Study design and setting
We performed a prospective cohort study at the University Hospital Basel (USB) ─ a tertiary care center in Switzerland with > 1700 abdominal and > 850 cardiac interventions per year (overall, ~ 38,000 surgical interventions per year). The present observational study was nested within an ongoing multicenter open-label cluster-randomized cross-over study on the efficacy of CHX versus PI skin antisepsis in preventing SSIs (i.e. PICASSo trial; ClinicalTrials.gov numbers, NCT03685604 and NCT03859375). The study data can be shared by the corresponding author upon request.
Patient selection
At the USB, inpatients aged ≥18 years receiving a non-emergency cardiac or abdominal surgery between April 15, 2019 and September 06, 2019 were eligible for study inclusion. We excluded patients who could not provide or declined the written informed consent for the nested study. The PICASSo trial protocol does not include an individual informed consent.
Study outcomes
Our predefined primary outcome was the reduction of microbial skin counts (that is, colony forming units [CFUs]) after two and three antiseptic paints, respectively. Secondary outcomes were (i) the proportion of patients with microbial skin counts of zero CFU after three antiseptic paints as compared to two antiseptic paints, and (ii) the proportion of patients with insufficient reduction of microbial skin counts after two antiseptic paints. We defined an insufficient reduction in microbial skin counts at the surgical site as the detection of > 5 CFUs and/or ≥ 1 pathogen(s) according to the National Healthcare Safety Network common commensals/pathogen list (version 9.2; www.cdc.gov). All outcomes were standardized per 25 cm2 skin area.
In a secondary analysis, we merged the prospectively collected SSI status within 30 days after surgery using a national SSI surveillance database [8, 9]. In this validated surveillance program, well-trained infection control practitioners ascertain SSIs by screening surgical patients for evidence of SSIs, and cases are double-checked by a board-certified infectious disease specialist. Standardized postdischarge SSI surveillance is conducted by telephone interviews and review of electronic medical records. SSIs are classified according to the Centers for Disease Control and Prevention definitions [10].
Study procedures
Patients routinely receive a whole body shower with CHX (CHX digluconate 40 mg/ml solution; Hibiscrub®, Mölnlycke Health Care AG, Schlieren, Switzerland) prior to cardiac or abdominal surgery. Furthermore, it is standard practice to clip hair prior to surgery ─ if deemed necessary. Presurgical antiseptic processes and management are in accordance with the recommendations of the World Health Organization (WHO) [4]. At the USB, it is standard of care to consecutively perform skin antisepsis for three times (approximately 3 min at a time) by using sterile gauzes. The routinely applied skin antiseptics are either CHX in alcohol (CHX digluconate 20 mg and propan-2-ol 0.7 ml; Softasept® CHX, B. Braun Medical AG, Sempach, Switzerland) or PI in alcohol (PI 0.9 mg and propan-2-ol 457.5 mg; Braunoderm®, B. Braun Medical AG, Sempach, Switzerland). The PICASSo trial did not affect the routine procedures for skin antisepsis as recommended by the WHO [4], apart from the regular randomized department-level switches from presurgical skin antisepsis with PI to CHX, or vice versa.
For the present study, well-instructed members of the surgical team obtained the skin swabs in the operating room under sterile conditions. We collected three skin swabs (sterile 0.9% sodium chloride premoistened swabs; FLOQSwab®, Copan Diagnostics Inc., Brescia, Italy) from the participant’s trunk. To standardize the skin area, we used sterilized metal templates with a window of 25 cm2, which was swabbed repeatedly horizontally and vertically in a uniform way under gentle pressure. The template was freely positionable at the surgical site (thorax or abdomen) as long as it did not interfere with the succeeding incision. Following collection of the first skin swab prior to skin antisepsis, and once the second and third application of PI or CHX had dried out, we obtained a second and third skin swab, respectively. We did not collect swab samples at later time points, as the surgical incision may follow directly after the drying of the third application of PI or CHX (that is, start of the at-risk period).
The first skin swab was neutralized for CHX to avoid bias by presurgical shower with CHX. Swabs 2 and 3 were neutralized before culture for either CHX or PI ─ depending on the applied antiseptic product. We used a standardized inactivation solution, which was tested for non-toxicity and which was microbiologically validated ─ containing either polysorbate 80 30 g/l, lecithin 3 g/l, L-histidine 1 g/l, sodium thiosulfate 5 g/l, saponine 30 g/l, trypticase soy broth 30 g/l, and distilled water 1 l for CHX skin antisepsis; or polysorbate 80 30 g/l, lecithin 3 g/l, L-histidine 1 g/l, sodium thiosulfate 5 g/l, trypticase soy broth 30 g/l, and distilled water 1 l for PI skin antisepsis (details can be requested from the corresponding author).
Data collection
A study physician collected the relevant clinical information during the screening visit and verified the respective data by use of electronic medical records. In the operation room, a trained study nurse recorded the applied antiseptic product and timing of skin swab collection.
Microbiological investigation
Skin swabs were immediately delivered to the in-house microbiological laboratory, where a study technician inoculated trypticase soy agar (TSA) plates (bioMérieux, Marcy l’Etoile, France) within 6 h during weekdays or within 72 h on weekends (samples were kept in the refrigerator before processing). Due to logistical reasons, we could not mask the study technician for the exposure status (that is, consecutive paint number; 1 to 3). We cultured an additional 1:10 sodium chloride diluted sample of the first swab to safeguard against unreliable results for samples with high microbial counts. We incubated the TSA plates for 2 days at 36 °C (±1 °C) and determined respective CFUs with a manual colony counter (Scan® 100, Intersciences, Saint Nom, France). We identified microbial skin isolates using the microflex™ LT MALDI-TOF mass-spectrometer system (Bruker Daltonics, Bremen, Germany).
Statistics
Based on a previous internal quality evaluation (unpublished data), we estimated a sample size of ≥228 patients in order to demonstrate superiority of three versus two paints at a clinically defined superiority margin of 2 log10(CFUs) difference (significance level of 5%, power of 90%). Our null hypothesis was that there was no difference between two and three paints of PI/CHX in reducing microbial skin counts (CFUs) at the surgical site.
With regards to the primary outcome, we compared the square root-transformed CFUs after paint 2 and 3 using a paired t-test. In a supplementary analysis of the primary outcome, we also compared the CFU distribution after paints 2 and 3 using the Wilcoxon signed rank test. We compared secondary outcomes (proportions) after paint 2 and 3 using a χ2-test (overall and stratified by PI/CHX application). In a secondary analysis, we compared CFU counts after paint 3 between patients with and without a subsequent SSI (within 30 days after index surgery) using the Wilcoxon rank sum test.
We fitted univariable logistic regression models (variables and categories provided in Table 3) to identify potential risk factors for insufficient reduction of microbial skin counts after paint 2. A low event-predictor ratio precluded a multivariable logistic regressions analysis. Results were considered significant at a P-value of ≤0.05. Analyses of secondary outcomes and subgroups were considered as hypothesis-generating. The study data were analyzed by a statistician (A.A.) using the R statistical software (www.r-project.org).
