Validation of a patient-specific hemodynamic computational model for surgical planning of vascular access in hemodialysis patients
2013; Elsevier BV; Volume: 84; Issue: 6 Linguagem: Inglês
10.1038/ki.2013.188
ISSN1523-1755
AutoresAnna Caroli, Simone Manini, Luca Antiga, Katia Passera, Bogdan Ene‐Iordache, Stefano Rota, Giuseppe Remuzzi, Aron Bode, Jaap Leermakers, Frans N. van de Vosse, Raymond Vanholder, Marko Malovrh, Jan H.M. Tordoir, Andrea Remuzzi, on behalf of the ARCH project Consortium,
Tópico(s)Dialysis and Renal Disease Management
ResumoVascular access dysfunction is one of the main causes of morbidity and hospitalization in hemodialysis patients. This major clinical problem points out the need for prediction of hemodynamic changes induced by vascular access surgery. Here we reviewed the potential of a patient-specific computational vascular network model that includes vessel wall remodeling to predict blood flow change within 6 weeks after surgery for different arteriovenous fistula configurations. For model validation, we performed a multicenter, prospective clinical study to collect longitudinal data on arm vasculature before and after surgery. Sixty-three patients with newly created arteriovenous fistula were included in the validation data set and divided into four groups based on fistula configuration. Predicted brachial artery blood flow volumes 40 days after surgery had a significantly high correlation with measured values. Deviation of predicted from measured brachial artery blood flow averaged 3% with a root mean squared error of 19.5%, showing that the computational tool reliably predicted patient-specific blood flow increase resulting from vascular access surgery and subsequent vascular adaptation. This innovative approach may help the surgeon to plan the most appropriate fistula configuration to optimize access blood flow for hemodialysis, potentially reducing the incidence of vascular access dysfunctions and the need of patient hospitalization. Vascular access dysfunction is one of the main causes of morbidity and hospitalization in hemodialysis patients. This major clinical problem points out the need for prediction of hemodynamic changes induced by vascular access surgery. Here we reviewed the potential of a patient-specific computational vascular network model that includes vessel wall remodeling to predict blood flow change within 6 weeks after surgery for different arteriovenous fistula configurations. For model validation, we performed a multicenter, prospective clinical study to collect longitudinal data on arm vasculature before and after surgery. Sixty-three patients with newly created arteriovenous fistula were included in the validation data set and divided into four groups based on fistula configuration. Predicted brachial artery blood flow volumes 40 days after surgery had a significantly high correlation with measured values. Deviation of predicted from measured brachial artery blood flow averaged 3% with a root mean squared error of 19.5%, showing that the computational tool reliably predicted patient-specific blood flow increase resulting from vascular access surgery and subsequent vascular adaptation. This innovative approach may help the surgeon to plan the most appropriate fistula configuration to optimize access blood flow for hemodialysis, potentially reducing the incidence of vascular access dysfunctions and the need of patient hospitalization. More than 940 patients per million population in Europe are affected by end-stage renal disease and live on chronic renal replacement therapy. Approximately 80% of these patients are treated chronically by hemodialysis (HD) (European Renal Association–European Dialysis and Transplantation Association (ERA-EDTA) Guidelines). The total number of patients on HD in Europe exceeds 500,000, and it increases annually at a constant rate of ∼7%.1.Grassmann A. Gioberge S. Moeller S. et al.ESRD patients in 2004: global overview of patient numbers, treatment modalities and associated trends.Nephrol Dial Transplant. 2005; 20: 2587-2593Crossref PubMed Scopus (397) Google Scholar Despite the major advances of HD procedure during the past three decades, the Achilles heel of this treatment is the vascular access (VA) required to connect patient's blood circulation to the artificial kidney. For safe and long-lasting VA, the native arteriovenous fistula (AVF), surgically created in the arm by anastomosis of an artery to a vein, is recommended by international guidelines (National Kidney Foundation Kidney Disease Outcomes Quality Initiative (NKF K-DOQI) Guidelines; ERA-EDTA European Best Practice Guidelines on HD).2.Tordoir J. Canaud B. Haage P. et al.EBPG on vascular access.Nephrol Dial Transplant. 2007; 22: ii88-117Crossref PubMed Scopus (639) Google Scholar However, short- and long-term AVF dysfunctions, including inadequate increase in blood flow volume (BFV) after surgery (nonmaturation), vein thrombotic occlusion, ischemic circulation in the distal arm and in the hand (steal syndrome), and massive increase in VA BFV with risk of cardiac failure, are among the major causes of morbidity and hospitalization in HD patients.3.Ikizler T.A. Himmelfarb J. Trials and trade-offs in haemodialysis vascular access monitoring.Nephrol Dial Transplant. 2006; 21: 3362-3363Crossref PubMed Scopus (7) Google Scholar,4.Allon M. Robbin M.L. Increasing arteriovenous fistulas in hemodialysis patients: problems and solutions.Kidney Int. 2002; 62: 1109-1124Abstract Full Text Full Text PDF PubMed Google Scholar Indeed AVF primary patency at 2 years after surgery was recently estimated to be ∼50%5.Field M. MacNamara K. Bailey G. et al.Primary patency rates of AV fistulas and the effect of patient variables.J Vasc Access. 2008; 9: 45-50PubMed Google Scholar, 6.Huijbregts H.J. Bots M.L. Wittens C.H. CIMINO study group et al.Hemodialysis arteriovenous fistula patency revisited: results of a prospective, multicenter initiative.Clin J Am Soc Nephrol. 2008; 3: 714-719Crossref PubMed Scopus (280) Google Scholar, 7.Fokou M. Teyang A. Ashuntantang G. et al.Complications of arteriovenous fistula for hemodialysis: an 8-year study.Ann Vasc Surg. 2012; 26: 680-684Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar and even lower in the United States.8.Nguyen T.H. Bui T.D. Gordon I.L. et al.Functional patency of autogenous AV fistulas for hemodialysis.J Vasc Access. 2007; 8: 275-280PubMed Google Scholar,9.Schinstock C.A. Albright R.C. Williams A.W. et al.Outcomes of arteriovenous fistula creation after the Fistula First Initiative.Clin J Am Soc Nephrol. 2011; 6: 1996-2002Crossref PubMed Scopus (147) Google Scholar Prediction and prevention of VA dysfunction are still open clinical challenges, with more than 90,000 procedures/year performed in Europe for revision or reoperation.10.Tordoir J.H. Keuter X. Planken N. et al.Autogenous options in secondary and tertiary access for haemodialysis.Eur J Vasc Endovasc Surg. 2006; 31: 661-666Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar Currently, the choice of location and type of the anastomosis is based, beyond evaluation of systemic factors, on blood vessel diameter only, following the indication that radial arteries <2mm will probably result in AVF failure (NKF K-DOQI Guidelines), and considering that distal VA more likely results in lower BFV and in a high incidence of early nonfunction,11.Tordoir J.H. Rooyens P. Dammers R. et al.Prospective evaluation of failure modes in autogenous radiocephalic wrist access for haemodialysis.Nephrol Dial Transplant. 2003; 18: 378-383Crossref PubMed Scopus (163) Google Scholar whereas proximal VA more likely results in very high BFV, increasing the risk of steal syndrome and cardiac failure.12.Tordoir J.H. Dammers R. van der Sande F.M. Upper extremity ischemia and hemodialysis vascular access.Eur J Vasc Endovasc Surg. 2004; 27: 1-5Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar,13.Wijnen E. Keuter X.H. Planken N.R. et al.The relation between vascular access flow and different types of vascular access with systemic hemodynamics in hemodialysis patients.Artif Organs. 2005; 29: 960-964Crossref PubMed Scopus (71) Google Scholar However, BFV enhancement following VA creation is influenced by the combination of geometrical factors (for example, vascular diameters and lengths along the arterial tree), vessel topology (for example, number and size of venous side branches), peripheral resistance, type of anastomosis, and vessel adaptation. An objective and reliable prediction of postoperative increase in BFV over time could be extremely relevant for planning the optimal AVF configuration, and may reduce the incidence of VA dysfunctions. A number of computational approaches have been proposed for the simulation of hemodynamics and vascular wall dynamics in complex vascular networks, which could potentially be used to predict BFV change after VA creation. Among them, 0D and 1D pulse-wave propagation methods allow to efficiently model BFV, pressure distributions, and wall displacements throughout vascular networks at low computational costs.14.Huberts W. Bosboom E.M. van de Vosse F.N. A lumped model for blood flow and pressure in the systemic arteries based on an approximate velocity profile function.Math Biosci Eng. 2009; 6: 27-40Crossref PubMed Scopus (21) Google Scholar,15.Reymond P. Merenda F. Perren F. et al.Validation of a one-dimensional model of the systemic arterial tree.Am J Physiol Heart Circ Physiol. 2009; 297: H208-H222Crossref PubMed Scopus (447) Google Scholar Recently, a pulse-wave propagation model16.Huberts W. Bode A.S. Kroon W. et al.A pulse wave propagation model to support decision-making in vascular access planning in the clinic.Med Eng Phys. 2012; 34: 233-248Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar based on boundary layer theory and approximated velocity profile17.Bessems D. Rutten M. van de Vosse F. A wave propagation model of blood flow in large vessels using an approximate velocity profile function.J Fluid Mech. 2007; 580: 145-168Crossref Scopus (110) Google Scholar,18.van de Vosse F.N. Stergiopulos N. Pulse wave propagation in the arterial tree.Annu Rev Fluid Mech. 2011; 43: 467-499Crossref Scopus (254) Google Scholar has been developed to accurately predict preoperative and postoperative BFV over different AVF types and locations with a fast computational approach. More recently, a sensitivity analysis has been carried out for the identification of parameters most sensitive for patient-specific BFV prediction.19.Huberts W. de Jonge C. van der Linden W.P. et al.A sensitivity analysis of a personalized pulse wave propagation model for arteriovenous fistula surgery. Part A: Identification of most influential model parameters.Med Eng Phys. 2013; 35: 810-826Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar,20.Huberts W. de Jonge C. van der Linden W.P. et al.A sensitivity analysis of a personalized pulse wave propagation model for arteriovenous fistula surgery. Part B: Identification of possible generic model parameters.Med Eng Phys. 2012; 35: 827-837Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar As these models enable to estimate changes in BFV only immediately after surgery, we extended this modeling approach to include a simulation of vessel wall remodeling and consequent hemodynamic changes that are responsible for the so-called access maturation.21.Manini S, Passera K, Huberts W et al. Computational model for simulation of vascular adaptation following vascular access surgery in hemodialysis patients. Comput Meth Biomech Biomed Eng (in press).Google Scholar Despite promising results, the potential clinical use of these computational tools needs assessment of their reliability and of the accuracy of model prediction in terms of BFV redistribution in arm vasculature. To this aim, a multicenter longitudinal clinical prospective study was conducted in patients with end-stage renal disease awaiting VA creation for HD treatment in the context of the EU-FP7 research project ARCH.22.Bode A.S. Caroli A. Huberts W. et al.Clinical study protocol for the ARCH project—computational modeling for improvement of outcome after vascular access creation.J Vasc Access. 2011; 12: 369-376Crossref PubMed Scopus (23) Google Scholar In the present study, we used an ad hoc–developed computer program to predict, on the basis of preoperative ultrasound (US) measurements, patient-specific potential changes in BFV that take place immediately after VA surgery and during the subsequent 6-week period (VA maturation). We then compared predicted postoperative results with data measured within the ARCH clinical study to assess the reliability of theoretical prediction and to provide evidence on the potential use of the computational model as a tool for planning VA surgery in HD patients. A total of 93 consecutive patients (mean age 62±16 (19–85) years, 31% women, 33% diabetic) were enrolled in four European HD centers: Maastricht University Medical Center, The Netherlands (n=39); Universzitetni Klinikni Center Ljubljana, Slovenia (n=10); Ospedali Riuniti di Bergamo, Italy (n=39); and Ghent University Hospital, Belgium (n=5). All patients underwent clinical and vascular US examinations preoperatively, and they were systematically followed up after VA surgery through clinical and US examinations for a period of 2 years.22.Bode A.S. Caroli A. Huberts W. et al.Clinical study protocol for the ARCH project—computational modeling for improvement of outcome after vascular access creation.J Vasc Access. 2011; 12: 369-376Crossref PubMed Scopus (23) Google Scholar During each US examination, brachial, radial, and ulnar artery BFV and size of major arm vessel diameters were assessed.22.Bode A.S. Caroli A. Huberts W. et al.Clinical study protocol for the ARCH project—computational modeling for improvement of outcome after vascular access creation.J Vasc Access. 2011; 12: 369-376Crossref PubMed Scopus (23) Google Scholar According to the study protocol, any AVF dysfunction occurring during the observation period was recorded and, in case of early end of the study, information on VA function at study end was collected. In 38 patients, AVF was created in the upper arm, and was either brachiocephalic (BC, n=28) or brachiobasilic (BB, n=10), whereas in 55 patients AVF was created in the lower arm and it was either radiocephalic (RC, n=54) or ulnarcephalic (UC, n=1). Anastomoses were side to end (S-E; 20 BC, 10 BB, 22 RC, and 1 UC), end to end (E-E; 28 RC), or side to side (S-S; 8 BC and 4 RC). Clinical data collected during the study documented that early AVF failure (defined as inability to use AVF for HD, or failure within 3 months of initial use) occurred in 20 out of 93 patients (21%). As shown in Figure 1, overall VA patency at 1 year averaged 66% (95% confidence interval (CI)=57–77%) and 56% (95% CI=44–70%) 2 years after surgery. Late VA failure was mainly due to stenosis occurring in the venous limb. VA patency was higher in male individuals (71% at 1 year and 60% at 2 years) than in female individuals (55% at 1 year and 49% at 2 years). Only two patients (both with BC S-E anastomosis) developed steal syndrome, probably because of high BFV in the draining AVF vein. Both patients developed concomitant high-output cardiac failure and underwent a successful banding procedure. Three additional patients with BC AVF developed high BFV in the draining vein, at a level that threatened the cardiac function. One of them was subjected to successful vein banding procedure, another developed vein stenosis 15 months after AVF surgery, which required reoperation, and the third patient died 7 months after AVF surgery probably because of heart failure. Other three patients died during follow-up owing to sudden death, liver complications, or intestinal ischemia. Data from 63 patients of the ARCH clinical study with newly created AVF, from patients at 40 days after surgery, were included in the current validation study, and divided into four groups based on AVF configuration:group 1: lower arm RC E-E AVF (n=22); group 2: lower arm RC S-E AVF (n=17); group 3: upper arm BC S-E AVF (n=19); and group 4: upper arm BB S-E AVF (n=5). Out of the 63 patients, 55 were needled and able to perform HD, whereas only 1 patient underwent unsuccessful VA cannulation. In the remaining patients, HD was not started owing to improved renal function. Demographic and clinical parameters, used to generate patient-specific vascular network models, are summarized in Table 1. BFV and the size of major blood vessels in the arm are reported in Table 2 and in Figure 2. BFV measured in brachial artery after VA surgery was lower in distal AVFs as compared with proximal AVFs. Among RC AVFs, E-E anastomosis resulted in higher BFV than S-E (Figure 2a). One interesting finding of our investigation was that brachial artery diameter did not significantly increase after surgery in diabetic patients with upper arm AVF (averaging 4.2±0.8mm and 4.4±0.6mm, preoperatively and 40 days postoperatively, respectively; paired t-test P=0.20; n=9). On the contrary, radial artery did increase after surgery in diabetic patients with distal VA (averaging 2.8±0.6mm and 4.2±1.2mm, preoperatively and 40 days postoperatively, respectively; paired t-test P<0.001; n=11).Table 1Sociodemographic and clinical characteristics of 63 patients with newly created AVF at 40 days after surgery, divided into four groups based on AVF configurationRC E-ERC S-EBC S-EBB S-EN2217195Age (years)55±2164±1566±1569±10Gender (females)3 (14%)4 (24%)8 (42%)2 (40%)AVF arm (right)5 (23%)4 (24%)3 (16%)2 (40%)Height (cm)171±10175±9171±7173±3Weight (kg)74±1483±1273±1477±15Systolic pressure (mmHg)143±20151±38145±26154±14Diastolic pressure (mmHg)aIn case of missing data, the number of available measured data is reported in square brackets.83±1584±1777±12 (17)78±8 (3)Cardiac output (ml/min)aIn case of missing data, the number of available measured data is reported in square brackets.ND5562±1172 [7]4621±1411 [15]6035±2296 [3]Cardiac frequency (b.p.m.)aIn case of missing data, the number of available measured data is reported in square brackets.74±1574±10 [12]68±8 [11]74±5 [3]Hematocrit (%)aIn case of missing data, the number of available measured data is reported in square brackets.33±534±532±6 [18]35±4Protein plasma concentration (g/dl)aIn case of missing data, the number of available measured data is reported in square brackets.6.6±0.7 [20]6.6±0.9 [12]6.0±1.1 [11]5.5±1.0 [3]Hypertension14 (64%)10 (59%)11 (58%)4 (80%)Diabetes5 (23%)6 (35%)6 (32%)3 (60%)Abbreviations: AVF, arteriovenous fistula; BB, brachiobasilic; BC, brachiocephalic; E-E, end to end; ND, not determined; RC, radiocephalic; S-E, side to end.Values are mean±s.d. for continuous variables or frequency (percentage) for gender, AVF arm, hypertension, and diabetes. Hypertension was defined as diastolic pressure+((systolic pressure-diastolic pressure)/3) ≥100mmHg.Individual data were used to generate patient-specific vascular network models for hemodynamic simulations.a In case of missing data, the number of available measured data is reported in square brackets. Open table in a new tab Table 2Measured vessel dimensions and arterial blood flow volumes in 63 patients with newly created AVF, followed up for 40 days after surgery, in comparison with predicted dataPre-op1 Day post-op40 Days post-opMeasuredPredictedMeasuredPredictedRC E-E BA flow (ml/min)65±42448±213341±128611±163584±147 RA flow (ml/min)21±20355±164274±110496±164503±123 Mid RA diameter (mm)2.7±0.64.0±0.83.0±0.64.7±1.04.1±0.4 Mid lower CV diameter (mm)2.6±0.8ND2.2±0.55.1±1.44.8±0.5RC S-E BA flow (ml/min)68±35268±142366±139533±172525±139 RA flow (ml/min)17±10197±164274±114404±150454±126 Mid RA diameter (mm)2.6±0.43.2±0.62.9±0.44.4±1.44.3±0.5 Mid lower CV diameter (mm)3.3±0.9ND2.5±0.65.2±1.84.9±0.8BC S-E BA flow (ml/min)49±20451±170561±216918±332942±321 Mid BA diameter (mm)4.2±0.74.5±0.54.3±0.74.9±0.75.1±0.8 Mid upper CV diameter (mm)3.3±1.1ND2.8±0.76.4±1.45.2±1.5BB S-E BA flow (ml/min)40±24NA810±3381338±5831310±474 Mid BA diameter (mm)3.9±0.84.3±0.73.9±0.74.8±1.05.7±1.2 Mid upper BV diameter (mm)4.6±2.0ND3.3±0.76.0±0.77.0±1.2Abbreviations: AVF, arteriovenous fistula; BA, brachial artery; BB, brachiobasilic; BC, brachiocephalic; BV, basilic vein; CV, cephalic vein; E-E, end to end; ND, not determined; post-op, postoperative; pre-op, preoperative; RA, radial artery; RC, radiocephalic; S-E, side to end.Values are mean±s.d. Vein diameters were computed as weighted average of long and short diameters.Preoperative diameters and blood flow volumes were used as data input for the patient-specific model in hemodynamic simulations. Open table in a new tab Abbreviations: AVF, arteriovenous fistula; BB, brachiobasilic; BC, brachiocephalic; E-E, end to end; ND, not determined; RC, radiocephalic; S-E, side to end. Values are mean±s.d. for continuous variables or frequency (percentage) for gender, AVF arm, hypertension, and diabetes. Hypertension was defined as diastolic pressure+((systolic pressure-diastolic pressure)/3) ≥100mmHg. Individual data were used to generate patient-specific vascular network models for hemodynamic simulations. Abbreviations: AVF, arteriovenous fistula; BA, brachial artery; BB, brachiobasilic; BC, brachiocephalic; BV, basilic vein; CV, cephalic vein; E-E, end to end; ND, not determined; post-op, postoperative; pre-op, preoperative; RA, radial artery; RC, radiocephalic; S-E, side to end. Values are mean±s.d. Vein diameters were computed as weighted average of long and short diameters. Preoperative diameters and blood flow volumes were used as data input for the patient-specific model in hemodynamic simulations. Preoperative vessel dimension and arterial BFV measured during US investigations were used as input parameters to predict postoperative diameters and BFVs at different time points after surgery. Predicted results were compared with measurements obtained by US examination at 1 day and 40 days after surgery. The results are reported in Table 2 and in Figures 3 and 4. In general, BFV prediction was more accurate at 40 days than immediately after surgery, regardless of AVF configuration (see Table 2). A good agreement was observed between measured and predicted brachial artery BFV in individual patients 40 days after surgery, for both proximal and distal AVF and for both AVF configurations, and this for the entire range of brachial artery BFV, from 2l/min. The Bland–Altman plot reported in Figure 5 indicates good agreement between predicted and measured brachial artery BFV in the whole simulation data set independently from brachial artery BFV. Furthermore, regression analysis between predicted and measured values of brachial artery BFV showed a high and statistically significant correlation for each of the four AVF configurations (R2 ranging from 0.77 to 0.96). As shown in Figure 6, a strong correlation was found between measured and predicted results in the whole patient population (R2=0.90, P<0.001). In the whole validation data set, a percent error of predicted versus measured brachial artery BFV of 3±19% (95% CI -2 to 8%) indicates a high precision of the prediction, with a root mean squared error of 19.5% as an index of prediction accuracy. Considering only patients with RC AVF, percent error of predicted versus measured radial artery BFV averaged 11±25% (95% CI -3 to 19%), with a root mean squared error of 25.4%. In addition to the previously described validation data set, we also simulated AVF function in 12 patients with AVF failure or nonfunction at 40 days after surgery. Mean predicted brachial artery BFV averaged 635±298 (range 415–1293) ml/min in patients with early AVF thrombosis (n=8) and 292±77 (range 183–355) ml/min in four patients with AVF nonmaturation. It is interesting to note that in all of these four patients, predicted brachial artery BFV was below the threshold of 400ml/min to perform HD.Figure 4Comparison between measured and predicted brachial artery blood flow volume (BFV) at 40 days after arteriovenous fistula (AVF) surgery in patients with end-stage renal disease and newly created upper arm side to end (S-E) AVF for hemodialysis treatment. Patients were divided into groups based on AVF configuration: brachiocephalic (BC, n=19) and brachiobasilic (BB, n=5). Letters D denote diabetic patients, for whom no adaptation algorithm was applied during simulation. Dashed lines denote thresholds routinely used in clinical practice to assess AVF nonmaturation (400ml/min) and high BFV threatening cardiac function (1500ml/min).View Large Image Figure ViewerDownload (PPT)Figure 5Bland–Altman plot showing agreement between measured and predicted brachial artery blood flow volume (BFV) at 40 days after arteriovenous fistula (AVF) surgery in 63 individual patients. Different symbols denote different AVF configurations (empty circle: radiocephalic (RC) end to end (E-E); full circle: RC side to end (S-E); full triangle: brachiocephalic (BC) S-E; empty triangle: brachiobasilic (BB) S-E).View Large Image Figure ViewerDownload (PPT)Figure 6Correlation between measured and predicted brachial artery blood flow volume at 40 days after AVF surgery in the group of 63 individual patients. (Empty circle: radiocephalic (RC) end to end (E-E); full circle: RC side to end (S-E); full triangle: brachiocephalic (BC) S-E; empty triangle: brachiobasilic (BB) S-E.) Regression line (solid) and 95% confidence intervals (dashed) were estimated using linear regression analysis.View Large Image Figure ViewerDownload (PPT) Despite the established indication to use native AVF for HD patients, the incidence of VA complications, such as nonmaturation and early failure, is high.4.Allon M. Robbin M.L. Increasing arteriovenous fistulas in hemodialysis patients: problems and solutions.Kidney Int. 2002; 62: 1109-1124Abstract Full Text Full Text PDF PubMed Google Scholar,6.Huijbregts H.J. Bots M.L. Wittens C.H. CIMINO study group et al.Hemodialysis arteriovenous fistula patency revisited: results of a prospective, multicenter initiative.Clin J Am Soc Nephrol. 2008; 3: 714-719Crossref PubMed Scopus (280) Google Scholar In the attempt to predict the best location and type of anastomosis, physical examination and US evaluation of brachial and radial artery and venous circulation are currently used.23.Malovrh M. Native arteriovenous fistula: preoperative evaluation.Am J Kidney Dis. 2002; 39: 1218-1225Abstract Full Text Full Text PDF PubMed Scopus (252) Google Scholar However, owing to the complex interplay of several factors, presurgery evaluation cannot reliably support the decision of the surgeon who is predominantly driven by experience and personal skill. The use of computational models to assist the surgeon in selecting the optimal AVF location and configuration could help perform more efficient planning of the AVF surgery. The results of our clinical and numerical investigations provide evidence that the patient-specific hemodynamic computational models that we used16.Huberts W. Bode A.S. Kroon W. et al.A pulse wave propagation model to support decision-making in vascular access planning in the clinic.Med Eng Phys. 2012; 34: 233-248Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar,21.Manini S, Passera K, Huberts W et al. Computational model for simulation of vascular adaptation following vascular access surgery in hemodialysis patients. Comput Meth Biomech Biomed Eng (in press).Google Scholar are accurate (deviation of predicted vs measured brachial artery BFV averaged 3±19%) and provide reliable prediction of the BFV distribution in the arm vasculature during VA maturation, suggesting that this computational approach is of potential use in surgery planning. Predicted results were reasonably accurate for different configurations and locations of the AVF. The predicted values have been obtained, at the patient-specific level, on the basis of demographic data, systemic parameters, clinical condition, and presurgery US measurements of AVF arm blood vessel diameters and BFVs. This data set can easily be provided during the evaluation usually performed in these patients before surgery. This study demonstrates that the problem of predicting VA BFV after maturation on the basis of a small set of preoperative evaluations can be successfully faced by a computer-based modeling approach that takes into account basic hemodynamic and biomechanical phenomena, properly fitted to the individual patient's characteristics. The level of accuracy achieved by our model suggests that other aspects, such as genetic or systemic risk factors, may have a secondary role in predicting postoperative BFV and that inaccuracies due to US measurements or the lack of complete patient-specific information about the entire vascular tree do not significantly affect the accuracy of the prediction. Widely accepted indications suggest that distal VA surgery would result in too low BFV whenever the radial diameter is <2mm. Our present data allowed us to investigate whether preoperative radial artery diameter and 40 days postoperative BFV are actually correlated. As shown in Figure 7, there is a poor correlation between radial artery diameter and VA BFV, estimated by measured BFV in the brachial artery, after VA maturation. These data clearly show that for distal VA, obtained either by E-E or S-E anastomosis, VA BFV cannot be simply predicted by the size of the radial artery. On the contrary, only the consideration of the entire vascular network in the arm allowed obtaining reliable predictions of actual BFV in distal and probably also in proximal VA. We actually identified a different behavior of vessel remodeling in diabetic patients subjected to proximal AVF. In these patients, we could successfully predict BVF after AVF surgery, neglecting completely the shear-induced vascular changes, indicating that medium-sized arteries in these patients (known to be affected by vessel wall calcification) do not remodel, probably
Referência(s)