Objectives: To measure morbidity and mortality rates following insertion of gastrostomy tubes in head and neck cancer patients. To determine evidence for any relationship between gastrostomy insertion technique and complication rates.
Design: A prospective cohort study and qualitative systematic review.
Setting: Multi-cancer networks in the South West of England, Hampshire and the Isle of White.
Participants: One hundred and seventy-two patients with head and neck cancer undergoing gastrostomy tube insertion between 2004 and 2005. Percutaneous endoscopic gastrostomy (PEG) was performed in 121 patients. Fifty-one patients had radiologically inserted gastrostomy (RIG). Twenty-seven studies reporting outcomes following 2353 gastrostomy procedures for head and neck cancer.
Main outcome measures: Post-procedure mortality, major and minor complications.
Results: In the present series, mortality rates were 1.0% (1/121) for PEG and 3.9% (2/51) for RIG. Overall major complication rates following PEG and RIG were 3.3% (4/121) and 15.6% (9/51) respectively. In our systematic review and meta-analysis of 2379 head and neck cancer patients, we observed fatality rates of 2.2% (95% CI 0.014-0.034) following PEG and 1.8% (95% CI 0.010-0.032) following RIG. Furthermore, major complication rates following PEG were 7.4% (95% CI 5.9-9.3%) and 8.9% (95% CI 7.0-11.2%) after RIG.
Conclusions: Procedure related mortality rates following gastrostomy in head and neck cancer patients are higher than those in mixed patient populations. Major complication rates following RIG in head and neck cancer patients are greater than those following PEG. Major complications following PEG in patients with head and neck cancer appear no worse than in mixed pathology groups. We have identified that RIG is associated with increased morbidity and mortality in patients who are ineligible for PEG. The serious nature of the complications associated with gastrostomy particularly in patients with head and neck cancer requires careful consideration by the referring physician.