Early remodeling of lower extremity vein grafts: Inflammation influences biomechanical adaptation
Article Outline
Background
The remodeling of vein bypass grafts after arterialization is incompletely understood. We have previously shown that significant outward lumen remodeling occurs during the first month of implantation, but the magnitude of this response is highly variable. We sought to examine the hypothesis that systemic inflammation influences this early remodeling response.
Methods
A prospective observational study was done of 75 patients undergoing lower extremity bypass using autogenous vein. Graft remodeling was assessed using a combination of ultrasound imaging and two-dimensional high-resolution magnetic resonance imaging.
Results
The vein graft lumen diameter change from 0 to 1 month (22.7% median increase) was positively correlated with initial shear stress (P = .016), but this shear-dependent response was disrupted in subjects with an elevated baseline high-sensitivity C-reactive protein (hsCRP) level of >5 mg/L. Despite similar vein diameter and shear stress at implantation, grafts in the elevated hsCRP group demonstrated less positive remodeling from 0 to 1 month (13.5% vs 40.9%, P = .0072). By regression analysis, the natural logarithm of hsCRP was inversely correlated with 0- to 1-month lumen diameter change (P = .018). Statin therapy (β = 23.1, P = .037), hsCRP (β = −29.7, P = .006), and initial shear stress (β = .85, P = .003) were independently correlated with early vein graft remodeling. In contrast, wall thickness at 1 month was not different between hsCRP risk groups. Grafts in the high hsCRP group tended to be stiffer at 1 month, as reflected by a higher calculated elastic modulus (E = 50.4 vs 25.1 Mdynes/cm2, P = .07).
Conclusions
Early positive remodeling of vein grafts is a shear-dependent response that is modulated by systemic inflammation. These data suggest that baseline inflammation influences vein graft healing, and therefore, inflammation may be a relevant therapeutic target to improve early vein graft adaptation.
Vein graft caliber is regulated by a complex interplay of hemodynamic stimuli, fluid-phase mediators, and vein wall biology. Final lumen dimension is ultimately a balance between the magnitude and direction of geometric remodeling. Using high-resolution duplex ultrasound (DUS) imaging, we have previously demonstrated that lower extremity vein grafts dilate significantly during the first month of implantation, although considerable heterogeneity exists despite similar anatomic and technical circumstances.1
The factors that regulate the adaptive remodeling of arterialized veins are incompletely understood. After implantation, the thin-walled vein is subjected to an acute increase in wall tension that induces a wall-thickening response.2 Early vein lumen enlargement (positive lumen remodeling) is conjectured to be a flow-induced, endothelium-dependent response to the acute increase in shear stress.1, 2, 3, 4 Many other factors, including pre-existing vein pathology, the extent of operative injury, and the magnitude of the subsequent cellular responses, influence the response to these hemodynamic stimuli and contribute to changes in lumen caliber.
Patients with advanced peripheral arterial disease (PAD) undergoing lower extremity bypass reconstructions have an inflammatory phenotype characterized by elevated levels of markers such as high-sensitivity C-reactive protein (hsCRP).5, 6, 7, 8, 9 We recently observed that among such patients, those with an elevated preoperative hsCRP (≥5 mg/L) level demonstrated an increased incidence of postoperative cardiovascular and vein graft related events.9 The mechanism underlying this observation is unknown, however.
Endothelial dysfunction is an early manifestation of atherosclerosis and linked to adverse clinical events. Numerous investigators have noted an association between CRP, reflecting systemic inflammation, and endothelial dysfunction in the brachial artery.10, 11 In patients undergoing coronary artery bypass grafting, CRP is independently and negatively associated with ex vivo acetylcholine-induced, endothelium-dependent relaxation of preimplantation vein rings.12 However, whether inflammation affects venous endothelial function in vivo and to what extent it may influence vein graft healing and changes in lumen caliber is currently unknown. In the present study, we sought to examine the hypothesis that systemic inflammation influences the early lumen remodeling of lower extremity vein bypass grafts.
Methods
Study design and population
This investigation was designed as a substudy of a larger prospective study examining the relationship between systemic inflammation and clinical outcomes in patients undergoing lower extremity bypass surgery with autogenous vein. Details of the parent study inclusion/exclusion criteria are published elsewhere.1, 9 Briefly, patients were eligible to participate in the study if they were undergoing primary or redo lower extremity arterial reconstruction with autogenous vein for either critical limb ischemia or lifestyle-limiting claudication.
Patients were excluded if a nonautogenous graft was used for any portion of the bypass or if the patient was unlikely to comply with the follow-up protocol. To ensure that the preoperative hsCRP measurement represented the patient's true baseline inflammatory status and was not reflective of an acute process, patients were excluded if there was evidence of local or systemic infection, including cellulitis, osteomyelitis, or deep space infection of the foot, or if they required operative débridement before bypass grafting. Other exclusion criteria included the use of immunosuppressive medication (ie, prednisone, cyclosporine), recent acute illnesses such as myocardial infarction or stroke, systemic illness such as malignancies, or major surgery ≤30 days of the index bypass.
We designed this prospective imaging substudy to characterize the patterns of vein graft remodeling and to examine the association between biomarkers of inflammation and measurements of structural change in the graft. All patients in the parent clinical study who were willing to comply with the imaging protocol were invited to participate in the substudy. The study protocol included DUS graft surveillance at 1, 3, 6, 9, and 12 months after surgery; at each of these time points, additional DUS measurements were obtained, as will be outlined. Patients were also asked to undergo high-resolution magnetic resonance imaging (MRI) of the grafts at 1, 6, and 12 months. Patients were not eligible for the MRI subgroup if they had metal implants or were claustrophobic or unable to lie in the gantry for the required imaging acquisition time.
All participants in this study provided written informed consent, and the protocol was approved by the Institutional Review Board of the Brigham and Women's Hospital (Boston, MA). Subjects were provided financial compensation for the additional time spent to acquire research imaging data.
The current report focuses on an analysis of early (≤1 month) remodeling patterns in this population. The cohort comprised 75 consecutive patients who consented to be in the imaging substudy and had at least one research DUS study in this early timeframe available for analysis. Of these, 62 subjects had DUS performed at the time of surgery, 51 had DUS data at 1 month, and 41 had a study DUS performed at both the time of surgery and at 1 month. Data from patients who had a single DUS at one time point were included for descriptive purposes to represent the population at that particular point in time. Patients who had DUS observations at both the time of surgery and 1 month were used to characterize remodeling changes. Among these 75 patients, 28 also had MRI of the graft at the 1-month time point. Study enrollment began in February 2004 and was closed for this analysis in December 2006.
Blood collection and assays for inflammatory markers
Plasma and blood were collected on the morning of the lower extremity bypass surgery by direct femoral vein puncture after the patient was anesthetized and before any surgical incision. Blood was collected into ethylenediaminetetraacetic acid and citrate Vacutainer (BD Diagnostics, Franklin Lakes, NJ) tubes and immediately iced. Tubes were spun at 3000g for 20 minutes at 4 °C in a refrigerated centrifuge. Baseline plasma samples were stored at −80 °C until analysis. Assays for hsCRP, serum amyloid A (SAA), and fibrinogen were performed in a core laboratory using validated, high-sensitivity assays, as previously described. 9 The upper limit of normal defined for the hsCRP assay in this laboratory is 5 mg/L, a level that we used for dichotomization in some of the end point analyses.
Operating room procedure
The vein graft configuration used for bypass grafting was left to the discretion of the surgeon and depended on the availability of ipsilateral great saphenous vein (GSV). In most patients, veins were tunneled superficially for ease of graft surveillance; however, if a portion of the vein was tunneled in an anatomic position, the index segment was chosen in a superficial portion of the conduit. At the completion of the vascular reconstruction, routine completion DUS imaging was performed to evaluate for flow disturbances or areas of stenosis. All intra-operative images were obtained on an ATL HDI 3000 scanner (Advanced Technology Laboratories, Bothell, Wash) with a 10-MHz transducer.
Index segment
Once the clinical DUS scan was completed, we selected a straight, 5-cm-long, superficial segment of graft to serve as the index segment for serial postoperative measurements of graft diameter. The index segment was selected far enough away from either anastomosis to minimize turbulent blood flow. Accurate spatial registration of the index segment for subsequent examinations was assured by carefully marking the distance between the index segment and the proximal anastomosis as well as another anatomic landmark (eg, scar or tibial tuberosity), and by placing metallic clips on nearby side branches. Schematic maps of the graft configuration and index segment location were made and shared among study personnel performing postoperative DUS scans and MRIs to insure consistent interrogation of the index segment over time.
The index segment lumen dimensions were then obtained using a series of M-mode images with a cross-sectional view of the vein graft, as previously described.1 Five lumen diameter measurements were taken at each of five 1-cm incremental locations along the index segment. These 25 measurements were then used to determine the mean lumen diameter of the index segment for that examination.
To determine vein graft stiffness, a pulse wave velocity (PWV) measurement of the vein graft was obtained, as previously reported.13 First, the length of the graft was measured using an umbilical tape, and this length was used for all subsequent PWV calculations.
Next, waveforms of the graft were obtained 1 cm distal to the proximal anastomosis and 1 cm proximal to the distal anastomosis. A time delay between the internal ultrasound scan electrocardiogram (ECG) trigger and the foot of the Doppler waveform was measured with the waveform and the ECG tracing simultaneously displayed. The PWV was subsequently calculated as the distance (in meters) between the proximal and distal measurement locations divided by the time difference (in seconds) of the QRS-to-onset of flow waveform at the proximal and distal ends of the graft. At least 10 individual time delays were measured at each location and averaged for this calculation.
Finally, a sagittal view of the index segment accompanied by a gated-Doppler waveform was recorded for determinations of volumetric flow and estimations of shear stress, as described below.
Magnetic resonance imaging data acquisition
To characterize wall thickness, the 28 eligible patients signed informed consent for high-resolution MRIs of their bypass grafts postoperatively. The MRIs were performed with a 1.5T GE Excite MR (GE Medical Systems Waukesha, Wis) system equipped with 4-G/cm gradients capable of a slew rate of 15-G/cm/ms. The body coil was used for radiofrequency transmission (maximum B1 250 mG), and a 5-inch circular coil was used for radiofrequency reception. An ECG-gated two-dimensional fast spin echo sequence with a double inversion recovery black blood module was used for T1-weighted (T1WI) and T2-weighted imaging (T2WI). For T2WI, a frequency-selective fat suppression pulse was used.
A 256 × 256 matrix with 10-cm field of view and four signal averages per phase encode with the “no phase wrap option” were used for both contrast weightings; 2.2-min acquisition/slice, depending on heart rate and with a 1 (about 1 second) and 2 (about 2 seconds) R-R TR period for T1- and T2WI, respectively. Effective echo times for T1- and T2WI are 14 and 58 ms, respectively. Echo train length was 8 for T1WI and 16 for T2WI, with 32-KHz and 16-KHz bandwidth, respectively. In the typical protocol, five 4-mm-thick slices orthogonal to the graft, spaced 10 mm apart, were acquired sequentially over 10 to 11 minutes per contrast weighting.
Postacquisition measurements were performed by direct planimetry using dedicated three-dimensional Vital Images 3.9 postprocessing software (Vital Images, Minnetonka, Minn).
Follow-up imaging protocol
Postoperative examinations were performed in the Brigham and Women's Hospital Vascular Laboratory. Subjects were examined at rest and in a supine position to allow normalization of heart rate and blood pressure. Clinical noninvasive graft evaluation includes measurement of ankle-brachial indices and DUS scans with peak systolic velocity maps. Once the surveillance protocol was completed, the aforementioned index segment was identified using the referenced external landmarks or clips placed at the time of surgery. M-mode diameter measurements, PWV, and time averaged flow were obtained and calculated as for the intraoperative imaging study. If an index segment developed a lesion requiring reintervention or the graft developed an occlusion, the patient would be removed from the imaging substudy. No graft occlusion developed in these patients.
Calculations
Calculations for wall stiffness were based on modifications of the Moens-Koerteweg formula (Equation 1),14 relating the elastic modulus of the vein graft to overall stiffness,
(1)
(2)
(3)Volumetric flow, Q, was calculated from commercially available software, (Brachial Analyzer, Medical Imaging Applications, Iowa City, Iowa), by integrating the area under the velocity spectral waveform and dividing by the time required to arrive at a time-averaged velocity. Mean flow was calculated by multiplying the time-averaged velocity by the mean area of the lumen (as obtained from the index segment M-mode measurements).
Mean shear stress was calculated according to the Hagen-Poiseuille formula τw = 4μQ/πri3, where τw is shear stress in dynes/cm2 and mean volumetric flow is Q. The viscosity of blood, μ, is assumed to be 0.035 poise. The lumen radius, ri, is in cm.
Statistical methods
All values are represented as either mean ± standard error of the mean or median and interquartile range (IQR), depending on their distribution. Comparisons of measures between individual time points between groups were made with the Student t test or Kruskal-Wallis nonparametric one-way analysis of variance.
Relationships between remodeling metrics and inflammatory marker levels were evaluated either by dichotomizing the population according to the preoperative plasma hsCRP levels9 or as a continuous variable with a natural logarithmic transformation. Log transformation was necessary to normalize its distribution. The Spearman rank correlation was used to compare hsCRP levels and relevant patient demographics.
Stepwise multivariable linear regression analysis15 was conducted to determine the independent contribution of variables within traditional cardiovascular risk factors, inflammation, and hemodynamic categories to percentage of lumen remodeling from 0 to 1 month. The covariates with P < 0.2 included in the final model were age, diabetes, statin use, initial shear stress, CRP risk group (elevated hsCRP >5 mg/L or ≤5 mg/L), or the natural logarithm of hsCRP (ln-hsCRP). C-reactive protein was entered into the model either as a categoric (high or low) or continuous variable, and both β-coefficients are presented.16 Colinearity between explanatory variables was assessed by calculating variance inflation factors. The β-coefficient, confidence intervals, and P values are presented. Values of P < .05 were considered significant. Statistical analyses were performed on Intercooled Stata 7.0, (Stata Corp, College Station, Tex).The authors had full access to and take full responsibility for the integrity of the data.
Results
Patient demographics and risk factors
The mean age for the study cohort was 68.8 ± 11.1 years. Of the 75 subjects, 42 (56.0%) were men, 58 (79.5%) were white, 38 (51.4%) had diabetes mellitus, 44 (59.5%) had known history of coronary artery disease, 5 (6.7%) had end stage renal disease, and 58 (77.3%) were taking a 3-hydroxy-3 methyl-glutaryl-coenzyme A (HMG-CoA) reductase inhibitor (statin).
Surgical procedures
The indication for the revascularization was critical limb ischemia (CLI), defined as rest pain or tissue loss, in 44 subjects (58.7%), claudication in 26 (36.0%), and repair of popliteal aneurysm in five (6.7%). Autogenous conduits included single-segment GSV (reversed or nonreversed) in 57 cases (76.0%), spliced great saphenous or small saphenous vein in 6 (8.0%), single-segment arm vein in 4 (5.3%), and spliced arm vein in 8 (10.7%). Inflow sites included the common femoral artery in 50 patients (66.7%), superficial femoral artery in 17 (22.7%), or popliteal artery in eight (10.7%). Distal anastomoses were performed to the popliteal artery in 39 patients (52.0%), to the tibial artery in 24 (32.0%), and to a pedal artery 12 (16.0%).
Inflammation and remodeling
The median hsCRP value for the entire cohort was 3.25 mg/L (IQR, 1.38-14.20). The median concentration was 0.94 mg/dL (IQR, 0.56-2.85) for SAA and 472.75 (IQR, 404.8-610.2) for fibrinogen. High-sensitivity CRP correlated with the presence of diabetes (R = 0.29, P = .0007), chronic kidney disease classification (R = 0.34, P = .0001), and critical limb ischemia (R = 0.35, P < .00001). In contrast with hsCRP, neither fibrinogen nor SAA were predictive of any imaging endpoints.
Vein graft morphometrics are summarized in Table I. For the entire cohort, the mean lumen diameter of the index segment was 3.77 ± 0.14 mm at implantation, and this increased to 4.67 ± 0.16 mm during the first month after surgery (P = .0001). The mean index segment lumen dilation (relative change) from the time of surgery to 1 month was 30.6% ± 32.3% (median, 22.7%; IQR, 7.1%-48.7%). No difference was found in the starting lumen diameters between patients in the low vs the elevated (>5 mg/L) hsCRP groups (3.83 ± 1.43 mm vs 3.71 ± .82 mm). During the first postoperative month, however, grafts in the low hsCRP group demonstrated greater median lumen enlargement, at 37.2% (40.9% ± 34.4%) vs 9.49% (13.5% ± 21.0%; P = .0072; Fig 1, A). Analyzing baseline hsCRP as a continuous risk variable, we observed a strong negative relationship between ln-hsCRP values and the percentage change in the index segment lumen diameter during this period (R = 0.40, P = .010, Fig 1, B).
Table I. Morphometrics of lower extremity vein grafts
| Vein graft metrics | High-sensitive CRP | P | |
|---|---|---|---|
| ≤5 mg/mL | Elevated | ||
| High-resolution ultrasound imaging, (N) | 24 | 16 | |
 Lumen diameter (surgery), mm ± SEM | 3.83±1.43 | 3.71±.82 | .696 |
| Lumen change (0-1 mon), % ± SEM | 40.88±7.0227 | 13.54±5.25 | .0072 |
| Shear stress (surgery), dynes/cm2 ± SEM | .63±3.56 | 24.93±3.00 | .566 |
| 2D magnetic resonance imaging, (N) | 11 | 17 | |
| Wall thickness (1 mo.), mm ± SEM | .75±.05 | .72±.04 | .685 |
| Elastic modulus (1 mo.), Mdynes/cm2 ± SEM | 25.05±6.51 | 50.44±12.22 | .069 |
Fig 1.
A, Mean index segment lumen percentage change dichotomizing the population by high-sensitivity C-reactive protein (hsCRP) levels of ≤5 mg/L vs elevated levels of >5 mg/L. Less outward lumen remodeling was demonstrated in those subjects with elevated hsCRP levels (P = .0072). Data are presented with the standard error of the mean. B, Linear regression of natural-logarithmic transformed hsCRP levels (ln-hsCRP) and the percentage lumen change from the time of surgery to 1 month (R = 0.371, P = .018).
The mean initial shear stress for the entire cohort was 26.7 ± 1.9 dynes/cm2 and decreased to 20.5 ± 2.3 dynes/cm2 during the first month of implantation (P = .05). No significant differences were found in the initial shear stress between grafts in the elevated hsCRP population vs those whose hsCRP was ≤5 mg/L. During the first month of implantation, initial shear stress positively correlated with lumen dilation (R = .36, P = .025, Fig 2, C). The association between initial shear stress and lumen dilation was relatively robust in subjects whose hsCRP level was ≤5 mg/L (R = .50, P = .015, Fig 2, B), but the association was not seen in patients with elevated hsCRP (R = .03, P = .926; Fig 2, C).
Fig 2.
A, Linear regression of the entire population comparing initial shear stress and mean index segment lumen dilation from 0 to 1 month (R = 0.387, P = .016). B, Subset of patients with high sensitivity C-reactive protein (hsCRP) ≤5 mg/L (R = 0.5, P = .014) and (C) subset of patients with elevated hsCRP >5 mg/L demonstrating loss of correlation between initial shear and lumen dilation (R = 0.103, P = .725).
A multivariable analysis was undertaken to assess variables contributing to early (≤1 month) vein graft remodeling (Table II). Significant variables included patient age (P = .016), hsCRP risk group (P = .006), initial shear stress (P = .003), and current statin therapy (P = .037). These four variables accounted for approximately 45% of the variability in lumen dilation between surgery and 1 month. Of importance, conduit type (ie, type or orientation of the vein), diabetes, or sex did not significantly add to the model. When modeling for the final 1-month lumen diameter (Table II), initial diameter, CRP risk group, and statin use were independently predictive.
Table II. Multivariable regression of variables influencing lumen diameter changes in vein grafts, 0 to 1 month
| Variable | β (95% CI) | P |
|---|---|---|
| Elevated CRP | −29.7(−50.5to−8.9) | .006 |
| Ln-CRPa | −10.8(−18.2to−3.6) | .002 |
| Statin | 23.1(1.5to44.2) | .037 |
| Initial shear | 0.84(0.32to1.4) | .003 |
| Age | 1.17(0.23to2.1) | .016 |
aModel created with either CRP as a dichotomized or natural log (ln) transformation variable. |
In a subset of 28 patients who agreed to participate in the MRI examinations, the vein graft wall thickness at 1 month after implantation was assessed by T2WI (Fig 3). No difference was found in either total wall thickness or total wall area between high vs low hsCRP risk groups (Table I). There was, however, a trend toward increased modulus of elasticity at 1 month in the high hsCRP group (50.44 ± 12.22 vs 35.05 ± 6.51 Mdynes/cm2, P =.069), suggestive of greater conduit stiffness.
Fig 3.
A, This image is a representative example of a high-resolution two-dimensional magnetic resonance T2-weighted image using black blood module of a human saphenous vein graft. B, The area inscribed by the two lines represents the T2-weighted wall area. The lumen area is 0.327 cm2, and the total vessel area is 0.566 cm2.
Discussion
This study demonstrates that in human lower extremity vein grafts, rapid lumen remodeling, which occurs within the first month after surgical implantation, is significantly modulated by systemic inflammation. Vein grafts in patients with baseline hsCRP values <5 mg/L had significantly more early lumen dilation than those within the elevated hsCRP group, and there was a negative linear association between ln-hsCRP and early lumen remodeling. Furthermore, within the elevated hsCRP group, the positive correlation between shear stress and outward lumen remodeling was not seen, suggesting a level of dysfunction in either shear stress-signal transduction or vessel response. Moreover, in the multivariable regression model, baseline hsCRP was independently and negatively associated with early diameter change. Notably, the significance of the hsCRP variable was not attenuated with the inclusion of traditional cardiovascular risk factors into the model. Conversely, use of statins, a class of agents known to reduce inflammation and improve endothelial function, was positively correlated with early lumen dilation.
Although preliminary, these results begin to integrate data from the domains of biomechanical, biochemical, and traditional cardiovascular risk factors to predict early changes in the vein graft lumen. A regression equation using four explanatory variables such as, lumen change = −10.9 (ln-hsCRP) + 23.1 (presence of statin) + 0.85 (initial shear stress) + 1.17 (age) − 72.5, accounts for 45% of the variance seen in early lumen caliber change, which is a substantial improvement over any one factor alone. To our knowledge, this is the first report of an in vivo link between inflammation and early lumen remodeling in peripheral bypass grafts. However, of the three markers of inflammation that were measured—CRP, SAA, and fibrinogen—only CRP had a negative association with vein graft lumen changes.
Inflammation and endothelial function
Patients with PAD have a pro-inflammatory phenotype that manifests as higher circulating levels of hsCRP compared with matched controls without PAD.17 This is confirmed in the present report, where the median hsCRP level was found to be 3.25 mg/L, which places this cohort in the highest risk category as defined for the Centers for Disease Control and Prevention.9, 18 Inflammation quantified by hsCRP has been shown to be associated in vivo with impaired brachial artery flow-mediated, endothelium-dependent, vasodilation in patients with PAD7 and blunted coronary endothelial vasoreactivity in patients with coronary artery disease.11 These observations suggest that inflammation impairs arterial endothelial function. In patients undergoing coronary artery bypass graft surgery, circulating plasma CRP concentrations result in venous endothelial dysfunction and impaired acetylcholine-induced, endothelium-dependent relaxation of preimplantation vein graft rings ex vivo (r −0.30, P = .02).12 Further investigations are needed to characterize endothelial function in vein bypass grafts and to determine its relationship to the inflammatory response, structural remodeling, and clinical outcomes.
In vivo measurements of remodeling
Despite the clinical importance of vein graft disease, relatively little attention has been given to in vivo measurements of human vein graft remodeling. Most reports to date have involved intravascular ultrasound examinations of aortocoronary bypass grafts, which have the advantage of adequate echo-differentiation between the vein graft adventitia and surrounding pericardial tissue so that wall thickness measurements may be obtained.19, 20, 21, 22 Collectively, these reports demonstrate that the mechanisms of saphenous vein graft narrowing in the coronary circulation are often a combination of negative remodeling and intimal hyperplasia. Although somewhat conflicting, these studies also demonstrate that saphenous vein grafts are capable of dilating to some degree in the face of an encroaching intimal hyperplastic lesion or atherosclerotic plaque. Consistent with these findings, Lau et al,23 using a noncontrast, ECG-gated, cardiac computed tomography scanner, demonstrated a decrease in total vessel diameter of >5% (defined as negative remodeling) in 62% of aortocoronary vein grafts. Notably, none of these studies used serial imaging beginning at the time of implantation. More effort is clearly needed to accurately document remodeling characteristics of both coronary and lower extremity vein grafts and to determine the extent maladaptive remodeling contributes to vein graft failure.23
Even fewer studies have focused on the structural changes of peripheral vein grafts. Fillinger et al4 demonstrated a positive correlation between shear stress and the percentage change in lumen caliber during the first year of implantation. Leotta et al24 observed negative lumen remodeling by three-dimensional lumen reconstructions of lower extremity saphenous vein graft revisions (patch angioplasties). There was an average loss of lumen cross-sectional area by 31% during a 35-week period. However, the average postoperative lumen diameter after patching was 7.5 mm, which was considerably larger than the unrevised vein grafts presented in our current report.24 Neither study was able to measure the vein graft wall, illustrating the technical difficulty in resolving the structure.
Limitations
Our characterization of vein graft remodeling currently focuses on the combined use of advanced noninvasive imaging tools, in particular the optimization of high-resolution MRI, to attain an accurate assessment of wall thickness, which has eluded us with conventional DUS technology. This will allow temporal assessment of wall structural and total vessel area rather than changes in lumen diameter. Limitations related to assumptions of the Moens-Korteweg equation, and therefore accuracy of the elastic modulus and stiffness calculations, have been detailed elsewhere.1, 13 Briefly, the Moens-Korteweg equation assumes a thin-walled, nontapered tube containing an ideal noncompressible liquid. The PWV is an integrated value over the entire vein graft and fails to account for heterogeneity in viscoelastic properties within the vein graft that may predispose it to local failure at susceptible sites. Despite these limitations, PWV is used extensively in the literature and is one of the most accurate noninvasive methodologies to assess stiffness.
Although this study was not sufficiently powered to detect differences in patency between vein grafts that did and did not undergo favorable lumen remodeling, it suggests that inflammation has a detrimental effect on the vein's ability to function as an arterial substitute, which may ultimately impact patency. Finally, limited sample size and missing observations might have introduced bias and confounded our results.
Postoperative wound complications accounted for most of the missing DUS data. Other reasons included one patient who died, two patients experienced early graft occlusion, and the remainder had scans that were uninterpretable, returned outside of their 1-month study window (2 weeks), or the study staff was unavailable. Of the many methods to handle missing values, including last observation carried forward, imputation (mean or regression), or list-wise omission,25 we favored the latter because no discernible differences existed in terms of cardiovascular risk factors and levels of inflammation between those patients with missing values and those with all observations present.
Conclusions
This study confirms that lower extremity vein grafts undergo significant lumen remodeling within the first month of implantation and that shear stress is strongly correlated with early changes in lumen caliber. We now report that baseline systemic inflammation, as reflected by hsCRP, appears to be an important modifier of this early hemodynamic response. This finding may have important clinical and therapeutic implications. Further effort is needed to characterize the temporal and spatial remodeling of the lumen and wall of peripheral vein grafts and to determine early remodeling signatures of healthy versus diseased grafts.
Author contributions
References
- Early biomechanical changes in lower extremity vein grafts–distinct temporal phases of remodeling and wall stiffness. J Vasc Surg. 2006;44:740–746
- . Stranger in a strange land: the pathogenesis of saphenous vein graft stenosis with emphasis on structural and functional differences between veins and arteries. Prog Cardiovasc Dis. 1991;34:45–68
- . Pathophysiology of vein graft failure: a review. Eur J Vasc Endovasc Surg. 1995;9:7–18
- . Vein adaptation to the hemodynamic environment of infrainguinal grafts. J Vasc Surg. 1994;19:970–978discussion 978-9
- . C-reactive protein and alpha 1-acid glycoprotein in monitoring of patients with acute arterial occlusion. Eur J Vasc Surg. 1991;5:641–645
- . Inflammatory mediators are associated with 1-year mortality in critical limb ischemia. J Vasc Surg. 2005;42:75–80
- Endothelial dysfunction in peripheral arterial disease is related to increase in plasma markers of inflammation and severity of peripheral circulatory impairment but not to classic risk factors and atherosclerotic burden. J Vasc Surg. 2003;38:374–379
- Increased preoperative c-reactive protein level as a prognostic factor for postoperative amputation after femoropopliteal bypass surgery for CLI. Ann Chir Gynaecol. 2001;90:19–22
- Elevated C-reactive protein levels are associated with postoperative events in patients undergoing lower extremity vein bypass surgery. J Vasc Surg. 2007;45:2–9discussion 9
- Effects of CRP infusion on endothelial function and coagulation in normocholesterolemic and hypercholesterolemic subjects. J Lipid Res. 2007;48:952–960
- . Elevated C-reactive protein levels and impaired endothelial vasoreactivity in patients with coronary artery disease. Circulation. 2000;102:1000–1006
- The association between saphenous vein endothelial function, systemic inflammation, and statin therapy in patients undergoing coronary artery bypass surgery. J Thorac Cardiovasc Surg. 2007;134:335–341
- Early adaptation of human lower extremity vein grafts: wall stiffness changes accompany geometric remodeling. J Vasc Surg. 2004;39:547–555
- . Properties of the arterial wall: theory. McDonald's blood flow in arteries. In: Fifth ed.. New York, NY: Hodder Arnold; 2005;p. 49–65
- . A handbook of statistical analysis using Stata. 3rd ed.. Boca Raton, FL: Chapman & Hall/CRC; 2004;
- . Using multivariable statistics. 2 ed.. Cambridge, MA: Harper & Row; 1989;
- . Comparison of usefulness of inflammatory markers in patients with versus without peripheral arterial disease in predicting adverse cardiovascular outcomes (myocardial infarction, stroke, and death). Am J Cardiol. 2005;96:1374–1378
- Markers of inflammation and cardiovascular disease: application to clinical and public health practice: A statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation. 2003;107:499–511
- Mechanisms of lumen narrowing of saphenous vein bypass grafts 12 months after implantation: an intravascular ultrasound study. Am Heart J. 2006;151:726–729
- Absence of focal compensatory enlargement or constriction in diseased human coronary saphenous vein bypass grafts (An intravascular ultrasound study). Circulation. 1996;93:683–690
- Intravascular ultrasound assessment of the presence of vascular remodeling in diseased human saphenous vein bypass grafts. Am J Cardiol. 1999;84:992–998
- Clinical, angiographic, and intravascular ultrasound characteristics of early saphenous vein graft failure. J Am Coll Cardiol. 2004;44:53–56
- Lumen loss in the first year in saphenous vein grafts is predominantly a result of negative remodeling of the whole vessel rather than a result of changes in wall thickness. Circulation. 2006;114:I435–I440
- . Remodeling in peripheral vein graft revisions: serial study with three-dimensional ultrasound imaging. J Vasc Surg. 2003;37:798–807
- . Statistical analysis with missing data. 2nd ed.. New York, NY: John Wiley; 2002;
Competition of interest: none.
Supported by funding from the National Heart, Lung, and Blood Institute (HL75771) to Drs Conte, Owens, and Rybicki, the Clinical Investigator Training Program, Harvard-MIT Division of Health Sciences and Technology to Dr Owens; NIBIB (K23-882) and the Whitaker Foundation to Dr Rybicki.
PII: S0741-5214(08)00021-9
doi:10.1016/j.jvs.2008.01.009
© 2008 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
