A Real-World Analysis of Steroid Prescribing in Surgical Patients With Septic Shock

Author(s):
Arnav Mahajan; Andrew Tran; Vanessa Ho; TianLi Du; Nimitt Patel; Patrick Maluso; Gary Jain; Justin Dvorak

Background:

Early administration of corticosteroids (CS) has been shown to have a survival advantage in septic shock (SS) patients on vasopressors in mixed medical-surgical intensive care units. However, surgeon use remains limited due to concerns for postoperative and infectious complications. There is a need to examine the implications of early CS utilization in surgical intensive care unit (SICU) patients.

Hypothesis:

1) Early CS (<24h, ECS) patients have a survival advantage over SICU patients in SS who received no steroids (NCS). 2)SICU patients in SS who receive ECS have a survival advantage over patients who received late CS (>24h, LCS). 3) ECS administration does not increase surgical complication rates.

Methods:

We identified patients who had a surgery followed by a diagnosis of SS and vasopressor use via TriNetX Analytics, a federated national aggregated platform of electronic health record data. We developed 1:1 propensity matched cohorts, balanced for demographics and APACHE score components. Two cohort sets were created: ECS vs. NCS and ECS vs. LCS. Our primary outcome was time to death, up to 1 year. Secondary outcomes include surgical complications (e.g. wound dehiscence), and infectious complications (e.g. surgical site infection). We report odds ratios and hazard ratios, with 95% CI and p-values.

Results:

We identified 6,494 matched patient pairs examining ECS vs. NCS. There was no significant difference in survival (HR: 1.02, 0.95-1.1, p=0.5). ECS patients had equivocal or improved postoperative and infectious outcomes compared to NCS patients (Table 1A). We identified 1,103 patient pairs comparing ECS vs. LCS. A survival benefit was seen in ECS patients (HR 0.55, 0.5-0.6, p<0.001). ECS only had a higher risk of wound dehiscence (OR: 1.8) (Table 1B).

Conclusions:

ECS demonstrated no survival benefit to NCS in SS patients. If the decision is made that the surgical patients in SS patient requires CS, our study suggests CS should be given within 24 hours to improve survival. ECS poses no statistically significant risk for postoperative or infectious complications against NCS and limited risk against LCS. We challenge notions that CS should be withheld due to concern for postoperative complication. Prospective studies are needed to elucidate survival benefit in surgical SS patients.