Persistent gastrocutaneous fistula: Factors affecting the need for closure

• Published in: Journal of pediatric surgery 2013-12, Vol.48 (12), p.2506-2510
• Main Authors: Wyrick, Deidre L, Bozeman, Andrew P, Smith, Samuel D, Jackson, Richard J, Maxson, R. Todd, Kelley, Karen R, Mathews, Donna L, Li, Jingyun, Swearingen, Christopher J, Dassinger, Melvin S
• Format: Article
• Language: English
• Subjects: Child; Child, Preschool; Cutaneous Fistula - epidemiology; Cutaneous Fistula - etiology; Cutaneous Fistula - surgery; Female; Follow-Up Studies; Fundoplication; Gastric Fistula - epidemiology; Gastric Fistula - etiology; Gastric Fistula - surgery; Gastrocutaneous fistula; Gastrostomy - methods; Gastrostomy tube; Humans; Incidence; Laparoscopy; Logistic Models; Male; Nissen fundoplication; Pediatrics; Postoperative Complications - epidemiology; Postoperative Complications - etiology; Postoperative Complications - surgery; Retrospective Studies; Risk Factors; ROC Curve; Surgery
• ISSN: 0022-3468; 1531-5037
• DOI: 10.1016/j.jpedsurg.2013.06.001

Abstract Purpose The occurrence of gastrocutaneous fistula (GCF) is a well-known complication after gastrostomy tube placement. We explore multiple factors to ascertain their impact on the rate of persistent GCF formation. Methods We retrospectively reviewed patient records for all gastrostomies (GT) constructed at our institution from 2007 to 2011. Association of GCF with method of placement, concomitant fundoplication, neurologic findings, duration of therapy, and demographics was evaluated using logistic regression. Results Nine hundred fifty patients had GTs placed, of which 148 patients had GTs removed and 47 (32%) of 148 required surgical closure secondary to persistent GCF. Laparoscopic and open procedures comprised 79 (53%) of 148 and 69 (47%) of 148, respectively. Seventeen (22%) patients in the laparoscopic group developed persistent GCF, compared to 30 (43%) in the open group ( P = 0.035, OR = 2.52). Seventy-one patients had concomitant Nissen fundoplication. Thirty-one (44%) developed GCF, compared to 16 (21%) without a Nissen ( P = 0.002, OR = 4.94). Patients with button in place for 303 days had persistent GCF incidence of 23%, compared to 45% at 540 days ( P < 0.001, OR = 3.51) and 50% at 850 days ( P = 0.011, OR = 4.51). Patients with device placed at 1.8 months of age were more likely to develop GCF compared to those with device placed at 8.9 months of age ( P = 0.017, OR = 2.35). Conclusion Open operations, concurrent Nissen and younger age at placement were all statistically significant factors causing persistent GCF.