Predictors for 30-day mortality and complications following radiologically inserted gastrostomies: a single centre, large cohort review

Abstract

Aim: To measure the 30-day mortality and delayed complication rates following radiologically inserted gastrostomy (RIG) placement and determine the predictive risk factors for 30-day mortality and delayed complications to aide pre-procedure informed consent.

Materials and methods: Retrospective analysis was undertaken of RIG insertions between July 2012 and August 2017 at a single tertiary centre, which included 373 patients (56% male; median age: 65 years, range: 19-92 years). Data were collected from electronic databases on patient demographics, RIG indication, all-cause mortality, complication rates, patient co-morbidities, and biochemical/haematological parameters. Multivariate analysis was performed to identify predictive risk factors for complications and mortality.

Results: The RIG procedural success rate was 97.9%. The overall 30-day mortality rate was 7.8%; associated with pre-procedural haemoglobin <130 g/l in men (p=0.030, odds ratio [OR] 23.38), white cell count >11×10⁹/l (p=0.001, OR 4.18), C-reactive protein >10 mg/l (p=0.003, OR 10.10) and international normalised ratio (INR) >1.2 (p=0.03, OR 4.63). Inpatient RIG referrals were associated with 10% 30-day mortality; compared to 1.1% for outpatients (p=0.028, OR 9.51). The incidence of immediate and delayed complications was 2.4% and 42.1%, respectively. Neuromuscular disease was associated with gastrostomy dislodgement (p=0.0001, OR 4.99) and fracture (p=0.0009, OR 13.45), cerebrovascular disease with gastrostomy dislodgement (p=0.009, OR 2.51), cardiovascular disease with sepsis 30-days post-RIG (p=0.02, OR 2.94), and diabetes mellitus with gastrostomy dislodgement (p=0.0001, OR 29.45), fracture (p=0.027, OR 5.63) and stoma site infections (p=0.0003, OR 7.16).

Conclusion: RIG 30-day mortality was significantly associated with inpatient procedures compared to outpatient, and a range of biochemical/haematological parameters that suggest infection pre-intervention. It is advised that the markers of infection and catabolism are investigated pre-intervention, which may reduce mortality and complication rates.