Severe postoperative infections following surgery for digestive system diseases

Abstract

BACKGROUND

Postoperative infection following digestive system surgery is a critical complication that leads to high rates of morbidity and mortality. Current studies on the incidence and, more importantly, the risk and microbiological factors related to these infections in patients undergoing surgery on the digestive system are lacking. Identifying such factors can aid in detecting patients who may be at post-operative infection risk and enable the modification of perioperative practices.

AIM

To describe the incidence and risk factors of severe postoperative infections, pathogens’ microbiological profile, and infection effects on clinical outcomes.

METHODS

This study is a retrospective cohort of patients who underwent digestive system surgery and attended the chosen hospital. This study enrolled 156 eligible patients. Severe postoperative infection was defined as an infectious complication that occurred within 30 days of surgery and was classified as Clavien-Dindo grade III or higher. The study included relevant demographic, clinical, and laboratory data as well as intraoperative and perioperative microbiological data. Risk factors were examined using univariate and multivariate logistic regression.

RESULTS

Among the 156 patients, 38 developed serious postoperative infections. The most common infections were intra-abdominal (52.6%), followed by surgical site (28.9%) and pulmonary infections (18.4%). Infections were dominated by Gram-negative bacteria (65.9%). Of these, Escherichia coli was responsible for 27.3% and Klebsiella pneumoniae for 18.2%. The following parameters were associated with an increased odds of infection: Diabetes and metabolic syndrome, preoperative serum albumin level < 35 g/L, surgical duration ≥ 240 minutes, and open surgical approach. Patients who developed supervening infections had longer inpatient stays (24.3 days vs 11.6 days, P < 0.001), and those who developed severe infections had a higher 30-day mortality rate (10.5% vs 0.8%, P = 0.004).

CONCLUSION

Digestive surgery leads to severe postoperative infection in a considerable number of patients and causes high morbidity and mortality. These findings support perioperative optimization and locally guided antimicrobial management.

Key Words: Postoperative infection; Digestive system surgery; Surgical site infection; Risk factors; Retrospective cohort study

Core Tip: This paper investigates the development of severe infections after surgery of the digestive system by analyzing factors such as patient age, comorbidities, and type of surgery. The study argues that the risk of developing infections after surgery can be mitigated by the timely recognition of symptoms and the initiation of appropriate antibiotic therapy. The study finds that the risk is greater in the case of patients with less immunity and patients who have undergone more complex surgeries. the individual patient and the complex nature of the surgery performed to correct a gastrointestinal problem should be the basis of treatment in order to achieve better results and reduce the incidence of post-operative infections.

INTRODUCTION

Surgical site infections (SSIs) and infections occurring after surgery comprise the majority of infections acquired in the course of patient care. SSIs account for approximately 20% of the total infections acquired in patients during care[1]. The infection rates of gastrointestinal tract surgeries are significantly higher than those of other branches of surgery. This difference is due to the contamination risk of the alimentary tract, referring to the progressive and cumulative colonization of the digestive system and colon by pathogenic microbes[2]. Some of the latest data in the multi-center surveillance of post-multiple surgeries show that the rate of SSIs in gastrointestinal tract surgery is between 5% and 30% based on the type of surgery, contamination level of the surgical wound, and other patient-related risk factors[3]. Among the various types of post-surgical infections, some of the worst infections that create a major clinical dilemma are those that require immediate surgery and vital organ support or that cause a major disutility to the patient. Moreover, these infections significantly contribute to the extension of patient hospitalization, cost of health care, and death cases during or after surgery[4].

In the last few decades, the microbiology related to post-operative infections after digestive surgery has evolved. The majority of abdominal SSIs are still caused by Gram-negative bacteria, among which Escherichia coli and Klebsiella pneumoniae are classic examples. The growing concern of multidrug-resistant (MDR) organisms is currently affecting the prognosis of patients related to this therapy. Empirical therapy has become more challenging because of these organisms[5]. Enterobacterales that produce the extended-spectrum beta-lactamases (ESBL) are increasingly recognized in most post-operative abdominal infections. Studies have shown that the colonization of these organisms also more than doubles the SSI rate in patients who receive the standard surgical antibiotic prophylaxis[6]. The resistance of these organisms highlights the need for collecting local data regarding the organisms and their susceptibilities to assist in developing locally appropriate treatment options.

SSIs after abdominal surgery are related to several risk factors, as part of the patient and procedure. A more recent systematic review by Marzoug et al[7] identified the following independent risk factors: Being a male, having a high body mass index (BMI), diabetes, smoking, American Society of Anesthesiologists (ASA) classification > 2, low albumin, longer duration of surgery, being an emergency surgery, and an open surgical approach. A recent nationwide study with 17353 patients who underwent gastrointestinal surgery confirmed that the contamination class of the procedure, procedure duration, and type of surgical approach were the main predictors of the development of SSI[8]. The Centers for Disease Control and Prevention (CDC) and World Health Organization have published standards on SSI prevention; however, several researchers have been unable to meet the “best practice” standards by following the available evidence[9].

Although numerous studies have focused on assessing the frequency and preventative measures of SSIs, the occurrences of Clavien-Dindo grade III and higher infections following surgical procedures involving the gastrointestinal system remain understudied. Severe infections occur along a spectrum and therefore require a specialized focus. These infections are indicative of a much worse clinical scenario and are associated with high rates of admission to the intensive care unit (ICU), reoperations, organ failure, and death[10]. Moreover, the factor profiles of severe infections may differ from those of less severe infection-related complications and therefore may require different approaches for prevention and management. This study aimed to conduct a retrospective study to assess the occurrence, risk factors, characteristics of pathogens, and the clinical course of patients who developed severe postoperative infections after undergoing surgical interventions involving the gastrointestinal system at a tertiary academic hospital.

MATERIALS AND METHODS

Study design and setting

This study was a single-center, retrospective cohort study conducted at the Ninth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. Ethical approval was obtained from the hospital’s Institutional Review Board, and the study was conducted in compliance with the Helsinki Declaration. Because the study was retrospective and utilized de-identified data, the informed consent requirement was waived by the Institutional Review Board. The study was reported in compliance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies.

Patients undergoing elective or emergency surgery for digestive system diseases from January 1, 2022 to December 31, 2024, were eligible for inclusion in the study. For final inclusion, patients had to meet the following criteria: They were 18 years or older, underwent major surgery of the gastrointestinal tract, hepatobiliary system, or pancreas, and had complete medical records, including an operative note and medical records for the 30-day postoperative period. Exclusion criteria were as follows: A patient underwent a purely diagnostic or non-definitive surgery (e.g., an exploratory laparotomy), had a pre-existing SSI as evidenced by positive cultures or clinical evidence of a systemic infection, had a non-infection-related surgery (e.g., intraoperative cardiac arrest or hemorrhage), and had incomplete medical records. Following post-study criteria, 156 patients were included in the final analysis.

Data collection

Two trained research assistants extracted data from the institution’s electronic medical record in compliance with a data collection form and standardized protocol. Variables of interest were grouped in the following categories. Demographic and baseline characteristics were captured with age, sex, BMI, smoking status, and alcohol use. The following comorbid conditions were included in the analysis: Diabetes mellitus, hypertension, chronic obstructive pulmonary disease, chronic liver disease, chronic kidney disease, and previous abdominal surgery. Preoperative clinical laboratory tests included complete blood count (white blood cell, hemoglobin, and platelets) and serum tests for albumin, total protein, prealbumin, C-reactive protein, procalcitonin, fasting blood glucose, glycosylated hemoglobin (HbA1c), total bilirubin, alanine aminotransferase, and creatinine, as well as coagulation tests. Nutritional risk was evaluated using the Nutritional Risk Screening (NRS) 2002, and the overall physical status was evaluated using the ASA classification.

The variables accounted for in the study included the primary surgical diagnosis, type of surgical procedure, surgical approach (open vs laparoscopic or minimally invasive), classification of the surgical wound based on the criteria established by the CDC (i.e., clean, clean-contaminated, contaminated, or dirty), procedural time, estimated intraoperative blood loss, necessity for intraoperative blood transfusion, placement of surgical drains, and surgical antibiotic prophylaxis, including the specific antibiotics, perioperative timing, and duration of prophylaxis. The variables of the infectious surgical complications of interest included the infection site, type of infection, surgical reintervention or percutaneous drainage, antimicrobial susceptibility testing, need for admission to the ICU, necessity for mechanical ventilation, total length of stay, and postoperative mortality.

Definitions and outcome measures

This study aimed to measure severe postoperative infections, which were captured as complications of infections that occurred within 30 days of surgery and met or exceeded a Clavien-Dindo grade III complication. All complications after surgery were graded using the Clavien-Dindo classification system[11]. Grade III complications are those requiring intervention by surgery, endoscopy, or radiology and are subdivided into grade IIIa, which does not require general anesthesia, and grade IIIb, which does require general anesthesia. Grade IV complications are life-threatening, requiring ICU intervention, and are subdivided into grade IVa (single organ dysfunction) and grade IVb (multiorgan dysfunction). Grade V represents death. Infectious complications of the surgical site are categorized as incisional and space SSIs by the CDC definitions[12]. Intra-abdominal infections include peritoneal cavity infections that are culture-positive, including the presence of intra-abdominal abscesses, infected anastomotic leaks and peritonitis, or infected peritoneal collections. Diagnoses of pulmonary infections were made using standard criteria and required the clinician to identify the infection, assess the radiographs, and obtain laboratory tests to confirm the diagnosis. Systemic infections were identified from blood cultures and the presence of infection in the bloodstream.

All specimens (wound swabs, intra-abdominal fluid, blood, sputum, and drain fluid cultures) were analyzed as per the standard microbiological practice. Specimens were cultured in MacConkey agar, blood, and chocolate agar media and incubated in both aerobic and anaerobic conditions for 24 hours to 48 hours. Bacterial cultures were identified using the MALDI-TOF MS method (Bruker Daltonics, Bremen, Germany). Antimicrobial susceptibility was determined using the broth microdilution technique on the VITEK 2 automated system (bioMérieux, Marcy l’Etoile, France) and was interpreted based on the Clinical and Laboratory Standards Institute breakpoints. MDR was defined as the acquired non-susceptibility of at least one agent in three or more antimicrobial classes. The production of ESBL was determined using the double-disc synergy test and confirmed using the combination disc test. Anaerobic culture results were extracted when available and interpreted simultaneously with the aerobic culture findings.

Statistical analysis

The wording of variables that are continuous (several possible values choosing from different ordered categories that are continuous e.g., time, speed, and temperature) or categorical (categorical variables that can assume a limited and usually fixed number of values) is standardized by their means and SD for normally distributed data, medians and interquartile ranges for non-normally distributed data, and n (%) (categorical variables). The analysis performed for continuous variables was the independent samples t-test or Mann-Whitney U test, and the χ² test or Fisher’s exact test were performed, as appropriate, for categorical variables to compare the severe infection and non-infection groups. Factors that were considered in multivariate analysis were those that scored a P < 0.1 in a preliminary univariate analysis. In these cases, a stepwise forward multivariate logistic regression analysis was performed to determine the effect of variable selection on the outcome of interest (i.e., severe postoperative infection). The adjusted odds ratios (ORs) with their 95% confidence intervals (CIs) were used to report the results of the multivariate analyses. The ability of the final model to discriminate between target states was determined by the area under the receiver operating characteristic curve, and the Hosmer-Lemeshow test was used to determine the calibration of the final model in a specified target state (i.e., severe postoperative infection). The threshold for determining statistical significance was set to a two-tailed P < 0.05. The statistical analyses were performed using the SPSS version 26.0 software (IBM Corporation, Armonk, NY, United States) and the R version 4.3.1 software (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Baseline characteristics

In total, 183 patients undergoing surgical treatment for digestive system diseases were screened. Subsequently, 12 patients with incomplete medical records, 8 with pre-existing active infections, 4 who died within 48 hours from non-infectious reasons, and 3 who had only diagnostic procedures were excluded; consequently, the final analysis included 156 patients. Of these, 38 patients (24.4%) had severe postoperative infections (the infection group), and the remaining 118 (75.6%) had no severe infections (the non-infection group). Table 1 presents the demographic and clinical characteristics of both groups.

Table 1 Baseline demographic and clinical characteristics of patients, n (%)/mean ± SD.

Variable	Infection group (n = 38)	Non-infection group (n = 118)	t/χ2	P value
Age (years)	63.2 ± 11.8	57.2 ± 12.1	2.70	0.008
Male sex	25 (65.8)	67 (56.8)	0.94	0.332
BMI (kg/m2)	24.8 ± 3.9	23.6 ± 3.5	1.78	0.077
Smoking	18 (47.4)	38 (32.2)	2.84	0.092
Diabetes mellitus	17 (44.7)	20 (16.9)	12.06	< 0.001
Hypertension	16 (42.1)	35 (29.7)	1.98	0.160
COPD	5 (13.2)	8 (6.8)	1.48	0.224
Chronic liver disease	8 (21.1)	14 (11.9)	1.98	0.160
Previous abdominal surgery	12 (31.6)	24 (20.3)	2.03	0.154
ASA classification
I-II	18 (47.4)	82 (69.5)	6.08	0.014
III-IV	20 (52.6)	36 (30.5)
NRS 2002 ≥ 3	22 (57.9)	40 (33.9)	6.85	0.009
Albumin (g/L)	32.8 ± 4.6	37.2 ± 4.1	5.62	< 0.001
Hemoglobin (g/L)	112.4 ± 18.6	124.8 ± 16.2	3.94	< 0.001
WBC (× 109/L)	7.8 ± 2.9	6.4 ± 2.1	3.17	0.002
CRP (mg/L, median, IQR)	12.4 (5.8-24.6)	4.2 (1.8-10.5)	3.861	< 0.001
HbA1c (%)	7.2 ± 1.8	5.9 ± 1.1	5.04	< 0.001
Fasting glucose (mmol/L)	7.6 ± 2.4	5.8 ± 1.6	5.20	< 0.001

¹Mann-Whitney U test.

BMI: Body mass index; COPD: Chronic obstructive pulmonary disease; ASA: American Society of Anesthesiologists; NRS: Nutritional Risk Screening; WBC: White blood cell; CRP: C-reactive protein; HbA1c: Glycosylated hemoglobin; IQR: Interquartile range.

The mean age of the population was 58.7 ± 12.4 years, among which 92 (59.0%) were male. The most frequent primary diagnoses were colorectal cancer (35.3%), gastric cancer (24.4%), and hepatocellular carcinoma (16.0%), as well as pancreatic diseases (12.2%) and other diseases of the digestive system (12.2%). Patients in the infection group were significantly older (63.2 ± 11.8 years vs 57.2 ± 12.1 years, P = 0.008), had a higher incidence of diabetes (44.7% vs 16.9%, P < 0.001), lower preoperative albumin levels (32.8 ± 4.6 g/L vs 37.2 ± 4.1 g/L, P < 0.001), and higher ASA classifications (P = 0.003). The baseline characteristics of both groups are presented in Table 1.

Operative characteristics

Table 2 presents the operative characteristics of both groups. The most frequently performed surgeries included colorectal resection (35.3%), gastrectomy (24.4%), hepatectomy (16.0%), and pancreaticoduodenectomy (8.3%). Patients in the infection group had longer operative times (285.6 ± 98.4 minutes vs 198.3 ± 78.6 minutes, P < 0.001), higher estimated blood loss (486.2 ± 312.5 mL vs 268.4 ± 198.7 mL, P < 0.001), and a higher percentage of intraoperative blood transfusions (42.1% vs 18.6%, P = 0.003). An open surgical approach was more common in the infection group (71.1% vs 46.6%, P = 0.008), and higher wound classifications (contaminated/dirty) were more common in infected patients (39.5% vs 16.1%, P = 0.002).

Table 2 Operative characteristics of patients with and without severe postoperative infections, n (%)/mean ± SD.

Variable	Infection group (n = 38)	Non-infection group (n = 118)	t/χ2	P value
Primary diagnosis			3.42	0.490
Colorectal cancer	15 (39.5)	40 (33.9)
Gastric cancer	10 (26.3)	28 (23.7)
Hepatocellular carcinoma	6 (15.8)	19 (16.1)
Pancreatic diseases	5 (13.2)	14 (11.9)
Other	2 (5.3)	17 (14.4)
Surgical approach			7.01	0.008
Open	27 (71.1)	55 (46.6)
Laparoscopic/MIS	11 (28.9)	63 (53.4)
Wound classification			9.27	0.002
Clean-contaminated	23 (60.5)	99 (83.9)
Contaminated/dirty	15 (39.5)	19 (16.1)
Emergency surgery	8 (21.1)	12 (10.2)	2.90	0.089
Operative duration (minute)	285.6 ± 98.4	198.3 ± 78.6	5.44	< 0.001
≥ 240 minutes	24 (63.2)	32 (27.1)	16.04	< 0.001
Blood loss (mL)	486.2 ± 312.5	268.4 ± 198.7	4.82	< 0.001
Blood transfusion	16 (42.1)	22 (18.6)	8.45	0.004
Drainage placement	30 (78.9)	72 (61.0)	3.98	0.046

MIS: Minimally invasive surgery.

The collected antimicrobial prophylaxis variables were reviewed descriptively; however, the retrospective records did not consistently document all elements required to judge appropriateness or intraoperative redosing; therefore, prophylaxis appropriateness was not entered into the prediction model.

Characteristics of severe postoperative infections

For the 38 patients who developed severe postoperative infections, the median time of infection development was 5 days (interquartile range: 3-8 days) after surgery. The distribution of infection types and Clavien-Dindo grades is presented in Table 3. The most common type of infection was intra-abdominal infections (n = 20, 52.6%), followed by SSIs (n = 11, 28.9%), and pulmonary infections (n = 7, 18.4%). In the 20 cases of intra-abdominal infections, we noted 12 intra-abdominal abscesses, 5 anastomotic leaks with secondary peritonitis, and 3 intra-abdominal fluid collections. In grading the severity of the infections, 16 patients (42.1%) were classified as having Clavien-Dindo grade IIIa infection, 12 as grade IIIb, 7 as grade IVa, 2 as grade IVb, and 1 as grade V. Intra-blood infections (sepsis) developed in 6 patients (15.8%), all of whom also had an intra-abdominal infection and/or a SSI.

Table 3 Distribution of severe postoperative infection sites and Clavien-Dindo grades (n = 38).

Infection site	n (%)	Grade IIIa/b	Grade IVa/b	Grade V
Intra-abdominal	20 (52.6)	14	5	1
Abscess	12 (31.6)	9	3	0
Anastomotic leak	5 (13.2)	2	2	1
Infected collection	3 (7.9)	3	0	0
Surgical site infection	11 (28.9)	9	2	0
Deep incisional	7 (18.4)	6	1	0
Organ/space	4 (10.5)	3	1	0
Pulmonary infection	7 (18.4)	5	2	0
Secondary bacteremia	6 (15.8)	-	-	-

Microbiological results

Of the 38 patients with severe infection, 35 were isolated for 44 pathogenic organisms (culture positive rate: 92.1%), with 9 comprising polymicrobial cultures. The microbiological analysis of these patients is presented in Table 4. Gram-negative organisms were the predominant pathogens, with 29 isolates (65.9%), followed by Gram-positive organisms, with 11 isolates (25.0%), and fungi, with 4 isolates (9.1%). Among Gram-negative organisms, the isolation of Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Enterobacter cloacae accounted for 12 (27.3%), 8 (18.2%), 5 (11.4%), and 4 (9.1%) isolates, respectively. The predominant isolates among Gram-positive organisms were Enterococcus faecalis, with 5 isolates (11.4%), and Staphylococcus aureus, with 4 isolates (9.1%), and the fungal isolates were all Candida species. No obligate anaerobic organisms were reported among the positive cultures, although routine culture procedures included anaerobic incubation. Of the Escherichia coli isolates, 5 (41.7%) and 2 (25.0%) Klebsiella pneumoniae were confirmed to produce ESBL. Methicillin-resistant Staphylococcus aureus was isolated in 1 of the 4 Staphylococcus aureus isolates.

Table 4 Distribution of pathogenic organisms isolated from severe postoperative infections (n = 44 isolates), n (%).

Pathogen	Isolates	%	ESBL+	MDR
Gram-negative bacteria	29	65.9
Escherichia coli	12	27.3	5 (41.7)	4 (33.3)
Klebsiella pneumoniae	8	18.2	2 (25.0)	2 (25.0)
Pseudomonas aeruginosa	5	11.4	-	2 (40.0)
Enterobacter cloacae	4	9.1	-	1 (25.0)
Gram-positive bacteria	11	25.0
Enterococcus faecalis	5	11.4	-	1 (20.0)
Staphylococcus aureus	4	9.1	-	1 (25.0)1
Other Gram-positive	2	4.5	-	0
Fungi	4	9.1
Candida albicans	3	6.8	-	-
Candida glabrata	1	2.3	-	-

¹Methicillin-resistant Staphylococcus aureus.

Percentages in the main % column are calculated using all 44 isolates as the denominator; multidrug-resistant percentages in parentheses are calculated using the number of isolates of each listed pathogen as the denominator. ESBL: Extended-spectrum beta-lactamase; MDR: Multidrug-resistant.

Risk factor analysis

In the univariate analysis, age, BMI, smoking, diabetes mellitus, ASA classification, NRS 2002 score, preoperative albumin, hemoglobin, white blood cell count, C-reactive protein, HbA1c, fasting glucose, surgical approach, wound classification, emergency surgery, operative duration, estimated blood loss, intraoperative blood transfusion, and drainage placement were significantly within the risk factor (P < 0.1) for severe postoperative infection. The related variables were incorporated into the multivariate logistic regression model. The multivariate results are presented in Table 5.

Table 5 Multivariate logistic regression analysis of risk factors for severe postoperative infection.

Variable	β	SE	Wald	Adjusted OR	95%CI	P value
Diabetes mellitus	1.138	0.415	7.52	3.12	1.38-7.06	0.006
Albumin < 35 g/L	1.054	0.408	6.68	2.87	1.29-6.39	0.010
Operative time ≥ 240 minutes	0.971	0.412	5.56	2.64	1.18-5.92	0.018
Open approach	0.880	0.424	4.31	2.41	1.05-5.54	0.038

OR: Odds ratio; CI: Confidence interval.

The multivariate logistic regression model found four independent risk factors for severe postoperative infection: (1) Diabetes mellitus (adjusted OR: 3.12, 95%CI: 1.38-7.06, P = 0.006); (2) Preoperative albumin < 35 g/L (adjusted OR: 2.87, 95%CI: 1.29-6.39, P = 0.010); (3) Operative time ≥ 240 minutes (adjusted OR: 2.64, 95%CI: 1.18-5.92, P = 0.018); and (4) Open surgical approach (adjusted OR: 2.41, 95%CI: 1.05-5.54, P = 0.038). Goodness of fit for the model was demonstrated using the Hosmer-Lemeshow test (χ² = 6.84, P = 0.554). An area under the receiver operating characteristic curve of 0.836 (95%CI: 0.768-0.904) from the receiver operating characteristic curve indicates a good level of discrimination.

Clinical outcomes

Table 6 compares the clinical outcomes of patients with and without severe postoperative infections. Patients in the infection group had longer total hospital stays, with an average of 24.3 ± 9.8 days vs 11.6 ± 5.2 days (P < 0.001). Postoperative hospital stays were longer for the infection group, averaging 19.8 ± 8.6 days compared with 8.4 ± 4.3 days for the non-infection group (P < 0.001). Fourteen and eight patients in the infection (36.8%) and non-infection (6.8%) groups, respectively, were admitted to the ICU (P < 0.001). Among the infection group patients, 10 patients (26.3%) underwent surgical re-interventions, comprising six percutaneous drainages and four surgical re-interventions. The 30-day infection mortality rate of the infection group was higher than that of the non-infection group (10.5% vs 0.8%, P = 0.004). The infection group had four deaths (one from septic shock and anastomotic leak and three from multiple organ failure and pulmonary sepsis), and the non-infection group had one death from a cardiac event. The 30-day readmission rate was higher in the infection group than in the non-infection group (21.1% vs 5.1%, P = 0.003).

Table 6 Comparison of clinical outcomes between infection and non-infection groups, n (%)/mean ± SD.

Variable	Infection group (n = 38)	Non-infection group (n = 118)	t/χ2	P value
Total hospital stay (days)	24.3 ± 9.8	11.6 ± 5.2	10.46	< 0.001
Postoperative stay (days)	19.8 ± 8.6	8.4 ± 4.3	10.72	< 0.001
ICU admission	14 (36.8)	8 (6.8)	21.02	< 0.001
ICU stay (days, median, IQR)	4.5 (2-8)	2 (1-3)	2.861	0.004
Mechanical ventilation	6 (15.8)	3 (2.5)	8.94	0.003
Re-intervention	10 (26.3)	2 (1.7)	22.84	< 0.001
Percutaneous drainage	6 (15.8)	1 (0.8)
Reoperation	4 (10.5)	1 (0.8)
30-day mortality	4 (10.5)	1 (0.8)	8.14	0.004
30-day readmission	8 (21.1)	6 (5.1)	9.04	0.003

¹Mann-Whitney U test.

ICU: Intensive care unit; IQR: Interquartile range.

DISCUSSION

Our retrospective cohort study of 156 patients who underwent surgical management of diseases of the digestive system found a postoperative infection rate of 24.4%. Recently published reports have found comparable values. A nationwide prospective cohort study by Yang et al[8] reported almost four times the number of cases (n = 17353) undergoing gastrointestinal surgery in 57 centers in China; the incidence of postoperative SSIs was approximately 5.5%. This value is based on the reports of all SSI grades. A study has reported 8% to 25% of severe SSIs, which necessitate surgical intervention or result in organ dysfunction; the range is defined by the mix of the cases and definitions used[13]. In our study, the severe infection rate is partially related to the referral policy of our center. Our center is a tertiary academic institution and receives a disproportionate number of advanced malignancies and patients with significant multimorbidity, explaining the increasing instances of SSIs in our study.

Our microbiological analysis identified Gram-negative bacteria in most pathogens (65.9%) isolated with Escherichia coli, with Klebsiella pneumoniae being the most often identified. These findings are consistent with those of existing studies describing the microbial ecology of abdominal SSIs. Following abdominal surgery, the polymicrobial infections found in a majority of the studies involved members of the Gram-negative and anaerobic bacteria, which are part of the endogenous gastrointestinal flora of the patient[2]. A major concern is the study’s findings that Escherichia coli and Klebsiella pneumoniae are (41.7% and 25.0%, respectively) ESBL-producing isolates. Dubinsky-Pertzov et al[6], the SSI rate in gastrointestinal surgery patients who received prophylactic antibiotics was significantly higher than that of those who were not colonized by ESBL-producing Enterobacterales. These findings also illustrate the need for antimicrobial surveillance in settings where resistant organisms are prevalent to make an informed decision on the antimicrobial agents to include in the treatment and prophylaxis of abdominal surgery patients. Based on the observed ESBL rates, MDR pathogens should be suspected particularly in patients with diabetes, hypoalbuminemia, prolonged or open operations, contaminated/dirty wounds, previous resistant-organism history, or early clinical deterioration. In lower-risk patients without these features, narrower-spectrum prophylaxis or empirical therapy may remain appropriate until culture results are available. Empirical regimens should be tailored to local resistance surveillance and de-escalated promptly after susceptibility results are available.

In our multivariate analysis, diabetes mellitus emerged as the strongest independent risk factor for the incidence of adverse outcomes, giving it an adjusted OR of 3.12. Existing studies support the findings of altered healing of surgical wounds in patients with diabetes, along with an increased risk of an infected surgical wound. Of particular interest are the immunological and vascular insults in diabetes that result in an altered response to infection because of reduced microvascular perfusion, impaired chemotaxis and phagocytosis of neutrophils, and an increased risk of an infected surgical wound[14]. In a study on 1418 patients undergoing hepatobiliary surgery, the presence of diabetes significantly increased the risk of SSIs with an OR of 6.2[15]. In addition, Marzoug et al[7], in their review, noted the presence of diabetes as a risk factor for various types of abdominal surgeries. Our data showed that patients with infection had significantly higher levels of HbA1c and fasting glucose, indicating that the absence of diabetes and control of hyperglycemia may be better predictors. These findings highlight the need for preoperative control of hyperglycemia in diabetic patients undergoing gastrointestinal surgery.

Preoperative hypoalbuminemia, defined as exhibiting serum albumin levels lower than 35 g/L, constituted the second most substantial independent risk factor in our study. Serum albumin is a recognized biomarker of nutritional status and predictor of postoperative morbidity in various surgical disciplines. Hypoalbuminemia represents malnutrition and a state of systemic inflammation, both of which are detrimental to the immune system and healing of wounds[16]. The NRS 2002, a nutrition-related assessment in which albumin is incorporated, revealed a score that was significantly higher in the infection group. A cross-sectional study of 132 patients undergoing gastrointestinal surgery reported that the development of SSIs was significantly associated with a serum albumin level below 3.5 g/dL[17]. Therefore, we endorse an early implementation of nutritional optimization, including oral nutritional supplements, the use of parenteral nutrition if appropriate, and perioperative immunonutrition, to mitigate the risk of severe infections. New recommendations have proposed nutrition-related screening and prompt solutions for patients undergoing significant gastrointestinal surgical interventions[18].

In our analysis, surgery durations over 240 minutes were identified as an independent risk factor for the development of severe postoperative infection. Lengthy surgery durations are among the strongest risk factors for SSIs. In a systematic review of prognostic models for SSIs in the setting of gastrointestinal surgery by McLean et al[19], the length of surgery was considered a predictor in 52% of the reviewed models. The relationship between infection and longer surgery duration has diverse reasons: Longer surgeries mean that wounds remain exposed for a long time to the environment, wounds become dry and cool, and prophylactic antibiotics become ineffective when they fall below the therapeutic concentration because of a lack of redosing and trauma to tissue[20]. Furthermore, a review by Seidelman et al[21] in Journal of the American Medical Association stated that surgery length is one of the most important risk factors for infection. The experience of the surgical team, surgical technique, and redosing antibiotics after the surgery exceed this drug’s half-life and may reduce this risk factor. The length of surgery is beyond the control of the surgical team.

In our study, compared with laparoscopic or minimally invasive methods, the open surgical method resulted in an adjusted OR of 2.41 and a significantly increased risk of serious postoperative infection. These findings corroborate the infection benefit of minimally invasive surgery. Yang et al[8] proved that open surgery independently contributed to the incidence of SSIs from an extensive cohort of patients undergoing gastrointestinal surgery in the case of laparoscopic surgery. The infection protective mechanism of laparoscopic surgery has diverse explanations; among the most important are less contamination to the surgical site due to a smaller incision, less injury and trauma to tissue, reduced tissue trauma and inflammatory stress, delay of postoperative ileus that slows the return of bowel function, and shorter exposure to the pathogens in the operating room[22]. As minimally invasive methods continue to be developed and become the standard of care for many gastrointestinal surgeries, their use should be promoted for oncological and functional benefits and, more importantly, for preventing infection.

The burden of severe postoperative infections on our patient group was significant, as infected patients had more than double the length of stay of uninfected patients, were more likely to require ICU admission and mechanical ventilation, and had a 30-day mortality rate of 10.5% (vs 0.8% for the uninfected group). These results are consistent with those of existing studies on the effects of postoperative infections. Calderwood et al[9] noted that SSIs increase direct and indirect costs of healthcare, length of hospitalization, and mortality in the perioperative period. Each instance of an SSI is estimated to increase perioperative costs by $20000-$30000. The combined impact of the increased costs of healthcare, significantly increased morbidity, and increased mortality provides a strong case for the design and establishment of a full-spectrum infection control program for surgical services. We did not perform patient-level cost extrapolation in the present cohort, and future studies should include direct medical costs, ICU costs, reintervention costs, and readmission-related expenses to better quantify the economic burden.

Many methods have been proposed to address the problem of increasing instances of SSIs after gastrointestinal surgery. The CDC recommends the bundle approach. This includes the proper selection and timing of surgical antibiotic prophylaxis, maintenance of perioperative normothermia, control of perioperative blood glucose levels, and reduction of surgery length. Additionally, the CDC recommends the preoperative skin antisepsis with chlorhexidine-alcohol[12]. The Enhanced Recovery After Surgery protocols comprise multiple practices and have reported a reduction, including early mobilization, early enteral feeding, and the use of less potent analgesia. The Enhanced Recovery After Surgery protocols reduce the rate of complications after surgery, especially those that are infectious[23,24]. Compared with the use of oral antibiotics or bowel preparation alone, the use of oral antibiotics with mechanical bowel preparation proves to have superior results in reducing SSIs in colorectal surgery[25]. The use of antimicrobial stewardship programs is significant in limiting the degree of antimicrobial resistance and postoperative infection[26].

This study has some limitations. For starters, retrospective studies may lead to issues with information bias and external confounding variables. We attempted to control some of these variables in our multivariate analyses; however, we are not certain of issues with confounding. Second, considering our sample size of only 156 and being a study with data from a single location, our findings had decreased statistical power and limited scope. Third, by only assessing Clavien-Dindo grade III complications or higher to identify severe postoperative infections, we may have omitted some of these cases because they were treated with antibiotics. Fourth, data were not collected for some relevant factors, such as the use of bowel management, perioperative control of temperature, and adherence to the surgical safety checklist. Finally, due to changing infections and surgery methods, time may have affected our results, as our study lasted for three years. In addition, the study period overlapped with the tail end of the coronavirus disease 2019 pandemic, and we could not fully account for staff shortages, isolation policies, changes in prophylaxis, or other infection-control measures. Moreover, we did not stratify risk factors by infection site because of the small subgroups, and neither did we perform a formal cost extrapolation nor a detailed assessment of prophylaxis appropriateness/empirical antimicrobial adequacy. To further strengthen our findings, large-scale, multi-site, prospective studies with a greater sample size are required to implement specific controls to help reduce severe postoperative infections from digestive surgeries.

CONCLUSION

Severe postoperative infections are a significant complication of surgeries involving the digestive tract, affecting approximately 25% of patients in the studied cohort. The most common pathogens are MDR, ESBL-producing, gram-negative bacilli, with increasing rates of resistance. The presence of diabetes mellitus, hypoalbuminemia, increased operative time, and the choice of open surgery increase the risk of severe infections. These infections increase the need for intensive care and result in longer hospital stays with a higher rate of 30-day mortality. The burden of severe postoperative infections following surgeries of the digestive tract can be reduced by optimizing blood sugar levels and preoperative nutrition, using advanced surgical techniques, and providing tailored perioperative care.