Patient selection

The patients in this study were a convenience sample as a sub-group of a larger study (NCT02174471) looking at myocardial structure of patients undergoing aortic valve replacement with or without additional procedures.

Ethical approval for this study was given by the NRES Committee, London (Ref No 07/H0715/101).

Patients scheduled for aortic valve replacement were approached and verbal and written consent to participate was obtained from those willing to take part in the study. No individual identifiable data is included in this submission. We aimed to include all patients from the parent study who had undergone first-time isolated AVR.

Exclusion criteria were age under 18 years, recent or ongoing infection, pregnancy, an immunosuppressive condition and concomitant use of immunosuppressive therapy.

Following recruitment, prior to surgery, the patient was venesected and blood samples centrifuged for 15 min at 3000 rpm and the serum stored in cryotubes (Thermoscientific, Mass) at − 80 °C for later analysis.

We chose to assay for antibodies to the extracellular polypeptides alpha-toxin (AT) and staphylococcal enterotoxin A (SEA), and to teichoic acid (TA), a major surface antigen of the staphylococcal organism; all are present in almost all strains of Staphylococcus aureus. A high proportion (86–95%) of the S. aureus isolates in clinical infections produce an anti-alpha-toxin antibody response (Granstrom et al., 1983a). In cases of serious staphylococcal infection, the levels of alpha-toxin have been demonstrated to be very high, suggesting that the antigen is highly immunogenic (Colque-Navarro et al., 1993). Teichoic acid is particularly expressed in case of long-standing staphylococcal infection, for example, deep-seated wound infection or endocarditis (Colque-Navarro et al., 1998). It was also decided to assay for SEA antibodies as this toxin is the most commonly produced enterotoxin in S. aureus strains (Kanclerski et al., 1996). These three antibodies are likely to be reliably expressed in those patients undergoing cardiac surgery that may go on to develop staphylococcal infections.

Antibody analysis

Anti-staphylococcal antibody analysis

The ELISA procedure for assaying alpha-toxin, teichoic acid, and staphylococcal enterotoxin A antibody levels has been detailed previously (Granstrom et al., 1983b; Colque-Navarro et al., 2000). Briefly, coating doses for the 96-well microtitration plates (Dynatech M-129B, Plochingen, Germany) with alpha-toxin, teichoic acid and SEA were established at 2.5 μg/mL, 1 μg/mL and 0.5 μg/mL, respectively. The working volume throughout the tests was 100 μL/well. The microtitration plates were coated with antigens diluted in phosphate-buffered saline (PBS), pH 7.4, and incubated overnight at 22 °C. The plates were washed and a serum dilution in PBS with Tween-20 0.05% v/v (PBS-T) of 1 in 1000 for α-toxin and SEA and 1 in 10,000 for teichoic acid was added to two coated wells. Positive and negative controls were included in each plate. The plates were incubated for 1 h at room temperature (20 °C). After washing the plates, alkaline phosphatase-conjugated goat anti-human antibody (Sigma) diluted in PBS-T was added to each well, and the plates were incubated for 2 h at room temperature. After the final wash, p-nitrophenyl- phosphate substrate (Sigma) was added. Titres were read when the positive controls reached previously established values at 405 nm on a Titertek Multiskan (Flow Laboratories, Irvine, Scotland) instrument.

The antibody levels were expressed as arbitrary units by using the reference line unit calculation method (Reizenstein et al., 1995).

EndoCAb analysis

Serum EndoCAb levels were measured with an ELISA using equimolar amounts of a lipopolysaccharide (LPS) from each of a selected Escherichia coli, Salmonella typhimurium, Klebsiella aerogenes, Pseudomonas aeruginosa rough mutant strain, lacking the LPS O-polysaccharide and part of the LPS outer core, but retaining the inner core structure. These were each complexed to polymyxin B, mixed in a cocktail in carbonate-bicarbonate buffer (pH 9.6), and the cocktail coated on 96-well polystyrene microtitre plates selected for optimal EndoCAb ELISA characteristics. Results are expressed as median units (MU) per millilitre, where 100 MU/mL is the median value of a population of 1000 healthy volunteers.

Test and control samples were diluted 1:200 with dilution buffer, and 100 μl of each sample to be assayed was added in triplicate to the wells of a pre-coated microtitre plate. The assay was standardised using a calibrated pooled-serum standard of a predetermined EndoCAb IgG concentration. An eight-point standard curve was constructed using doubling dilutions of the calibrated serum, which gave a range of IgG EndoCAb of 12.25 to 784 MU, and 100 μl of each added in triplicate to the microtitre plate. Following the addition of the samples and standards, the plate was covered and incubated for 1 h at 37 °C.

The plates were then washed three times with phosphate-buffered saline solution-polysorbate (Tween) buffer (sodium chloride, 0.138 M; phosphate, 0.01 M; pH 7.4 containing 0.10% [v/v] polyoxyethylene sorbitan monolaurate [Tween 20]) and finally blotted dry.

An alkaline phosphatase goat anti-human IgG antibody (Sigma-Aldrich Chemical Co; Poole, UK) was diluted 1:500 with dilution buffer, and 100 μl of conjugated antibody was added to each well; the plates were covered and incubated for 1 h at 37 °C.

After incubation, the plates were washed three times with Tween buffer. Substrate (100 μl), comprising 1 mg/mL disodium p-nitrophenyl phosphate in 0.05 M sodium bicarbonate; 0.05 M sodium carbonate 2:1 (v/v) containing 0.002 M magnesium chloride, was added to each plate and incubated at room temperature in the dark for approximately 30 min.

The plate was then read at 405-nm wavelength using an automated plate reader (Anthos Labtech; Salzburg, Austria). A standard curve was constructed at 405 nm against EndoCAb concentration, and the result of test samples was deduced from the curve. Results were rejected if the EndoCAb concentration of the control sample varied > 1 SD from its assigned value or the correlation coefficient for the standard curve was < 0.98.

VZV antibody analysis

These were assayed using a semi-quantitative test which enabled the positioning of the antibody levels relative to the other samples being tested (Biomerieux, Basingstoke, UK). The assay principle combines a two-step enzyme immunoassay sandwich method with a final fluorescent detection––enzyme-linked fluorescent assay (ELFA).

The solid-phase receptacle (SPR) acts as the solid phase, as well as the pipetting device. The reagents are ready-to-use and dispensed in the sealed reagent strips. The instrument performs all of the assay steps automatically, with the reaction medium being cycled in and out of the SPR several times.

After a preliminary wash step and dilution in the sample diluent, the sample is incubated in the SPR. Any anti-VZV antibodies present in the sample will bind to the VZV antigen coating on the interior of the SPR. Unbound components are removed in the subsequent wash step. Anti-human IgG antibodies conjugated with alkaline phosphatase will bind to any human IgG bound to the interior of the SPR. Washing removes unbound conjugate. During the final detection step, the substrate (4-methyl-umbelliferyl phosphate) is cycled in and out of the SPR. The bound conjugate enzyme present catalyses the hydrolysis of the substrate into a fluorescent product (4-methyl-umbelliferone), which is measured at 450 nm. The intensity of the fluorescence is proportional to the concentration of anti-VZV IgG antibodies in the sample. This value is automatically compared to the calibration curve and a result printed.

The result for each sample is expressed as the ‘test value’ (TV), a ratio of the ‘relative fluorescence value’ (RFV) of the sample to that of the stored value of the standard.

TV=RFVofsampleRFVofstandardstoredonsystem

Pre-operative risk scoring

EuroSCORE 2 values were obtained for each patient along with demographic information including age, gender, height, weight, body mass index, left ventricular ejection fraction, peak gradient across the aortic valve, presence of diabetes, presence of hypertension, NYHA class and glomerular filtration rate. All patients underwent routine screening prior to surgery for evidence of colonisation with methicillin resistant Staphylococcus aureus.

Diabetes was considered present if the patient had a pre-operative requirement for oral hypoglycaemic or insulin therapy. Left ventricular ejection fraction (LVEF), peak gradient across the aortic valve and estimated aortic valve area (AVA) were determined by standard 2D echocardiography performed within 6 weeks prior to surgery.

Outcomes

The primary outcome measure was post-operative length of stay in hospital. Secondary outcomes included length of stay on ICU, development of infection and mortality at 30 days. Peri-operative outcome data was collected regarding length of cardiopulmonary bypass, aortic close-clamp time, length of stay on the ICU, and on the development of infection, as well as survival up to 30 days. We used the Centre for Disease Control and Prevention (CDC) and National Healthcare Safety Network (NHSN) guidance to define the presence of a healthcare-associated infection (HAI) (Horan et al., 2008). Attributing whether or not there was a significant post-operative infection was determined by examining the following characteristics: We noted for all patients the clinical judgement by staff directly caring for the patient, direct observation of patients and wounds, and blood and microbiology laboratory results, whether samples were sent for culture of pathogenic organisms, whether any positive result was obtained, whether an antimicrobial agent was started and whether the clinical teams determined that there was a (HAI).

All patients received standard institutional antibiotic prophylaxis consisting of 1 g flucloxacillin and 1.5 mg/kg gentamicin.

Statistical analysis

Data are presented as absolute numbers and percentages for categorical variables and median and interquartile range for continuous variables. Continuous variables were compared by t tests and Mann-Whitney U tests whereas dichotomous variables were compared using Fisher’s exact test. Probabilities are two-tailed.

Given an average length of post-operative stay in our institution for uncomplicated valve surgery of 7 days, and a stay of 10 days or greater being considered prolonged (Bennett-Guerrero et al., 1997), we estimated that, at 80% power, with a 5% type II error rate, a sample size of 21 in each group would be sufficient to detect a 3-day increase in length of stay. This would require us to examine 84 patients in order to have sufficient numbers in each quartile.

After losing 32 patients from the initial cohort, we performed post hoc power calculations based on the demonstrated reduction in length of stay from 10 to 6 days. With 15 subjects in each quartile group, we are able to detect the change in length of stay with 80% power and 5% type II error rate. In a previous study comparing pre-operative EndoCAb levels, significant differences in outcome were demonstrated in a cohort of 60 patients (Hamilton-Davies et al., 1997).