study suggest that we should test the first blood glucose level even earlier, e.g., within 30 min of the procedure, rather than at 1 h, because the expected blood glucose effects from glucagon can have an onset within 30 s. The instances of hypoglycaemia might have been rebound from immediate hyperglycaemia that went undetected because of the timing of blood glucose testing. Given the decreased variability of blood glucose levels in the no glucagon group, there have been some attempts by the interventional radiologists at our institution to do the procedure without the use of glucagon, or with use of alternative agents (e.g., hyoscine butylbromide, or Buscopan).
There are several limitations to the current study. As the practice of monitoring blood sugars post-procedure gradually became established, several children received glucagon but did not have sugars tested and others did not have dextrose running pre-, per- or post-procedure. The former issue has been addressed by incorporating blood glucose checks into the computerized order entry sets. The variability in timing of blood glucose testing post-procedurally (on arrival to the post-anaesthetic recovery area and on arrival to the ward) is a limitation because these arrival times were influenced by many unrelated variables. It is now specified in the computerized orders to test blood sugar at 1 h and 4 h post-procedure. We acknowledge the first blood glucose test may have missed some hyperglycaemias or even hypoglycaemias given the rapid onset of the effect of glucagon on blood sugar. The symptoms of hypoglycaemia are typically not visually apparent , and many in this study were infants or otherwise non-verbal. Thus, unless a blood glucose test was performed at the precise time of hypoglycaemia, the jitteriness, irritability and crying that serve as clinical signs of hypoglycaemia in a non-verbal child might have been discounted as symptoms of post-procedure recovery. Therefore the incidence of hypoglycaemia might have been underdetected. Additionally, in the clinical efforts to ensure that the child did not become hypoglycaemic, the intervention to run dextrose-containing IV fluids will automatically have corrected some potential hypoglycaemias.
This study has raised our awareness of hypoglycaemia and educated us to the prevalence of hyperglycaemia post-glucagon. It has prompted us to introduce the following safety steps to protect against hypoglycaemia: (1) a requirement for dextrose-containing IV fluids during and post gastrostomy tube insertion, (2) confirmation at time-out of the plan to use glucagon and confirmation that there is a dextrose infusion running and (3) post-procedure orders for blood glucose monitoring at 1 h and 4 h following the gastrostomy tube insertion. The relative stability in blood sugar values for children who did not receive glucagon has been the impetus to avoid the use glucagon with gastrostomy/gastrojejunostomy tube insertions. Future prospective research could explore the effectiveness of agents other than glucagon, as well as more frequent, earlier and time-specified blood sugar monitoring before, during and after the procedure.
We gratefully acknowledge the advice and direction provided by Dr. Sanjay Mahant in the completion of the study.
Conflicts of interest