Onal

Sepsis is currently defined as an infection with organ dysfunction. Sepsis and septic shock are important global health problems causing significant mortality and morbidity rates.[1-3] Fast and proper management of septic shock is important in terms of patient survival. Proper management includes early initiation of the antibiotics after the microbial cultures/testing, as well as prompt infectious source control.[4] Infectious source control in intra-abdominal infections comprises the rapid diagnosis of a specific infectious site and its control via a surgical intervention such as drainage of an abscess, necrotic tissue debridement, or removal of an infected device. Several studies have underscored the significance of timely source control interventions. Bloos et al. conducted a comprehensive study demonstrating that delays in surgical source control are significantly associated with increased mortality rates in patients with sepsis. Their findings revealed that every hour of delay in surgical intervention was linked to a 1% rise in 28-day mortality, with notably higher mortality rates observed when surgical source control was delayed beyond 6 hours.[5] Similarly, Azuhata et al. investigated the impact of early initiation of surgical source control in patients with gastrointestinal perforation and septic shock. Their results emphasized the critical importance of prompt surgical intervention, with survival rates declining substantially as the time to initiation of surgery increased, reaching 0% for delays exceeding 6 hours.[6] These studies collectively highlight the imperative of promptly identifying and implementing appropriate source control measures to improve outcomes in patients with sepsis and septic shock. This aligns with recommendations from the 2021 Surviving Sepsis Campaign guideline, which emphasizes the need for expedited source control interventions to rapidly identify or exclude specific anatomical diagnoses of infection requiring emergent source control, implementing any necessary interventions as soon as medically and logistically practical.[7] Herein, we aimed to evaluate the epidemiology of septic shock associated with intra-abdominal infections (IAI) as well as associated mortality and efficacy of early source control (during the first 12 h or afterward) in a tertiary-care educational hospital.

This study was performed in an 1800+-bedded tertiary-care educational university hospital located in a city with a population of 4.3 million in 2017.[8]
Patients with septic shock (sepsis+hypotension+adrenergic agent+arterial lactate level of >2mg/dL)[2] and IAI and consulted by Infectious Diseases consultants during night shifts between the December 1^st, 2013, and October 15^th, 2022, in our center were recorded prospectively and analyzed retrospectively. The patients were evaluated following the first visit on days 3, 14, and 30. Septic shock definition was considered to be sepsis with hypotension requiring vasopressors to maintain a mean arterial blood pressure above 65 mm Hg despite adequate fluid resuscitation.[1] Systemic inflammatory response syndrome (SIRS), Quick Sequential Organ Failure Assessment (qSOFA), and SOFA (Sequential Organ Failure Assessment) scores with suspected infection were used for the definition of sepsis.[2,7] SIRS was defined by several clinical variables, including temperature >38°C or <35°C, heart rate >90 beats/min, respiratory rate >20 breaths/min or PCO₂ <32 mmHg, and WBC >12000 cells/mm³ or <4000 cells/mm³. After that, 2 or more points increase in SOFA score or qSOFA score was used for the definition of sepsis. For qSOFA, the following data were used: 1 point for each systolic arterial blood pressure ≤100 mm Hg, respiratory rate above 21 breaths per minute, or altered mental status.
Case assessment forms included data related to the demographical and clinical findings of septic shock patients, as well as microbiological culture results, surgical operation schedules, and mortality rates. Clinical assessment and radiological or surgical operation reports determined underlying intra-abdominal infectious foci.
The inclusion criteria were:
-    Age ≥ 18-year-old (adult patients were included)
-    Meeting the criteria of the septic shock described above as well as a final diagnosis of IAI, which was defined as a complicated IAI infection that extended beyond the hollow viscus of origin into the peritoneal space and was associated with either abscess formation or peritonitis, whereas, uncomplicated infection involved intramural inflammation of the gastrointestinal tract and had a substantial probability of progressing to complicated infection if not adequately treated.[9]
The exclusion criteria were:
-    The presence of an infectious source other than IAI
-    IAI patients from other hospitals are referred to our center without exact intervention time records, and patients are referred from our center to other centers due to the need for more available beds in intensive care units (ICU).
Microbiological analysis. Peripheral/catheter blood cultures were inoculated into aerobic and anaerobic culture bottles (BacT/ALERT, BioMérieux, Durham, USA), and an automated microbial detection system (BacT/ALERT 3D, BioMérieux, Durham, USA) was used. Signal-positive blood culture samples, together with other samples like urine, intra-abdominal fluid/pus, etc, were inoculated into 5% blood sheep agar and Eosin Methylene-blue Lactose Sucrose (EMB) agar (BioMerieux, France). MALDI-TOF mass spectrometry (VITEK MS, BioMérieux, France) was used for microbial identification. Antibiotic sensitivity tests were performed using the VITEK2 (BioMérieux, France) system and evaluated according to EUCAST criteria.[10] Carbapenem minimum inhibitory concentration (MIC) levels were determined by gradient tests (E test, BioMérieux, France). An antimicrobial regimen was considered to be appropriate when the isolated pathogens were found to be sensitive to the empirically started antibiotics according to the antibiotic susceptibility test results.
Ethics. The local Institutional Review Board approved the study (21-6.IT/63 on 25/06/2021).
Statistical analysis. SPSS 25.0 program (Statistical package for the social sciences) was used for the statistical analysis. A comparison of categorical values between the two groups was performed via the Chi-square test. Student t-test was performed for the numerical values of the independent groups.
Binary logistic regression was performed using the enter method. Day 30 mortality rate was the dependent variable, and the variables with a p<0.05 in univariate analysis were used as covariates, i.e., source control, source control within the first 12 hours, lactate>3 mg/dL. A p-value of less than 0.05 was considered significant.

General characteristics. There were a total of 390 patients (mean age 67.26 ±0.73 years and 44.6% female) fulfilling the study inclusion criteria. A total of 166 (42.6%) cases were aged ≥65-year-old. Mean leucocyte, CRP, procalcitonin, and lactate levels at the time of septic shock diagnosis were 18621 ± 1738/mm³, 143.05 ± 6.09 mg/L, 34.47 ± 4.81 µg/L, and 11.94 ± 0.95 mg/dL, respectively.
Comorbidities were recorded as 122 patients (31.3%) having malignancy, 103 patients (26.4%) having hypertension, 97 patients (24.9%) having diabetes mellitus, 39 patients (10%) having congestive heart failure, and 27 patients (6.9%) having chronic renal failure.
Table 1 summarizes the underlying IAI foci. The most common intra-abdominal problem was peritonitis without perforation (20%).

Table 1. Underlying intra-abdominal problems of the study cohort.

Etiological agents. Microbiological etiology was elucidated in 187 (47.9%) of 143 IAI cases (Table 1, 2). The most common pathogen was E.coli (73/187, 39%, Table 2), and multiple pathogens were isolated from 66 of 187 microbial culture-positive cases. Carbapenem-resistance in Gram-negative pathogens was recorded in 42 of 187 microbiologically-confirmed IAI cases (22.5%), and 22 of the 42 (52.4%) were carbapenem-resistant K.pneumoniae. Overall carbapenem-resistance among Gram-negative bacteria was 31.3% (42/134), and overall extended-spectrum beta-lactamase positivity among Gram-negative bacteria was 48.5% (65/134). Yeasts yielded in 15% of the etiologically confirmed cases, and they were the 4th most common etiological agent.

Table 2. Pathogens isolated from the study cohort.

Mortality and associated factors. Overall mortality was 42.5% (166/390) on day 3, while day 14 and day 30 mortality rates were 63.3% (247/390) and 71.3% (278/390), respectively. In females, 30-day mortality was significantly higher than in males (138/174 vs 140/216, p=0.002), but there was no significant difference in mortality between cases aged ≥ 65 years vs. others. (Table 3).

Table 3. Univariate analysis of independent variables on mortality (1 month).

In terms of laboratory findings during the first evaluation of the septic shock case, 30-day mortality among the cases with lactate levels>3 mg/dL was significantly higher (230/298-77% vs 48/92-52%; p<0.001). There was no significant difference in the mean levels of leucocytes (19550 ± 2331/mm³ vs. 16316 ± 1776/mm³, p=0.401), CRP levels (146.7 ± 7.28 mg/L vs. 133.9 ± 11.14 mg/L, p=0.342), and procalcitonin levels (33.17 ± 5.79 μg/L vs. 37.11 ± 8.77 μg/L, p=0.704) between patients who had 30-day mortality and those who survived.
The mean SOFA score was 10.33 ± 0.16 in 374 patients, while the median qSOFA score was 3 in 244 out of 390 patients. The mean SOFA score among cases with 30-day mortality was 10.93 ± 0.18 versus 8.84 ± 0.28 among survived cases (p<0.001). The mortality rate among the patients with qSOFA ≥2 was significantly higher than others (p=0.016). The overall univariate analysis for 30-day mortality is in Table 3.
Antimicrobial therapy/etiology versus mortality. 30-day mortality in cases with elucidated bacterial etiology vs. others did not change significantly (Table 3). Comparing 30-day mortality in cases with Gram-negative etiology versus all others, we observed rates of 98/134 vs. 180/256 (p=0.559). When we compared, 30-day mortality in cases with carbapenem-resistant Gram-negative etiology vs. all others was 86% (36/42 vs. 242/348, p=0.029), and in cases with ESBL-producing Gram-negative bacteria all others was 63.1% (41/65 vs. 237/325; p=0.109). However, 30-day mortality was significantly higher in those with carbapenem-resistant Gram-negative etiology vs. carbapenem-sensitive Gram-negative etiology.
In 36 of the 187 microbiological culture-positive patients, the antimicrobial regimen did not cover the isolated pathogens at the first visit, and 30-day mortality among them vs. others did not differ significantly (23/36-63.9% vs 113/151-74.8%, p=0.185).
Two hundred seven of 287 patients received antifungal-containing treatment, and there was no significant difference in mortality rates among them vs others (207/287-72% vs 71/103-69%, p=0.539). Furthermore, 42 cases had microbiological culture positivity for carbapenem-resistant Gram-negative pathogens, and 14 of them received a colistin-containing regimen with a 30-day mortality of 85.7% (12/14 colistin-receiving cases vs. 24/28 not colistin receiving cases, p=1.000).
When overall 390 cases were considered, 30-day mortality was not significantly different among cohorts in antifungal-containing regimens vs. others (207/287-72% vs 71/103-69%, p=0.539) and in colistin-containing regimens vs. others (39/52-75% vs 239/338-71%, p=0.524). There was no statistically significant difference in mortality among the patients with polymicrobial etiology versus culture-positive cases for single pathogen (50/66-76% vs 86/121-71%, p=0.492).
Impact of Source Control. Source control by surgical or percutaneous operation was performed in 123 of 390 cases (31.5%), and 30-day mortality was significantly lower in cases that were performed source control at any time during the septic shock period vs. others (65/123-%53 vs. 213/267-80%, p<0.001, Table 3). Additionally, in 44 out of 123 cases (35.7%), source control was performed within the first 12 hours. The 30-day mortality rate was significantly lower in this group compared to others (those in whom source control was performed after the first 12 hours + source control was not performed anytime during the septic shock period-24/44 or 55% vs 254/346 or 73%, p=0.009, refer to Table 3). However, 30-day mortality did not differ in those source controls performed during the first 12 h versus any other time during the septic shock period (24/44-55% vs. 41/79-57%, p=0.854).
In 77 of 123 cases (62.6%), invasive surgical operation/laparotomy was performed, while percutaneous source control was performed in 46 cases (37.4%). Mortality rates among the invasive surgical operation subgroup were statistically higher than the percutaneous source control group (52/77-67.5% vs 13/46-28.3%, p<0.001). The mean SOFA scores did not change significantly (9.80±0.52) in the invasive source controlled group vs. 10.11±0.71 in the percutaneous source controlled group (p=0.752).
30-day mortality was found to be similar among cases with intra-abdominal abscess vs. others (21/35-60% vs. 257/355-72.4%, p=0.122) and among cases with organ or intestinal perforation vs. others (33/40-82.5% vs 244/349-69.9%, p=0.096). Invasive surgical operations were performed more commonly among cases with organ or intestinal perforation in source controlled group (35/37-94.6% vs. 42/86-48.8%, p<0.001), and percutaneous source control was performed more commonly among cases with intra-abdominal abscess (19/22-86.4% vs 27/101-26.7%, p<0.001) in source-controlled group. Univariate analyses of independent variables on 30-day mortality are shown in Table 3.
Multivariate analysis. In logistic regression analysis, female gender (p<0.001, odds ratio (OR)=2.943, 95%CI=1.714-5.054), diabetes mellitus (p=0.014, OR=2.284, 95%CI=1.179-4.424), carbapenem-resistant Gram-negative etiology (p=0.011, OR=4.386, 95%CI=1.398-13.759), SOFA≥10 (p<0.001, OR=3.036, 95%CI=1.802-5.114), lactate >3 mg/dl (p<0.001, OR=2.764, 95%CI=1.562-4.891) and lack of source control (p=0.001, OR=2.796, 95% CI =1.523-5.133) were significantly associated with 30-day mortality (Table 4).

Table 4. Results of logistic regression analysis.

Herein, we analyzed the outcomes of 390 SS with IAI. The 30-day mortality rate was 71.3%. Source control by surgical or percutaneous operation was performed in 31.5%, and in 35.7% of them, source control was performed during the first 12 hours. On the other hand, female gender, diabetes mellitus, carbapenem-resistant Gram-negative etiology, SOFA≥10, lactate >3 mg/dl, and lack of source control were significantly associated with 30-day mortality in logistic regression analysis.
Source control in patients with septic shock is vital and recommended by the guidelines. However, the evidence regarding optimal timing for these interventional approaches is relatively limited.[5-7] Thus, this paper is one of the unique studies that emphasizes the impact of source control together with the optimal duration of surgical intervention, especially for the subgroup of septic shock patients with IAI.
In 2003, the Turkish Ministry of Finance, which is responsible for the payback of over 90% of the population's health expenditures, eased a new budget application instruction for regulating the usage of parenteral antibiotics inside and outside of the hospitals. The instruction took effect on March 1^st, 2003. According to this instruction, the payback of extended-spectrum antimicrobials (vancomycin, teicoplanin, meropenem, imipenem, antifungals, etc.) has been restricted without prior approval of infectious diseases specialist (IDS). Hence, all septic shock cases that were considered to be starting extended-spectrum antibiotics or consulted to Infectious Diseases Consultants during night shifts were included in the study.[11-13]
Martinez et al.[14] performed a multicenter observational study in Spain with 3663 severe sepsis or septic shock patients. Of these, 1234 cases had the abdominal site of infection source, with 788 (67.2%) patients requiring source control included. They reported that the crude ICU mortality rate was lower for those who underwent source control (21.2% vs 25.1%; p=0.010), but the source control after 12 hours was not found statistically significant in terms of mortality (27.6% vs 26.8%; p=0.789).[14] Kim et al. also performed a multicenter observational study with a total number of 2250 septic shock patients visiting 11 different Emergency Departments. They reported that 28-day mortality was significantly lower in patients who underwent source control (p<0.001). However, no significant association was noted between the duration of source control after 6 hours or 12 hours and mortality.
Additionally, out of the 2,250 patients, 46.6% met the criteria for Sepsis-3 septic shock, and 26.8% underwent source control; nevertheless, source control conducted after 6 or 12 hours showed no significant association with 28-day mortality in Sepsis-3 septic shock, with adjusted hazard ratios of 1.309 (95% CI: 0.612–2.797, p=0.488) and 1.344 (95% CI: 0.612–2.951, p=0.462).[15]
The guidelines of the Infectious Diseases Society of America (IDSA) about the diagnosis and management of complicated IAI in adults and children recommends (B-II) an appropriate source control procedure for nearly all patients with IAI.[16] This guideline also recommends (B-II) an urgent approach for hemodynamically stable patients without acute organ failure. Intervention may be delayed for as long as 24 hours with appropriate antimicrobial therapy plus close clinical monitoring. In contrast, it recommends (B-II) that patients with diffuse peritonitis should undergo an emergency surgical procedure as soon as possible.[16] Furthermore, the guideline recommends (B-II) that a percutaneous approach of abscesses and other well-localized fluid collections, if feasible, is preferable to surgical drainage.[16] The percutaneous approach of selected cases may cause fewer physiologic alterations with more acceptable mortality and morbidity rates and may eliminate or reduce the need for open techniques.[17,18] Marshall et al. also reviewed that source control represents a key component of success in the therapy of sepsis, which includes drainage of infected fluids, debridement of infected soft tissues, removal of infected devices or foreign bodies, and definite measures to correct anatomic derangement. Thus, they emphasized that appropriate source control should be part of the systematic checklist in sepsis with Grade E recommendation.[19] Hecker et al. reviewed that an IAI source could be detected in 66% of all surgical patients with sepsis, and the 28-day mortality rate increased from 26.7% to 42.9% in patients with inadequate initial source control.[20] Unlikely the literature, 30-day mortality was found to be as high as 71.3% (53% in the source-controlled group versus 80% in the non-source-controlled group) in our study. In comparison to a multicenter study in our country reporting 6.9% of patients meeting SEPSIS-III criteria for septic shock with a 75.9% mortality rate, our study found a similar 30-day mortality (71.3%) with a focus on Sepsis-3 septic shock patients, suggesting the potential influence of patient selection on the elevated mortality.[21]
Van de Groep et al. performed a retrospective study with a total number of 353 critically ill patients with IAI, and they concluded that the persistence of organ failure on day 14 was associated with inadequacy of source control while the clinical outcomes were similar between the patients having a surgical versus percutaneous initial approach to source containment.[22] In our study, the mortality was significantly lower in the group source control was performed (65/123 vs 213/267, p<0.001). These data suggest that the source of infection is one of the most important factors in septic shock patients in terms of management and prognosis. Aligned with existing literature, effective source control, encompassing interventions such as surgical procedures or percutaneous drainage, is essential for eradicating the infectious focus and decreasing mortality rates. We believe that source control plays a pivotal role in managing intraabdominal septic shock, highlighting the critical need to promptly address the infection's source for effective management. Despite its significance in the larger picture, source control executed within the initial 12 hours did not emerge as a protective factor against mortality. This nuanced result sheds light on the complexities of timing and effectiveness in source control, highlighting that achieving source control within a specific time frame may not necessarily confer a survival advantage in septic shock patients with intraabdominal infection. Further investigation and analysis are imperative to understand the intricacies of this relationship and to refine our approach to source control in septic shock scenarios.
Jung et al. reported that procalcitonin kinetics failed to predict the outcome in 101 consecutive cases that had perioperative abdominal infection with septic shock.[23] Wang et al. showed that the most accurate index was the SOFA score for predicting the outcome of patients with sepsis caused by abdominal cavity infection when the threshold value was 9.50. They found that sensitivity and specificity were 81.2% and 83.5%, respectively.[24] In our study, we also found that there was no statistically significant difference in the mean values of CRP, procalcitonin, and leucocyte levels among the mortal cases vs. others. On the other hand, SOFA score ≥10 (OR=3.036, 95%CI=1.802-5.114, p<0.001) was significantly associated with 30-day mortality in logistic regression analysis. Conversely, we believe that additional data are necessary to demonstrate the prognostic value of these biomarkers and scoring systems.
The results of the multicentered cohort study by Arvaniti et al.[25] involving 2337 critically ill elderly adults with intra-abdominal infection shed light on crucial risk factors impacting mortality such as late-onset hospital-acquired intra-abdominal infection, diffuse peritonitis, sepsis/septic shock, source control failure, and underlying conditions like liver disease, congestive heart failure, diabetes mellitus, and malnutrition. Our study also revealed distinctive correlations, underscoring that factors like female gender, diabetes mellitus, carbapenem-resistant Gram-negative etiology, SOFA score (≥10), elevated lactate levels (>3 mg/dl), and inadequate source control were notably associated with poorer outcomes. These observations stress the necessity for customized interventions targeting these specific risk elements to enhance patient outcomes in IAI cases.
Our study has several limitations, such as the mortality rates were recorded as all-cause mortality (autopsy could not been performed), repeated source control procedures could not be investigated, and timing of "12 hours" was considered to be a cut-off time for source control. However, to our knowledge, this is one of the unique and the largest studies performed solely in septic shock cases and concluding that surgery performed after 12 h may also contribute to survival and management of these groups of patients and that showed no difference between early source control vs. source control anytime during the septic shock episode. In addition, this is the first study from our country regarding the source control in septic shock patients with IAI.

Importance of Source Control in the Subgroup of Intra-Abdominal Infections for Septic Shock Patients: Analysis of 390 Cases