The swinging pendulum: a national perspective of nonoperative management in severe blunt liver injury.

BACKGROUND

Despite a shift toward nonoperative management (NOM) of blunt liver trauma, severe injuries continue to require operative management. Our objective was to examine current trends of NOM for severe blunt liver injury from a national perspective.

METHODS

Patients with blunt liver injury with Abbreviated Injury Scale (AIS) score of 4 or greater and no other major solid organ injury or pelvic fracture were identified in the National Trauma Data Bank 2002 to 2008. Attempted NOM was defined as no surgery in 6 hours or less. Failed NOM was defined as surgery in greater than 6 hours. Cox regression evaluated the association of NOM outcome with 30-day mortality after controlling for injury severity and center. Logistic regression was used to define independent predictors of failed NOM. Annual attempted and failed NOM rates were compared during the study period.

RESULTS

A total of 3,627 patients were identified with a median Injury Severity Score (ISS) of 29 (interquartile range, 20-38) and 20% mortality. Early operative management occurred in 20%, while initial NOM occurred in 73% of the patients. Of these, 93% had successful NOM, and 7% had failed NOM. Failed NOM was an independent predictor of mortality (hazard ratio, 1.7; 95% confidence interval, 1.1-2.6; p = 0.01). Increasing age, male sex, increasing ISS, decreasing Glasgow Coma Scale (GCS) score, hypotension, and hepatic angioembolization were independent predictors of failed NOM. The rate of attempted and failed NOM increased during the study period (p < 0.01).

CONCLUSION

NOM for isolated severe blunt liver injury is increasing nationally with similar increment in failure. Failed NOM was associated with higher mortality. Several predictors of failed NOM were identified including age, sex, ISS, GCS, and hypotension. These factors may allow for better patient selection and improved outcomes.

LEVEL OF EVIDENCE: Therapeutic study, level IV; prognostic/epidemiologic study, level III.