Postoperative Feeding Regimens After Laparoscopic Gastrostomy Placement

Abstract

Purpose:

The objective of this study was to evaluate postoperative feeding regimens after laparoscopic gastrostomy placement and their effect on outcomes.

Methods:

Children 18 years of age or younger, who underwent laparoscopic gastrostomy placement at a tertiary-care academic children's hospital between January 2014 and October 2016, were reviewed. Data collected included patient characteristics, postoperative feeding regimen, and clinical outcomes. Statistical analysis was performed using Chi-square, Fisher's exact, and Wilcoxon Rank-Sum tests.

Results:

We reviewed the medical records of 270 children that underwent laparoscopic gastrostomy placement by 15 pediatric surgeons. The median age was 2.7 (interquartile range [IQR], 0.7–9.6) years, and 50% (n = 136) were male. The median body mass index was 15.5 (IQR, 14.0–17.5). Complications within 90 days included: granulation tissue (34%), leakage (17%), dislodgement (14%), and skin and soft-tissue infection (9%). Two patients returned to the operating room, 1 for a dislodged tube, and another for a volvulus within 10 days of gastrostomy tube placement. A subset analysis of outpatients that underwent elective laparoscopic gastrostomy placement showed variation in the day of initial feeds (0–2 postoperative days [POD]), method of initial feeds (continuous versus bolus) and choice of initial feeds (Pedialyte versus formula/breast milk). There was a significant difference in median hospital length of stay for early versus late initiation of feeds (POD 0: 2.1 days versus POD ≥1: 3.1 days, P < .01) without a difference in postoperative complications.

Conclusion:

There is substantial variation in the postoperative feeding regimen after laparoscopic gastrostomy. Initiation of early postoperative feeds may result in decreased length of stay without increasing complications.

Introduction

Laparoscopic gastrostomy tube placement was introduced in the 1990s and is one of the most common procedures performed in children. In 2006, an estimated 11,000 gastrostomy tubes were placed in children <18 years of age in the United States.¹ Gastrostomy feeds has become the method of choice for providing nutrition to children who require long-term nutritional support.^2,3 A study by Cameron et al. identified gastrostomy tube placement as being one of the most pressing areas in the field of pediatric surgery for comparative effective research and standardization of care due to the high cost resulting from the degree of variation and the frequency of the procedure in the pediatric population.²

In some practices, feeds are traditionally started 24 hours after gastrostomy placement due to concern for high gastric residuals that may lead to aspiration and to decrease the risk of peritonitis.⁴ However, this practice is not supported by high-level evidence. Studies in children after percutaneous endoscopic gastrostomy tube placement have shown that it is safe to initiate feeds within hours of placement even in patients in a critical care unit.⁵ Yet, despite decades of experience, and a growing body of literature, no definitive consensus regarding optimal practice has been described. There is no consensus regarding the decision of when to initiate feeds, what type of feeds to initiate, or the rate of increase.

The objective of this study was to evaluate postoperative feeding regimens after laparoscopic gastrostomy placement and their effect on outcomes.

Materials and Methods

Following approval by the Institutional Review Board at Texas Children's Hospital (H-39576), medical records of children 18 years of age or younger, who underwent a laparoscopic gastrostomy tube placement between January 2014 and October 2016, were retrospectively reviewed. Patients were identified based on Current Procedural Terminology codes (43653-Laparoscopic Surgical Gastrostomy). Patients who underwent concurrent bowel surgeries (i.e., Malone procedure, cecostomy, Ladd procedure, posterior sagittal anorectoplasty, or ileostomy closure) or had a clinical indication necessitating delays in feeds (i.e., respiratory distress, intubation, or intraoperative complication) were excluded. Data collected included patient characteristics (age, gender, weight, body mass index [BMI], race, and preoperative diet), operative and perioperative details (procedure time, concurrent procedures, type of gastrostomy, hospital length of stay, hospital costs, and complications), and details regarding postoperative feeding regimen (timing of initiation of feeds, method of initial feeds, and type of initial feeds).

To analyze the postoperative feeding regimen, we performed a subset analysis of outpatients that underwent elective laparoscopic gastrostomy placement and compared complications, total direct variable hospital costs, perioperative costs, and hospital length of stay by day of initiation of feeds. A second subset analysis was performed on outpatients undergoing elective laparoscopic gastrostomy placement that were already nasogastric tube (NGT) dependent for feeds. This cohort of patients was analyzed to compare complications, total direct variable hospital costs, and hospital length of stay by type of feeds initiated after laparoscopic gastrostomy placement.

Data analyses were performed using STATA software version 14.2 (STATACorp, LLC, College Station, TX). Patient characteristics and outcomes are described descriptively using counts and percentages for categorical variables and as median with interquartile range (IQR) for continuous variables. Comparative analysis was performed using the Wilcoxon rank test and Fisher's exact test or Chi-square test, as appropriate. A P < .05 was used to determine significance.

Results

Patient characteristics

There were 270 children who underwent laparoscopic gastrostomy tube placement by 15 pediatric surgeons. The median age was 2.7 years (IQR, 0.7–9.6), and 50% (n = 136) were male. The median weight was 11.4 kg (IQR, 6.7–21.4), with a median BMI of 15.5 (IQR, 14.0–17.5). Based on the American Society of Anesthesiologists (ASA) physical status classification, the majority (76%) of patients was an ASA class 3 (severe systemic disease, i.e., not incapacitating). Preoperatively, 47% (n = 115) of patients were NGT dependent for feeds and 35% (n = 85) of patients were taking a regular diet. Fourteen percent of the patients had previous abdominal surgery (Table 1). The indications for gastrostomy placement included: failure to thrive, inability to tolerate oral intake, dysphagia, oral aversion, and anticipation of cancer/chemotherapy.

Table 1.

Patient Characteristics

	All patients (n = 270)	Inpatients (n = 114)	Same-day admit (n = 156)	P
Age (years)^a	2.7 (0.7–9.6)	1.0 (0.3–8.9)	5.1 (1.3–9.8)	<.01
Male	50% (136)	49% (56)	51% (80)	.73
Weight (kg)^a	11.4 (6.7–21.4)	7.2 (4.6–19.3)	13.3 (8.8–23.5)	<.01
BMI (kg/m²)^a	15.5 (14.0–17.5)	15.9 (14.2–17.6)	15.4 (13.9–17.2)	.20
ASA				<.01
2	13% (35)	9% (10)	16% (25)
3	76% (205)	71% (80)	80% (125)
4	10% (29)	20% (23)	4% (6)
Race				.09
African American	14% (33)	21% (18)	10% (15)
Hispanic	35% (80)	28% (24)	39% (56)
Asian	6% (14)	7% (6)	6% (8)
White	44% (101)	44% (37)	45% (65)
Preoperative diet				<.01
NGT dependent	47% (115)	63% (64)	36% (51)
Regular diet	35% (85)	25% (25)	42% (60)
Puree/nectar	17% (41)	11% (11)	20% (30)
TPN	2% (4)	2% (2)	1% (2)
Previous abdominal surgery	14% (38)	12% (14)	15% (24)	.45

Reported as median (IQR).

ASA, American Society of Anesthesiologists; BMI, body mass index; IQR, interquartile range; NGT, nasogastric tube; TPN, total parenteral nutrition.

Operative and perioperative details

The median operative time in patients undergoing isolated laparoscopic gastrostomy tube placement was 43 minutes (IQR, 35–56). A total of 69 patients (26%) underwent concurrent procedures at the time of laparoscopic gastrostomy placement, of which 39% (n = 27) were abdominal procedures (fundoplication, herniorrhaphy, or appendectomy). Nonabdominal concurrent procedures included circumcisions, tracheostomies, or myringotomies. A MIC-KEY^® low-profile gastrostomy button (Halyard Health, Alpharetta, GA) was placed in 97% of patients at the time of surgery.

Postoperatively, 73% (n = 198) of patients were admitted to the floor and 23% (n = 62) were admitted to an intensive care unit. The surgical service was the primary postoperative admitting service in 49% (n = 132) of the patients.

Complications within 90 days included: granulation tissue (34%), leakage from gastrostomy site (17%), dislodged gastrostomy tubes (14%), and skin and soft-tissue infections (9%). Two patients returned to the operating room within 90 days. One patient had a dislodged gastrostomy tube replaced under endoscopic guidance, and 1 developed a volvulus within 10 days of gastrostomy requiring an exploratory laparotomy. Additionally, after discharge, 11% (n = 31) of patients returned to the emergency department within 90 days. The most common reasons for return to the emergency department were dislodged gastrostomy tubes (55%, n = 17) and skin and soft-tissue infections (23%, n = 7) (Table 2).

Table 2.

Operative and Perioperative Details

	All patients (n = 270)	Inpatients (n = 114)	Same-day admit (n = 156)	P
Postoperative admission location				<.01
NICU	13% (36)	32% (36)	0% (0)
CVICU	1% (2)	2% (2)	0% (0)
PICU	9% (24)	17% (19)	3% (5)
PCU	4% (10)	4% (4)	4% (6)
Floor	73% (198)	46% (53)	93% (145)
Primary admitting team				<.01
Surgery	49% (132)	4% (5)	81% (127)
Other	51% (138)	96% (109)	19% (29)
Concurrent procedure	26% (69)	33% (38)	20% (31)	.04
Procedure time (minutes)^a	42.5 (35–56)	48 (38–64)	41 (33–53)	.56
Concurrent abdominal Procedure	10% (27)	8% (9)	12% (18)	.62
Fundoplication	77% (21)	4% (5)	89% (16)
Hernia repairs	26% (7)	4% (4)	17% (3)
Laparoscopic appendectomy	4% (1)	1% (1)	0% (0)
Type of gastrostomy placed				.94
MIC-KEY button	97% (261)	96% (110)	97% (151)
Mini-ONE button	2% (5)	2% (2)	2% (3)
Pezzar	0% (1)	1% (1)	0% (0)
Mic-Gastrojejunostomy tube	1% (3)	1% (1)	1% (2)
Postoperative gastrostomy study	7% (19)	9% (10)	6% (9)	.61
ED visit (within 90 days)	11% (31)	14% (16)	10% (15)	.22
Dislodged gastrostomy tube	55% (17)	69% (11)	40% (6)
Skin and soft-tissue infection	23% (7)	25% (4)	20% (3)
Other	23% (7)	13% (2)	33% (5)
Complications (within 90 days)				.11
Granulation tissue	34% (91)	25% (29)	40% (62)
Leakage	17% (47)	23% (26)	13% (21)
Dislodged gastrostomy tube	14% (39)	20% (23)	10% (16)
Skin and soft-tissue infection	9% (23)	10% (11)	8% (12)
Return to operating room	1% (2)	0% (0)	1% (2)

Reported as median (IQR).

Mini-ONE button (Applied Medical Technology, Brecksville, OH).

CVICU, cardiovascular intensive care unit; ED, emergency department; IQR, interquartile range; NICU, neonatal intensive care unit; PCU, progressive care unit; PICU, pediatric intensive care unit.

Subset analysis of outpatients that underwent elective laparoscopic gastrostomy placement

There were 156 outpatients identified that underwent elective laparoscopic gastrostomy placement and same-day admission. Their median age was significantly higher at 5.1 years (IQR, 1.3–9.8) compared with the total population (P < .01). Their median weight was 13.3 kg (IQR, 8.8–23.5), with a median BMI of 15.4 (IQR, 13.9–17.2), and 51% (n = 80) of patients were male. Based on ASA classification, the patients were generally healthier than the total population (P < .01). Preoperatively, 36% (n = 51) of patients were NGT dependent for feeds and 42% (n = 60) of patients were taking a regular diet. This was also significantly different from the total population (P < .01). Fifteen percent of the same-day admission cohort had previous abdominal surgery (Table 1). The location and primary service responsible for the postoperative care of these patients also significantly differed compared with the total population. The same-day admission cohort was more frequently admitted to the floor (93%, P < .01) and primarily managed by the surgical service (81%, P < .01) (Table 2).

Analysis of the postoperative feeding regimen of the same-day admission cohort demonstrated variation in the timing, method, and type of initial feeds. The majority (87%) of patients had feeds initiated on postoperative day (POD) 1. However, feeds were also initiated on POD 0 in 8% (n = 13), and POD 2 or later in 4% (n = 7) of patients. Continuous feeds were initiated in 61% (n = 75) of patients, whereas 39% (n = 47) of patients were initiated on bolus feeds. The type of feeds initiated also varied, with 21% (n = 33) of patients starting on formula or breast milk, whereas the remaining 79% (n = 123) were started on Pedialyte (Table 3).

Table 3.

Postoperative Feeding Regimen

	Same-day admit (n = 156)
Day of initiation of feeds
POD 0	8% (13)
POD 1	87% (136)
POD 2+	4% (7)
Method of initial feeds
Continuous	61% (75)
Bolus	39% (47)
Initial feeds
Formula/breast milk	21% (33)
Pedialyte	79% (123)

POD, postoperative day.

Comparative analysis of outpatients that underwent elective laparoscopic gastrostomy placement based on postoperative feeding regimen

Comparative analysis of outpatients that underwent elective laparoscopic gastrostomy placement and same-day admission demonstrated a significant difference in hospital length of stay and total hospital direct variable costs for early versus late initiation of feeds. Median hospital length of stay was significantly shorter at 2.1 days (IQR 1.5–2.5) with initiation of feeds on POD 0 compared with 3.1 days (IQR 2.3–3.5) when feeds were initiated on POD 1 or later (P < .01). Additionally, total hospital direct variable costs were significantly lower at $3200 (IQR, 2600–3700) with initiation of feeds on POD 0 compared with $4600 (IQR, 3900–5700) when feeds were initiated on POD 1 or later (P < .01). There was no difference in postoperative complications or perioperative costs based on the timing of initial feeds (Table 4).

Table 4.

Comparative Analysis for Same-Day Admissions

	All same-day admit			NGT dependent+same-day admit
	POD 0 (n = 13)	POD ≥1 (n = 143)	P	Pedialyte (n = 41)	Formula/breast milk (n = 10)	P
Leakage	3	18	.39	7	1	1.00
Granulation tissue	7	55	.28	21	2	.09
Skin/soft-tissue infection	0	12	.60	5	0	.57
Return to operating room	0	2	1.00	0	0
Nausea or vomiting during advancement of feeds	3	17	.33	4	1	1.00
Total direct variable costs (dollars, thousands)^a	3.2 (2.6–3.7)	4.6 (3.9–5.7)	<.01	4.4 (3.5–5.6)	3.7 (3.3–4.6)	.10
Perioperative costs (dollars, thousands)^a	1.6 (1.5–1.9)	1.8 (1.5–2.3)	.32	1.7 (1.5–2.1)	1.9 (1.5–1.9)	.65
Hospital length of stay (days)^a	2.1 (1.5–2.5)	3.1 (2.3–3.5)	<.01	3.0 (2.3–3.5)	2.2 (2.1–2.4)	.02

Reported as median (IQR).

IQR, interquartile range; NGT, nasogastric tube; POD, postoperative day.

Further comparative analysis of outpatients undergoing elective laparoscopic gastrostomy placement that were already NGT dependent for feeds also demonstrated differences in hospital length of stay and total hospital direct variable costs based on type of feeds initiated (formula/breast milk versus Pedialyte). Median hospital length of stay was significantly shorter at 2.2 days (IQR, 2.1–2.4) with initiation of formula or breast milk feeds compared with 3.0 days (IQR, 2.3–3.5) when feeds were initiated with Pedialyte (P = .02). Additionally, total hospital variable costs were lower at $3700 (IQR, 3300–4600) with initiation of formula or breast milk feeds compared with $4400 (IQR, 3500–5600) when feeds were initiated with Pedialyte (P = .10). There was no difference in postoperative complications or perioperative costs based on the type of feeds initiated in this cohort of NGT-dependent patients (Table 4).

Discussion

While the drivers of increasing healthcare costs are multifactorial, variation in care is known to be a major contributor.^6,7 For example, in colorectal surgery, Krell et al. analyzed determinants of variation in hospital length of stay after 20,000 adult colorectal resections using multivariate regression modeling and found that the differences in hospital length of stay are frequently due to “practice style differences.”⁸ Standardized perioperative care after colorectal, hepatobiliary, and cardiac procedures has been shown to decrease morbidity and shorten time to hospital discharge in the adult population.⁹ This, in turn, has led to the promulgation and adoption of enhanced postsurgical recovery protocols in adults. While there are less data in children, emerging efforts aimed at decreasing variability of care through protocol implementation have been initiated at several major pediatric centers, and subsequent studies have demonstrated that decreasing variation in care results in improved outcomes and improved efficiency.^10,11 In children undergoing appendectomies, Skarda et al. found that their hospital length of stay could be reduced from 40 to 23 hours for simple appendicitis and from 134 to 94 hours for perforated appendicitis with the implementation of a standardized pathway.¹² However, despite increasing evidence supporting protocols after appendectomies, a national survey of pediatric surgeons found that there was tremendous variation in how perforated appendicitis is managed and only 17% endorsed using a formal clinical pathway which highlights the challenges of implementing protocols.¹³

In this study, we identified the variability in the postoperative management of patients undergoing laparoscopic gastrostomy placement at a single institution with 15 pediatric surgeons. In an attempt to decrease other potential confounders we concentrated our analysis of the postoperative feeding regimen to outpatients that underwent elective laparoscopic gastrostomy placement and same-day admission. This resulted in a cohort of patients that were admitted to the floor 93% of the time and managed primarily by the surgical service 81% of the time. The assumption was made that this population of patients would have fewer factors that would potentially extend their hospitalization and the feeding regimen would more likely be dictated by the surgical service. Within this cohort of patients, our results still demonstrated variation in the timing of initiation of feeds, the method of initial feeds, and the choice of initial feeds. On further comparative analysis of the effects of the variation in the timing of initiation of feeds within this cohort, we demonstrated significant decrease in hospital length of stay and total direct variable hospital costs with early initiation of feeds without any difference in postoperative complications.

To further analyze the effects of variation in postoperative management after laparoscopic gastrostomy, we narrowed the cohort of outpatients that underwent elective laparoscopic gastrostomy placement and same-day admission to those that were also already NGT dependent for feeds. The assumption was that these patients were already tolerating formula or breast milk feeding through an NGT in the outpatient setting and were then scheduled for elective laparoscopic placement of a gastrostomy. In this cohort of patients, we found that 80% (n = 41) were still initially started on Pedialyte after laparoscopic placement of the gastrostomy. Comparative analysis of the effects of the variation in the choice of initial feeds within this cohort demonstrated a significant decrease in hospital length of stay and a trend toward less total direct variable hospital costs with initiation of formula or breast milk rather than Pedialyte, without any difference in postoperative complications.

These results add additional retrospective data to the existing literature that demonstrates the safety in initiating early postoperative feeds in children undergoing gastrostomy placement.^4,5,14,15 With continually rising healthcare costs and an increasing emphasis on delivering value-based care, these data support the potential opportunity to standardize the postoperative management after laparoscopic gastrostomy placement with an emphasis on early initiation of feeds.

Our study did have several limitations. This was a retrospective review of patients who underwent laparoscopic gastrostomy placement at a single institution. As such, data may be unavailable for review in the medical records and there may be inherent biases based on the practice patterns of providers. Additionally, we may not have consistently been able to distinguish from the retrospective nature of the review the true indications for the postoperative feeding regimen, whether it was a choice based on practice patterns or if there indeed was a clinical indication for the choice. A prospective study may provide additional valuable insight into the potential for standardized postoperative management after laparoscopic gastrostomy placement.

These data suggest that early initiation of feeds with formula or breast milk after laparoscopic gastrostomy placement may decrease hospital length of stay and direct variable hospital costs associated with the postoperative care of these patients. Based on these results, prospective evaluation of early initiation of postoperative feeds with formula or breast milk after laparoscopic gastrostomy placement through a standardized protocol is warranted.

Footnotes

Acknowledgment

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Disclosure Statement

No competing financial interests exist.

References

Goldberg

, Barton

, Xanthopoulos

, Stettler

, Liacouras

. A descriptive study of complications of gastrostomy tubes in children. J Pediatr Nurs, 2010; 25:72–80.

Cameron

, Graham

, Milliren

, et al. Quantifying the burden of interhospital cost variation in pediatric surgery: Implications for the prioritization of comparative effectiveness research. JAMA Pediatr, 2017; 171:e163926.

Gauderer

, Ponsky

, Izant

Jr . Gastrostomy without laparotomy: A percutaneous endoscopic technique. 1980. Nutrition, 1998; 14:736–738.

Bechtold

, Matteson

, Choudhary

, Puli

, Jiang

, Roy

. Early versus delayed feeding after placement of a percutaneous endoscopic gastrostomy: A meta-analysis. Am J Gastroenterol, 2008; 103:2919–2924.

Stein

, Schulte-Bockholt

, Sabin

, Keymling

. A randomized prospective trial of immediate vs. next-day feeding after percutaneous endoscopic gastrostomy in intensive care patients. Intensive Care Med, 2002; 28:1656–1660.

Fisher

, Wennberg

, Stukel

, Gottlieb

, Lucas

, Pinder

. The implications of regional variations in Medicare spending. Part 2: Health outcomes and satisfaction with care. Ann Intern Med, 2003; 138:288–298.

Fisher

, Wennberg

, Stukel

, Gottlieb

, Lucas

, Pinder

. The implications of regional variations in Medicare spending. Part 1: The content, quality, and accessibility of care. Ann Intern Med, 2003; 138:273–287.

Krell

, Girotti

, Dimick

. Extended length of stay after surgery: Complications, inefficient practice, or sick patients?. JAMA Surg, 2014; 149:815–820.

Brescia

, Tomassini

, Berardi

, et al. Development of an enhanced recovery after surgery (ERAS) protocol in laparoscopic colorectal surgery: Results of the first 120 consecutive cases from a university hospital. Updates Surg, 2017. [Epub ahead of print]; DOI: 10.1007/s13304-017-0432-1.

10.

Emil

, Taylor

, Ndiforchu

, Nguyen

. What are the true advantages of a pediatric appendicitis clinical pathway?. Am Surg, 2006; 72:885–889.

11.

Skarda

, Schall

, Rollins

, et al. A dynamic postoperative protocol provides efficient care for pediatric patients with non-ruptured appendicitis. J Pediatr Surg, 2015; 50:149–152.

12.

Skarda

, Schall

, Rollins

, et al. Response-based therapy for ruptured appendicitis reduces resource utilization. J Pediatr Surg, 2014; 49:1726–1729.

13.

Chen

, Botelho

, Cooper

, Hibberd

, Parsons

. Current practice patterns in the treatment of perforated appendicitis in children. J Am Coll Surg, 2003; 196:212–221.

14.

Brown

, Miedema

, King

, Marshall

. Safety of early feeding after percutaneous endoscopic gastrostomy. J Clin Gastroenterol, 1995; 21:330–331.

15.

McCarter

, Condon

, Aguilar

, Gibson

, Chen

. Randomized prospective trial of early versus delayed feeding after percutaneous endoscopic gastrostomy placement. Am J Gastroenterol, 1998; 93:419–421.