br METHODS br Study design setting

METHODS
Study design, setting, and participants
From September 2014 to December 2015, we conducted a prospec-tive cohort study of adult patients, with gynecological or gastrointesti-nal malignancies, who Necrosulfonamide were admitted to the hospital for elective abdominal and pelvic surgical procedures. This single-center study was conducted at the Instituto Nacional de Cancerología, a 135-bed public tertiary-care, teaching cancer hospital in Mexico City, with an annual mean of 7,500 hospital discharges and 3,500 surgeries per year.
Written informed consent was obtained from all participants enrolled in the study. The study was approved by the Institutional Review Board (014/016/INI) (CEI/906/14).
Patients and follow-up
At admission, patients who agreed to participate were screened for ESBL-E coli through a rectal swab. Data collected included the fol-lowing: length of hospital stay, gender, age, weight, height, obesity (body mass index ≥ 30 kg/m2), smoking status, alcohol use disorder (as defined by the Diagnostic and Statistical Manual of Mental Disor-ders, fifth edition [DSM-V]), diabetes, high blood pressure, hospital admissions, antibiotic use, chemotherapy and radiotherapy within 3 months, central venous catheter placement, ileostomy, colostomy, gastrostomy, blood cell count, Model for End-Stage Liver Disease, type and duration of surgery, surgical blood loss, reoperation, paren-teral nutrition, blood transfusion, postsurgical drainages, and the use of antibiotic prophylaxis. Information on the characteristics 
of patients and neoplasia is provided (Table 1). Additional rectal swabs were obtained on days 5, 10, 15, 20, and 25 of the hospital stay.
During the length of hospital stay, study subjects were evaluated every other day by 1 of the investigators. Information on clinical evo-lution, wound appearance, drains, and antibiotic use was collected. Investigators and attending physicians were blinded to the rectal swab result. Before discharge, study participants were instructed on SSI symptoms and when to contact a physician. Surgical patients had weekly appointments during the first 2 postoperative weeks, and at least 1 additional appointment 30-40 days after surgery. If a compli-cation occurred, the number of appointments varied.
The primary outcome variable was SSI, defined by US Centers for Disease Control and Prevention (CDC) criteria.10 Pneumonias, urinary tract infections, and bloodstream infections were also evaluated as outcomes variables.
All patients were evaluated for at least 30 days after surgery for SSI or any other health-related infection. SSI surveillance was con-ducted on a regular basis. The Instituto Nacional de Cancerología has a broad hospital program with postdischarge surveillance of up to 30 days or up to 1 year if prosthetic material is inserted. All surgical patients are registered in a database. Infection-control nurses see the patients and follow their postoperative course 2-3 times per week until discharge. All cultures are reviewed daily at the Microbiology Laboratory, along with readmission records. This information is cross-checked, and between 30-45 days after surgery (up to 1 year if prosthetic material is inserted) the medical chart is reviewed to determine whether an infection has occurred.
Microbiology analysis
A rectal swab was obtained at the time of admission and every 5 days until discharge or on postoperative day 25. Only 1 rectal swab was obtained per time point from each patient. All samples were plated in blood, MacConkey, and chocolate agar and incubated at 37°C for 24 h. Species identification was performed with the BD Phoe-nix 100 Automated System (Franklin Lakes, NJ). ESBL-producing microorganisms were confirmed by means of the double disc diffu-sion method.
Fig 1. Flow chart of patients included in gastric pits study.
Study sample
Recruitment of patients was performed by means of convenience sampling. Sample size was calculated using Epi Info software version 6 (CDC, Atlanta, GA), with an historical 37% prevalence of ESBL-producing bacteria7 taken as expected frequency, with an acceptable error margin of 5%, and a confidence level of 80%. The estimated sam-ple size to detect ESBL-PE colonization was 174 patients, considering a potential dropout of 20%.