Objective: To establish the incidence and long-term prognosis of iatrogenic gas embolism.

Methods: This was a prospective inception cohort. We included all consecutive adults with proven iatrogenic gas embolism admitted to the sole referral academic hyperbaric center in Paris. Treatment was standardized as one hyperbaric session at 4 ATA for 15 min followed by two 45-min plateaus at 2.5 then 2 ATA. Inspired fraction of oxygen was set at 100% during the entire dive. Primary endpoint was 1-year mortality. All patients had evaluation by a neurologist, visual field tested by Goldman kinetic perimetry and brain MRI or CT scan at 6 months and 1 year.

Results: From January 1993 to August 2004, 125 of 4,727,496 hospitalizations had proven iatrogenic gas embolism. The crude mortality was 25/119 (21%) at 1 year. Cardiac arrest at time of accident and ICU admission, and SAPS II of 33 or more were independent prognostic factors of 1-year mortality (OR = 4.39, 95% CI 1.46-12.20 and OR = 6.30, 1.71-23.21, respectively). Among ICU survivors, independent predictors of 1-year mortality were age (OR = 1.07, 1.01-1.14), Babinski sign (OR = 6.58, 1.14-38.20) and acute kidney failure (OR = 8.09, 1.28-51.21). Focal motor deficits (OR = 12.78, 3.98-41.09) and Babinski sign (OR = 6.76, 2.24-20.33) on ICU admission, and duration of mechanical ventilation of 5 days or more (OR = 15.14, 2.92-78.52) were independent predictors of long-term sequels.

Conclusions: Gas embolism complicates 2.65 per 100,000 hospitalizations, and is associated with high mortality and morbidity. Babinski sign on ICU admission is associated with poor prognosis.