Despite the increased incidence of sepsis and septic shock, recent studies have consistently shown a decrease in sepsis mortality over time, owing to advances in medical technology [1, 13]. However, the overall in-hospital mortality rate due to sepsis was quite high at 29.0% in some recent studies in Korea [9, 14]. Similarly, our study showed that the overall in-hospital mortality rate with sepsis was 28.1%, and the mortality rate for sepsis caused by IAI was 23.3%. Moreover, the mortality rate of all patients with sepsis treated in the ICU and the mortality rate of patients treated for sepsis due to IAI in the ICU were even higher at 35.1% and 28.8%, respectively.
There have been many studies on tools for predicting the progress and prognosis of the disease, which are still underway [15–17]. One of the most important factors affecting the mortality and clinical outcomes in sepsis is comorbidity, and the CCI can quantify a patient's comorbidity. In this time of aging populations, the CCI of patients is increasing. The CCI is widely used as a tool to predict the mortality rate of patients with sepsis due to IAI and determine their prognosis in advance [14, 18, 19]. In our study, the mean CCI in the non-survivor group was higher than that in the survivor group, with significant differences (5.9 ± 3.0 vs. 5.1 ± 2.3, p = 0.011). In addition, CCI (p = 0.039, OR = 1.13 [1.00–1.26]) was an independent risk factor for mortality prediction in the univariate logistic regression analysis.
Multicenter research and efforts are being conducted worldwide to improve the clinical outcomes of patients with sepsis. For instance, the Surviving Sepsis Campaign released new guidelines for the treatment of sepsis and a new updated “Hour-1 bundle” for sepsis treatment [20, 21]. Immediate resuscitation, initial early screening, antibiotic treatment, and source control are imperative to reduce the social burden and improve patient outcomes [14]. In addition, initial lactate levels and initial SOFA scores can be used to predict clinical progress and prognosis of sepsis. A large number of studies have already shown that, the higher the initial lactate level and initial SOFA score, the higher the mortality rate in patients with sepsis [22, 23]. In our study, initial lactate level and SOFA score were also independent risk factors in predicting the mortality rate in patients with sepsis caused by IAI (p < 0.001).
Many studies have shown that patients with septic shock have higher mortality rates than those without septic shock [9, 24, 25]. There were more patients with septic shock in the non-survivor group, and the study of predicting risk factors can interpret that patients with septic shock have a 1.94 times higher risk of mortality than those without septic shock (p = 0.029).
In the treatment of patients with sepsis due to IAI, the usage of antibiotics and source control is very important. Adequate and swift antibiotic administration and coverage from the time of recognition of sepsis are vital [26–29]. Although no significant difference was noted, the non-survivor group showed a tendency of delayed antibiotic administration in our study compared with that in the survivor group (217.5 ± 422.8 min vs. 171.4 ± 238.8 min, p = 0.495). Sepsis caused by IAI due to biliary sepsis, intestinal perforation, postoperative leakage, and intra-abdominal abscess can be controlled using open laparotomy, percutaneous transhepatic biliary drainage, and percutaneous catheter drainage insertion. Ultimately, the prognosis of patients with sepsis depends on the source controls, and the faster the source control, the lower is the mortality rate [30, 31]. Our study showed a significantly higher number of survivors in the source control group (46.8% vs. 27.0%, p = 0.007). In addition, patients who did not perform source control have a 2.38-fold higher association with mortality than patients who did (p = 0.008, OR = 2.38 [1.26 ~ 4.51]).
In several studies on IAI, E. coli and K. pneumoniae were the most common gram-negative causative pathogens. In gram-positive pathogens, most of them were E. faecalis and E. faecium [5, 8]. In our study, the same pathogens were detected and the proportion was similar, as previously mentioned. The mortality rate was higher in patients whose causative pathogens were not identified. However, there was no significant difference in our study (p = 0.087). Additionally, some studies on IAI have shown that MDR is an independent risk factor for mortality [5, 8, 32]. However, the impact of MDR on mortality was not significant in our study (p = 0.136).
Organ dysfunction is a useful prognostic indicator for mortality in patients with sepsis, and the mortality rate increases significantly as the number of organ dysfunctions increases [33–35]. In one study by Takeshi et al., patients with three (23.5%) and four or more organ dysfunctions (38.9%) had over two and four times the ICU mortality rates, respectively, compared with that of single organ dysfunction (8.9%). In addition, the hazard ratios were 1.6, 2.0, and 2.7 in 2-, 3-, and 4 or more organ dysfunctions, respectively, showing an increasing trend as the number of organ dysfunctions increases [33]. In our study, the mortality rate was more than two-fold higher in patients with three organ dysfunctions (55.6%) and four or more organ dysfunctions (50.0%) than that of patients with single organ dysfunction (24.7%). The ORs were also 4.07 (p < 0.001) and 2.76 (p = 0.028) in patients with three and four or more organ dysfunctions, respectively. Similarly, the dysfunction of three or more organs in our study could be considered as an independent risk factor for mortality.
The impact of each organ dysfunction on mortality and the proportion of organ dysfunction occurring in patients with sepsis varies between the studies. In addition, in several studies of organ dysfunction in patients with sepsis, the most frequent cause of respiratory and cardiovascular system dysfunction is organ dysfunction [13, 33]. In contrast to the above-mentioned studies, respiration (36.5%), renal (36.1%), coagulation (34.2%), cardiovascular (25.6%), CNS (19.6%), and liver (18.3%) were observed in patients with sepsis caused by IAI in our study. And in patients with sepsis, respiratory and cardiovascular dysfunctions have a higher mortality rate than other organ dysfunctions [13, 33–35]. By contrast, our study showed the mortality rate of each organ dysfunction was highest in patients with CNS dysfunction (44.2%, p = 0.013), followed by coagulation (44.0%, p < 0.001), renal dysfunction (38.0%, p = 0.024), respiration (37.5%, p = 0.03), cardiovascular dysfunction (35.7%, p = 0.183), and liver dysfunction (27.5%, p = 0.845).
Microvascular thrombosis and ischemia are the most important processes in sepsis, causing tissue damage and multiple organ dysfunctions [36, 37]. In addition, the coagulation cascade may be abrogated owing to the aberrant expression of cytokines and tissue factors in response to systemic inflammation. As the coagulation system is deranged, sepsis-induced coagulopathy occurs, and manifestations of bleeding increase [36–38]. Coagulopathy in patients with sepsis is highly associated with mortality [39, 40]. Moreover, after adjusting for age, sex, CCI, and lactic acid level, coagulatory dysfunction (p < 0.001, OR = 2.99 [1.63–5.48]) had the greatest impact on mortality among the organ dysfunctions in our study.
This study has several limitations. First, this was a retrospective study based on medical records conducted in multiple institutions. Each hospital differed in its treatment policy, level, and system, including its facilities. As this study did not only focus on sepsis caused by IAI, the data may also be inadequate. To overcome these limitations, a prospective study of patients with sepsis caused by IAIs may be needed.