Back to Annual Meeting Program

Objective: Due to their relatively infrequent occurrence, it has been historically difficult to predict which patient with a diaphragmatic hernia (DH) will go on to either obstruct or strangulate and result in a surgical emergency. Given that patients with DH tend to be older and have multiple comorbidities, avoiding such an emergent situation would be beneficial. Our goal was to define potential comorbidities which could predict the likelihood of developing a complicated diaphragmatic hernia.
Methods: Using the NSQIP database, we identified all diaphragmatic hernias and grouped them by ICD9M code as either uncomplicated (553.3-without obstruction or gangrene) or complicated (551.3-with gangrene,552.3-with obstruction). Preoperative comorbidities, operative time, length of hospitalization and perioperative mortality and morbidity were compared between the two groups using chi-square and independent t-test as appropriate. A multivariate regression analysis was used to analyze potential factors contributing to obstruction or strangulation. Logistic regression was used to select correlates of 30 day mortality that were subsequently weighted and integrated into a scoring system based on the number of comorbidities.
Results: We identified 4778 patients, of which 4059 (85%) had an uncomplicated DH and 719 (15%) who had a complicated DH. The mean age for uncomplicated DH was statistically less than for a complicated DH (62.2±14.1 vs 66.9±15.2, p <.001). Independent predictors associated with a complicated DH included dyspnea at rest (AOR 2.9), partially or totally dependent functional status (AOR 4.4 and 7.1), CHF (AOR 4.3), history of MI (AOR 7.97) and >10% weight loss (AOR 1.82). Active smoking, alcohol consumption, dyspnea at exertion and use of steroids had no significant association. Risk stratification based on the number of preoperative comorbid factors demonstrated a step-wise increase in the rate of complicated DH: 12.1% (0-2 comorbidities), 21.5% (3-4 comorbidities), and 38.5% (≥5 comorbidities). Analysis of perioperative outcomes revealed that the mortality rate in the complicated DH group is much higher when compared to that of the uncomplicated DH group (5.1% vs 0.7%, P <.001). The same is true for the reoperation rate (6.3% vs 3.2%, p< .001) and length of stay (9.1+9.6 d, vs 4.1+6.7 d, p <.001).
Conclusion: This tool provides a simple, accurate and easily applicable method for predicting a complicated DH. Of note, history of a prior MI and dependent functional status most strongly predicted a complicated diaphragmatic hernia. Our findings suggest that if patients with uncomplicated DH are discovered and have a high score, elective repair should be sought expeditiously to avoid a life-threatening emergency. More studies are needed to further evaluate the timing of the progression from diagnosis of an uncomplicated DH to complication.