Characterization of Early and Late Deaths for Surgical Patients with Sepsis

Author(s):
Anahita Jalilvand; Whitney Kellett; Holly Baselice; Megan Ireland; Wendy Wahl; Jon Wisler

Background:

It is unclear whether timing of mortality following sepsis impacts causes of death (COD) in critically-ill surgical patients. The primary objective of this study was to characterize differences in COD by timing of mortality for patients admitted to the surgical ICU (SICU) with sepsis. 

Hypothesis:

We hypothesized that early mortalities would demonstrate a distinct pattern of COD compared to later deaths.

Methods:

A single-institution retrospective analysis of patients admitted to the SICU with sepsis was performed (2014-2019, n=1401). Inpatient mortalities (n=325) were reviewed to determine COD and classified by primary etiology (cardiac, respiratory, intra-abdominal not amendable to source control, multisystem organ failure with source control (MSOF), neurologic, musculoskeletal, or secondary infections). Early mortality was defined within 7 days of admission, and later deaths occurred beyond one week from admission. Baseline characteristics and COD were compared between mortality cohorts. A p value <0.05 was considered statistically significant.

Results:

Eighty-eight patients experienced an early death, accounting for 27% of all inpatient mortalities. Compared to later deaths, early mortalities had higher baseline serum lactates, comparable admission SOFA scores, but were less likely obese (20% vs 38%, p=0.001). The distribution of COD between groups was significantly different (p=0.002), and early deaths were more often from cardiac causes (12% vs 7%), or musculoskeletal (8% vs 2%) and intra-abdominal etiologies (36% vs 23%) not amenable to source control. A larger proportion of late mortalities were due to respiratory (22% vs 9%, p=0.002) or secondary infections (6% vs 1%, p=0.002).

Conclusions:

Most deaths following admission to the SICU occurred later in the hospitalization. These mortalities exhibited a distinct pattern of COD from early deaths, with a higher preponderance of respiratory and secondary infections in the later time point. Further studies should focus on the contribution of post-sepsis immunosuppression and chronic critical illness to late sepsis mortality.