All experiments were performed in accordance with the Declaration of Helsinki, after ethical approval (number #452/21, Local Independent Ethics Committee of the University of Ulm) and obtaining written informed consent of the patients. The inclusion criteria were the ability to provide written informed consent, full legal age, and planned surgical aortic valve replacement with extracorporeal circulation. Exclusion criteria were revision procedures, active malignant disease with ongoing radio-/chemotherapy, and/or immunomodulatory medication. Blood was drawn on admission (A), which was usually the day before surgery, during the surgery (labeled ‘OR’, 45 min after the initiation of the extracorporeal circulation), and at 24, 48, and 120 h after the end of surgery (± 10%). Blood was drawn by peripheral venipuncture, through an existing peripheral venous catheter (A), from the arterial and the venous branch of the extracorporeal circulation (OR), from peripheral arterial catheters (if available, normally 24 h and 48 h after OR), or peripheral venous catheters (normally 120 h after OR). Samples from venous and arterial branches of the extracorporeal circulation showed no significant differences (data not shown), only data from the arterial branch is reported. Moreover, age- (± 10%) and sex-matched healthy volunteers (HVs) were included in the present study. HVs were recruited by announcements on public boards at Ulm University Hospital and study facilities. The inclusion criteria were no fever in the last seven days, no preexisting acute or malignant diseases, no immunomodulatory medication, no instable chronic preexisting diseases, the ability to provide written informed consent, and full legal age. The HVs were taking the following medications at the time of blood sampling: One HV was taking levothyroxine, one HV was taking pantoprazole and valsartan/hydrochlorothiazide, one HV was taking rivaroxaban and bisoprolol, and one HV was taking candesartan.

Sodium, potassium, ionized calcium, and glucose were determined using a standard blood gas analyzer (ABL 800 Flex, Radiometer GmbH, Krefeld, Germany). Differential blood count (EDTA anticoagulated blood) and global coagulation parameters (activated partial thromboplastin time and the international normal ratio from citrate anticoagulated plasma) were determined using a standard hematology (Sysmex CN 2000, Sysmex, Kobe, Japan) and coagulation (BCS XP, Siemens, Marburg, Germany) analyzer, respectively, each according to the respective manufacturer’s standard protocol. C-reactive protein (CRP) was measured by a turbidimetric assay on a cobas c system (Roche, Penzberg, Germany). Procalcitonin (PCT), interleukin-6 (IL6), creatine phosphokinase-MB, and troponin-T were quantified by an electrochemiluminescence immunoassay on a cobas 8000/e 801 system (Roche). Urea, alanine transaminase, and creatine kinase were measured by a photometric assay on a Cobas c system (Roche) (all from lithium-heparin anticoagulated plasma). All the above-listed parameters were analyzed in cooperation with the Department of Clinical Chemistry of the University Hospital Ulm. For the analysis of myeloperoxidase (MPO) and matrix metalloproteinase 9 (MMP9), citrate anticoagulated whole blood was centrifuged for 10 min at 400 × g (adapted from (20)). The samples were stored at –80°C until further use. MPO was quantified using a LEGENDplex Human Vascular Inflammation Panel 1-S/P (#740809, Biolegend, San Diego, USA), while MMP9 was quantified by a standard enzyme-linked immunosorbent assay (#DY911, R&D Systems, Minneapolis, USA), in accordance to the instructions of the manufacturers.

A total of 10 µL citrate anticoagulated blood was diluted with 5 µL phosphate-buffered saline (PBS) with calcium and magnesium (PBS^+/+, #14040-091, Gibco, Thermo Fisher Scientific, Waltham, USA) and, if indicated in the figures, incubated with pharmacological inhibitors of 1 µM iloprost (#SML1651, Sigma-Aldrich, Steinheim, Germany), 1 mM ropivacaine (#ZYA1821/-22, Fresenius Kabi, Bad Homburg v. d. Höhe, Germany), or 2 µL anti-CD62P (#304904, Biolegend) or the respective buffer control for 10 min in a light-protected water bath at 37°C. Subsequently, the mixture was stimulated with either 1 µM PAF (#18779, Cayman Chemical Company, Ann Arbor, Michigan, USA) or buffer control in a total volume of 30 µL and incubated for 15 min at 37°C. The PAF concentration was chosen based on previous data to ensure maximal stimulation (13). Thereafter, for phenotype analysis, staining was performed with antibodies (all from BioLegend) against CD10 (PE-Cy7, 120 ng/mL, #312214; clone HI10a; isotype #400126), CD11b (APC, 600 ng/mL, #101212; clone M1/70; isotype #400612), and CD66b (APC-Cy7, 1 µg/mL, #305126; clone G10F5; no isotype available). For the analysis of the phagocytic activity, ROS generation, and PNC formation, 100 µL/mL fluorescent microspheres (Fluoresbrite BB Carboxylate 0.50 Micron Microspheres, Polysciences, Inc., Warrington, USA), 5 µM CellROX Deep Red (#C10422, Thermo Fisher Scientific), and an antibody against CD61 (PerCP, 2 µg/mL, #336410; clone VI-PL2; isotype #400148), respectively, were added as previously described (21). The samples were made up to a total volume of 50 µL by adding PBS^+/+ and were incubated for 15 min in a light-protected water bath at 37°C. The stained and stimulated blood was transferred to 950 µL BD FACS Lysing Solution (#349202, BD Biosciences, San Jose, USA) and incubated for 30 min at room temperature in the dark. Following centrifugation at 340 × g for 5 min, the supernatant was discarded and the cells were resuspended in 100 µL PBS without calcium or magnesium (PBS ^−/− , #14190-094, Thermo Fisher Scientific) but with 1% bovine serum albumin (BSA, #A8022, Sigma Aldrich) Finally, the samples were maintained at 4°C in the dark until measurement.

For flow cytometric analysis, PNCs were identified as CD61⁺-neutrophils as previously described (13, 21). For analysis by light microscopy (Axio Imager M1, Carl Zeiss Microscopy GmbH, Jena, Germany), 250 µl of citrate anticoagulated blood was diluted with 250 µL PBS^+/+ and stimulated with either PBS^+/+ as buffer control or 1 µM PAF. Blood smears were stained with the ‘Hemacolor Rapid staining of blood smear - staining set for microscopy’ (#111661, Merck, Darmstadt, Germany). For each sample, a minimum of 50 intact neutrophils per specimen were analyzed. Each neutrophil with at least one thrombocyte in direct proximity was considered to be a PNC (13, 21, 22).

Granulocytes mainly consisting of neutrophil granulocytes (hereafter referred to as neutrophils) were purified by Ficoll separation, dextran sedimentation, and hypotonic lysis as previously described (14, 18). In brief, 9 mL of citrate anticoagulated blood was centrifuged for 10 min at 400 × g. After removing the plasma, the remaining blood cells were mixed with 0.9% sodium chloride (#1312813, Fresenius Kabi) up to a total volume of 20 mL, which was subsequently layered on ficoll (#17144003, Cytiva Sweden AB, Uppsala, Sweden) and centrifuged for 30 min (400 × g) followed by dextran sedimentation. Isolated granulocytes (purity usually > 95% as indicated by flow cytometry) were adjusted to a final concentration of 1 × 10⁶ cells/mL and resuspended in Hank’s balanced salt solution with calcium and magnesium (HBSS^+/+, #14025050, Thermo Fisher Scientific, Darmstadt, Germany) containing 15 mM 2-(4-(2-hydroxyethyl)-1-piperazinyl)-ethanesulfonic acid (HEPES, #7365-45-9, Sigma). Cells were stained with 50 nM bis(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC₄(3), #D8189, Merck, for measuring the MP) and 1 µM SNARF 5-(and-6)-carboxy-SNARF-1 (SNARF, #C1272, Invitrogen Thermo Fisher Scientific, Dreieich, Germany, for measuring pH_i) in HBSS^+/+ containing HEPES and maintained in a light-protected water bath at 37°C. After 20 min, the cells were centrifuged (5 min, 340 × g, room temperature) and resuspended in PBS^+/+, followed by another incubation period of 10 min with 50 nM DiBAC₄(3) before stimulation and measurement (14, 18). The granulocytes were stimulated with either 1 µM PAF or the respective buffer control as indicated. Cells were analyzed after 1, 5, and 10 min. The graphs report the cellular response at the time of the maximum for the respective parameter (1 min for MP, 5 min for pH_i, 10 min for forward scatter (FSC)). Of note, the FSC is only a brief indicator for cellular size, because it more likely reflects a change in cellular shape as discussed before (13, 23).

A total of 10 µL citrate anticoagulated whole blood from healthy human volunteers (HVs, aged 21 – 23 years, otherwise inclusion criteria see above) was exposed to the drugs listed in Supplementary Table 4 or PBS^+/+ as buffer control for 10 min in a light-protected water bath at 37°C. The drug doses were determined by values reported in the literature and increased approximately 10- to 30-fold to briefly explore their effects on granulocytes and platelets. Subsequently, the samples were stained with anti-CD61 and anti-CD11b (as described above) and simultaneously exposed to 1 µM PAF for 30 min at 37°C in a light-protected water bath. The stained and stimulated blood was transferred to 950 µL BD FACS Lysing Solution for 30 min at room temperature in the dark. Following centrifugation at 340 × g for 5 min, the supernatant was discarded, the cells were resuspended in PBS^–/– containing 1% BSA, and the samples were maintained in the dark at 4°C until analysis by flow cytometry.

To analyze platelet activation, 50 μL of citrate-anticoagulated blood was diluted with 562.5 µL HBSS^+/+ as previously described (21). Subsequently, 10 µL of this diluted blood was added to 40 µL PBS^+/+ including prior added stimuli (either buffer control or 1 µM PAF) and 1µL anti-CD61 and anti-CD62P (FITC, 8 µg/mL, #304904; isotype #400108, BioLegend). Following incubation for 10 min in a light-protected water bath at 37°C, 950 µL HBSS⁺⁺ were added to the sample followed by immediate flow cytometry analysis. Platelets were identified by the properties of FSC, side scatter (SSC), and CD61 expression and were gated as described before (21).

Doublets were removed by plotting the FSC area versus the height followed by analyzing linearity of both parameters. Neutrophils were identified based on their FSC and SSC area properties. The spillover between the fluorescence channels was corrected by a compensation matrix. For all antigens, appropriate isotype controls and single staining controls were performed (data not shown). For all experiments, a minimum of 3000 neutrophils or platelets were recorded using a BD FACSLyric (BD Biosciences) within a predefined stopping timer of 120 seconds.

The flow cytometry data were analyzed using the custom-written, python-based flow cytometry analytics software ‘BFlow’ (BFlow Project, www.bflow.science, last accessed 05 April 2024). All data is presented as median with bars indicating the interquartile range, for example, median (25^th percentile | 75^th percentile), if not indicated otherwise. Figure 1 was partially created with biorender.com. Data analysis was performed with licensed versions of Microsoft Excel 2019 (Microsoft, Redmond, USA) and GraphPad Prism 10 (GraphPad Software Inc., San Diego, USA). The study is intended to be an exploratory monocentric approach to monitor neutrophil function in perioperatively in patients with cardiac surgery. Therefore, several assumptions, including the justifications provided, were made. As most biological data are typically assumed to be normally distributed, and given that neither theoretical nor empirical evidence suggests otherwise, we performed a visual inspection of the present dataset to confirm this assumption. Consequently, a parametric approach was used for the statistical analysis. However, the resulting findings should be interpreted with these assumptions in mind, as they may overestimate the significance of the reported results. Furthermore, it could be argued that a correction for multiple testing would have been necessary; such a correction was not performed due to the exploratory nature of this study and the disparate nature of the parameters analyzed. A sample size calculation and effect size estimation were not performed because, to the authors’ knowledge, no information regarding the biological relevance of changes in the investigated parameters is currently available. For the statistical analysis comparing HVs with patients on admission, an unpaired t test was performed. The statistical description for patients during their course of stay was conducted by setting the admission time point as reference, the other time points were compared by applying an ordinary one-way analysis of variance (ANOVA). Further statistical testing was conducted as described in the figure legends. A p value < 0.05 was considered to be significant and marked with *, **, ***, and ****, indicating < 0.05, < 0.01, < 0.001, and <0.0001, respectively.

Fourteen patients were included as indicated in Figure 1 , of whom 11/14 had preexisting arterial hypertension, 9/14 had hyperlipidemia, and 5/14 had preexisting type 2 diabetes mellitus. The duration of the operation including anesthesiology was 300 (270; 348) min and excluding anesthesiology (incision – suture) was 234 (173; 254) min. The duration of the extracorporeal circulation in total was 132 (123; 153) min and the ischemia time was 89 (77; 100) min. Access to the operation site was prepared by traditional sternotomy in 8/14 cases, by partial (upper) sternotomy in 3/14 cases, and by minimally invasive anterolateral thoracotomy in 3/14 cases. Replacement of the aortic valve was conducted by implanting biological valves in 9/14 cases and mechanical valves in 5/14 cases. An additional procedure was required in 7/14 patients, such as supracoronary ascending aortic replacement and/or bypass procedures. Apart from two patients requiring revision (one due to asymptomatic pericardial effusion treated by inferior pericardiotomy, one due to third-degree atrioventricular block treated by implantation of a pacemaker), no major complications were observed. The median stay in intensive care was 3 (2; 5) days postoperatively and the total length of the hospital stay was 9 (7; 11) days postoperatively. Of the 14 patients, 12 were discharged home, 1 was transferred to another hospital, and 1 was referred to a rehabilitation facility.

Basic monitoring of inflammation revealed largely unchanged leukocyte count, IL6, and PCT levels during surgery ( Figure 1 ; Supplementary Table 1 ), but were increased postoperatively with peaks around the first or second postoperative day. By contrast, the MMP9 and MPO as markers of humoral inflammation were significantly elevated during surgery but returned to baseline 24 h post surgery ( Figure 1 ). On a cellular level, erythrocytes and thrombocytes decreased slightly as expected during surgery ( Supplementary Tables 1 , 2 ).

The elevation of the markers of humoral inflammation was clearly preceded by changes in the neutrophil phenotype ( Figure 2 ). In detail, CD10 downregulation and CD11b and CD66b upregulation showed maximal changes intraoperatively. While CD10 remained decreased during the first two postoperative days, CD11b and, to a lesser extent, CD66b returned close to baseline levels one day postoperatively. Of note, neutrophil PAF-induced responsiveness with regard to CD10 but not to CD11b nor CD66b was impaired during the postoperative observation period. CD11b expression on neutrophils was significantly reduced by the addition of ropivacaine but not iloprost or an antibody against CD62P in neutrophils with or without additional exposure to PAF ( Table 1 ). Of note, CD10 and CD11b largely did not correlate with conventional biomarkers postoperatively ( Supplementary Table 3 ).

In an exploratory in vitro screening of relevant drugs during anesthesia and surgery, no pharmacological agent reduced CD11b expression in a relevant manner ( Table 2 ). By contrast, protamine and the combination of protamine and heparin increased CD11b expression ( Table 2 ).

Subsequently, changes in neutrophil physiology with respect to the PAF-induced responses of the MP and pH_i were assessed. In this regard, PAF-induced depolarization remained stable during surgery but was severely impaired on the first two postoperative days ( Figure 3 ). By contrast, neither baseline levels nor PAF-induced changes of pH_i and cellular shape were recorded perioperatively ( Supplementary Figures 1 , 2 ).

Neutrophil function regarding PAF-induced ROS generation remained stable despite a small but nonsignificant reduction during surgery ( Figure 3 ). Phagocytic activity in resting cells was unaltered perioperatively. However, PAF-stimulated neutrophils exhibited reduced phagocytic activity intraoperatively ( Figure 3 ).

Surprisingly, PNC formation was strongly impaired intraoperatively but not pre- or postoperatively ( Figure 4 ). This phenomenon could be confirmed by two independent techniques (flow cytometry and conventional light microscopy). Moreover, PNC formation was also massively restricted despite the presence of PAF ( Figure 4 ). In accordance, the PAF-elicited increase in CD62P-positive platelets was significantly diminished intraoperatively ( Figure 4 ). Of note, in PNCs induced by PAF, the number of platelets attached to neutrophils did not differ significantly intraoperatively (data not shown). PNC formation was diminished by ropivacaine, iloprost, and an antibody against CD62P ( Table 1 ). In blood from HVs that was stimulated with PAF, the addition of drugs relevant to anesthesia and surgery did not reduce PNC formation in a relevant manner, despite the addition of high doses of heparin ( Table 2 ).

In a more detailed analysis, the differences between PNCs and PMNs (neutrophils without attached platelets) were investigated. With regard to variations in the phenotype, PMN and PNCs did not differ regarding CD10, CD11b, or CD66b, despite an increase in CD11b in PNCs intraoperatively in comparison to PMN ( Figure 5 , Supplementary Figures 3 , 4 ). However, PNCs exhibited a strong augmentation in ROS generation and phagocytic activity before, during, and after cardiac surgery ( Figure 6 ).