Usefulness of Bacterial Culture of Drainage Fluid for Predicting Surgical Site Infection After Crohn’s Disease Surgery

Abstract Aim Early detection of surgical site infection (SSI) allows for appropriate management after Crohn's disease (CD) surgery. The aim of this study was to evaluate the usefulness of bacterial culture of postoperative drainage fluid after CD surgery. Methods This study included 110 patients with CD who underwent surgery with bowel resection between January 2010 and March 2020 at Osaka University Hospital. Patients with only perianal surgery or incomplete records were excluded. Risk factors for SSI were evaluated in the context of clinical findings, including bacterial culture of postoperative drainage fluid, and bacterial species related to SSI were also examined. Results Of 110 patients, 18 (16.4%) developed SSI. Organ/space SSI developed in six, and a positive bacterial culture of drainage fluid (D‐Posi) was found in five (83.3%). Of 104 patients without organ/space SSI, 31 (29.8%) were D‐Posi (P = .027). Similarly, 68.8% with incisional SSI were D‐Posi, whereas 26.6% without incisional SSI were D‐Posi (P = .0021). Multivariate analysis revealed that D‐Posi was an independent risk factor in both organ/space and incisional SSI. Bacterial examination showed that Pseudomonas aeruginosa and Enterococcus faecalis were significantly detected in patients with SSI. Conclusion This study suggests the usefulness of postoperative drainage fluid bacterial culture for early diagnosis of SSI after CD surgery.

complications may not only delay resumption of medication but also trigger recurrence and complicate subsequent surgeries because of adhesions. 7,8 If complications are detected in the early phase, exacerbations could be avoided and prognosis improved. 7,8 For these reasons, accurate predictors of complications are needed.
Surgical site infection (SSI) is one of the postoperative complications, and the rate of SSI after CD surgery is higher than after colorectal cancer surgery. 4,[9][10][11] Several SSI risk factors have been identified in patients with CD, including anemia, a longer duration of surgery, and higher intraoperative lactate level. 4,11 Although these are indirect risk factors for SSI, a direct factor is bacterial contamination of the surgical site. 12 Few reports, however, have described the association between positive bacterial culture of lavage or drainage fluid and SSI in gastrointestinal surgery. [12][13][14] To our knowledge, no studies have evaluated the role of positive bacterial culture of drainage fluid (D-Posi) after CD surgery.
We hypothesized that D-Posi could predict SSI in the early phase after CD surgery. The aims of this study were to evaluate the clinical impact of D-Posi in detecting SSI and the characteristics of detected bacterial species after CD surgery.

| Patients
A total of 245 consecutive patients who underwent CD surgery be-

| Perioperative management and surgical procedure
Prophylactic antibiotics with second-generation cephalosporin (cefmetazole) were intravenously administered within 30 min before skin incision, repeatedly at 3-h intervals during surgery, and postoperatively twice a day until postoperative day (POD) 2. Drainage tubes were placed in a Morrison's pouch or Douglas' pouch. Before March 2012, we performed closed passive drainage using surgeon's choice of a SILASCON duple drain (Kaneka Medical Products, Japan) or closed active drainage using a BLAKE silicone drain (Ethicon, Cincinnati, OH). After April 2012, closed active drainage was used in all cases.
A wound protector was used during surgery, and intraperitoneal lavage was performed before wound closure. All surgical staff changed gloves after bowel anastomosis. The peritoneum was closed by a running suture with 3-0 Vicryl (Ethicon). The muscular fascia was closed by a knotted suture with 1-PDS PLUS (Ethicon).
The wound was washed with 200 mL of saline before closure of the skin by a buried suture with 4-0 PDS plus (Ethicon). Board-certified surgeons performed all surgeries.

| Detection method of bacterial species
Bacterial culture of drainage fluid was performed on the same day a sample was submitted. The sample was smeared for detecting the type by gram staining, and then coated on both Agar's and Ringer's medium. Results of the gram staining were available the same day.
The bacterial species were examined with matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry using colonies on Agar medium. On the day following sample submission, D-Posi status and bacterial species were identified. The antibiotic sensitivity was identified at 2 d after sample submission. If the amount of bacteria was limited, this step sometimes took a few more days and up to 1-2 wk for very low counts associated with using Ringer's medium.  and "Others" together with D-Posi was defined as "Combination." Diagnostic accuracy was judged using the diagnostic OR (DOR), which is not affected by disease frequency and incidence of complications. Statistical significance was defined as P < .05. The ratios were calculated as follows:

| Patient characteristics
Data for 110 patients were analyzed and the characteristics of the 84 men and 26 women are shown in Table 1. The median age at surgery was 39.5 y, and 36 patients (32.7%) were D-Posi. The APR accounted for 10 patients (9.1%). The median duration of surgery was 203.5 min, and the median intraoperative bleeding was 145 mL.
Seven patients (6.4%) received a blood transfusion. A total of 17 patients (15.5%) received preoperative antibiotics for high inflammatory response with penetrating disease.

| SSI risk factors
Abdominoperineal resection has been reported as a risk factor for incisional and organ/space SSI, leading to assess also the clinical impact of D-Posi with exclusion of APR. 11 Of 100 patients with APR excluded, overall SSI developed in 14 patients: 10 with incisional SSI only, two with organ/space only, and two with both. Eleven (78.6%) with overall SSI were D-Posi, whereas 20 (23.3%) without overall SSI were D-Posi (P = .0004). We found that 75% with incisional SSI were D-Posi and 22.0% without incisional SSI were D-Posi (P = .002).
Similarly, 75% with organ/space SSI were D-Posi, and 29.2% without organ/space SSI were D-Posi (P = .091). Multivariate analysis identified D-Posi as an independent risk factor in overall SSI and incisional SSI also among the patients with APR excluded.

| Treatment details for the case of D-Posi
Of 110 patients, 36 were D-Posi, 19 (52.8%) of whom received only

| Comparison of predictive and diagnostic accuracy
Next, we evaluated the predictive and diagnostic ability of D-Posi.
In overall SSI, the AUC for D-Posi was 0.736, and the DOR was 7.81, and the AUC was higher than the "Others" group (AUC = 0.733).
Values for the "Combination" were the highest, with an AUC of 0.819 and DOR of 21.3 ( Figure 2A and Table 5). In organ/space SSI, the AUC for D-Posi was 0.768, and the DOR was 11.8 ( Figure 2C and Table 5), while D-Posi was not a strong predictor for incisional SSI compared to "Others" and "Combination," with an AUC of 0.711 and DOR of 6.07 ( Figure 2B and Table 5). The negative predictive value was above 90% in all groups.

| Characteristics of detected bacteria
We investigated which bacteria were associated with SSI risk. In univariate analysis, Pseudomonas aeruginosa and Enterococcus faecalis were significantly associated with overall SSI (P = .0024 and .017, respectively). Pseudomonas aeruginosa was frequently detected in patients with incisional and organ/space SSI Klebsiella pneumoniae and E. faecalis were often frequently detected in patients with organ/space SSI (Table S1). The comparison of detected bacteria between SSI and D-Posi were also performed. Of 16 patients with incisional SSI, seven (43.8%) had the same organism detected from both samples. Among six patients with organ/space SSI, four (66.6%) had the same organism in both samples (Table S2) (Table S3).      were the same between drainage fluid and of SSI bacterial culture.

| D ISCUSS I ON
This suggested that sensitivity results in drainage fluid could guide the choice of an antibiotic.
The present study had several limitations. First, these results were based on a single-center retrospective cohort study, including some selection bias that could not be avoided. Second, the impact of preoperative medicines, such as biologics except anti-TNFα therapy, immune suppressants, and 5-aminosalicylic acid were not evaluated in this study. A larger study that includes these factors is needed.
In conclusion, this study confirmed the clinical usefulness of D-Posi for SSI and demonstrated the association of bacteria species with CD surgery. Bacterial culture of drainage fluid represents an easy, noninvasive, and inexpensive tool to perform appropriate management of SSI.

ACK N OWLED G M ENT
The authors thank the staff at Osaka University Hospital.

D I SCLOS U R E S
The study protocol was approved by the Institutional Review Board of Osaka University Hospital (# 15028).
Informed consent was obtained from all patients before the surgery.
This research was not preregistered in an independent, institutional registry (N/A).
This research was not an animal study (N/A).