[Identifying phenotypes of surgical site infection in patients after gastrointestinal surgery].

Objective: Surgical site infection (SSI) is a common health care-related infection after gastrointestinal surgery. Once SSI occurs after surgery, it can significantly prolong the postoperative hospital stay, increase the cost burden of patients and society, and even endanger the life safety of patients. The purpose of this study was to investigate the clinical phenotypes of gastrointestinal surgery, identify the clinical characteristics of SSI, and provide reference for the prevention of SSI after gastrointestinal surgery. Methods: This is a multicenter prospective cohort study that collected clinical data from all adult patients undergoing gastrointestinal surgery from March 2021 to February 2022 at 42 hospitals in China, including baseline and perioperative characteristics. Based on the variables associated with SSI, latent class analysis (LCA) was used to explore the population characteristics of SSI. Results: In total, 16 087 patients were included in the study, of whom 345 (2.1%) developed SSI. LCA analysis revealed that patients undergoing gastrointestinal surgery were classified into four clinical phenotypes, including α (3851), β (1538), γ (6387), and δ (4311). Type α had minimal abnormality on related system functions (ASA score > 2: 4.5% [173/3851]), and mainly underwent appendix surgery (98.9% [3808/3851]). The postoperative SSI incidence of type α was 0.4% (16/3,851), which belonged to the group of SSI low risk. The abnormality of system functions of type β (ASA score > 2: 17.4% [268/1538]) was worse than that of type α. Type β mainly underwent stomach surgery (72.4% [1113/1538]), and its incidence of postoperative SSI was 1.2% (18/1538), belonging to the group of SSI medium risk. The ASA score of type γ (ASA score > 2: 18.0% [1148/6387]) was comparable to that of type β. Type γ mainly received colorectal surgery (colon surgery: 40.1% [2562/6387]; rectal surgery: 33.6%[2143/6387]), and its incidence of postoperative SSI was 1.7% (106/6387), belonging to the group of SSI medium risk. Type δ (ASA score > 2: 23.5%[1015/4311]) was the most serious type with the highest proportion of open surgery. Type δ mainly underwent small intestine (54.0%[2327/4311]) and stomach surgery (32.3% [1392/4311]) and had the highest incidence of SSI (4.8% [205/4311]) and the highest mortality rate (0.6% [24/4311]), belonging to the group of SSI high risk. Compared with type α and β, the median length of hospital stay (α, β, γ, and δ: 5.0 days, 9.6 days, 13.0 days, and 16.0 days, P<0.001) and postoperative hospital stay (α, β, γ, and δ: 4.0days, 6.0days, 8.3 days, and 10.0 days, P<0.001) of type γ and δ were significantly increased, and the median medical costs (α, β, γ, and δ: 14 178.7 yuan, 39 514.2 yuan, 62 893.0 yuan and 57 266.6 yuan, P<0.001) were also significantly increased. Conclusion: LCA analysis elucidated four clinical phenotypes of patients undergoing gastrointestinal surgery. Type α had a low risk of SSI. Type β and γ had a medium risk of SSI, and type δ had a high risk of SSI.