69 A comparison of outcomes following either fundoplication or gastro-jejunal tube insertion in children.
Sesi Hotonu, Bruce Jaffray
The Great North Children's Hospital, Newcastle upon Tyne, United Kingdom



Aims: Children with medically resistant gastro-oesophageal reflux disease (GORD) may receive fundoplication or gastro-jejunal (GJ) feeding.  We compared the effect of these interventions on survival, need for revision and subsequent hospital admission.

Methods: All children undergoing fundoplication or GJ since 2012 were identified. Outcomes of interest: mortality, revision, readmission, duration of stay. Predictor variables: choice of procedure, age, sex, neurological status (normal/ impaired), cardiac status (normal/abnormal), chronic respiratory disease (present/absent) and presence of tracheostomy.

Statistical analysis: continuous variables were compared with t-test, categorical variables with Χ2 test.  The effect of predictor variables was assessed using logistic regression, and survival compared using Kaplan-Meier graphs and the Log Rank test.  p <0.05 was significant.

Results: 173 children underwent fundoplication or GJ. Fundoplication: 102 (70 boys) mean age 4.6 years (range 0.04-17.9), GJ 71 (37 boys) mean age 6 (range 0.4-18.7). Median follow up: fundoplication 3 years (range 0.5-6.6), GJ 2 years (0.06 - 5.7).

16 children died. (9%). Fundoplication: 7 (6.9%) GJ: 9 (12.5%) (Χ2 = 1.6, p = 0.1). Mortality was strongly associated with neurological impairment: 2 (2%) of 88 neurologically normal children died while 14 (17%) of 85 neurologically impaired died (Χ2 = 10.3, p<0.001). Neurological impairment increased the odds of death by 8 (2-38), p = 0.006. Neither choice of operative procedure, nor any other predictor was significant (Figure).

Significantly more GJ than fundoplications were: i) readmitted (94% vs 56% p = 0.02), ii) readmitted more often (Mean number of readmissions, GJ: 11 fundoplication: 4.7 p = 0.02), iii) for longer (86d vs 36d p <0.001) iv) were revised (83% vs 25%) p <0.001.  The median number of revisions for GJ was 2.

Conclusions: There is no difference in mortality between interventions, which is related to neurology. GJ is associated with more revisions, re-admissions and longer hospital stay.