TY - JOUR
AU - Ziani,, Barbara
AB - Abstract View largeDownload slide View largeDownload slide OBJECTIVES Our goal was to identify anatomical and physiological factors that could predict the amount of cerebrospinal fluid (CSF) drainage in patients undergoing elective endovascular repair of descending thoracic and thoracoabdominal aortic disease. METHODS All consecutive elective endovascular procedures performed for descending thoracic or thoracoabdominal aortic disease between January 2015 and December 2017 were included in the study. Routine use of CSF drainage was established in all patients. The goal of drainage was to reach a spinal fluid pressure of 10–12 mmHg by draining in 5–15-ml aliquots. The number of visible intercostal and lumbar segmental arteries (SAs) was evaluated before and after endovascular repair. The covering ratio of SAs was calculated as covered preoperative SAs/total preoperative SAs. RESULTS Twenty-four consecutive patients were included in the final analysis. The indication for the intervention was a descending thoracic aneurysm in 13 cases, a thoracoabdominal aneurysm in 4 cases and a chronic type B dissection in 7 cases. The procedure performed was thoracic endovascular aortic repair in 20 cases and fenestrated endovascular aneurysm repair in 4 cases. None of the patients developed spinal cord ischaemia. The mean volume of CSF drained was 46 ml. The mean length of aortic coverage was 231 mm. The mean number of total preoperative SAs and of covered preoperative SAs was, respectively, 22 and 9. The volume of CSF drained was significantly correlated with all these variables (length of aortic coverage, total visible SAs and covered SAs), but the most powerful correlation was individuated with the covering ratio of SAs. CONCLUSIONS Our findings suggest that the percentage of intercostal and lumbar SAs covered by placement of a stent graft can predict the volume of CSF drained in patients undergoing elective endovascular repair of descending thoracic and thoracoabdominal aortic disease. Spinal arteries, Cerebrospinal fluid drainage, Endovascular repair, Thoracic aortic disease INTRODUCTION Prevention of spinal cord ischaemia (SCI) after endovascular repair of aortic disease requires an understanding of the anatomy and physiology of the vascularization of the spinal cord [1]. Computed tomographic angiography (CTA) is the modality of choice for imaging spinal cord vascularization [2]. Among measures against postoperative SCI, cerebrospinal fluid (CSF) drainage has been one of the most extensively studied [3–7]. However, the current literature lacks studies dealing primarily with analysis of CSF drainage as a consequence of changes occurring in spinal cord vascularization after endovascular procedures. The aim of our study was to identify anatomical and physiological factors that could predict the volume of CSF drainage in patients undergoing elective endovascular repair for descending thoracic and thoracoabdominal aortic disease. METHODS All procedures performed were in accordance with the ethical standards of the Helsinki Declaration and its later amendments or comparable ethical standards. Institutional review board ethical approval was obtained from the parent institution. Informed consent was obtained from all individual participants included in the study. Study design Data regarding endovascular aortic procedures performed at our institution are prospectively collected in an institutional database and were retrospectively reviewed from January 2015 to December 2017 for inclusion in the study. Inclusion criteria were (i) availability of preoperative and postoperative CTA images of adequate quality (slice thickness <1 mm, inclusion of the complete aorta) and (ii) elective endovascular repair of descending thoracic or thoracoabdominal aortic disease. The following definitions were adopted: total preoperative segmental arteries (SAs)—the number of visible SAs within the entire aortic segment before endograft implantation; total postoperative SAs—the number of visible SAs within the entire aortic segment after endograft implantation; covered preoperative SAs—the number of visible SAs within the stented aortic segment before endograft implantation; and covered postoperative SAs—the number of visible SAs within the stented aortic segment after endograft implantation. The covering ratio of SAs was calculated to allow for better comparison of the normalized % of SAs covered and was defined as covered preoperative SAs/total preoperative SAs. We investigated 2 different sets of covariates that could determine the volume of CSF drained. The first set included variables that appraised either directly or indirectly the extent of SAs covered: length of aortic coverage [defined as the length of the stented aortic segment from the origin of the left subclavian artery (LSA) to the native aortic bifurcation]; total preoperative SAs, covered preoperative SAs; and covering ratio of SAs. The second set included variables that could theoretically influence the volume of CSF drained: the use of vasopressors; previous repair of an abdominal aortic aneurysm (AAA); type of disease (aneurysm versus dissection); lumbar versus intercostal SA coverage; hypogastric artery (HGA) and/or LSA coverage. Patients and procedures All procedures were performed in the operating room with the patients under general anaesthesia. The endografts used included the Relay-Plus (Terumo Aortic, Inchinnan, Renfrewshire, UK) and the Valiant Captivia (Medtronic, Minneapolis, MN, USA). Deployment criteria included (i) a minimum of 20 mm of aortic neck for proximal and distal endograft anchoring; and (ii) a minimum of 30 mm of endograft modular overlap for sealing in multiple-module devices. Procedure planning for aneurysm cases involved oversizing the endograft by 10–20% more than the intended aortic landing zone diameter, whereas no oversizing was done for the dissection cases. After successful deployment of the endograft and balloon apposition (at the level of the proximal, distal and component seal zones), completion angiography was performed to assess for technical success, which was defined as the absence of any type of endoleak, adequate endograft fixation in proximity to branch vessels without occlusion of the branch vessel itself and acceptable endograft configuration within the aorta. Cerebrospinal fluid drainage and neurological evaluation Routine prophylactic CSF drainage using the same standardized protocol was established in all patients. Lumbar drains were placed with the patient in the operating room before induction of general anaesthesia by a neurosurgeon (in the subarachnoid space between approximately L2 and L4). Mean arterial blood pressure was maintained intraoperatively and postoperatively at 85–90 mmHg with the use of vasopressors as needed. The goal of drainage was to achieve a spinal fluid pressure of 10–12 mmHg (intraoperatively and postoperatively). This goal was accomplished by draining the fluid in 5–15-ml aliquots utilizing a Buretrol system until the target spinal fluid pressure was achieved. CSF drainage was always accomplished by gravity alone. The volume of CSF drained was monitored, and there was no volume limit threshold at which CSF drainage was stopped. The lumbar catheter was occluded at 72 h after the index operation and subsequently removed in the absence of a neurological deficit (i.e. the patient remained neurologically stable). If bloody CSF was noted, the lumbar drain was clamped, all antiplatelet/anticoagulants drugs were stopped while a complete coagulation panel was ordered to identify and correct any possible coagulative disorder, and a non-contrast computed tomographic scan of the head was obtained to evaluate for extra-axial haemorrhage. Postoperative neurological evaluation was performed hourly. SCI was defined as any new lower extremity motor or sensory deficit, or both, in the absence of any documented intracerebral hemispheric events. A patient who was fully ambulatory preoperatively must have been able to bear his or her own weight without assistance to be considered neurologically intact. SCI was considered transient (versus permanent) when a clear deficit was documented and then was fully resolved at the time of discharge. When a patient’s symptoms had improved but his or her functional status was not completely restored to preoperative levels, the complication was considered permanent. Severity of SCI was classified according to the modified version of the Tarlov paraplegia scoring scale [8]. Analysis of visible intercostal and lumbar segmental arteries CTA-visible intercostal and lumbar SAs were quantified in the descending thoracic and thoracoabdominal aortic segment before and after the intended implantation of the stent graft using previously described methods [9, 10]. The scan volume was checked by means of preliminary anatomical survey imaging (topogram) to assess for inclusion of the complete aorta together with the outlets of the supra-aortic trunks down to the hip joint. Image post-processing was carried out using a dedicated workstation (3mensio, Pie Medical Imaging, Maastricht, Netherlands). Spinal arterial anatomy was assessed using axial and multiplanar reformations. In these models, each level of the spine was examined for visible posterior intercostal and lumbar SAs bilaterally. A visible SA was defined by good contrast opacification with a diameter ≥2 mm and a contrast density of 30–70 Hounsfield units [10]. Superimposition of osseous landmarks (i.e. the vertebral bodies) was used to accurately overlap preoperative and postoperative CTA to identify the stented aortic segment with the pertinent visible SAs. Image assessment was performed in consensus by 2 investigators (M.D. and B.Z.) in every case. To assess for reproducibility of our study findings, intraobserver and interobserver variabilities were evaluated by counting visible intercostal and lumbar SAs in 10 preoperative CTA scans. Intraobserver variability revealed almost perfect agreement, with a proportion in agreement of 0.93 and a k value of 0.88 [95% confidence interval (CI) 0.83–0.93; P < 0.001]. Interobserver variability revealed substantial agreement, with a proportion in agreement of 0.89 and a k value of 0.78 (95% CI 0.67–0.90; P < 0.001). Statistical analyses Data regarding preoperative characteristics, intraoperative features and postoperative outcomes were examined. Continuous variables are presented as mean ± standard deviation and range. Dichotomic/categorical variables are presented as absolute numbers and percentages. For the first set of variables, Spearman’s rho was used to calculate correlation coefficients. Given that the variables were approximate measures of the same entity, only the most significant predictive variable was used in a sequential multiple linear regression to avoid possible multicollinearity. The explained variability measure increment (ΔR2) was then assessed for each variable of the second set when it was added to the model. Only variables with a large ΔR2 were included in the final equation, the goal being to design the most parsimonious linear model that could predict the volume of CSF drained. The most significant predictive variable was plotted against the volume of CSF drained (binary classified as low ≤40 ml vs high ≥40 ml) using receiver operating characteristic curve analysis and the ‘point closest to the top-left part’ method to identify the cut-off value with the highest sensitivity and specificity. Possible differences of preoperative characteristics between the groups were analysed using the Fisher’s exact test for categorical variables and the Mann–Whitney U-test for continuous variables. All statistical tests were 2-sided, and P-values ≤0.01 were considered statistically significant since no correction for multiple testing was performed. Data were analysed using SPSS 22.0. (IBM, Armonk, NY, USA). RESULTS As shown in Table 1, we identified 24 consecutive patients (17 men, mean age 72 ± 9 years) who met inclusion criteria for the final analysis. Indication for intervention was a descending thoracic aneurysm in 13 cases, a thoracoabdominal aneurysm in 4 cases and a chronic type B dissection in 7 cases. The technical success rate was 100%; all of the patients completed the operation neurologically intact and none developed SCI in the postoperative period or after discharge. The LSA was never excluded, either because of a proximal stent graft landing below its origin (12 cases) or a carotid-to-LSA bypass/transposition when its origin was covered by the stent graft (12 cases). Two patients had unilateral exclusion of the HGA. The use of vasopressors to achieve the target mean arterial blood pressure in the postoperative period was necessary in 7 cases (29%). There were no in-hospital deaths. Complications related to lumbar drains occurred in 3 cases (1 spinal headache, 1 CSF leak not requiring intervention, 1 CSF leak requiring intervention). Table 1: Preoperative characteristics with comparison between groups (<40 ml vs ≥40 ml) Preoperative characteristics Total, n/total (%) P-value Patients (gendera) 24 (17/7) 0.64 Ageb (years) 72 ± 9 (50–87) Hypertension 22/24 (92) 1.00 Diabetes mellitus 3/24 (12.5) 0.25 Dyslipidaemia 17/24 (71) 0.65 Smoking 10/24 (42) 0.67 Coronary artery disease 9/24 (38) 0.41 Chronic heart failure 11/24 (46) 0.68 Chronic obstructive pulmonary 8/24 (33) 1.00 Chronic kidney disease 3/24 (16.7) 0.58 Type of disease 0.05  Aneurysm 17/24 (71)   Descending thoracic aneurysms 13/24   Thoracoabdominal aneurysms 4/24  Chronic type B dissection 7/24 (29) 0.05 Previous AAA repair 8/24 (33) 0.36  Endovascular repair 6/24  Open surgery 2/24 Preoperative characteristics Total, n/total (%) P-value Patients (gendera) 24 (17/7) 0.64 Ageb (years) 72 ± 9 (50–87) Hypertension 22/24 (92) 1.00 Diabetes mellitus 3/24 (12.5) 0.25 Dyslipidaemia 17/24 (71) 0.65 Smoking 10/24 (42) 0.67 Coronary artery disease 9/24 (38) 0.41 Chronic heart failure 11/24 (46) 0.68 Chronic obstructive pulmonary 8/24 (33) 1.00 Chronic kidney disease 3/24 (16.7) 0.58 Type of disease 0.05  Aneurysm 17/24 (71)   Descending thoracic aneurysms 13/24   Thoracoabdominal aneurysms 4/24  Chronic type B dissection 7/24 (29) 0.05 Previous AAA repair 8/24 (33) 0.36  Endovascular repair 6/24  Open surgery 2/24 a Total (m/f). b Continuous variables are presented as mean ± SD (range). AAA: abdominal aortic aneurysm; f: female; m: male; SD: standard deviation. Table 1: Preoperative characteristics with comparison between groups (<40 ml vs ≥40 ml) Preoperative characteristics Total, n/total (%) P-value Patients (gendera) 24 (17/7) 0.64 Ageb (years) 72 ± 9 (50–87) Hypertension 22/24 (92) 1.00 Diabetes mellitus 3/24 (12.5) 0.25 Dyslipidaemia 17/24 (71) 0.65 Smoking 10/24 (42) 0.67 Coronary artery disease 9/24 (38) 0.41 Chronic heart failure 11/24 (46) 0.68 Chronic obstructive pulmonary 8/24 (33) 1.00 Chronic kidney disease 3/24 (16.7) 0.58 Type of disease 0.05  Aneurysm 17/24 (71)   Descending thoracic aneurysms 13/24   Thoracoabdominal aneurysms 4/24  Chronic type B dissection 7/24 (29) 0.05 Previous AAA repair 8/24 (33) 0.36  Endovascular repair 6/24  Open surgery 2/24 Preoperative characteristics Total, n/total (%) P-value Patients (gendera) 24 (17/7) 0.64 Ageb (years) 72 ± 9 (50–87) Hypertension 22/24 (92) 1.00 Diabetes mellitus 3/24 (12.5) 0.25 Dyslipidaemia 17/24 (71) 0.65 Smoking 10/24 (42) 0.67 Coronary artery disease 9/24 (38) 0.41 Chronic heart failure 11/24 (46) 0.68 Chronic obstructive pulmonary 8/24 (33) 1.00 Chronic kidney disease 3/24 (16.7) 0.58 Type of disease 0.05  Aneurysm 17/24 (71)   Descending thoracic aneurysms 13/24   Thoracoabdominal aneurysms 4/24  Chronic type B dissection 7/24 (29) 0.05 Previous AAA repair 8/24 (33) 0.36  Endovascular repair 6/24  Open surgery 2/24 a Total (m/f). b Continuous variables are presented as mean ± SD (range). AAA: abdominal aortic aneurysm; f: female; m: male; SD: standard deviation. The target spinal pressure was achieved in all of the patients, and the mean volume of CSF drained over 72 h was 46 ml (range 0–200 ml). The mean length of the aortic coverage was 231 mm (range 100–386 mm). The mean number of total preoperative SAs and of covered preoperative SAS was, respectively, 22 (range 13–30) and 9 (range 4–15). All of the patients had SAs visualized in the region of stent graft placement on postoperative CTA. As shown in Table 2, after the endovascular intervention, the mean number of total SAs decreased to 21 (range 13–30) and the mean number of covered SAs decreased to 8 (range 4–15). However, on postoperative CTA, the reduction in the number of visible (total and covered) SAs was not different for patients draining low volumes of CSF compared with patients draining high volumes of CSF (Table 3). Table 2: Intraoperative variables and postoperative outcomes Total TEVARa 20/24 (83) FEVARa 4/24 (17) HGAs excludeda 2/48 (4) LSAs excludeda 0/24 (0) Vasopressor usea 7/24 (29) Length of aortic coverageb (mm) 231 ± 76 (100–386) Any type of endoleaka 0/24 (0) Spinal cord ischaemiaa 0/24 (0) CSF drainage-related complicationsa 3/24 (12.5) Volume of CSF drainedb (ml) 46 ± 57 (0–200) Total preoperative SAsb (n) 22 ± 4 (13–30) Covered preoperative SAsb (n) 9 ± 3 (4–15) Total postoperative SAsb (n) 21 ± 4 (13–30) Covered postoperative SAsb (n) 8 ± 3 (4–15) Total TEVARa 20/24 (83) FEVARa 4/24 (17) HGAs excludeda 2/48 (4) LSAs excludeda 0/24 (0) Vasopressor usea 7/24 (29) Length of aortic coverageb (mm) 231 ± 76 (100–386) Any type of endoleaka 0/24 (0) Spinal cord ischaemiaa 0/24 (0) CSF drainage-related complicationsa 3/24 (12.5) Volume of CSF drainedb (ml) 46 ± 57 (0–200) Total preoperative SAsb (n) 22 ± 4 (13–30) Covered preoperative SAsb (n) 9 ± 3 (4–15) Total postoperative SAsb (n) 21 ± 4 (13–30) Covered postoperative SAsb (n) 8 ± 3 (4–15) a Categorical variables are presented as n/total (%). b Continuous variables are presented as mean ± SD (range). CSF: cerebrospinal fluid; FEVAR: fenestrated endovascular aneurysm repair; HGA: hypogastric artery; LSA: left subclavian artery; SAs: segmental arteries; SD: standard deviation; TEVAR: thoracic endovascular aortic repair. Table 2: Intraoperative variables and postoperative outcomes Total TEVARa 20/24 (83) FEVARa 4/24 (17) HGAs excludeda 2/48 (4) LSAs excludeda 0/24 (0) Vasopressor usea 7/24 (29) Length of aortic coverageb (mm) 231 ± 76 (100–386) Any type of endoleaka 0/24 (0) Spinal cord ischaemiaa 0/24 (0) CSF drainage-related complicationsa 3/24 (12.5) Volume of CSF drainedb (ml) 46 ± 57 (0–200) Total preoperative SAsb (n) 22 ± 4 (13–30) Covered preoperative SAsb (n) 9 ± 3 (4–15) Total postoperative SAsb (n) 21 ± 4 (13–30) Covered postoperative SAsb (n) 8 ± 3 (4–15) Total TEVARa 20/24 (83) FEVARa 4/24 (17) HGAs excludeda 2/48 (4) LSAs excludeda 0/24 (0) Vasopressor usea 7/24 (29) Length of aortic coverageb (mm) 231 ± 76 (100–386) Any type of endoleaka 0/24 (0) Spinal cord ischaemiaa 0/24 (0) CSF drainage-related complicationsa 3/24 (12.5) Volume of CSF drainedb (ml) 46 ± 57 (0–200) Total preoperative SAsb (n) 22 ± 4 (13–30) Covered preoperative SAsb (n) 9 ± 3 (4–15) Total postoperative SAsb (n) 21 ± 4 (13–30) Covered postoperative SAsb (n) 8 ± 3 (4–15) a Categorical variables are presented as n/total (%). b Continuous variables are presented as mean ± SD (range). CSF: cerebrospinal fluid; FEVAR: fenestrated endovascular aneurysm repair; HGA: hypogastric artery; LSA: left subclavian artery; SAs: segmental arteries; SD: standard deviation; TEVAR: thoracic endovascular aortic repair. Table 3: Differences in visualization of intercostal and lumbar segmental arteries (total and covered) before and after endovascular repair SAs Total Low volume of CSF drained (<40 ml) High volume of CSF drained (≥40 ml) P-value Total preoperative SAsa (n) 22 ± 4 (13–30) 24 ± 3 (18–30) 21 ± 4 (13–29) Total postoperative SAsa (n) 21 ± 4 (13–30) 23 ± 3 (18–30) 20 ± 4 (13–27) 1 Covered preoperative SAsa (n) 9 ± 3 (4–15) 7 ± 2 (4–13) 11 ± 3 (8–15) Covered postoperative SAsa (n) 8 ± 3 (4–15) 6 ± 2 (4–13) 10 ± 3 (8–15) 1 SAs Total Low volume of CSF drained (<40 ml) High volume of CSF drained (≥40 ml) P-value Total preoperative SAsa (n) 22 ± 4 (13–30) 24 ± 3 (18–30) 21 ± 4 (13–29) Total postoperative SAsa (n) 21 ± 4 (13–30) 23 ± 3 (18–30) 20 ± 4 (13–27) 1 Covered preoperative SAsa (n) 9 ± 3 (4–15) 7 ± 2 (4–13) 11 ± 3 (8–15) Covered postoperative SAsa (n) 8 ± 3 (4–15) 6 ± 2 (4–13) 10 ± 3 (8–15) 1 a Continuous variables are presented as mean ± SD (range). CSF: cerebrospinal fluid; SAs: segmental arteries; SD: standard deviation. Table 3: Differences in visualization of intercostal and lumbar segmental arteries (total and covered) before and after endovascular repair SAs Total Low volume of CSF drained (<40 ml) High volume of CSF drained (≥40 ml) P-value Total preoperative SAsa (n) 22 ± 4 (13–30) 24 ± 3 (18–30) 21 ± 4 (13–29) Total postoperative SAsa (n) 21 ± 4 (13–30) 23 ± 3 (18–30) 20 ± 4 (13–27) 1 Covered preoperative SAsa (n) 9 ± 3 (4–15) 7 ± 2 (4–13) 11 ± 3 (8–15) Covered postoperative SAsa (n) 8 ± 3 (4–15) 6 ± 2 (4–13) 10 ± 3 (8–15) 1 SAs Total Low volume of CSF drained (<40 ml) High volume of CSF drained (≥40 ml) P-value Total preoperative SAsa (n) 22 ± 4 (13–30) 24 ± 3 (18–30) 21 ± 4 (13–29) Total postoperative SAsa (n) 21 ± 4 (13–30) 23 ± 3 (18–30) 20 ± 4 (13–27) 1 Covered preoperative SAsa (n) 9 ± 3 (4–15) 7 ± 2 (4–13) 11 ± 3 (8–15) Covered postoperative SAsa (n) 8 ± 3 (4–15) 6 ± 2 (4–13) 10 ± 3 (8–15) 1 a Continuous variables are presented as mean ± SD (range). CSF: cerebrospinal fluid; SAs: segmental arteries; SD: standard deviation. As shown in Table 4, the volume of CSF drained correlated significantly with all of the aforementioned variables (length of aortic coverage, total preoperative SAs, covered preoperative SAs and the covering ratio of SAs). However, the covering ratio of SAs showed the most powerful correlation and was used for univariate regression. The only variable with a large effect size was then identified as the use of vasopressors and subsequently introduced in the multivariable regression (Table 5). Consequently, we developed a linear model whose derived equation could predict the volume of CSF drained if the covering ratio of SAs and the use of vasopressors were known, which is shown in the Central Image. Table 4: Correlations for the primary outcome measures with the volume of cerebrospinal fluid drained Total preoperative SAs Covered preoperative SAs Length of aortic coverage (mm) Covering ratio of SAs Spearman’s rho Volume of CSF drained Correlation coefficient 0.594 0.603 0.693 0.731 P-value (2-tailed) 0.002 <0.001 <0.001 <0.001 Sample size (N) 24 24 24 24 Total preoperative SAs Covered preoperative SAs Length of aortic coverage (mm) Covering ratio of SAs Spearman’s rho Volume of CSF drained Correlation coefficient 0.594 0.603 0.693 0.731 P-value (2-tailed) 0.002 <0.001 <0.001 <0.001 Sample size (N) 24 24 24 24 CSF: cerebrospinal fluid; SAs: segmental arteries. Table 4: Correlations for the primary outcome measures with the volume of cerebrospinal fluid drained Total preoperative SAs Covered preoperative SAs Length of aortic coverage (mm) Covering ratio of SAs Spearman’s rho Volume of CSF drained Correlation coefficient 0.594 0.603 0.693 0.731 P-value (2-tailed) 0.002 <0.001 <0.001 <0.001 Sample size (N) 24 24 24 24 Total preoperative SAs Covered preoperative SAs Length of aortic coverage (mm) Covering ratio of SAs Spearman’s rho Volume of CSF drained Correlation coefficient 0.594 0.603 0.693 0.731 P-value (2-tailed) 0.002 <0.001 <0.001 <0.001 Sample size (N) 24 24 24 24 CSF: cerebrospinal fluid; SAs: segmental arteries. Table 5: Multivariable regression for covering ratio of segmental arteries and use of vasopressors Predictor B (95% CI) SE B β P-value Step 1  Intercept −65.1 (99.6 to −30.7) 16.1 0.001  Covering ratio of segmental arteries 284.3 (203.6–364.9) 38.9 0.84 0.000 Step 2  Intercept −74.7 (−101.7 to −47.6) 13.0 0.000  Covering ratio of segmental arteries 275.8 (213.4–338.3) 30.0 0.81 0.000  Use of vasopressorsa 44.0 (21.2–66.8) 11.0 0.35 0.001 Predictor B (95% CI) SE B β P-value Step 1  Intercept −65.1 (99.6 to −30.7) 16.1 0.001  Covering ratio of segmental arteries 284.3 (203.6–364.9) 38.9 0.84 0.000 Step 2  Intercept −74.7 (−101.7 to −47.6) 13.0 0.000  Covering ratio of segmental arteries 275.8 (213.4–338.3) 30.0 0.81 0.000  Use of vasopressorsa 44.0 (21.2–66.8) 11.0 0.35 0.001 R2 = 0.708 for step 1 and ΔR2 = 0.127 for step 2. a Coded 0: no vasopressors; 1: vasopressors. B: unstandardized coefficients; CI: confidence interval; SE: standard error; β: standardized coefficients. Table 5: Multivariable regression for covering ratio of segmental arteries and use of vasopressors Predictor B (95% CI) SE B β P-value Step 1  Intercept −65.1 (99.6 to −30.7) 16.1 0.001  Covering ratio of segmental arteries 284.3 (203.6–364.9) 38.9 0.84 0.000 Step 2  Intercept −74.7 (−101.7 to −47.6) 13.0 0.000  Covering ratio of segmental arteries 275.8 (213.4–338.3) 30.0 0.81 0.000  Use of vasopressorsa 44.0 (21.2–66.8) 11.0 0.35 0.001 Predictor B (95% CI) SE B β P-value Step 1  Intercept −65.1 (99.6 to −30.7) 16.1 0.001  Covering ratio of segmental arteries 284.3 (203.6–364.9) 38.9 0.84 0.000 Step 2  Intercept −74.7 (−101.7 to −47.6) 13.0 0.000  Covering ratio of segmental arteries 275.8 (213.4–338.3) 30.0 0.81 0.000  Use of vasopressorsa 44.0 (21.2–66.8) 11.0 0.35 0.001 R2 = 0.708 for step 1 and ΔR2 = 0.127 for step 2. a Coded 0: no vasopressors; 1: vasopressors. B: unstandardized coefficients; CI: confidence interval; SE: standard error; β: standardized coefficients. Finally, receiver operating characteristic curve analysis for the prediction of a low volume of CSF drainage (<40 ml) showed that, for a covering ratio 0.385–0.465, sensitivity and specificity values oscillated between 0.857–1 and 0.931–1, respectively. Indeed, all of the patients with a covering ratio ≤0.4 had a low volume of CSF drained (0–40 ml), whereas all of the patients with a covering ratio >0.4 had a high volume of CSF drained (60–200 ml). DISCUSSION SCI is the most feared and most dramatic complication of descending thoracic and thoracoabdominal aortic procedures that has not been completely eliminated by endovascular repair [11]. The risk of SCI increases with the extent of the aneurysm, the length of the stent graft used, pre-existing aortic reconstruction and occlusion of the LSA or the HGA [3–5, 12–15]. Furthermore, recent studies have shown that the total number of intercostal and lumbar SAs sacrificed is a more powerful predictor of the risk of paraplegia than is the loss of any specific one [12, 13]. These clinical findings are complemented by laboratory studies suggesting that an extensive longitudinally continuous collateral network must exist to explain preservation of spinal cord perfusion when intercostal and lumbar SAs are interrupted [1, 14, 15]. To reduce the risk of SCI, current guidelines for the management of descending thoracic aorta disease state that the selective use of prophylactic CSF drainage with thoracic endovascular aortic repair (TEVAR) should be considered in patients with planned extensive aortic coverage (≥200 mm) and/or previous AAA repair [16]. The main findings of our analysis were that the percentage of visible and perfused intercostal and lumbar SAs covered (expressed as covering ratio) showed the most powerful correlation and could predict the volume of CSF drained. All of the patients with a covering ratio ≤0.4 had a low volume of CSF drained (0–40 ml), whereas all of the patients with a covering ratio >0.4 had a high volume of CSF drained (60–200 ml). Although our findings could suggest that mandatory CSF drainage may have a protective role in preventing neurological complications in that none of our patients developed SCI, our analysis might also indicate that only a specific subset of patients probably received substantial benefit from prophylactic CSF drainage. Indeed, previous findings suggested that a volume of CSF drained <50 ml is possibly inadequate to have a meaningful treatment effect in most patients [17]. Based on current guidelines, all of the patients in the high-drainage group but half of the patients in the low-drainage group would have received prophylactic CSF drainage. However, we could not establish any relationship between the amount of CSF drained and the occurrence of SCI or between the ratio of covered SAs and the occurrence of SCI. Thus, when considering prophylactic placement of lumbar drains, whether preoperative assessment of the expected covering ratio of SAs may help identify those patients in whom this procedure would result in low volumes of CSF drainage that may then be safely avoided still represents an unresolved issue. We found that the majority of intercostal and lumbar SAs were still visible in the region of stent graft coverage, a finding also present in previous studies [8–10, 14]. Furthermore, on postoperative CTA, the reduction in the number of visible (total and covered) SAs was not different for patients draining low volumes of CSF compared with patients draining high volumes of CSF. Indeed, even if endovascular repair inevitably occludes direct inflow to intercostal and lumbar SAs in the region of stent graft placement, the distal segments of these vessels are seen to remain patent through collateral pathways. Even if retrograde perfusion through the collateral network is able to restore blood flow within covered intercostal and lumbar SAs, spinal cord perfusion during the early postoperative period after endovascular repair could actually be reduced. It seems reasonable to assume that the higher the percentage of SAs covered (i.e. the covering ratio), the more likely an ischaemic insult to the spinal cord is to occur. We hypothesized that beyond a critical point, this ischaemic insult would result in spinal cord swelling (as indicated by a rise in spinal fluid pressure), and subsequent compensation eventually occurred by means of CSF drainage. This hypothesis is also supported by animal studies that showed that a time interval of 2–4 days is required for compensation of spinal cord perfusion [18]. In this regard, careful analysis of medical records indicated that in all of the patients requiring high volumes of CSF drainage, more than half of the volume was drained in the first 24 h after the index operation. Thus, it seems reasonable that CSF drainage may be crucial during this period for restoration of the fragile spinal cord perfusion postoperatively, in particular in those patients with a high proportion of intercostal and lumbar SAs that were overstented. The 2 major recognized medical interventions to prevent and treat SCI in patients undergoing descending thoracic and thoracoabdominal aortic procedures are arterial pressure augmentation and CSF drainage [19, 20]. However, overall, the available literature regarding prophylactic CSF drainage in patients treated by endovascular means is only of moderate quality, and the type of patient who could benefit the most from CSF drainage has not yet been clearly defined [13]. Indeed, current guidelines are mainly based on data from open thoracic aortic repair and the efficacy of CSF drainage still remains a controversial issue because 2 systematic reviews of the available evidence have failed to establish prophylactic CSF drainage as a method of preventing SCI in patients who have TEVAR [19, 21]. As a result, despite different institutional protocols [22, 23], pre-emptive CSF drainage is recommended as a prophylactic procedure in patients at high risk of SCI during TEVAR [24]. A major shortcoming of current guidelines is the lack of a method that accounts for interindividual variability in spinal cord vascular supply distribution [25]. Indeed, the length of the thoracic aorta exclusion associated with neurological complications and the subsequent need for prophylactic protective measures differ across different studies [3–6]. In our study, the anatomical location of the SAs that were overstented did not influence the amount of CSF drained in our study, and the covering ratio may permit us to overcome the interindividual variability in SAs distribution along the spinal cord. Furthermore, when a history of previous AAA repair is considered, our results indicate that the proposed model is still valid in patients with previous aortic intervention. In our cohort, 36% (n = 8) of the patients had a previous AAA repair, but the proposed model was still valid for these individuals. Indeed, only those patients with a covering ratio >0.4 (n = 4) had a high volume of CSF drained, whereas those patients with a covering ratio ≤0.4 (n = 4) had a low volume of CSF drained. This result may indicate that, with an appropriate time delay, the collateral network can compensate for SAs exclusion after AAA repair. However, no firm conclusion can be drawn at this time about which time interval should be adopted in order to enable the collateral network to compensate for SAs loss. Care should be always taken when considering secondary endovascular repair in patients with previous open surgery. Although the potential benefit of CSF drainage may be large, the procedure itself is invasive and not without risks, and the utility of preoperative lumbar drain placement should also be weighed against the risk of related complications, which is not negligible. Intuitively, the size of the catheter used, the spinal fluid pressure needed for drainage, the amount of CSF drained and the duration of catheter insertion could all impact the incidence of CSF drainage-related complications. However, a recent systematic review and meta-analysis did not find any correlation between any of the previous factors and the incidence of CSF drain-related complications [26]. In our cohort, the need for the use of vasopressors in the postoperative period to achieve the target mean arterial blood pressure predicted an increase in the volume of CSF drained of about 45 ml. Coverage of the LSA may also induce SCI by reducing spinal cord blood flow; the clinical significance of this source of collateral perfusion to the spinal cord has been clearly demonstrated [27, 28]. In this regard, in our study LSA revascularization was used routinely when the LSA needed to be covered; as a result, all of the patients included in the study had patent LSA. Thus, our findings may indicate that adequate perfusion of the spinal cord is strictly dependent on other factors, first haemodynamic stability and LSA patency. Accurate preoperative knowledge of the arterial supply to the spinal cord is a prerequisite for stratifying and decreasing the risk of SCI. The ability to discriminate between arterial supply and venous drainage is a clear advantage of magnetic resonance angiography; moreover, proximity to skeletal structures and the body mass of the patient are irrelevant [29]. However, millimetre-sized arteries are nowadays well within the detection capabilities of CTA [30]. The fact that, in most centres, CTA is the imaging modality of choice for most patients with descending thoracic and thoracoabdominal aortic disease makes magnetic resonance angiography less practical for preoperative planning/sizing of endovascular treatment in the real-life clinical environment since the data set required for analysis of spinal cord vascular anatomy is already available. Limitations The retrospective comparison limited the variables available for analysis, and some unidentified confounding bias might have influenced the results. Also, the small sample size and the selection of patients from a single centre may have limited the generalizability of our results. Finally, individual assessment of CTA was not performed for each type of disease. Although this approach was due to the limited number of patients available for evaluation, it should be borne in mind that endovascular repair is the technique of choice for patients with a variety of aortic diseases and our study population can be considered representative of the wide spectrum of patients potentially likely to benefit from endovascular procedures. CONCLUSIONS Our findings suggest that the percentage of intercostal and lumbar SAs covered by stent graft placement can predict the volume of CSF drained in patients undergoing elective endovascular repair of descending thoracic and thoracoabdominal aortic disease. If confirmed by further prospective studies, this finding may lead to the application of the covering ratio to selected patients who require placement of prophylactic lumbar drains. Conflict of interest: none declared. REFERENCES 1 Etz CD , Kari FA , Mueller CS , Silovitz D , Brenner RM , Lin HM. The collateral network concept: a reassessment of the anatomy of spinal cord perfusion . J Thorac Cardiovasc Surg 2011 ; 141 : 1020 – 8 . Google Scholar Crossref Search ADS PubMed 2 Melissano G , Civilini E , Bertoglio L , Calliari F , Campos Moraes Amato A , Chiesa R. Angio-CT imaging of the spinal cord vascularisation: a pictorial essay . Eur J Vasc Endovasc Surg 2010 ; 39 : 436 – 40 . Google Scholar Crossref Search ADS PubMed 3 Bisdas T , Panuccio G , Sugimoto M , Torsello G , Austermann M. Risk factors of spinal cord ischemia after endovascular repair of thoracoabdominal aortic aneurysms . J Vasc Surg 2015 ; 61 : 1408 – 16 . Google Scholar Crossref Search ADS PubMed 4 Buth J , Harris PL , Hobo R , van Heps R , Cuypers P , Duijm L. Neurologic complications associated with endovascular repair of thoracic aortic pathology: incidence and risk factors. A study from the European Collaborators on Stent/Graft Techniques for Aortic Aneurysm Repair (EUROSTAR) registry . J Vasc Surg 2007 ; 46 : 1103 – 11 . Google Scholar Crossref Search ADS PubMed 5 Feezor RJ , Martin TD , Hess PJ , Daniels MJ , Beaver TM , Klodell CT et al. Extent of aortic coverage and incidence of spinal cord ischemia after thoracic endovascular aneurysm repair . Ann Thorac Surg 2008 ; 86 : 1809 – 14 . Google Scholar Crossref Search ADS PubMed 6 Bobadilla JL , Wynn M , Tefera G , Acher CW. Low incidence of paraplegia after thoracic endovascular aneurysm repair with proactive spinal cord protective protocols . J Vasc Surg 2013 ; 57 : 1537 – 42 . Google Scholar Crossref Search ADS PubMed 7 Khan NR , Smalley Z , Nesvick CL , Lee SL , Michael LM 2nd . The use of lumbar drains in preventing spinal cord injury following thoracoabdominal aneurysm repair: an update systematic review and meta-analysis . J Neurosurg Spine 2016 ; 25 : 383 – 93 . Google Scholar Crossref Search ADS PubMed 8 Kasprzak PM , Gallis K , Cucuruz B , Pfister K , Janotta M , Kopp R. Editors’ choice—Temporary aneurysm sac perfusion as an adjunct for prevention of spinal cord ischemia after branched endovascular repair of thoracoabdominal aneurysms . Eur J Vasc Endovasc Surg 2014 ; 48 : 258 – 65 . Google Scholar Crossref Search ADS PubMed 9 von Tengg-Kobligk H , Böckler D , Jose TM , Ganten M , Kotelis O , Nagel S et al. Feeding arteries of the spinal cord at CT angiography before and after thoracic aortic endografting . J Endovasc Ther 2007 ; 14 : 639 – 49 . Google Scholar Crossref Search ADS PubMed 10 Chang CK , Chuter TAM , Reilly LM , Ota MK , Furtado A , Bucci M et al. Spinal arterial anatomy and risk factors for lower extremity weakness following endovascular thoracoabdominal aortic aneurysm repair with branched stent-grafts . J Endovasc Ther 2008 ; 15 : 356 – 62 . Google Scholar Crossref Search ADS PubMed 11 Cheng D , Martin J , Shennib H , Dunning J , Muneretto C , Schueler S et al. Endovascular aortic repair versus open surgical repair for descending thoracic aortic disease: a systematic review and meta-analysis of comparative studies . J Am Coll Cardiol 2010 ; 55 : 986 – 1001 . Google Scholar Crossref Search ADS PubMed 12 Gallis K , Kasprzak PM , Cucuruz B , Kopp R. Evaluation of visible spinal arteries on computed tomography angiography before and after branched stent-graft repair for thoracoabdominal aortic aneurysm . J Vasc Surg 2017 ; 65 : 1577 – 83 . Google Scholar Crossref Search ADS PubMed 13 Kari FA , Saravi B , Krause S , Puttfarcken L , Scheumann J , Förster K et al. New insights into spinal cord ischaemia after thoracic aortic procedures: the importance of the number of anterior radiculomedullary arteries for surgical outcome . Eur J Cardiothorac Surg 2018 ; 54 : 149 – 56 . Google Scholar Crossref Search ADS PubMed 14 Meffert P , Bischoff MS , Brenner R , Siepe M , Beyersdorf F , Kari FA. Significance and function of different spinal collateral compartments following thoracic aortic surgery: immediate versus long-term flow compensation . Eur J Cardiothorac Surg 2014 ; 45 : 799 – 804 . Google Scholar Crossref Search ADS PubMed 15 Eagleton MJ , Shah S , Petkosevek D , Mastracci TM , Greenberg RK. Hypogastric and subclavian artery patency affects onset and recovery of spinal cord ischemia associated with aortic endografting . J Vasc Surg 2014 ; 59 : 89 – 95 . Google Scholar Crossref Search ADS PubMed 16 Riambau V , Bockler D , Brunkwall J , Cao P , Chiesa R , Coppi G et al. Editor’s choice—Management of descending thoracic aorta diseases. Clinical practice guidelines of the European Society for Vascular Surgery (ESVS) . Eur J Vasc Endovasc Surg 2017 ; 53 : 4 – 52 . Google Scholar Crossref Search ADS PubMed 17 Fedorow CA , Moon MC , Mutch WAC , Grocott HP. Lumbar cerebrospinal fluid drainage for thoracoabdominal aortic surgery. Rationale and practical considerations for management . Anesth Analg 2010 ; 111 : 46 – 58 . Google Scholar Crossref Search ADS PubMed 18 Griepp EB , Di Luozzo G , Schray D , Stefanovic A , Geisbusch S , Griepp RB. The anatomy of the spinal cord collateral circulation . Ann Cardiothorac Surg 2012 ; 1 : 350 – 7 . Google Scholar PubMed 19 Dijkstra ML , Vainas T , Zeebregts CJ , Hooft L , van der Laan MJ. Editor’s choice—Spinal cord ischaemia in endovascular thoracic and thoraco-abdominal aortic repair: review of preventive strategies . Eur J Vasc Endovasc Surg 2018 ; 55 : 829 – 41 . Google Scholar Crossref Search ADS PubMed 20 Maurel B , Delclaux N , Sobocinski J , Hertault A , Martin-Gonzalez T , Moussa M et al. The impact of early pelvic and lower limb reperfusion and attentive perioperative management on the incidence of spinal cord ischemia during thoracoabdominal aortic aneurysm endovascular repair . Eur J Vasc Endovasc Surg 2015 ; 49 : 248 – 54 . Google Scholar Crossref Search ADS PubMed 21 Wong CS , Healy D , Canning C , Coffey JC , Boyle JR , Walsh SR. A systematic review of spinal cord injury and cerebrospinal fluid drainage after thoracic aortic endografting . J Vasc Surg 2012 ; 56 : 1438 – 47 . Google Scholar Crossref Search ADS PubMed 22 Keith CJ Jr , Passman MA , Carignan MJ , Parmar GM , Nagre SB , Patterson MA et al. Protocol implementation of selective postoperative lumbar spinal drainage after thoracic aortic endograft . J Vasc Surg 2012 ; 55 : 1 – 8 . Google Scholar Crossref Search ADS PubMed 23 Arnaoutakis DJ , Arnaoutakis GJ , Beaulieu RJ , Abularrage CJ , Lum YW , Black JH III. Results of adjunctive spinal drainage and/or left subclavian artery bypass in thoracic endovascular aortic repair . Ann Vasc Sur 2014 ; 28 : 65 – 73 . Google Scholar Crossref Search ADS 24 Mazzeffi M , Abuelkasem E , Drucker CB , Kalsi R , Toursavadkohi S , Harris DG et al. Contemporary single-center experience with prophylactic cerebrospinal fluid drainage for thoracic endovascular aortic repair in patients at high risk for ischemic spinal cord injury . J Cardiothorac Vasc Anesth 2018 ; 32 : 883 – 9 . Google Scholar Crossref Search ADS PubMed 25 Kari FA , Saravi B , Krause S , Puttfarcken L , Wittmann K , Förster K et al. Spinal ischaemia after thoracic endovascular aortic repair with left subclavian artery sacrifice: is there a critical stent graft length? Eur J Cardiothorac Surg 2018 ; 53 : 385 – 91 . Google Scholar Crossref Search ADS 26 Rong LQ , Kamel MK , Rahouma M , White RS , Lichtman AD , Pryor KO et al. Cerebrospinal-fluid drain-related complications in patients undergoing open and endovascular repairs of thoracic and thoraco-abdominal aortic pathologies: a systematic review and meta-analysis . Br J Anaesth 2018 ; 120 : 904 – 13 . Google Scholar Crossref Search ADS PubMed 27 Huang Q , Chen XM , Yang H , Lin QN , Qin X. Effect of left subclavian artery revascularisation in thoracic endovascular aortic repair: a systematic review and meta-analysis . Eur J Vasc Endovasc Surg 2018 ; [Epub ahead of print]. https://doi.org/10.1016/j.ejvs.2018.07.018. 28 Fukui S , Tanaka H , Kobayashi K , Kajiyama T , Mitsuno M , Yamamura M et al. Development of collaterals to the spinal cord after endovascular stent graft repair of thoracic aneurysms . Eur J Vasc Endovasc Surg 2016 ; 52 : 801 – 7 . Google Scholar Crossref Search ADS PubMed 29 Melissano G , Chiesa R. Advances in imaging of the spinal cord vascular supply and its relationship with paraplegia after aortic interventions. A review . Eur J Vasc Endovasc Surg 2009 ; 38 : 567 – 77 . Google Scholar Crossref Search ADS PubMed 30 Boll DT , Bulow H , Blackham KA , Kajiyama T , Mitsuno M , Yamamura M et al. MDCT angiography of the spinal vasculature and the artery of Adamkievicz . AJR Am J Roentgenol 2006 ; 187 : 1054 – 60 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Coverage of visible intercostal and lumbar segmental arteries can predict the volume of cerebrospinal fluid drainage in elective endovascular repair of descending thoracic and thoracoabdominal aortic disease: a pilot study
JF - European Journal of Cardio-Thoracic Surgery
DO - 10.1093/ejcts/ezy371
DA - 2019-04-01
UR - https://www.deepdyve.com/lp/oxford-university-press/coverage-of-visible-intercostal-and-lumbar-segmental-arteries-can-DcU46CX0FM
SP - 646
VL - 55
IS - 4
DP - DeepDyve
ER -