Predicting reactive gastrostomy-tube placement after head and neck free flap reconstruction

The objectives of the study are: 1) to analyze the factors associated with post-operative placement of gastrostomy tube (G-tube) in head and neck reconstruction patients, and 2) to create a simple scoring system that can predict G-tube dependence.

Patients who underwent head and neck free tissue transfer at the author's institution from January 2015-May 2021 were identified. Data on patient characteristics, clinical outcomes, and G-tube placement were collected. Two regression models were developed, a comprehensive model including all significant clinical variables and a practical model including the most predictive variables.

We identified 525 head and neck free tissue transfer patients, of whom 63 (12%) required G-tube placement. The comprehensive model revealed a ROC curve AUC of 0.897 and identified the following significant variables: age (OR 1.04, 1.01-1.08 95% CI), CCI (OR 1.35, 1.14-1.62 95% CI), primary lesion site: oropharynx (OR 4.58, 1.68-12.3 95% CI), flap harvested: ALT (OR 3.69, 1.46-9.33, 95% CI), tracheostomy (OR 4.04, 1.41-13.4 95% OR), and bilateral neck dissection (OR 2.33, 0.99-5.39 95% CI). Based on these variables, a simple 9-point scoring system was created to assess the risk of post-operative G-tube dependence and could accurately make predictions in up to 92.3% of patients.

Risk factors for G-tube dependence were identified using the largest single-center database of head and neck reconstruction patients. This is the first report of the association of ALT free flaps and G-tube dependence. The proposed scoring system could be implemented prospectively to help determine prophylactic G-tube placement.