Intra-abdominal infections (IAIs) is a common infectious disease in ICU (1, 2) with a high morbidity and mortality (3, 4). Compared with other infections, IAIs are more likely associated with septic shock and acute kidney injury (1, 5). Patients in ICU often have various underlying diseases. There are many factors, such as age, nutritional status, chronic underlying diseases, sepsis, organ failure, surgical intervention or infection removal and antibiotic therapy, may all play an important role in identifying patients' prognosis and assessing severity of diseases timely (3, 6–11). With the development of surgery or drainage methods, and adjustment in bundle treatment of sepsis and antibiotic treatment strategies (12), the abovementioned factors might not be applicable to the present conditions to predict mortality rate. Treatment of ICU-IAIs is very complicated and most IAIs patients have poor prognosis. Early identification of specific clinical feature and potential risk factors that may improve their survival is vital to clinicians, however, the existing literature is rarely reported about it.

To this end, we conducted a retrospective clinical study and analyzed the data of ICU-IAIs patients during the past 8 years in ICU of Nanfang Hospital, Southern Medical University to figure out the clinical characteristics and explore the relevant risk factors of 28-day mortality. Furthermore, we compared the differences in mortality between the previous 4 years and the next 4 years.

We retrospectively analyzed the data of patients diagnosed with IAIs who were admitted to ICU of Nanfang Hospital, Southern Medical University from January 1, 2011 to December 31, 2018. We searched for relative information on the database of Nanfang Hospital from January 1, 2011 to December 31, 2018. The search was limited to “abdominal infection” or “peritonitis” or “intestinal fistula” or “anastomotic fistula” or “digestive tract perforation” and patients who have been admitted to ICU. The inclusion criteria for ICU-IAIs patients were listed as follows: (1) Age older than 18 years; (2) Meeting the criteria for IAI according to the 2005 International Sepsis Forum Consensus Conference (13). Patients who stayed in ICU for <24 h were excluded. By reviewing and collecting the relative data in the case system and tracking for the 28-day mortality, we figured out potential risk factors that would predict for the prognosis of the patients.

The research protocol was approved by the Ethical Committee of Nanfang Hospital, Southern Medical University (No.: NFEC-2019-162), and all research work was in compliance with the Declaration of Helsinki and the rules of Nanfang Hospital of Southern Medical University on clinical research.

The diagnostic criteria of IAIs are based on the definition in the 2005 International Sepsis Forum Consensus Conference (13). Those criteria are clinical manifestations (including abdominal pain and systemic inflammatory response syndromes such as fever, tachycardia, and shortness of breath) that match the signs and symptoms of IAI and the laboratory examination of peritoneal specimens meets the criteria for infection or IAI confirmed by surgery or microbiologic culture of peritoneal specimens. According to the Surgical Infection Society's (SIS) definition (14, 15), healthcare or hospital-associated IAIs (HA-IAIs) are defined as follows: patients who have been hospitalized for at least 48 h during the previous 90 days; patients residing in a skilled nursing or long-term care facility during the previous 30 days; patients who have received intravenous infusion therapy, wound care, or renal replacement therapy within the preceding 30 days; patients who have received several days of broad-spectrum antimicrobial therapy within the previous 90 days; patients who have post-operative infections; and patients known to have been colonized by or previously infected with a resistant pathogen. Patients not meeting those criteria were classified as CA-IAIs.

Underlying chronic diseases include the following conditions: chronic obstructive pulmonary disease; chronic heart failure (NYHA grade III-IV); metastatic cancer (metastasis confirmed by surgery or imaging methods); hematological malignancies (lymphoma, acute leukemia or multiple myeloma); liver cirrhosis; chronic renal failure (chronic renal insufficiency requires maintenance hemodialysis or combined with serum creatinine level > 300 μmol/L); immunosuppressive status (receiving corticosteroids within the past 6 months: Prednisolone equivalent ≥0.3 mg/kg per day for at least 1 month, severe malnutrition, congenital humoral or cellular immune deficiency); chemotherapy/radiotherapy status (received treatment within the past 6 months); human immunodeficiency virus (Human Immunodeficiency Virus, HIV) infection (HIV positive combined with clinical complications such as pneumocystis elvis pneumonia, Kaposi's sarcoma, lymphoma, tuberculosis or toxoplasmosis, etc.); diabetes.

Septic shock was regarded as patients with sepsis who needed vasopressor drugs to maintain the mean arterial pressure higher than 65 mmHg as well as the serum lactate level over 2 mmol/L after fluid resuscitation (16).

Initial antibiotic therapy (IAT) failure was defined as the initial antibiotic treatment cannot completely cover the organisms isolated from the intra-abdominal specimen.

There were 665 patients eligible for the present study; 112 patients did not meet the criteria of IAIs, 13 patients were peritonitis without IAIs and 18 were younger than 18 years old and 46 patients were hospitalized in ICU for <24 h. Finally, the remaining 476 patients were entered for analysis (Figure 1).

In our retrospective study over 8 years, 51.9% of patients had underlying chronic diseases. The most common chronic diseases were metastatic cancer, diabetes, and immunosuppressive status. The median scores of APACHE-II and SOFA are 15[11,20] and 5[3,7], respectively. The ICU mortality, 28-day mortality, and overall hospital mortality of IAIs were 12.4, 16.0, and 16.4%, respectively. The 28-day mortality of patients with septic shock was 30.9%. And ICU mortality, hospital mortality, and 28-day mortality are lower than the EPIC-II study (3) and the AbSeS study (17), we considered that demographic data, treatment strategy adjustments, and heterogeneity of ICU may be the reasons for the differences, and the severity of disease may be milder than the AbSeS study (17), as the SOFA score were lower in our study.

There have been a large number of literature studies on the risk factors of IAIs, and many factors have been confirmed to be related to the mortality of IAIs patients. These risk factors are usually included clinical and physiological characteristics, infection severity, surgical intervention, microbial factors, antibiotic treatment status, and progression of diseases (15). Risk factors for mortality of IAIs include age, underlying chronic diseases, loss of consciousness, septic shock, admission to ICU, hypoalbuminemia, peak level of serum PCT and lactate concentration, high MPI score, high APACHE-II score, and high SOFA score, diffuse peritonitis, enterococci and fungi or drug-resistant bacteria infection (2, 9, 10, 18–21). In addition, hematological malignancies, liver cirrhosis, mechanical ventilation, renal replacement therapy and high Simplified Acute Physiology Score (SAPS) II scores have been confirmed to be independent risk factors of mortality for patients with ICU-IAIs (3). Enterococcal isolation is associated with mortality in elderly patients in ICU-IAIs (22), high APACHE-II, kidney injury, cardiovascular insufficiency, low hematocrit, and low body temperature are related to the death of patients with fecal peritonitis (11). Multivariate COX analysis identified underlying chronic diseases, high SOFA score, low hematocrit, and receiving more fluids within 72 h in ICU as independent risk factors for 28-day mortality in the present study. That is not surprising, underlying chronic diseases (18, 23–26) and SOFA score (27, 28) have been confirmed by numerous studies as a risk factor for mortality in patients with IAIs.

The relationship between hematocrit and patient prognosis is rarely mentioned (11, 29). In our study, the hematocrit at the time of ICU admission was related to the mortality of the patient, but the reason for this result is not clear to us. In severe IAIs, there are many factors that lead to a decrease in hematocrit. Loss of blood due to infection or surgery, and fluid resuscitation are common factors. If a patient is combined with septic shock, massive fluid resuscitation can lead to dilutional anemia, so that oxygen-carrying capacity decreases, which in turn causes insufficient tissue oxygen supply and ultimately affect the patient's prognosis. Whether infusion of red blood cells improves the prognosis of patients with ICU-IAIs remains to be further studied.

Timely removal of the source of infection, usually by surgery, is an important measure for the management of IAIs, which may affect the prognosis of patients. In our study, we found that the surgical intervention rate of the survival group (89.8%) was higher than that of the non-survival group (68.4%). The reason may be that patients in non-survival group had more underlying chronic diseases and were so critically ill that it is not suitable for them to receive surgical treatment. This indirectly indicates that the removal of infection is a crucial measure for the management of IAIs, which may affect the prognosis of patients, but surgery is not the only method, since surgical intervention was not an independent risk factor for mortality in the statistical analysis. Surprisingly, we found that the timing of surgical intervention (the time from the establishment of a clinical diagnosis to the operation) has no significant impact on the mortality of patients. This is different from the results of several earlier studies (9, 30), and is similar with the recent GenOSept study (11). The timing of surgical intervention in the GenOSept study was the time from the onset of peritonitis symptoms to the operation. In addition to surgical operations, the measures to control infections also included percutaneous drainage. For patients who are critically ill and intolerant for major surgery, percutaneous drainage of abscesses or effusions under ultrasound positioning is preferable to surgical intervention (14). It is too arbitrary to judge whether surgical intervention or infection source control is delayed simply judged by the time between onset and surgery, and sometimes the surgeon needs to decide the timing of surgery according to the patient's basic condition and clinical manifestations.

Early goal-directed therapy (EGDT) is a 6-h resuscitation program for patients with sepsis and has been included in the guidelines of surviving sepsis campaign (12, 31). Subsequent three studies consistently showed that EGDT did not reduce the mortality of patients with sepsis when compared with the conventional treatment program, in contrast, patients received more fluid resuscitation and cost more (32–34). Recent studies have shown that positive fluid balance is an independent risk factor for increased mortality in patients with sepsis (35). Another study also showed that prognosis of patients is related with fluid balance at any phase in the treatment course (36). Similarly, our study found that the 28-day mortality of ICU-IAIs patients was significantly related to the positive fluid balance within 72-h after ICU admission, which was detected to be an independent risk factor for mortality. Fluid resuscitation will promote tissue edema while restoring tissue perfusion, which is not helpful to improve tissue oxygen metabolism. This may be the reason why excessive fluid resuscitation increases the mortality of patients. In addition, fluid resuscitation on the basis of abdominal infection and surgical trauma can promote the exudation of tissues in the abdominal cavity, thereby causing abdominal hypertension or even causing abdominal compartment syndrome in severe cases, finally leading to insufficient perfusion of abdominal organs, for example, gastrointestinal tract, kidney, liver, etc. the uncontrolled vicious circle may cause sequential organ dysfunction, and increase the risk of death. In the comparison of the first 4 years (2011–2014) and the last 4 years (2015–2018), we found that the patients in the two periods were similar in terms of disease severity and underlying disease, but ICU mortality, hospital mortality and 28-day mortality of patients in the later period (2015–2018) were significantly lower than those in the previous period (2011–2014). Specifically, the daily fluid balance for the first 3 days was a risk factor for the mortality of patients. The lower positive fluid balance was significantly related to the reduced mortality, which further confirmed that the positive fluid balance is a risk factor for increased mortality in patients with sepsis (35).

The standard antibiotic treatment of IAIs should cover gram-negative enterobacteria, aerobic streptococci and intestinal anaerobes. In critically ill patients, pathogens are more complicated and assessment is more difficult. Additional antibiotics may be needed to cover uncommon drug resistant or opportunistic pathogens (14, 37). When this principle is not followed or the initial antibiotic treatment cannot effectively cover the pathogens, the mortality rate of IAIs is significantly increased (26, 38–40). Our study found that failure of initial antibiotic therapy (IAT) is a risk factor associated with mortality, but we have not been able to confirm that failure of initial antibiotic therapy is an independent risk factor for mortality in ICU-IAIs.

This study has some limitations. First of all, this is a single-center retrospective observational study. The patients with IAIs in the study may not fully represent the situation in other regions. Secondly, ICU-IAIs refer to IAIs admitted to the ICU. There are both severe IAIs and IAIs combined with other critical illnesses. When referring to the conclusions of this study, we need to recognize that the actual situation of our included data. Third, it is difficult to distinguish whether the isolated microorganism is pathogenic or contaminated, as this is a retrospective study, so the definition of IAT failure may be biased. Fourth, fluid overload maybe to bring about the problem of intra-abdominal hypertension. During the study, we found 35 patients with high risk factors underwent intra-abdominal pressure monitoring. These patients were selected, and there was bias, and these small amounts of data are difficult to help us to determine the relationship between intra-abdominal pressure changes and fluid load. Finally, a considerable number of patients were transferred from other hospitals. The time onset of IAIs is sometimes difficult to define. Therefore, it is difficult to determine the time from the onset of symptoms to the operation. In this way, we did not analyze the relationship between surgery delay (defined as the time from the onset of symptoms to the surgery over 24 h) and 28-day mortality.

The 28-day mortality of ICU-IAIs was 16.0%. Underlying chronic diseases, high SOFA score, low hematocrit, and receiving more fluids within 72 h in ICU were independent risk factors for 28-day mortality. Comparing the first 4 years (2011–2014) and the last 4 years (2015–2018), the early use of antibiotics, the optimization of IAT strategies, and the restriction of positive fluid balance are related to the decline in mortality of IAIs in the last 4 years (2015–2018).

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by the Ethical Committee of Nanfang Hospital, Southern Medical University (No. NFEC-2019-162). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

XL analyzed the data and wrote the article. LL collected the data. SO helped to analyzed the data and reviewed the tables. ZZ reviewed and modified the article. ZC designed the study. All authors contributed to the article and approved the submitted version.

This work was funded by the Natural Science Foundation of China (81871604) and the Natural Science Foundation of Guangdong Province (2017A030313590 and 2016A030313561).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.