Postoperative Staphylococcus aureus Infections in Patients With and Without Preoperative Colonization

Key Points Question What is the cumulative incidence of Staphylococcus aureus surgical site infections (SSIs) and bloodstream infections (BSIs) in Europe, and what factors are associated with an increased risk of SSIs and BSIs? Findings In a cohort study of 5004 surgical patients, the weighted cumulative incidence of S aureus SSIs and BSIs was 1.23%. Preoperative S aureus carriage, mastectomy or neurosurgery, higher body mass index, and having nonremovable implants in the body were independently associated with S aureus SSIs and BSIs. Meaning Staphylococcus aureus SSIs and BSIs are important postoperative complications, and future interventions aimed at prevention of these infections should focus on at-risk surgical patient groups to achieve a higher efficacy.

eMethods.Supplemental Methods eTable 1. Number of study participants per country and European sub-region eTable 2. Baseline characteristics of the weighted study population and original source population eTable 3. Weighted cumulative incidence of SA SSI/BSI within 90 days post-surgery by surgery type eTable 4. Unweighted cumulative incidence of SA SSI/BSI by preoperative SA colonization status eTable 5. Methicillin-susceptibility of colonizing strains eTable 6. ST types of the isolates from the sub-cohort (N=346) eTable 7. Unweighted risk factor analysis for SA SSI/BSI eTable 8. Weighted risk factor analysis for SA SSI/BSI (keeping preoperative SA decolonization in multivariable model) eTable 9. Weighted Fine and Gray model for SA SSI/BSI eTable 10.Weighted risk factor analysis for SA SSI/BSI (complete case analysis) eFigure 1. Subject selection and number of isolates for the MLST analysis eFigure 2. Cumulative incidence function for SA SSI/BSI (unweighted data) eReferences This supplementary material has been provided by the authors to give readers additional information about their work.

Description of collected data
Based on the literature and clinical reasoning, we collected the following variables as they were considered potential etiological factors for S. aureus SSI or postoperative BSI (SA SSI/BSI) or the competing event death without SA SSI/BSI: current preoperative SA colonization status; past history of SA colonization or infection; sex; body weight and height (to calculate the body mass index [BMI] as the weight in kilograms divided by height in meters squared); age; presence of comorbidities (to calculate the Charlson comorbidity index (CCI); 1 the American Society of Anesthesiologist's (ASA) physical status classification score; 2 presence of any non-removeable implant in the body prior to surgery; receipt of preoperative decolonization treatment; receipt of immunosuppressive medication within two weeks prior to surgery; presence of surgical drain in the body after surgery; and type of surgery.None of these potential etiological factors were considered to be intermediates in the occurrence relation of other factors with the outcome SA SSI/BSI, though they could have been confounders of those occurrence relations.We also collected microbiological data, mortality data and data on the development of SSI and other infections.

Definitions:
Immunosuppressive medication was defined as any immunosuppressive/ antineoplastic/ chemotherapy medication or systemic corticosteroids, given in a dose equivalent to ≥2 weeks of daily prednisolone 20mg once daily.Sites were inquired about the decolonization strategies they employ, and the study definition was based on this information.Preoperative decolonization treatment was defined as any medication given to the patient preoperatively for the indication preoperative decolonization (as reported by the participating site), or the receipt of any (or combination) of the following medication prior to surgery (start date of decolonization should have been equal to or preceded the date of surgery): nasal mupirocin, fusidic acid (was used by some sites as an equivalent for mupirocin, and the indication for use should not have been treatment of an infection), naseptin, octenisan, octenidine, and/or chlorhexidine.

Description of the weighting methods
As reported in the manuscript, S. aureus carriers and non-carriers were enrolled into the study cohort in a 2:1 ratio.However, this ratio was approximately 1:3 in the overarching source population.Because the source population was a random sample of the general population, it was the population that we wanted to make inference on.For this reason we aimed to recreate the source population by weighting the study cohort subjects with the inverse probability of their inclusion in the study cohort. 3After conducting multiple imputation for missing values in the source population (correctly screened), we fitted a logistic regression model using predictor variables available for the source population, to estimate the probability of inclusion.After this, we took the inverse of this probability as the weight.The predictor variables are listed in the table below, as well as their distributions in the source population, study cohort, and weighted study population.There is good agreement between the source and weighted population.The calculated weights were used for the incidence calculations and in the risk factor analyses in the accompanying manuscript.The cumulative incidences were bootstrapped.10,000 bootstrap samples of the study cohort were made, after which the bootstrap samples were inflated using the weights.In each weighted bootstrap sample, the cumulative incidence of SA SSI/BSI for each surgery type was calculated.Using the sequence of 10,000 cumulative incidences for each surgery type, the median cumulative incidence with 95% CI (2.5 th and 97.5 th percentile) was derived.

Characteristic
Abbreviations: BSI, bloodstream The cumulative incidences were bootstrapped.10,000 bootstrap samples of the study cohort were made.In each bootstrap sample, the cumulative incidence for SA carriers and non-carriers was calculated.The sequence of 10,000 cumulative incidences for SA carriers and non-carriers was used to derive the median cumulative incidence with 95% CI (2.5 th and 97.5 th percentile).
Abbreviations: BSI, bloodstream infection; CI, confidence interval; IQR, interquartile range; No., number; SSI, surgical site infection.The follow-up time was 90 days for both SA colonized and non-colonized subjects.However, the cumulative risk for non-colonized subjects did not change after day 56.

eFigure 1 . 2 .
Subject selection and number of isolates for the MLST analysis eFigure Cumulative incidence function for SA SSI/BSI (unweighted data)

Level Source population (correctly screened) Study cohort Weighted population
Number of study participants per country and European sub-region Baseline characteristics of the weighted study population and original source population Baseline characteristics of the weighted study population and original source population Weighted cumulative incidence of SA SSI/BSI within 90 days post-surgery by surgery type Abbreviations: ASA.American Society of Anesthesiologist's; BMI, body mass, index; IQR, interquartile range; SA, S. aureus.eTable 1. © 2023 Troeman DPR et al.JAMA Network Open.eTable 2. a Included: open cardiac surgery, implantable cardioverter defibrillator (ICD) implantation, peripheral artery bypass surgery, and central artery reconstructive surgery.bIncluded: hip prosthesis and knee prosthesis surgery.cIncluded: craniotomy, laminectomy, and spinal fusion surgery.The variables immunosuppressive medication; preoperative decolonization treatment, and CCI were not collected in the original source population.Abbreviations: ASA, American Society of Anesthesiologist's; BMI, body mass index; CCI, Charlson comorbidity index; IQR, interquartile range; SA, S. aureus.eTable3.

analysis for the association with SA SSI/BSI Univariable analysis for the association with death without SA SSI/BSI Multivariable analysis for the association with SA SSI/BSI e
Weighted risk factor analysis for SA SSI/BSI (keeping preoperative SA decolonization in multivariable model) Risk factors with a p-value ≤ 0.157 in either univariable analyses, were selected for the multivariable analysis.Abbreviations: ASA, American Society of Anesthesiologist's; BMI, body mass index; BSI, bloodstream infection; CCI, Charlson Comorbidity Index; CI, confidence interval; HR, hazard ratio; SA, S. aureus; SSI, Surgical site infection.Weighted Fine and Gray model for SA SSI/BSI Weighted risk factor analysis for SA SSI/BSI (complete case analysis) Colonization status prior to surgery based on S. aureus screening of the nose, throat and perineum; b per 1-year increase in age [range 18-99]; c per 1-point increase in BMI [range 13.5-65.8];d per 1-point increase in the CCI [range 0-12].
Subjects with multiple isolates (colonizing or infecting) were included more than once if they had at least 2 isolates with different STs (then they were counted in twice in the table above).Subjects with multiple isolates of the same ST were only counted in once.Abbreviations: ST, sequence type.©2023TroemanDPRet al.JAMA Network Open.eTable 7. Unweighted risk factor analysis for SA SSI/BSI e Risk factors with a p-value ≤ 0.157 in either univariable analyses, were selected for the multivariable analysis.Abbreviations: ASA, American Society of Anesthesiologist's; BMI, body mass index; BSI, bloodstream infection; CCI, Charlson Comorbidity Index; CI, confidence interval; HR, hazard ratio; SA, S. aureus; SSI, Surgical site infection.©2023TroemanDPR et al.JAMA Network Open.eTable 8. e © 2023 Troeman DPR et al.JAMA Network Open.eTable 9. © 2023 Troeman DPR et al.JAMA Network Open.eTable 10. a e Risk factors with a p-value ≤ 0.157 in either univariable analyses, were selected for the multivariable analysis.Abbreviations: ASA, American Society of Anesthesiologist's; BMI, body mass index; BSI, bloodstream infection; CCI, Charlson Comorbidity Index; CI, confidence interval; HR, hazard ratio; SA, S. aureus; SSI, Surgical site infection.© 2023 Troeman DPR et al.JAMA Network Open.