https://orcid.org/0000-0003-1742-5841
Abstract
Spinal cord ischaemia (SCI) remains the most devastating complication after thoraco-abdominal aortic aneurysm (TAAA) repair. Its early detection is crucial if therapeutic interventions are to be successful. Cerebrospinal fluid (CSF) is readily available and accessible to microdialysis (MD) capable of detecting metabolites involved in SCI [i.e. lactate, pyruvate, the lactate/pyruvate ratio (LPR), glucose and glycerol] in real time. Our aim was to evaluate the feasibility of CSF MD for the real-time detection of SCI metabolites.
In a combined experimental and translational approach, CSF MD was evaluated (i) in an established experimental large animal model of SCI with 2 arms: (a) after aortic cross-clamping (AXC, N = 4), simulating open TAAA repair and (b) after total segmental artery sacrifice (Th4–L5, N = 8) simulating thoracic endovascular aortic repair. The CSF was analysed utilizing MD every 15 min. Additionally, CSF was collected hourly from 6 patients undergoing open TAAA repair in a high-volume aortic reference centre and analysed using CSF MD.
In the experimental AXC group, CSF lactate increased 3-fold after 10 min and 10-fold after 60 min of SCI. Analogously, the LPR increased 5-fold by the end of the main AXC period. Average glucose levels demonstrated a 1.5-fold increase at the end of the first (preconditioning) AXC period (0.60±0.14 vs 0.97±0.32 mmol/l); however, they decreased below (to 1/3 of) baseline levels (0.60±0.14 vs 0.19±0.13 mmol/l) by the end of the experiment (after simulated distal arrest). In the experimental segmental artery sacrifice group, lactate levels doubled and the LPR increased 3.3-fold within 30 min and continued to increase steadily almost 5-fold 180 min after total segmental artery sacrifice (P < 0.05). In patients undergoing TAAA repair, lactate similarly increased 5-fold during ischaemia, reaching a maximum at 6 h postoperatively. In 2 patients with intraoperative SCI, indicated by a decrease in the motor evoked potential of >50%, the LPR increased by 200%.
CSF is widely available during and after TAAA repair, and CSF MD is feasible for detection of early anaerobic metabolites of SCI. CSF MD is a promising new tool combining bedside availability and real-time capacity to potentially enable rapid detection of imminent SCI, thereby maximizing chances to prevent permanent paraplegia in patients with TAAA.
INTRODUCTION
The incidence of spinal cord ischaemia (SCI) following extensive (i.e. Crawford extent I, II, III) thoraco-abdominal aortic aneurysm (TAAA) repair is—despite the clinical application of various modern neuroprotective adjuncts—still significant, with up to 20% of patients suffering permanent paraplegia [1, 2] after both open and endovascular TAAA repair—with a significant impact on short- and long-term survival [3].
Early diagnosis of acute SCI is essential to avoid irreversible spinal cord damage because the severity and reversibility of the neurological impairment depend on immediate detection and rapid institution of the limited therapeutic options available—first and foremost the optimization of haemodynamics and oxygenation supported by enforced cerebrospinal fluid (CSF) drainage and continuous (para)spinal perfusion (and neuro-)monitoring [1, 4].
Because the pathogenesis of SCI is complex and multifactorial, elaborate experimental and translational clinical research is paramount to progress in the field. For rapid diagnosis of SCI, real-time biomarkers are a potentially powerful tool, particularly in the early postoperative period when spinal cord perfusion is highly vulnerable to unstable haemodynamics and when haemodynamic management is most challenging.
Previous studies investigating clinically relevant biomarkers during aortic repair focused on proteomic profiling of CSF [5]. However, these markers are released into the CSF long after the onset of acute SCI, when irreversible damage has already occurred (12–48 h postoperatively), rendering them unfavourable for clinical use.
Although reliable detection of early ischaemia markers is difficult with conventional analytic methods [6], tissue metabolism can be monitored easily by microdialysis (MD). MD is capable of detecting biomarkers in any fluid in real time [7, 8], including markers of cell damage and ischaemia [e.g. lactate, pyruvate, the lactate/pyruvate ratio (LPR), glutamate, glucose and glycerol]. Distressed tissue with inadequate oxygen and substrate delivery reacts by switching its metabolism to the anaerobic pathway with lactate production proportional to the amount of energy produced [9, 10]. LPR ≥25 indicates the onset of anaerobic metabolism together with high lactate and low glucose levels, suggesting disturbances in energy metabolism [6]. Normal CSF levels of glycerol vary between 8.6 and 25 μmol/l [11]. MD has previously been established in neurointensive care as a cerebral perfusion monitoring method in cardiac surgery [12]. Guidelines [13, 14] recommend routine CSF drainage for open TAAA repair, and its use in extensive thoracic endovascular aortic repair is increasing. CSF is therefore readily available and can be easily utilized for analysis of SCI markers.
This bifid translational study focused on the feasibility of CSF MD with real-time capacity for analysis of ischaemic metabolites. The study comprised 2 arms: (i) an established experimental large animal model reliably inducing controlled SCI (a) via aortic cross-clamping (AXC) and (b) through consecutive open segmental artery sacrifice (SAS) followed by 3 h of consecutive hypoperfusion of the spinal cord.The second, translational arm addressed the clinical feasibility of this test in patients undergoing open TAAA repair in a high-volume aortic reference centre.
MATERIALS AND METHODS
Ethics statement
Animal experiments were approved by the Institutional Animal Care and Use Committee in accordance with the local Saxon Veterinary Office and the Principles of Laboratory Animal Care [15] and human investigations were approved by the local ethics committee (IRAS project ID: 143355. REC reference: 14/EE/000) (Fig. 1).
Set-up of the 3 groups: 2 experimental groups (left) with aortic cross-clamping and SAS to provoke spinal cord ischaemia and 1 group of patients with TAAA repair as a translational approach (right). In all groups, microdialysis (with real-time capacity) of cerebrospinal fluid was performed to detect metabolic markers of spinal cord ischaemia. AXC: aortic cross-clamping; SAS: segmental artery sacrifice; TAAA: thoraco-abdominal aortic aneurysm.
Experimental set-up: anaesthesiological protocol
Surgery was performed on 12 juvenile female pigs (36–47 kg). The animals were sedated using intramuscular administration of 15 mg/kg ketamine, 5 mg/kg diazepam and 0.03–0.05 mg/kg fentanyl. After endotracheal intubation, anaesthesia was maintained with propofol (1–2 mg/kg, Pfizer, New York, NY, USA). A volume-controlled mode with 50% oxygen at 20 breaths per min and a tidal volume adjusted to 8 ml/kg was used for mechanical ventilation. For direct blood pressure monitoring, blood gas analyses and withdrawal of reference blood samples for microsphere analysis were established via the left femoral and right subclavian arteries. Body temperature (37±1.5°C) was monitored using a rectal probe. A central venous catheter for pressure monitoring was inserted into the left jugular vein using a 7-Fr, 20-cm multilumen catheter for additional infusions and blood sampling. To prevent spinal injury and to ensure correct placement, a cerebrospinal drain was inserted at the level of L3/4; the MD catheter (IV M67, membrane length 10 mm, M Dialysis®, Stockholm, Sweden) at Th12/L1 was inserted under angiographic guidance. The MD catheter was continuously flushed at a flow rate of 2 µl/min.
Aortic cross-clamp group
SCI was triggered by AXC of the proximal descending aorta (AXC group). The experimental set-up, including the time points for MD and microsphere application, is illustrated in Fig. 2.
Time points (T) for microsphere injection and microdialysis measurements of the 2 experimental groups, 1 with aortic cross-clamping (n = 4), left and 1 with SAS (n = 8). AXC: aortic cross-clamping; SAs: segmental arteries; SAS: segmental artery sacrifice.
A left lateral thoracotomy was performed at the 4th–5th intercostal space. The proximal descending aorta was mobilized in the regular manner. Finally, purse-string sutures were placed at the left atrial appendage and a catheter was inserted for microsphere measurements. After aortic preparation, AXC was applied at the proximal descending aorta for 10 min followed by 10 min of reperfusion (recovery phase). Then AXC was repeated for 60 min and again followed by a second reperfusion phase of 60 min. After the second recovery phase, the anaesthetized pigs were euthanized in deep sedation. The spinal cord was removed en bloc immediately after euthanasia.
Segmental artery sacrifice group
SCI was achieved by consecutive surgical SAS from Th4 to L5 (SAS group). A left lateral thoracotomy at the 4th and 8th intercostal spaces, as well as a lateral-to-midline laparotomy, was performed to facilitate access to the thoracic and abdominal aortas, respectively. Afterwards, all segmental arteries (SAs) arising directly from the aorta were mobilized. Finally, purse-string sutures were placed at the left atrial appendage and a catheter was inserted for microsphere measurements.
Intervention
Microspheres were injected and CSF samples for MD analysis were collected at 6 time points throughout the experiment (T1–T6) (Fig. 2). After exposure, all thoracic SAs (Th4–Th13) were sacrificed (clip occlusion), followed by injection of microspheres (T2). Subsequently, all lumbar SAs were clipped, again followed by microsphere injection (T3) for regional perfusion analysis. For evaluation of blood flow redirection and potential compensatory changes, microspheres were injected at 10 min (T4), at 60 min (T5) and at 180 min (T6) after occlusion of all SAs. Sedation of the animals was deepened prior to their euthanasia. The spinal cord was immediately removed en bloc for microsphere and histological analysis, and the aorta was examined for completeness of SA occlusion.
Microsphere measurements (regional tissue perfusion)
At each time point (detailed in Fig. 2), 3 million fluorescent microspheres (Dye-Trak®, Triton Technology, San Diego, CA, USA) were injected into the left atrium (as previously described [4]). To account for slightly variable haemodynamics, reference blood was drawn at each time point via a catheter in the left subclavian artery. The fluorescence intensities of the microspheres were measured using Synergy H1 Hybrid Multi-Mode Reader (BioTek, Winooski, VT, USA).
Microdialysis
CSF dialysates were continuously collected at 15-min intervals during the entire experiment. The expected metabolite recovery at the used flow rate of 2 µl/min has been reported to be ∼30% [16]. At the end of the experiment, the concentrations of the dialysates (lactate, pyruvate, LPR, glucose, glycerol) were analysed immediately for all human samples. These marker metabolites were measured by colorimetric methods with an automated analyser (ISCUSflex Microdialysis Analyzer, M Dialysis AB).
Clinical feasibility study
Ethics and patients
For translation into clinical practice, a human pilot study was performed at Barts hospital as part of the translational research of the Barts Health National Health Service Trust, BioResource, supported by the National Institute of Health Research Barts Biomedical Research Centre (IRAS project ID: 143355. REC reference: 14/EE/000). Patients gave consent preoperatively and were recruited to a prospective observational study with no intended change beyond usual clinical care. This is a non-interventional, feasibility and descriptive study. Patients were included between February and May 2019. Inclusion criteria were TAAA requiring elective extent II repair; age 18–80 years; preoperative insertion of a cerebrospinal drain catheter and maintenance for 48–72 h. Exclusion criteria were a preoperative major neurological event (12 months prior to the operation) and preoperative SCI or an active pathological process affecting the brain–blood barrier (viral/bacterial meningitis, encephalitis).
Operative technique
The open Crawford extent II repair comprised the replacement of the entire descending thoracic aorta and extended into the infrarenal aortic segments. The operative technique is detailed in the Supplementary Material. All the patients were incorporated into the spinal protocol pathway with continuous CSF drainage (drain inserted preoperatively) to maintain a subarachnoid pressure ≦10 mmHg; the drainage system remained in place for 72 h. We utilized moderate hypothermia and distal aortic perfusion with LHB, segmental AXC and selective SA reimplantation directed by motor evoked potential (MEP) monitoring (5/6 patients), permissive hypertension 85–100 mmHg, perfusion of visceral and renal vessels (pulsatile) and paraspinal collateral network near-infrared spectroscopy [17].
Microdialysis
CSF dialysate samples were collected ex vivo at 12 different time points (Fig. 3) and analysed directly for lactate, pyruvate, LPR, glucose, glycerol and glutamate by colorimetric methods with an automated analyser (ISCUSflex Microdialysis Analyzer, M Dialysis AB). With the probes used (M67, M Dialysis AB, membrane length 10 mm) and a flow rate of 0.3 µl/min, the expected metabolite recovery was 80–90% [16]. We used the term ‘near real time’ because for each sample the measuring time was 30 s and the throughput time was 90 s.
Time points for microdialysis analysis of cerebral spinal fluid in human patients. DTA: descending thoracic aorta; SA: segmental artery.
Statistical analyses
The statistical programme SPSS (SPSS Statistics for Windows, Version 17.0, 2008; SPSS, Chicago, IL, USA) was used. Continuous variables are expressed as mean ± standard deviation. Q–Q plots verified the normal distribution of measurements. Continuous dependent variables were analysed using a general linear model and within-subjects comparisons for repeated measures. Regional blood flow analyses for the spinal cord tissue were performed in groups by merging the segments (Th8–L2 and L3–S). Statistical significance was set at a P-value of <0.05 for 2-tailed testing.
RESULTS
Experimental model
Aortic cross-clamp group
In the AXC group, 4 pigs were available for full analysis: CSF dialysate was analysed for lactate, glycerol, glucose and the LPR using MD (Fig. 4A–D). After 10 min of AXC (preconditioning period), the mean lactate level in the CSF dialysate had increased three-fold compared to baseline (0.41 ± 0.23 vs 1.28 ± 0.68 mmol/l). After a 10-min reperfusion period for cardiopulmonary recovery and the subsequent 60 min of AXC (main period simulating distal arrest), the mean lactate level in the CSF dialysate progressed to an almost 10-fold increase compared to baseline (0.41 ± 0.23 vs 3.90 ± 0.68 mmol/l; P = 0.01). Details for additional data points are summarized in Supplementary Material, Table S1.
Microdialysis analysis of cerebrospinal fluid dialysate samples obtained during and after aortic cross-clamping in 4 animals. (A) Cerebrospinal fluid dialysate lactate level in mmol/l; (B) lactate/pyruvate ratio; (C) glucose level in mmol/l and (D) glycerol level in mmol/l. Red area represents ischaemic time during AXC. Overall comparison of all markers at different time points showed significant differences (P < 0.001). Red dotted line in (B) indicates the estimated threshold where anaerobic metabolism begins (pending cell damage). AXC: aortic cross-clamping.
Analogously, the LPR showed a five-fold increase by the end of the main AXC period of simulated distal arrest (11.67 ± 6.03 vs 61.33 ± 24.19), and the mean glycerol level reached a three-fold increase (81.75 ± 9.64 vs 265.75 ± 59.63) at this point.
Average glucose levels demonstrated a 1.5-fold increase at the end of the first (preconditioning) AXC period (0.60 ± 0.14 vs 0.97 ± 0.32 mmol/l); however, they decreased below (to 1/3 of) baseline levels (0.60 ± 0.14 vs 0.19 ± 0.13 mmol/l) by the end of the experiment (after simulated distal arrest). Overall comparison of all metabolic markers at the described time points showed significant differences (P < 0.001; comparisons of each marker category at different time points revealed no significant difference, most likely due to the small numbers and the associated extensive standard deviations).
Complementary microsphere measurements for assessment of regional spinal cord perfusion showed a reduction during AXC from the baseline of 0.80 ± 0.75 to 0.13 ± 0.16 (preconditioning AXC) vs 0.46 ± 0.63 (AXC simulating distal arrest) for Th8–L2 and also for L3–S3 [0.67 ± 0.92 (baseline) vs 0.36 ± 0.09 (first AXC) vs 0.32 ± 0.55 (second AXC)], none of which reached statistical significance (P = 0.80).
Segmental artery sacrifice group
CSF dialysate from all 8 animals could be analysed utilizing MD (details shown in Fig. 5A–D). The dialysate lactate level doubled 30 min after complete SA sacrifice (0.73 ± 0.40 vs 1.46 ± 0.64; P < 0.001) and increased 2.8-fold compared to baseline after 3 h of ischaemia (0.73 ± 0.40 vs 2.00 ± 0.66; P < 0.001). Analogously, the LPR continuously increased after complete SA sacrifice from 3.3-fold at 30 min to 3.5-fold at 60 min and to 4.9-fold at 180 min, all compared to baseline [19.50 ± 21.31 (baseline) vs 64.38 ± 54.54 (30 min; P < 0.001) vs 68.71 ± 48.23 (60 min) vs 95.38 ± 50.54 (180 min; P < 0.001)]. Glucose and glycerol levels in the CSF dialysate declined but did not reach statistical significance (Supplementary Material, Table S2). The regional spinal cord perfusion decreased significantly after complete SAS [Th8–L2: 0.38 ± 0.28 (baseline) vs 0.04 ± 0.03 (180 min after SAS); P < 0.001; L3–S3:0.56 ± 0.46 (baseline) vs 0.12 ± 0.10; P < 0.001].
Microdialysis analysis of cerebrospinal fluid samples during and after SAS of 8 animals. (A) Cerebrospinal fluid levels of lactate (mmol/l); (B) lactate/pyruvate ratio; (C) glucose level in mmol/l and (D) glycerol level in mmol/l. Red area represents ischaemic time after complete SAS. Red dotted line in (B) indicates the estimated threshold where anaerobic metabolism begins (pending cell damage). SAs: segmental arteries.
Clinical feasibility study
Of the 6 consecutive patients (5 men) undergoing TAAA repair (mean age: 50 ± 14.6 years), 67% had a chronic aortic dissection, type B, but none had previous endovascular repair. For preoperative, perioperative and postoperative details of the patients, see Supplementary Material, Table S3. The mean AXC time was 305.8 ± 41.49 min; reimplantation of SAs between T9 and T11 was performed in 83%. Two patients showed a perioperative MEP decrease (changes >50%) from baseline; 1 patient died on the second postoperative day of multiorgan failure and the other patient experienced transient neurological impairment. Magnetic resonance imaging of the spine showed severe spinal canal stenosis but no SCI.
Microdialysis analysis of human cerebral spinal fluid samples and translational comparison
MD revealed a relevant trend in the concentration of the biomarkers examined intraoperatively during TAAA repair (Fig. 6A–D). Lactate levels in the CSF (Fig. 6A) increased steadily, reaching a maximum at the end of the procedure [0.67 ± 0.42 (baseline) vs 3.35 ± 1.80]. On average, the CSF glycerol level increased, reaching a maximum at 6 h postoperatively [16.83 ± 8.5 (baseline) vs 203.83 ± 248.82] (Fig. 6D). A similar increase could be seen in the experimental AXC group 60 min after the second cross-clamp (Fig. 4D). The detected increase in the CSF glucose level until the end of the procedure [1.78 ± 1.09 (baseline) vs 5.28 ± 1.99] was similar to the increase in the experimental AXC group (Fig. 4C and Supplementary Material, Table S4).
Microdialysis analysis of human cerebral spinal fluid dialysate samples during and after thoraco-abdominal aortic aneurysm repair of 6 patients. (A) Cerebral spinal fluid lactate level in mmol/l; (B) the lactate/pyruvate ratio; (C) glucose level in mmol/l and (D) glycerol level in mmol/l. Red area represents cumulative ischaemic time during aortic cross-clamping. AXC: aortic cross-clamping; SA: segmental artery.
DISCUSSION
This study utilized MD as a novel method for detection of SCI metabolites in CSF dialysate in (i) an established experimental large animal SCI model reliably simulating (a) distal aortic arrest during open TAAA repair and (b) endovascular repair through total SAS. Subsequently—in a translational clinical approach (ii)—we investigated for the first time changes in lactate, the LPR, glucose and glycerol levels in human CSF dialysate to clinically detect SCI in near real time.
The acquired data for the first time provide insights into the feasibility of using CSF MD to detect acute metabolic changes and the consecutive release and distribution of anaerobic metabolites in the CSF as a response to simulated SCI. In the experimental arm, we could detect a noticeable increase in CSF lactate levels and the LPR during AXC and after SAS. We observed similar changes in 6 patients during and after routine open TAAA repair in the translational study arm. However, all observations need to be interpreted with caution due to the limited number of animals and patients.
No reliable readily accessible SCI biomarkers are available for clinical use. Although several studies have focused on serum markers [18, 19]—analogous to troponin or creatine kinase for myocardial infarction—the release and accumulation of potential early biomarkers of acute SCI in the bloodstream seem slow or delayed; hence, the detection of SCI markers in blood is challenging. Other complex approaches were only partly successful or not easily applicable for routine clinical use, e.g. the analysis of volatile organic compounds by gas chromatography in exhaled air, reflecting the chemical composition of the bloodstream, in a large animal model of MEP-guided SAS that non-specifically indicated metabolic stress and oxidative damage during ischaemia-reperfusion [19]. A clinical study conducted by the Hannover group prospectively utilized CSF gas-tension analysis, thereby demonstrating glucose variability intraoperatively in patients who developed paraplegia, suggesting a disruption of the blood–brain barrier [20].
Focusing on CSF comprises 2 major advantages: the CSF continuously interacts with spinal cord tissue across minor barriers and is therefore predestined to reflect metabolic changes, and it is readily available (supported by the current guidelines) in most aortic procedures, baring a high risk of SCI [13, 14].
However, several attempts to detect biochemical markers in CSF for early SCI detection in patients with TAAA failed to reach clinical significance. A recent meta-analysis of 14 studies (321 patients) found that the levels of members of the S100B protein family, neuron-specific enolase and glial fibrillary acidic protein, increased in patient CSF long after SCI had occurred [21]. For example, the levels of these proteins increased many hours after the tissue was irreversibly damaged, e.g. S100 levels in the CSF increased in a paraplegic patient after 24 h—too late for any therapeutic option [18].
Therefore, detecting sensitive anaerobic early-release metabolites—assumed to be rapidly released into the CSF—holds a realistic chance for the rapid translation into clinical implementation of this novel approach. Previously, cerebral MD was utilized in neurocritical care patients for real-time assessment of cerebral metabolism, demonstrating a strong correlation between CMD and neurological outcome [22, 23]. A significant correlation of morbidity and mortality in this setting has been shown for glucose, the LPR and glycerol [24]. Glycerol is a sensitive and reliable marker of cellular injury after ischaemia resulting from the phospholipase-activated degradation of cell membranes [23, 25]. Lactate is a universally recognized but unspecific marker of ischaemia. The LPR has been demonstrated to be an extremely sensitive indicator of tissue ischaemia, hypoxia and cellular mitochondrial failure [23, 25]. An abnormal LPR even predicted a rise in intracranial pressure hours in advance: in an analysis of 802 samples, an abnormal LPR (<25) at normal intracranial pressure levels indicated a significant risk of the patient developing intracranial hypertension within the next 3 h (odds ratio 9.8; P < 0.001) [24].
Waelgaard et al. [10], in an experimental pig model, demonstrated that during haemorrhagic shock the organ and tissue metabolism progresses through 2 phases: the first phase is characterized by a progressive decrease in partial oxygen pressure and an increase in partial CO2; however, no significant changes occur in lactate, LPR and glycerol levels. In the second phase, triggered by progredient decline in blood flow and oxygen extraction, the transition from solely aerobic to mixed aerobic and anaerobic metabolism leads to a solely anaerobic metabolism characterized by significant elevations of lactate, LPR and glycerol levels complemented by a marked decrease in tissue pH, bicarbonate and glucose.
In our experimental model with AXC, we experienced the first phase at 10 min of AXC and the second phase during the 60 min of AXC with a significant increase in lactate and LPR. Furthermore, in the clinical pilot arm, 2 patients with loss of the MEP signal >50% CSF dialysate analysis showed an early LPR increase above 25. An LPR ≥25 indicates the onset of anaerobic metabolism together with high lactate and low glucose levels, suggesting severe disturbances in energy metabolism [6].
Previous studies of MD focused on brain injury during deep hypothermic circulatory arrest. Mavroudis et al. [26] used a neonatal pig model undergoing deep hypothermic circulatory arrest and evaluated cerebral mitochondrial bioenergetics by MD, demonstrating that the LPR—as an indicator of cerebral ischaemia—and glycerol levels, as a marker of neuronal death, were increased during and after rewarming. In their study, glycerol levels and the LPR increased, whereas glucose levels in brain tissue decreased significantly. Post-mortem analyses confirmed an increase in the production of mitochondrial reactive oxygen species as a measure of mitochondrial dysfunction, hence anaerobic metabolism. These results are supported by Salazar et al., who also investigated deep hypothermic circulatory arrest with and without selective cerebral perfusion. In piglets without selective cerebral perfusion, lactate levels and the LPR significantly increased and glucose declined. MD served as a reliable and sensitive indicator of inadequate cerebral oxygenation.
Clinical implications
Early detection of SCI is crucial in preventing permanent paraplegia—the most devastating complication after open and endovascular TAAA repair. CSF is available perioperatively and postoperatively for analysing early anaerobic metabolites of SCI in CSF dialysate, utilizing MD as a promising new diagnostic tool. MD combines bedside availability and real-time capacity to enable rapid detection of imminent SCI, allowing for immediate therapeutic measures to maximize chances to prevent permanent paraplegia in patients with TAAA.
Limitations
This is the first utilization of MD for CSF analysis in an established large animal model and the first use in a clinical pilot study. The small number of patients restricts statistical analyses in terms of significance. Furthermore, the overall diverse nature of the patients (e.g. chronic aortic dissection type B, Marfan syndrome, degree of urgency) precludes concerted interpretations. Only 1 patient developed a neurological deficit; in this case stenosis of the spinal canal (thus possibly traumatic injury) might have been causative. Furthermore, clinical thresholds—indicative of pending early (intraoperative) or delayed (postoperative) SCI after TAAA repair—might significantly differ in human CSF. Reliable clinical application of this promising new tool therefore requires thorough evaluation, ideally in a prospective multicentre randomized controlled trial. Therefore, we will evaluate this new technology independently in a subcohort of patients enrolled in the PAPAartis (‘Paraplegia Prevention in Aortic Aneurysm Repair by Thoracoabdominal Staging with “Minimally Invasive Segmental Artery Coil-Embolization”: A Randomized Controlled Multicentre Trial’; NCT03434314) trial, the largest publicly funded clinical multicentre randomized controlled trial on paraplegia prevention currently enrolling patients undergoing open and endovascular TAAA repair worldwide. Additionally, the correlation of lactate and glucose levels in the peripheral blood and in the CSF might be subject to confounding results (Supplementary Material, Fig. S1).
Because we augmented our established large animal model by inserting a CSF catheter for MD analysis, technical problems in the first experiments led to only a small number of pigs in the AXC group available for limited statistical analysis. Furthermore, extending MD analysis to a chronic pig model only allows for correlation of changes in CSF metabolites by MD and neurological outcome. Of note, Reinstrup et al. [25] investigated the interstitial concentration of the most important metabolic markers in the brain using MD in awake and anaesthetized patients and found differences in all values; only the LPR reached statistical significance.
SUPPLEMENTARY MATERIAL
Supplementary material is available at EJCTS online.
Funding
The human pilot study was partly funded by the Barts Health National Health Service Trust, BioResource, supported by the National Institute of Health (IRAS project ID: 143355. REC reference: 14/EE/00). The investigations on CSF metabolites are a subproject of the PAPAartis (‘Paraplegia Prevention in Aortic Aneurysm Repair by Thoracoabdominal Staging with “Minimally Invasive Segmental Artery Coil-Embolization”: A Randomized Controlled Multicentre Trial’) randomized controlled trial funded by the European Union Horizon 2020 research programme under grant agreement 733203 and the German Research Foundation (DFG) under grant number ET 127/2-1 and ET 127/1-1 (Project No.: 278040814).
Conflict of interest: none declared.
Author contributions
Urszula D. Simoniuk: Investigation; Visualization; Writing—original draft. Josephina Haunschild: Conceptualization; Formal analysis; Methodology; Project administration; Resources; Visualization; Writing—review & editing. Konstantin von Aspern: Conceptualization; Methodology; Resources; Software; Visualization. Michael Boschmann: Investigation; Methodology; Resources. Lars Klug: Investigation; Methodology. Zara Khachatryan: Data curation; Investigation; Visualization. Edoardo Bianchi: Investigation; Resources. Susann Ossmann: Investigation; Methodology; Project administration. Aung Y. Oo: Project administration; Resources. Michael A. Borger: Methodology; Resources; Supervision. Christian D. Etz: Conceptualization; Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Software; Supervision; Writing—review & editing.
Presented at the 33rd Annual Meeting of the European Association for Cardio-Thoracic Surgery, Lisbon, Portugal, 3–5 October 2019.
REFERENCES
ABBREVIATIONS
- AXC
Aortic cross-clamping
- CSF
Cerebrospinal fluid
- LPR
Lactate/pyruvate ratio
- MD
Microdialysis
- MEP
Motor evoked potential
- SAs
Segmental arteries
- SAS
Segmental artery sacrifice
- SCI
Spinal cord ischaemia
- TAAA
Thoraco-abdominal aortic aneurysm
Author notes
Urszula D. Simoniuk and Josephina Haunschild contributed equally to this study.
