In situ exploration of endothelial function in vasoplegic syndrome after cardiac surgery with cardiopulmonary bypass: protocol for a single-centre prospective cohort study – the Vasoshock study Authors: Nicolas Ferry, Domitille Renard, André Gillibert, Thomas Clavier, Fabienne Tamion, Adel Abedi, Jean Selim, Jeremy Bellien, Emmanuel Besnier Published: 2026 ABSTRACT Abstract AbstractIntroductionVasoplegic syndrome (VS) is a severe complication after cardiac surgery with cardiopulmonary bypass (CPB), characterised by vasodilation, hypotension and vasopressor dependence. Endothelial dysfunction, dysregulated nitric oxide signalling, inflammation and glycocalyx injury may contribute. Combining flow-mediated dilation (FMD), sublingual side-stream dark-field (SDF) imaging and biomarkers may improve postoperative vascular phenotyping. The primary objective is to compare allometrically corrected brachial artery FMD between patients with VS and patients without VS 48 hours after CPB weaning.Methods and analysisThis prospective, single-centre observational cohort study will enrol 124 adults aged 18–80 years undergoing cardiac surgery with CPB (62 VS and 62 non-VS). VS is defined clinically as norepinephrine bitartrate ≥0.05 µg/kg/min for at least 4 hours to maintain mean arterial pressure ≥65 mm Hg after CPB weaning, following haemodynamic optimisation and exclusion of a predominant low-cardiac-output phenotype or another cause of shock. FMD, SDF imaging and biomarkers will be assessed at 4–8 and 48 hours after CPB weaning. The primary comparison will use multiple linear regression for corrected FMD at 48 hours with prespecified covariate adjustment. Secondary and exploratory analyses will assess SDF-derived parameters, biomarkers and associations across modalities.Ethics and disseminationThe protocol was approved by the Comité de Protection des Personnes Sud-Méditerranée III (CPP Sud-Méditerranée III; 2023-A02079-36). For this non-interventional routine-care study, French law requires prior oral and written information and absence of opposition rather than written consent; absence of opposition will be documented. Results will be published in peer-reviewed journals and presented at scientific conferences.Trial registration numberNCT06318689.Keywords: Cardiac surgery, Anaesthesia in cardiology, Vasoconstrictor Agents, Cardiovascular imaging Strengths and limitations of this study Brachial artery flow-mediated dilation (FMD) is measured with a standardised conduit-artery ultrasound protocol, allometric scaling and adjustment for the eliciting shear stimulus. Brachial artery flow-mediated dilation (FMD) is measured with a standardised conduit-artery ultrasound protocol, allometric scaling and adjustment for the eliciting shear stimulus. The primary analysis compares corrected FMD at 48 hours after cardiopulmonary bypass weaning using prespecified covariate adjustment. The primary analysis compares corrected FMD at 48 hours after cardiopulmonary bypass weaning using prespecified covariate adjustment. Sublingual side-stream dark-field measurements use repeated bedside recordings with automated quality control. Sublingual side-stream dark-field measurements use repeated bedside recordings with automated quality control. Vasoplegic syndrome is defined pragmatically following haemodynamic optimisation, without systematic cardiac output assessment, which may allow some residual misclassification. Vasoplegic syndrome is defined pragmatically following haemodynamic optimisation, without systematic cardiac output assessment, which may allow some residual misclassification. The single-centre observational design and recruitment when a trained investigator is available may limit generalisability and introduce selection bias. The single-centre observational design and recruitment when a trained investigator is available may limit generalisability and introduce selection bias. Introduction Vasoplegic syndrome (VS) is characterised by a significant reduction in systemic vascular resistance despite normal or elevated cardiac output, leading to tissue hypoperfusion and potential progression to shock.1 2 VS is frequently encountered in patients undergoing cardiac surgery with cardiopulmonary bypass (CPB), with reported incidence rates varying from 10% to 60% depending on patient characteristics and definitions used.3–6 Clinically, VS manifests as profound hypotension refractory to fluid resuscitation and requiring vasopressors,7 8 often leading to increased morbidity, prolonged intensive care unit (ICU) stay and higher mortality rates.3 4 9 The pathophysiology of VS is complex and multifactorial. CPB exposes circulating blood to artificial surfaces, abnormal shear stress, temperature variations and ischaemia-reperfusion events. These factors trigger a systemic inflammatory response, activating leukocytes, platelets and endothelial cells.10 11 This inflammatory response promotes the expression of inducible nitric oxide synthase (iNOS), leading to excessive nitric oxide (NO) production.12 While NO derived from endothelial nitric oxide synthase (eNOS) is crucial for maintaining vascular homeostasis, excessive NO release via iNOS can cause profound vasodilation and vascular hyporesponsiveness, contributing to VS. Another critical component is the endothelial glycocalyx, a specialised carbohydrate-rich layer lining the luminal surface of endothelial cells. This structure is essential for mechanotransduction, converting shear stress into biochemical signals that regulate NO production through eNOS in normal conditions.13–15 Cardiac surgery with CPB has been associated with glycocalyx degradation.16 Given its central role in endothelial homeostasis, disruption of the glycocalyx may impair vascular reactivity and other vascular functions, leading to adverse consequences such as increased capillary permeability, microvascular perfusion heterogeneity and further endothelial dysfunction. Macrovascular endothelial vasoreactivity can be assessed non-invasively using flow-mediated dilation (FMD) of the brachial artery.17 Under standardised conduit-artery conditions, particularly with distal cuff occlusion for 5 min, FMD is considered a measure of endothelium-dependent vasoreactivity with a substantial nitric oxide contribution.18 FMD has been widely used to assess endothelial function in cardiac surgery. Previous studies have demonstrated, using FMD, that endothelial function is impaired after continuous-flow CPB compared with both off-pump surgery and pulsatile-flow CPB.19 Additionally, preoperative atorvastatin administration for 3 weeks showed a protective effect on endothelial function.20 While FMD provides valuable information on conduit-artery endothelial vasoreactivity, it does not directly assess glycocalyx-related alterations or the microvascular bedside phenotype. Sublingual side-stream dark-field (SDF) imaging allows in vivo, non-invasive visualisation of the microcirculation and provides parameters such as the perfused boundary region (PBR), a glycocalyx-related parameter reflecting red blood cell (RBC) penetration into the endothelial surface layer.21 22 Changes in PBR, capillary density and relative capillary blood volume may provide complementary insights into the postoperative vascular phenotype associated with VS. To complement these functional assessments, circulating biomarkers will also be measured. Interleukin-6 (IL-6), a pro-inflammatory cytokine, is typically elevated during acute inflammatory responses, including in the postoperative setting.23 Syndecan-1, a structural component of the endothelial glycocalyx, is released into the circulation on glycocalyx degradation, and serves as a marker of endothelial injury.24 Finally, E-selectin is an endothelial adhesion molecule upregulated in response to inflammatory stimuli, reflecting endothelial activation and dysfunction.25 By integrating conduit-artery FMD with complementary SDF-derived glycocalyx-related and microvascular parameters, as well as relevant biomarkers (IL-6, Syndecan-1, E-selectin), this study aims to provide a multimodal postoperative assessment of endothelial function in VS, to characterise postoperative vascular phenotypes and to explore relationships between these modalities in order to generate mechanistic hypotheses for future targeted studies. Objectives of the study The primary goal of this study is to compare brachial artery FMD, corrected using allometric scaling, between patients developing VS and patients without VS at 48 hours after cardiac surgery with CPB, within a broader multimodal assessment of postoperative endothelial function. The secondary goals are to compare early postoperative FMD (4–8 hours), SDF-derived parameters and endothelial biomarker levels between patients with VS and patients without VS, and to explore the relationships between these complementary modalities and postoperative clinical outcomes. Methods and analysis Study design and setting This study is a single-centre, prospective, observational cohort conducted in the cardiac surgical ICU of the University Hospital of Rouen, France. Participant enrolment began in February 2024 after Comité de Protection des Personnes approval. The study was registered on ClinicalTrials.gov. The study was initially planned for a 2-year recruitment period. A 12-month extension has been approved by the ethics committee because of the limited number of eligible patients, investigator availability constraints and technical issues. Eligible adult patients undergoing cardiac surgery with CPB will be enrolled during periods when a trained investigator is available to perform bedside vascular measurements. Patients will be classified postoperatively as VS or non-VS. Recruitment will continue until the planned number of patients in each group is reached, within the constraints of investigator availability. Patient selection Inclusion criteria Age 18–80 years. Age 18–80 years. Hospitalised in the cardiac surgical ICU after cardiac surgery with CPB (eg, coronary artery bypass grafting, valvular surgery, combined procedures). Hospitalised in the cardiac surgical ICU after cardiac surgery with CPB (eg, coronary artery bypass grafting, valvular surgery, combined procedures). Inclusion feasible within 4–8 hours after the end of CPB (availability of a trained investigator for the measurements). Inclusion feasible within 4–8 hours after the end of CPB (availability of a trained investigator for the measurements). Non-inclusion criteria Suspected/confirmed sepsis at inclusion. Suspected/confirmed sepsis at inclusion. Heart transplantation. Heart transplantation. Mechanical circulatory support (extracorporeal membrane oxygenation, catheter-based microaxial flow pump, left ventricular assist device, others). Mechanical circulatory support (extracorporeal membrane oxygenation, catheter-based microaxial flow pump, left ventricular assist device, others). Patients with specific systemic disorders known to significantly alter the microcirculation, including connective tissue diseases (eg, systemic lupus erythematosus, systemic sclerosis), systemic vasculitides or haemoglobinopathies. Patients with specific systemic disorders known to significantly alter the microcirculation, including connective tissue diseases (eg, systemic lupus erythematosus, systemic sclerosis), systemic vasculitides or haemoglobinopathies. Contraindications to the tools used (eg, bilateral upper limb arteriopathy or neuropathy, bilateral axillary lymph node dissection). Contraindications to the tools used (eg, bilateral upper limb arteriopathy or neuropathy, bilateral axillary lymph node dissection). Individuals deprived of liberty by administrative or judiciary decision or placed under legal protection/guardianship. Individuals deprived of liberty by administrative or judiciary decision or placed under legal protection/guardianship. Unavailability of an investigator trained in the exploration techniques (FMD, SDF imaging). Unavailability of an investigator trained in the exploration techniques (FMD, SDF imaging). Patient opposition to participation. Patient opposition to participation. Post-inclusion exclusion from primary endpoint analysis Patients requiring unplanned reoperation before the 48-hour assessment will be excluded from the primary endpoint analysis, because V2 endothelial measurements may be strongly affected by the new surgical insult and may not reflect the initial postoperative vascular phenotype. Patients requiring unplanned reoperation before the 48-hour assessment will be excluded from the primary endpoint analysis, because V2 endothelial measurements may be strongly affected by the new surgical insult and may not reflect the initial postoperative vascular phenotype. Definition of VS While the standard definition of VS relies on low systemic vascular resistance, systematic direct assessment of cardiac output and systemic vascular resistance is not feasible in all patients within routine postoperative care. Therefore, a pragmatic clinical definition will be used in this study. VS will be defined as the need for norepinephrine bitartrate at a dose ≥0.05 µg/kg/min for at least 4 hours to maintain a mean arterial pressure (MAP) ≥65 mm Hg after CPB weaning, following haemodynamic optimisation according to routine ICU practice. Haemodynamic optimisation includes assessment of intravascular filling using echocardiography when available and/or fluid responsiveness tests, together with routine evaluation of perfusion status and postoperative myocardial function. Patients will not be classified as having VS if vasopressor requirement is judged to be primarily explained by a predominant low-cardiac-output phenotype, ongoing major bleeding, tamponade, uncontrolled arrhythmia, need for mechanical circulatory support or another obvious alternative cause of shock. This assessment will be based on routinely available clinical, biological, haemodynamic, inotropic and echocardiographic data when performed. The minimum duration of 4 hours was chosen to reduce misclassification due to transient hypotensive episodes, treatment delays or perioperative arrhythmias. The selected dose threshold and duration were designed to identify clinically relevant vasoplegia while remaining consistent with the heterogeneous but comparable vasopressor cutoffs reported in studies of post-cardiac surgery vasoplegia.3 9 26 27 Anaesthesia and intraoperative management under standard care Anaesthesia induction and maintenance will follow local protocol, using balanced techniques to ensure haemodynamic stability and adequate depth of anaesthesia. Induction will be achieved with a target-controlled infusion (TCI) of propofol (effect-site concentration 1.5–3 µg/mL) or intravenous etomidate (0.2–0.5 mg/kg), combined with a neuromuscular blocker (cisatracurium 0.2 mg/kg). Maintenance of anaesthesia will involve TCI of propofol aiming for a Patient State index (PSI) between 25 and 50 (SedLine, Masimo, France), TCI of remifentanil 1–4 ng/mL based on Minto model and cisatracurium 0.1 mg/kg/hour. At induction, patients will also receive dexamethasone (0.1 mg/kg), ketamine (0.2 mg/kg) and lidocaine (1.5 mg/kg), followed by continuous infusions of ketamine (0.2 mg/kg/hour) and lidocaine (1.5 mg/kg/hour). Ketamine will be stopped at the end of surgery, whereas lidocaine will be maintained at 1.33 mg/kg/hour for the first 24 postoperative hours. This institutional protocol is consistent with French postoperative pain-management guidelines.28 Standard monitoring includes invasive arterial pressure, electrocardiography, pulse oximetry, capnography, central venous access (jugular or subclavian), bladder or rectal temperature monitoring and PSI. Anticoagulation will be achieved with intravenous heparin (10 000 IU/m²±5000 IU) to maintain an activated clotting time of at least 400 s prior to the initiation of CPB. Neutralisation of heparin will be performed using protamine with a 1:1 ratio for each unit administered within the last hour and a 0.6:1 ratio for each unit after the first hour. CPB will be performed using a membrane oxygenator and a roller pump (Stockert S5 and Inspire 6 or 8, LivaNova, France), with an indexed blood flow between 2.4 and 3.0 L/min/m² to maintain a MAP between 40 and 80 mm Hg and an arterial oxygen delivery (aDO2) above 280 mL/min/m². Haemoglobin will be maintained above 70 g/L, with a higher threshold used if aDO2 remains <280 mL/min/m² despite adequate pump flow. CPB management will follow institutional protocols consistent with the European Association for Cardio-Thoracic Surgery (EACTS) guidelines.29 Myocardial protection will be conducted using either histidine-tryptophan-ketoglutarate based crystalloid (Perisoc, SERB, France) or 4°C potassium-esmolol-magnesium enriched blood cardioplegia. Surgery will be performed at a body temperature (≥35°C). On completion of the surgical procedure, patients will be weaned from CPB with inotropic or vasopressor support if clinically indicated. In the event of hypotension (MAP ≤65 mm Hg) outside the CPB period, vasopressors such as ephedrine, phenylephrine or norepinephrine will be administered, with fluid resuscitation as needed. Intraoperative hypertension will be managed with titrated doses of nicardipine (1 mg) or urapidil (10 mg), according to predefined thresholds agreed on by the anaesthesia and surgical teams. Antibiotic prophylaxis will include cefazolin 2 g (with 1 g every 4 hours and an additional 1 g added to the CPB priming solution) or, in case of allergy, vancomycin 30 mg/kg or teicoplanin 12 mg/kg. Postoperatively, patients will be transferred under mechanical ventilation to the cardiac surgical ICU. Sedation with propofol will be continued until haemodynamic stability and normothermia are achieved, and bleeding is controlled. Extubation will be performed in awake patients with spontaneous breathing, provided they meet these criteria: pressure support ≤10 cm H2O, PEEP (positive end-expiratory pressure) ≤5 cm H2O and absence of respiratory distress. For patients ventilated >48 hours, a spontaneous breathing trial will be required (pressure support 7–8 cm H2O with zero PEEP), and extubation will be performed if no signs of intolerance (eg, tachypnoea, hypoxemia, agitation, tachycardia or hypertension) occur within 30 min. Supplemental oxygen may be administered to maintain SpO2 between 92% and 96%. Haemodynamic optimisation in the ICU will include fluid infusion guided by echocardiography, continuous blood pressure monitoring, clinical and biological perfusion assessment, and additional cardiac output monitoring when clinically indicated (eg, FloTrac, Edwards Lifesciences, France). Vasopressors and inotropes will be titrated based on clinical needs to maintain a MAP ≥65 mm Hg and, when cardiac output monitoring is available, a cardiac index ≥2.2 L/min/m², together with echocardiographic assessment and other perfusion metrics such as ScvO2, mottling and lactate levels. Postoperative analgesia during the first 48 hours will include paracetamol 1 g every 6 hours (orally or intravenously), a continuous nefopam infusion of 80–120 mg/day, non-steroidal anti-inflammatory drugs in the absence of contraindications (ketoprofen every 8 hours at a dose of 100 mg, or 50 mg in patients weighing <70 kg) and intravenous morphine titration followed by as-needed doses. FMD and data processing FMD will be assessed in accordance with the 2019 expert consensus recommendations.17 Imaging of the brachial artery will be performed using a Philips Affinity 70G system (Philips, France) equipped with a 3–12 MHz high-resolution linear-array transducer. After a 1 min baseline recording, a pressure cuff placed on the forearm will be inflated to 50 mm Hg above the patient’s systolic blood pressure for 5 min to induce distal ischaemia. Rapid cuff deflation will then trigger reactive hyperaemia, with subsequent continuous recording for 3 min to capture shear stimulus, peak dilation and return towards baseline. Continuous recordings will be acquired using a dedicated acquisition device (HD video recorder HVO-550MD, SONY, France). Blood flow velocity will be simultaneously recorded using pulsed-wave Doppler at the centre of the artery lumen, with a fixed insonation angle of 60°, ensuring alignment with flow direction. All ultrasound recordings will be analysed offline using automated wall-tracking software (Cardiovascular Suite, version 4.2.1, Quipu srl, Pisa, Italy), providing continuous edge detection of the arterial walls at a sampling rate of 30 Hz and synchronised velocity data. Offline analyses will be performed by trained operators blinded to VS status and clinical outcomes. Two operators will undergo formal, standardised training prior to study initiation and will perform all acquisitions. An articulated arm will secure the ultrasound probe in a fixed position during measurements to minimise movement artefacts. Whenever feasible, FMD acquisition will be performed during a clinically stable period with no major changes in vasoactive support during the preceding 30 min. Norepinephrine dose, inotropic support, invasive ventilation status and ongoing propofol sedation will be recorded at each visit. The technical validity of FMD recordings will be assessed during offline video analysis. If image quality, probe stability or the Doppler signal is insufficient for analysis, the corresponding FMD measurement time point will be recorded as not available. No systematic repeat acquisition is planned, because failed recordings may only be identified during post hoc video analysis and repeated cuff occlusion may influence subsequent vascular responses. Allometric correction of FMD Since baseline arterial diameter significantly influences the magnitude of FMD, potentially confounding between-subject comparisons, an allometric scaling approach will be applied to correct for differences in baseline artery size. As previously described,30 31 we will model the relationship between baseline and post-hyperaemic diameters using the following log-log linear regression model: The allometric exponent β will be estimated from the entire study cohort, assuming a common scaling exponent across patients and visits. Allometrically adjusted FMD values will then be computed for each measurement as follows: This allometrically scaled index will be used as the corrected FMD endpoint for between-group comparisons. It is derived from the same baseline-to-peak diameter response and has the same directional interpretation as conventional FMD (higher values indicate greater dilation), but it is not formally a percentage and is not numerically interchangeable with percentage-point differences. Conventional FMD (%) will therefore also be reported descriptively to support interpretability. Adjustment for shear stimulus Postoperative haemodynamic changes, particularly reduced cardiac output, may attenuate reactive hyperaemia, resulting in lower shear stress and potentially underestimating FMD responses independently of endothelial function. To account for interindividual differences in hyperaemic exposure, cumulative shear stimulus will be quantified as the area under the shear-rate curve from cuff release to the time of peak dilation. This shear stimulus parameter will be included as a prespecified covariate in FMD models to reduce the risk that between-group differences simply reflect differences in reactive hyperaemia, including postoperative haemodynamic influences such as reduced cardiac output, rather than conduit-artery endothelial vasoreactivity itself. We will not directly normalise FMD to shear stress, thereby avoiding additional ratio-induced bias. Sublingual SDF imaging Glycocalyx-related microvascular assessment Sublingual glycocalyx-related and microcirculatory parameters will be assessed at the bedside using an SDF camera (CapiScope HVCS, KK technology, Honiton, UK), with recordings analysed using GlycoCheck software, V.5.2 (Microvascular Health Solutions, Salt Lake City, Utah, USA). This system has been used and described in previous studies.21 22 The system provides real-time assessment of microvascular morphology and glycocalyx barrier properties through automated analysis of sublingual video sequences. After acquisition, the software analyses over 3000 vascular segments from 4 to 25 µm in diameter, automatically extracting the following parameters: PBR (µm): estimates the penetration depth of RBCs into the luminal part of the endothelial glycocalyx. Higher PBR values are generally interpreted as reflecting deeper RBC penetration into the endothelial surface layer and reduced glycocalyx barrier properties. PBR (µm): estimates the penetration depth of RBCs into the luminal part of the endothelial glycocalyx. Higher PBR values are generally interpreted as reflecting deeper RBC penetration into the endothelial surface layer and reduced glycocalyx barrier properties. Capillary density (µm/µm²): corresponds to the number of capillary segments identified within the analysed field, for vessels with diameters between 4 and 25 µm. Capillary density (µm/µm²): corresponds to the number of capillary segments identified within the analysed field, for vessels with diameters between 4 and 25 µm. Absolute capillary blood volume (µm³): represents the total capillary blood volume, calculated as the product of the number of capillary segments, their individual length and their segment-specific cross-sectional area. Absolute capillary blood volume (µm³): represents the total capillary blood volume, calculated as the product of the number of capillary segments, their individual length and their segment-specific cross-sectional area. Relative capillary blood volume: defined as the ratio between capillary blood volume in true capillaries (4–7 µm) and larger feeding vessels (10–25 µm), providing an estimate of functional microvascular density and perfusion distribution. Relative capillary blood volume: defined as the ratio between capillary blood volume in true capillaries (4–7 µm) and larger feeding vessels (10–25 µm), providing an estimate of functional microvascular density and perfusion distribution. For each patient, at least three valid independent recordings will be obtained from different sublingual regions to account for local variability and improve reproducibility.22 All acquisitions will be performed by trained operators following manufacturer recommendations, during a clinically stable period whenever feasible. The GlycoCheck software provides real-time feedback on retained sequences and automated quality control, including checks related to focus, illumination, motion stabilisation and artefact rejection. Additional sequences will be acquired during the visit, when needed and feasible, until three valid recordings are available. If fewer than three valid recordings are obtained, SDF-derived variables for that visit will be recorded as not available. Focus on the PBR Among these microcirculatory parameters, PBR is considered the main SDF-derived glycocalyx-related parameter in the present study, whereas other SDF-derived parameters will be considered complementary exploratory descriptors. PBR will be interpreted as an indirect in vivo marker of glycocalyx-related barrier alterations rather than as a direct measurement of glycocalyx thickness. Biomarkers Whole blood will be collected into ethylenediaminetetraacetic acid tubes at each visit through a venous or arterial catheter. Samples will be centrifuged immediately for 15 min at 2000×g and 4°C. Plasma aliquots will be stored at −20°C for up to 2 weeks and then at −80°C until analysis. IL-6, Syndecan-1 and E-selectin will be measured using enzyme-linked immunosorbent assays (Roche Diagnostics, Meylan, France). Clinical data Demographics, intraoperative variables (eg, CPB duration, aortic clamping duration), ICU data (eg, weight, diuresis, fluid balance, vasopressor dosing, haemodynamic monitoring), clinical outcomes (eg, acute kidney injury (AKI), ICU length of stay) and body composition assessed using bioelectrical impedance analysis (BIA) with the body composition monitor (Fresenius Medical Care, Bad Homburg, Germany) will be recorded prospectively. Data collection timeline Visit 1 (V1), 4–8 hours after CPB:FMD measurement.SDF imaging (PBR, capillary density, relative capillary blood volume).BIA.Blood sampling will be performed as part of routine patient management, allowing biomarker sampling (IL-6, Syndecan-1, E-selectin) and measurement of other biological data of interest (eg, blood gas analysis, N-terminal pro-B-type natriuretic peptide, liver function tests). The measurement context will also be documented, including vasoactive and inotropic support, invasive ventilation status and ongoing sedation.Clinical data assessment (eg, weight, diuresis, intravenous fluid therapy volumes and types, blood pressure, heart rate, body composition). Visit 1 (V1), 4–8 hours after CPB: FMD measurement. FMD measurement. SDF imaging (PBR, capillary density, relative capillary blood volume). SDF imaging (PBR, capillary density, relative capillary blood volume). BIA. BIA. Blood sampling will be performed as part of routine patient management, allowing biomarker sampling (IL-6, Syndecan-1, E-selectin) and measurement of other biological data of interest (eg, blood gas analysis, N-terminal pro-B-type natriuretic peptide, liver function tests). The measurement context will also be documented, including vasoactive and inotropic support, invasive ventilation status and ongoing sedation. Blood sampling will be performed as part of routine patient management, allowing biomarker sampling (IL-6, Syndecan-1, E-selectin) and measurement of other biological data of interest (eg, blood gas analysis, N-terminal pro-B-type natriuretic peptide, liver function tests). The measurement context will also be documented, including vasoactive and inotropic support, invasive ventilation status and ongoing sedation. Clinical data assessment (eg, weight, diuresis, intravenous fluid therapy volumes and types, blood pressure, heart rate, body composition). Clinical data assessment (eg, weight, diuresis, intravenous fluid therapy volumes and types, blood pressure, heart rate, body composition). Visit 2 (V2), 48 hours after CPBFMD will be repeated.SDF imaging will be repeated.BIA will be repeated.Blood sampling will be repeated.Clinical data assessment will be repeated, including vasoactive and inotropic support, invasive ventilation status and ongoing sedation at the time of measurement. Visit 2 (V2), 48 hours after CPB FMD will be repeated. FMD will be repeated. SDF imaging will be repeated. SDF imaging will be repeated. BIA will be repeated. BIA will be repeated. Blood sampling will be repeated. Blood sampling will be repeated. Clinical data assessment will be repeated, including vasoactive and inotropic support, invasive ventilation status and ongoing sedation at the time of measurement. Clinical data assessment will be repeated, including vasoactive and inotropic support, invasive ventilation status and ongoing sedation at the time of measurement. Endpoints The primary endpoint is the comparison of corrected brachial artery FMD between patients with VS and patients without VS at 48 hours after CPB weaning. Key secondary endpoints include early postoperative corrected FMD (V1) and change in corrected FMD between V1 and V2, together with the following components of the multimodal endothelial assessment: Sublingual microcirculatory parameters measured by SDF imagingPBR.Capillary density.Relative capillary blood volume. Sublingual microcirculatory parameters measured by SDF imaging PBR. PBR. Capillary density. Capillary density. Relative capillary blood volume. Relative capillary blood volume. Plasma levels of biomarkersIL-6.Syndecan-1.E-selectin. Plasma levels of biomarkers IL-6. IL-6. Syndecan-1. Syndecan-1. E-selectin. E-selectin. Routine laboratory parametersIncluding arterial and venous blood gases: lactate, CO2 gap, central venous oxygen saturation (ScvO2). Routine laboratory parameters Including arterial and venous blood gases: lactate, CO2 gap, central venous oxygen saturation (ScvO2). Including arterial and venous blood gases: lactate, CO2 gap, central venous oxygen saturation (ScvO2). BIA, including total body water, extracellular and intracellular water and overhydration status. BIA, including total body water, extracellular and intracellular water and overhydration status. Clinical outcome parametersAKI within 48 hours (defined using KDIGO (Kidney disease : improving global outcomes) classification and need for renal replacement therapy).Total fluid balance and type of fluids administered during the first 48 hours.Sedation and mechanical ventilation duration.ICU length of stay.Mortality. Clinical outcome parameters AKI within 48 hours (defined using KDIGO (Kidney disease : improving global outcomes) classification and need for renal replacement therapy). AKI within 48 hours (defined using KDIGO (Kidney disease : improving global outcomes) classification and need for renal replacement therapy). Total fluid balance and type of fluids administered during the first 48 hours. Total fluid balance and type of fluids administered during the first 48 hours. Sedation and mechanical ventilation duration. Sedation and mechanical ventilation duration. ICU length of stay. ICU length of stay. Mortality. Mortality. Sample size calculation and statistics Sample size calculation From prior data showing a postoperative FMD of approximately 1.4%,32 and using the original planning SD of 0.6%, an absolute between-group difference of 0.35 percentage points, corresponding to an approximately 25% relative difference, was prespecified as a detectable difference for this exploratory study rather than as an established clinically meaningful minimum difference. At α=0.05 and 90% power, 62 patients per group, 124 total, are required to detect this difference at a single time point. Statistical analyses All analyses will be performed in R. Two-sided tests will use α=0.05. Continuous variables will be visually inspected (histograms/density, Q-Q plots) and transformed when appropriate (eg, logarithmic transformation for skewed biomarkers). Patient characteristics will be described using standard parameters (mean, SD, median, IQR, extremes for quantitative variables; counts and percentages for categorical variables). Primary analysis A multiple linear regression model will estimate the adjusted between-group difference in corrected FMD (VS vs non-VS) at 48 hours (V2), with adjustment for the following prespecified covariates: age, sex, surgical category, chronic kidney disease, diabetes, preoperative arterial hypertension, preoperative statin therapy and CPB duration. The covariate set was prespecified and restricted to clinically relevant preoperative and intraoperative variables to limit overfitting. Because FMD reflects a shear-mediated response, FMD models will additionally adjust for the shear stimulus at the corresponding visit (shear-rate area under the curve up to peak dilation). FMD will be analysed in its corrected/allometrically scaled form. Secondary analyses The same between-group multivariable framework used for the primary endpoint will be applied at V2 to PBR, IL-6, Syndecan-1 and E-selectin, using appropriate transformations for skewed variables when required. The same between-group multivariable framework used for the primary endpoint will be applied at V2 to PBR, IL-6, Syndecan-1 and E-selectin, using appropriate transformations for skewed variables when required. Between-group differences at V1 (post-CPB) for corrected FMD, PBR, IL-6, Syndecan-1 and E-selectin will also be evaluated using the same covariate set; for FMD, the corresponding V1 shear stimulus will be included. Between-group differences at V1 (post-CPB) for corrected FMD, PBR, IL-6, Syndecan-1 and E-selectin will also be evaluated using the same covariate set; for FMD, the corresponding V1 shear stimulus will be included. Additional exploratory conditional models will evaluate V2 corrected FMD, PBR, IL-6, Syndecan-1 and E-selectin while adjusting for the corresponding V1 value and the same covariates; for FMD, the V2 shear stimulus will be included. Missing V1 covariate values may be imputed as described in the missing-data section; V2 outcomes will not be imputed. A complete-case version of these conditional models will be reported as a sensitivity analysis in the online supplemental appendix. Additional exploratory conditional models will evaluate V2 corrected FMD, PBR, IL-6, Syndecan-1 and E-selectin while adjusting for the corresponding V1 value and the same covariates; for FMD, the V2 shear stimulus will be included. Missing V1 covariate values may be imputed as described in the missing-data section; V2 outcomes will not be imputed. A complete-case version of these conditional models will be reported as a sensitivity analysis in the online supplemental appendix. Sensitivity analyses will include alternative adjustment sets and exclusion of patients receiving inotropic support at classification and/or measurement. Sensitivity analyses will include alternative adjustment sets and exclusion of patients receiving inotropic support at classification and/or measurement. Exploratory associations between endothelial markers (corrected FMD, PBR, IL-6, Syndecan-1, E-selectin) and clinical outcomes (eg, AKI, ICU/hospital length of stay, vasopressor requirement duration) will be modelled with the same covariate set and labelled as hypothesis-generating. Exploratory associations between endothelial markers (corrected FMD, PBR, IL-6, Syndecan-1, E-selectin) and clinical outcomes (eg, AKI, ICU/hospital length of stay, vasopressor requirement duration) will be modelled with the same covariate set and labelled as hypothesis-generating. Correlation analyses Pairwise correlations will be computed separately at V1 and V2 across endothelial, microcirculatory and biomarker measures. Two correlation heatmaps, one for V1 and one for V2, will be presented. Multiplicity will be addressed separately within the V1 and V2 exploratory correlation matrices using the Benjamini-Hochberg procedure, and false-discovery-rate-adjusted p values will be reported alongside raw p values. These analyses will be interpreted as exploratory and hypothesis-generating. Exploratory temporal analyses (V1→V2) Exploratory models will assess whether early markers (eg, Syndecan-1 at V1, reflecting glycocalyx injury) are associated with parameters at 48 hours (PBR or FMD at V2), adjusted for group and covariates. These analyses will be hypothesis-generating; p values will be reported descriptively. If many comparisons are undertaken, FDR-adjusted p values may be provided in an online supplemental appendix. Missing data Values analysed as outcomes, including FMD, SDF-derived parameters and biomarker concentrations, will not be imputed. Primary and secondary analyses will be conducted using available data for the outcome under consideration. For conditional V2 models in which the corresponding V1 value is used as a covariate, missing V1 covariate values may be imputed using chained equations, whereas V2 outcomes will not be imputed. FMD measurement time points without analysable shear-stimulus area under the curve (AUC) will be recorded as not available for FMD, and shear-stimulus AUC will not be imputed. The imputation model will include study group, the corresponding non-imputed V2 measure as an auxiliary variable, model covariates, baseline characteristics and clinical/biological data collected before or at V1. A complete-case version of these conditional models will be reported as a sensitivity analysis. Data availability at V1 and V2 will be summarised in a study flow diagram for each measure. Reasons for missing V2 measurements, including death, unplanned reoperation, clinical instability, technical failure or investigator unavailability, will be reported when available. Baseline, intraoperative and early postoperative characteristics of patients with and without available V2 primary endpoint data will be compared descriptively, and the plausibility of the missing-at-random assumption will be discussed. Additional exploratory descriptive analyses If emerging patterns warrant it, further descriptive or exploratory analyses (eg, subgroup summaries, alternative transformations, additional visualisations) may be undertaken and clearly identified as exploratory. Quality assurance and monitoring As an observational study, risks are minimal. The principal investigator will oversee data quality and protocol adherence. Any amendments will be communicated to the ethics committee and trial registry. Ethics and dissemination The protocol has been approved by the Comité de Protection des Personnes Sud-Méditerranée III (CPP Sud-Méditerranée III; 2023-A02079-36), with approval effective from 5 February 2024. This is a non-interventional study conducted in routine care (category 3 research involving the human person under French law). For this category, prior individual information and absence of opposition are required rather than written consent. The CPP approved the protocol, including this information and non-opposition procedure. Patients, or the appropriate representative where legally applicable, receive oral and written information before inclusion and are asked whether they object to participation; absence of opposition is documented. The study was registered on ClinicalTrials.gov. Results will be published in peer-reviewed journals and presented at conferences. Patient and public involvement Patients and the public were not involved in the design, conduct, reporting or dissemination plans of this study. The protocol was developed by the clinical and research team based on the scientific rationale, feasibility constraints and routine postoperative care pathways in cardiac surgical intensive care. Expected contributions This study is expected to provide a structured description of postoperative vascular alterations associated with VS after cardiac surgery with CPB. By combining conduit-artery endothelial vasoreactivity assessment, glycocalyx-related bedside parameters and inflammatory/endothelial biomarkers at two postoperative time points, it may help refine postoperative vascular phenotyping in vasoplegia and support future hypothesis-driven studies. Supplementary material Footnotes Funding: This work was supported by the French Society of Anaesthesia and Critical Care (SFAR) to cover study operating costs. Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2026-122692). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. References Supplementary Materials