Article ID: CJ-18-0592
Background: Inflammation and platelet activation have been shown to be involved in acute thromobogenicity following venous occlusive conditions. The aim of this study was to identify the association of baseline platelet count and neutrophil/lymphocyte ratio (NLR) with venous stent failure.
Methods and Results: Patients who underwent technically successful iliocaval venous stent placement with available baseline complete blood count and follow-up stent patency data were selected (n=50). Stent failure was defined as >50% stenosis or occlusion at follow-up angiography, contrast-enhanced CT, MRI or duplex US. Median patient age was 49.5 years (range, 13–76 years), and 62% were female. Median follow-up time was 10.2 months (range, 0.1–76.4 months). Stent failure occurred in 13 patients (26%) after a median of 1.2 months (range, 1 day–76.4 months). On multivariable-adjusted Cox modeling, baseline platelets (HR, 2.28; P=0.004) and WBC count (HR, 2.03; P=0.013) were significantly associated with stent failure on follow-up; neutrophils (HR, 16.10; P=0.050); and NLR (HR, 12.19; P=0.050) had borderline significance. Compared with patients without stent failure, those with early, but not late, stent failure had higher baseline platelets (P=0.031) and neutrophils (P=0.025), and NLR (P=0.026).
Conclusions: Baseline platelet count and NLR are associated with early but not late failure of iliocaval venous stents. This suggests different pathophysiologic mechanisms and a role for both platelet activation and inflammatory mechanisms in early rather than late stent thrombosis. Future research is needed to better explain this novel finding.
Stents are increasingly used to maintain patency in chronic and acute occlusive venous disease, while in-stent re-stenosis and thrombosis remain the main complication limiting the durability of venous stenting.1–3 Although mechanisms of stenosis of arterial stents are extensively studied,4 such mechanisms are poorly understood in veins.5 As such, prophylactic guidelines after venous stenting rely on extrapolated data from arterial interventions and mechanisms implied from venous thromboembolic disease.6,7 Inflammatory mechanisms along with platelet activation have been shown to be involved in venous thrombus propagation.8 Neutrophil/lymphocyte ratio (NLR) and platelet count, as markers of inflammatory activation, have been extensively studied as predictors of poor outcome in cardiovascular diseases and cancer.9–11 The aim of this study was to assess whether pre-procedure platelet count and NLR can predict venous stent thrombosis.
Patients who underwent technically successful iliocaval venous stent placement for venous occlusion between March 2009 and April 2017 with available baseline complete blood count (CBC) and follow-up stent patency data were selected (n=58). Five patients were excluded due to surgical venous intervention prior to percutaneous stenting, stenting of a cryopreserved vein, presence of an indwelling hemodialysis catheter inside the stented vein, or unknown procedural details. Three patients with compressive venous occlusion due to metastatic tumor were excluded from analyses. None of the included subjects had active infection or were taking corticosteroids in the 1 month prior to stent placement. Fifty patients were ultimately evaluated in the study.
Endovascular TechniqueVenous stenting was performed in a dedicated angiography suite under conscious sedation or general anesthesia. Briefly, after establishment of the appropriate percutaneous ultrasound (US)-guided venous access and localization of the stenosis on angiography and/or intravascular US (IVUS), serial balloon angioplasty followed by stent placement covering the entire lesion was performed (Figure 1). In cases of venous occlusion, recanalization was performed typically using hydrophilic guidewires with a straight or angled 4-Fr or 5-Fr catheter. The number and size of the deployed stents were determined by the extent of the lesion, and the anticipated native vessel diameter on venography, respectively. Extensive venous thromboses were treated with catheter-directed thrombolysis (CDT) or rheolytic thrombectomy prior to angioplasty and stent placement. Peri-procedural anticoagulation with unfractionated (40–80 U/kg) or low-molecular-weight heparin (1 mg/kg) immediately before balloon angioplasty and stenting was routinely performed.
Venography in a 41-year-old woman with May–Thurner syndrome showing (A) complete occlusion of the left common iliac vein with cross-pelvic collateralization, (B) recanalized with balloon angioplasty and (C) stent placement with a 16×90-mm Wallstent. (D) Completion venography shows brisk venous flow through the stented segment and decreased filling of collateral veins.
Follow-up visits typically were done at 4–6 weeks and then every 6–12 months with clinical examination and color Doppler US or computed tomography (CT) venography. Symptomatic recurrence or restenosis detected on Doppler US (>50% in-stent restenosis with inflow obstruction) were further evaluated on venography. In-stent or adjacent segment restenoses were treated with balloon angioplasty, thrombolysis, thrombectomy or stenting as appropriate.
Baseline Blood CountCBC including platelet count was done prior to each intervention per routine pre-procedure workup. A differential white blood cell (WBC) count ≤2 weeks prior to the procedure was available in a subpopulation of patients. In cases of extensive lower extremity thrombosis requiring CDT or thrombectomy, the values prior to these procedures were recorded.
Definition of TermsTechnical success was defined as achieving antegrade venous flow with <30% residual stenosis on completion venography after stent placement.12 Stent failure was defined radiographically as >50% stenosis or occlusion at follow-up angiography, contrast-enhanced CT or magnetic resonance imaging or duplex US. Early stent failure was defined as occurring ≤30 days after stent placement. Acute stent recoil was defined as decrease in stent minimum luminal diameter relative to corresponding luminal diameter with balloon inflation.13 Post-stenting antiplatelet therapy was defined as any use of high-dose antiplatelet medication (i.e., 325 mg aspirin daily) or regular use of lower dose (i.e., daily dose of 81 mg aspirin, 75 mg clopidogrel or both) any time after stent placement.
Definition of OutcomesPrimary outcome was defined as stent failure on follow-up. Time to outcome was defined as time to the first imaging with the diagnosis of stent failure, or the time to the most recent follow-up.
Statistical AnalysisData are expressed as median±IQR and n (%) for numerical and categorical variables, respectively. Medians were compared between groups using the Mann-Whitney U-test. Uni- and multivariable Cox proportional hazard models were used to evaluate predictability of peripheral blood count for stent failure on follow-up. To assess independence of the blood cell count and ratios from potential confounding factors in predicting the risk of venous stent failure, the multivariable models were constructed and adjusted for CDT in the 2 weeks prior to stenting, the total length of stented vein, extension of the stent complex to the common femoral vein (CFV), post-intervention antiplatelet medication use and the blood cell count or ratio of interest. Interaction terms between covariates were examined on logistic and linear regression analyses, and the multivariable models were additionally adjusted for the interaction terms between stent length and extension of the stent complex to the CFV, CDT before stent placement, and antiplatelet medication use, and the blood cell count or ratio of interest, as appropriate. Proportional hazard assumption in Cox proportional hazard models was assessed using Schoenfeld’s global test, and variables with significant time interaction in the univariable model were considered time varying14 in all models. Z-standardized values of the numerical variables were applied in the regression models to make the obtained hazard ratios able to be compared with categorical variables. In this method, hazard ratios are calculated per 1-SD change in the numerical values. All analyses were conducted using Stata Intercooled (IC) for Mac version 14.2, and 2-tailed P<0.05 was considered statistically significant.
Table 1 lists the baseline and follow-up subject characteristics. Median age was 49.5 years, and 62% of the subjects were female. Etiology of venous occlusion included acute deep vein thrombosis (DVT; n=9), chronic DVT (n=18) and May-Thurner syndrome (MTS; n=23).15 Twelve patients with MTS presented with acute venous thrombosis in the 2 weeks prior to stent placement. Twenty-two patients (44%) underwent CDT ≤2 weeks after stenting. Median dilated stent diameter was 15 mm. Of the subjects with available balloon inflation and deflation imaging (n=27), stent recoil (≥30% of balloon-inflated diameter) immediately after stent placement was seen in 6 (22%). Varying degrees of residual in-stent stenosis after stent placement was seen on completion venography in 11 subjects (25%), of whom 5 (12%) had >50% residual stenosis. Prior to stent placement, 21 patients (42%) were anticoagulated with warfarin, enoxaparin or a factor Xa inhibitor for ≥2 weeks, and 10 patients (20%) were on antiplatelet therapy with aspirin or clopidogrel. After stent placement, 48 patients (96%) were on anticoagulation, and 30 patients (60%) were on antiplatelet prophylaxis with aspirin, clopidogrel or both. Median length of the stented vein segment was 12±13 cm.
| Characteristics | Data |
|---|---|
| Age (years) | 49.5±26 |
| Sex | |
| Male | 19 (38) |
| Female | 31 (62) |
| Etiology of venous occlusion | |
| MTS | 23 (46) |
| Acute DVT | 9 (18) |
| Chronic DVT | 18 (36) |
| CDT ≤2 weeks prior to stenting | 22 (44) |
| Baseline hypercoagulable state | 5 (10) |
| Total length of stented segment (cm) | 12±13 |
| Dilated stent diameter (mm) | 15±4 |
| Acute stent recoil† | |
| ≥30% | 6 (22) |
| <30% | 21 (78) |
| Stenting extent | |
| Unilateral | 36 (72) |
| Bilateral | 7 (14) |
| IVC only | 7 (14) |
| Extension of stent to IVC | 35 (71) |
| Extension of stent to CFV | 13 (27) |
| Post-stenting residual stenosis‡ | 11 (25) |
| Post-stenting residual in-stent stenosis grade | |
| ≥50% | 5 (11) |
| <50% | 38 (86) |
| Not assessable | 1 (3) |
| Stent inflow patency | |
| ≥50% | 39 (89) |
| <50% | 5 (11) |
| Antiplatelet medication | |
| Before stenting | 10 (20) |
| After stenting | 30 (60) |
| Anticoagulation§ | |
| Before stenting | 21 (42) |
| After stenting | 48 (96) |
| Baseline blood cell count | |
| Platelets (×106/μL) | 213±106 |
| WBC (×106/μL) | 6.9±2.6 |
| Neutrophils (%)¶ | 67.1±10.2 |
| Lymphocytes (%)¶ | 20.5±8 |
| Monocytes (%)¶ | 9.5±3.6 |
| NLR¶ | 3.4±2.3 |
| PLR¶ | 128.8±120.5 |
| Stent malfunction on follow-up | |
| Yes | 13 (26) |
| No | 37 (74) |
| Total follow-up time (months) | 7.8±16.6 |
| Time to dysfunction (months)†† | 1.2±6.6 |
Data given as median±IQR or n (%). †Based on subjects with available balloon inflation and post-deflation imaging (n=27). ‡Based on subjects with assessable post-stenting completion venography (n=44). §Warfarin, enoxaparin or a factor Xa inhibitor. ¶Based on differential cell counts available in a subgroup (n=19). ††Patients with stent malfunction on follow-up (n=13). CDT, catheter-directed thrombolysis; CFV, common femoral vein; DVT, deep vein thrombosis; IVC, inferior vena cava; MTS, May-Thurner syndrome; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio; WBC, white blood cells.
Median pre-procedure platelet and WBC counts were 213±106×106/μL and 6.9±2.6×106/μL, respectively, with a median NLR and platelet/lymphocyte ratio (PLR) of 3.4±2.3 and 128.8±120.5, respectively.
Stent Failure on Follow-upMedian follow-up time was 7.8 months. Stent failure occurred in 13 patients (26%) at a median of 1.2 months (range, 1 day–76 months; 1- and 2-year primary patency, 72.5% and 67.5%, respectively; Figure 2). Early stent failure (≤1 month) occurred in 6 cases, and 7 subjects had late stent failure (>1 month; Supplementary Table 1). Two patients were lost to follow-up after detection of primary stent failure. One 63-year-old woman with MTS was found to have stent re-thrombosis on follow-up CT 20 days after the initial stent placement. The patient died 3 months later for unknown cause with no further venous intervention. Of the remaining subjects, primary stent failure was treated with repeat CDT (n=7), balloon angioplasty (n=10) and re-stenting (n=8). Two subjects had secondary occlusion of the stents 6 and 22 months after the initial stent placement; 7 subjects had secondary stent patency 41 days–52 months after the initial stenting. Secondary status of the stent was not clear in 1 case, although the patient continued to stay symptom free until the last follow-up at 32 months after stenting.
Overall primary and secondary patency after iliocaval venous stenting (n=50).
Table 2 lists the results of uni- and multivariable Cox regression analysis. In the univariable Cox regression model, the total length of stented vein (standardized HR, 1.77; P=0.019), baseline platelet count (standardized HR, 2.23; P=0.001) and WBC count (standardized HR, 1.94, P=0.005), and a high neutrophil differential count (standardized HR, 19.83; P=0.022), were positively associated with an increased risk of stent failure on follow-up. In contrast, post-stenting antiplatelet medication use (HR, 0.18; P=0.011), and higher lymphocyte (standardized HR, 0.10; P=0.028) and monocyte (standardized HR, 0.18; P=0.048) differential counts were associated with a lower risk of stent failure. Baseline WBC count was higher in patients presenting with acute DVT and in those undergoing CDT ≤2 weeks after stenting in patients both with and without subsequent stent failure (Supplementary Tables 2,3; Supplementary Figure). NLR (standardized HR, 9.29; P=0.026) and PLR (standardized HR, 3.56; P=0.037) were significantly associated with increased risk of stent failure in the univariable model.
| Univariate model | Multivariable model | |||
|---|---|---|---|---|
| HR (95% CI) | P-value | HR (95% CI) | P-value | |
| Clinical factors | ||||
| Age (per 17.5-year increase) | 0.81 (0.48–1.39) | 0.450 | ||
| Sex (female vs. male) | 1.35 (0.40–4.49) | 0.627 | ||
| Etiology of venous occlusion (compared with chronic DVT) | 0.85 (0.27–2.68) | 0.778 | ||
| Acute DVT | 1.78 (0.36–8.87) | 0.482 | ||
| MTS with acute presentation | 0.78 (0.13–4.69) | 0.786 | ||
| MTS without acute presentation | 1.94 (0.43–8.73) | 0.385 | ||
| CDT ≤2 weeks before stenting | 0.33 (0.09–1.23) | 0.100 | ||
| Baseline hypercoagulable state | 0.78 (0.10–6.06) | 0.813 | ||
| Procedure-related factors | ||||
| Total length of stented vein (per 9.7-cm increase) | 1.77 (1.10–2.86) | 0.019 | ||
| Dilated stent diameter (per 4.9-mm increase) | 0.66 (0.31–1.39) | 0.271 | ||
| IVC extension | 2.16 (0.47–9.90) | 0.319 | ||
| CFV extension | 2.74 (0.86–8.70) | 0.088 | ||
| Presence of post-stenting in-stent residual stenosis | 0.85 (0.18–4.09) | 0.839 | ||
| Post-stenting in-stent residual stenosis grade (≥50% vs. <50%) | 2.31 (0.48–11.16) | 0.297 | ||
| Stent inflow patency (≥50% vs. <50%) | 0.86 (0.11–6.86) | 0.884 | ||
| Pre-intervention antiplatelet medication | 0.33 (0.04–2.57) | 0.290 | ||
| Post-intervention antiplatelet medication | 0.18 (0.05–0.67) | 0.011 | ||
| Laboratory data | ||||
| Platelets (per 106.8×106/μL increase)† | 2.23 (1.39–3.58) | 0.001 | 2.28 (1.29–4.01) | 0.004 |
| WBC (per 3.6×106/μL increase)† | 1.94 (1.22–3.09) | 0.005 | 2.03 (1.16–3.56) | 0.013 |
| Neutrophils (per 9.2% increase)‡ | 19.83 (1.54–254.91) | 0.022 | 16.10 (1.00–260.16) | 0.050 |
| Lymphocytes (per 8.3% increase)‡ | 0.10 (0.01–0.78) | 0.028 | 0.09 (0.008–1.03) | 0.053 |
| Monocytes (per 3.3% increase)‡ | 0.18 (0.03–0.99) | 0.048 | 0.01 (0.00–3.87) | 0.138 |
| NLR (per 3.7-unit increase)‡ | 9.29 (1.30–66.56) | 0.026 | 12.19 (1.00–147.85) | 0.050 |
| PLR (per 125.1-unit increase)‡ | 3.56 (1.08–11.77) | 0.037 | 4.46 (0.80–24.80) | 0.088 |
Multivariable model is adjusted for use of CDT ≤2 weeks before stenting, total length of stented vein, extension of the stent complex to CFV, post-intervention antiplatelet medication use and the cell count/ratio of interest. Additionally adjusted for the interaction term between stent length and extension of the stent complex to CFV, between CDT before stent placement and any antiplatelet use and between CDT before stent placement and cell count/ratio of interest in the multivariable model. ‡Subpopulation with available differential cell count (n=19; no. events=5). Abbreviations as in Table 1.
In the multivariable model, higher baseline platelet count (standardized HR, 2.28; P=0.004) and WBC count (standardized HR, 2.03; P=0.013) significantly predicted stent failure on follow-up after adjusting for potential confounding factors including CDT ≤2 weeks before stenting, the total length of stented vein, extension of the stent complex to the CFV, and post-intervention antiplatelet medication use. Neutrophil count (standardized HR, 16.10; P=0.050) and NLR (standardized HR, 12.19; P=0.050) had borderline significance, and PLR did not predict venous stent thrombosis (standardized HR, 4.46; P=0.088) after adjusting the models for several potential confounders.
Table 3 lists the blood cell counts according to early (≤1 month), late (>1 month) or any stent failure, and no stent failure on follow-up. Baseline platelet count was significantly higher in early stent failure (262.5±53×106/μL; P=0.031) compared with no stent failure (198±99×106/μL). Patients with late stent failure also had higher platelet count at baseline (255±104×106/μL) compared with those with no stent failure, but the difference did not reach statistical significance (P=0.056). In total, baseline platelet count was significantly higher in subjects with both early and late stent failure (255±62×106/μL; P=0.007) compared with those without stent failure.
| Blood cells | Early failure (≤1 month) (n=6) |
Late failure (>1 month) (n=7) |
All failure (n=13) |
No failure (n=37) |
|---|---|---|---|---|
| Platelets (×106/μL) | 262.5±53 (0.031) | 255±104 (0.056) | 255±62 (0.007) | 198±99 |
| WBC (×106/μL) | 6.6±10.9 (0.806) | 8.5±4.0 (0.205) | 7.1±4.0 (0.304) | 6.9±1.9 |
| Neutrophils (%)† | 86.5±5 (0.025) | 68.7±14.3 (0.441) | 76.3±15.3 (0.061) | 66.3±8.4 |
| Lymphocytes (%)† | 6.5±3 (0.026) | 20.5±6.3 (0.488) | 14.7±12.5 (0.071) | 20.8±6 |
| Monocytes (%)† | 6.5±3 (0.074) | 7.4±4.3 (0.346) | 7.4±1.3 (0.084) | 10±2.4 |
| NLR† | 14.1±7.3 (0.026) | 3.3±2.2 (0.614) | 5.2±7.1 (0.096) | 3.3±1.6 |
| PLR† | 428±422.6 (0.112) | 151.8±94.6 (0.801) | 182.1±64.9 (0.459) | 120.8±120.5 |
Data given as median±IQR (P-value vs. no failure). †Subgroup with available differential cell counts: early failure (n=2), late failure (n=3), no failure (n=14). Abbreviations as in Table 1.
In the subgroup of patients with available baseline differential cell counts, subjects with early stent failure compared with no stent failure had a higher proportion of neutrophils (86.5±5% vs. 66.3±8.4%, P=0.025), a lower proportion of lymphocytes (6.5±3% vs. 20.8±6%, P=0.026), and higher NLR (14.1±7.3 vs. 3.3±1.6, P=0.026). No significant difference between groups was observed for PLR. All patients with stent failure on follow-up had baseline platelet count >190×106/μL. All patients with early stent failure (100%) had NLR >5, compared with only 2/14 (14.3%) of those with no stent failure (Figure 3).
Baseline (prior to stenting) platelet count (n=50), neutrophil/lymphocyte ratio (n=19) and platelet/lymphocyte ratio (n=19) in patients with early (≤1 month), late (>1 month) and no stent failure on follow-up after iliocaval venous stenting. Red lines in the platelet and neutrophil/lymphocyte ratio graphs suggest cut-offs for categorizing outcome. *P<0.05. NS, non-significant.
Although reported failure rates of stents are higher for veins compared with those in arteries, particularly coronary arteries,16–18 to the best of our knowledge no study has evaluated the differences between underlying mechanisms of stent failure in arteries compared with veins. Early stent patency and 1- and 2-year primary patency rates in the present study were similar to those in previous reports.16,17 In the present study, baseline higher platelet and WBC counts, and NLR were associated with higher risks of early stent failure on follow-up independent of other potential confounding factors. The association between NLR and venous stent failure overall had borderline significance (P=0.050) after adjusting for multiple factors, but was significant for early stent failure (P=0.026). Borderline significance seen in the overall stent failure rate may reflect loss of power with multivariable adjustment vs. factors specific for early venous stent failure. The early re-thrombosis rate, and the association with inflammation WBC count noted in this cohort, suggest that there are inflammation factors at least partially responsible for early venous stent failure. Interestingly, a correlation was seen between venous wall enhancement on baseline CT and venous stent failure, further suggesting an association between venous inflammation and underlying factors leading to venous stent thrombosis.19
There is growing evidence that WBC count, differentials and ratios are predictive for a variety of cardiovascular and disease states. Platelets were found to be the primary component of acute arterial stent thrombosis in an experimental porcine study using 111-Indium-labeled platelets.20 In a post-hoc analysis of the ADAPT-DES coronary stent trial, of 8,665 patients undergoing pre-procedure dual antiplatelet therapy with aspirin and clopidogrel, a significantly positive association was observed between platelet count and baseline (pre-treatment) P2Y12 receptor reactivity.21 The authors concluded that higher platelet count and higher platelet reactivity after starting antiplatelet therapy, measured in P2Y12 reactivity units (PRU), have independent and synergistic effects on thrombotic events after coronary stenting with drug-eluting stents. The study did not distinguish between early and late thrombosis, however. In another study of patients undergoing percutaneous coronary intervention (PCI), those with a previous history of stent thrombosis had increased platelet aggregation, a reduced antiplatelet effect of aspirin, and increased levels of immature platelets, a surrogate of increased platelet turnover.22
Platelets have also been shown to play a role in venous thromboembolic disease as well. In an experimental model of DVT using a mouse model, von Bruhl et al showed that a close interaction between platelets and the immune system plays a crucial role in venous thrombosis.8 They described a cross-talk mechanism between platelets, neutrophils and monocytes for initiation and propagation of i.v. thrombosis after reduction of blood flow by approximately 80% in the inferior vena cava. They noted a massive platelet- and endothelium-dependent neutrophil and monocyte accumulation along the venous endothelium prior to the development of DVT.8 According to their study, neutrophils amplify and propagate venous thrombosis by formation of neutrophil extracellular traps and by initiating factor XII-dependent coagulation.8 Notably, post-stenting antiplatelet therapy has additionally been shown to be significantly associated with stent patency after iliocaval venous stenting in a retrospective population study, further suggesting a link between platelet activity and venous stent patency.23
In the present study neutrophil differential count and NLR predicted early but not late venous stent thrombosis. NLR is recognized as a reproducible biomarker of systemic inflammation, and is a well-studied predictor of outcome in cardiovascular disease as well as cancer.9–11,24 In a study comparing patients with previous PCI admitted with ST-elevation myocardial infarction due to stent thrombosis or rupture of atheromatous plaque, higher NLR was associated with higher odds of stent thrombosis and in-hospital mortality.25 Few studies have evaluated the predictive ability of NLR for venous thrombosis, and to the best of our knowledge, no study has assessed the predictive ability of NLR specifically for venous stent thrombosis. In 1 case-control study of venous thromboembolism (VTE), patients with cerebral venous thrombosis (CVT) and healthy controls had an increased risk of provoked CVT with higher PLR. Neither NLR nor PLR, however, were associated with overall increased risk of VTE or CVT.26 In another pilot study of patients undergoing arthroplasty, a significant association was observed between higher NLR and increased risk of in-hospital VTE.27 In a population cohort study, although a single baseline measurement of NLR was not a predictor of long-term risk of VTE, higher baseline NLR was associated with an increased risk of mortality in those who had VTE.28
We observed a baseline platelet count >190×106/μL in all patients (100%) with early or late stent failure, compared with 21/39 (53.8%) without stent failure. Moreover, all patients with early stent failure had NLR >5, compared with 2/39 (5.1%) with no stent failure. No patient with late stent failure reached this NLR level. Data describing baseline platelet count and NLR cut-offs to predict venous stent thrombosis risk are lacking, and future studies with higher sample sizes are needed for validation. A national study of the US population noted a mean NLR of 2.15 with slightly higher values in non-Hispanic white people (2.24), diabetic people (2.34), those with cardiac disease (2.44), obese individuals (2.21) and those who smoked (2.22).29 The cut-off predicting venous stent thrombosis in the present study was higher than these average, suggesting a significant role for inflammation in early venous stent thrombosis. Interestingly, the present results suggest different underlying mechanisms between early (≤1 month) and late (>1 month) venous stent thrombosis. In the coronary arteries, mechanical and procedure-related factors have been proposed to contribute to early stent failure, vs. neointimal hyperplasia and in-stent stenosis mainly contributing to late stent thrombosis.30 The present results emphasize that platelets and the innate inflammatory response also may play critical roles in early thrombosis of venous stents, whereas other factors may underlie late stent thrombosis.
Study LimitationsThis study carries limitations. The relatively small sample size may limit the generalizability of the results. And due to its retrospective nature, the study may be prone to selection bias. Furthermore, the nature of stent failure could not be definitively classified as either thrombosis or re-stenosis on angiography alone in cases of non-occlusion. In addition, platelet aggregation quality, repeat platelet count and NLR after stent placement were not available to evaluate potential fluctuations and their relative predictive ability for early and late stent thrombosis. A prospective study with intravascular assessment using IVUS, optical coherence tomography or biopsy with histology will further clarify the possible mechanisms underlying venous stent thrombosis and re-stenosis.
We have demonstrated the predictive value of baseline platelet and inflammatory WBC differential counts for predicting early venous stent thrombosis independent of potential confounding factors. This finding may be important in developing future practice guidelines and post-procedural patient management. Prospective studies with larger sample sizes are needed to explore and validate this novel finding. Further understanding of the underlying biological mechanisms of venous stent thrombosis will also be crucial for optimizing this procedure.
We appreciate Dr. Keng Wei Liang’s assistance with reviewing venography and US images.
The authors declare no conflicts of interest.
The study was approved by the local Institutional Review Board, and a consent waiver was obtained for the retrospective review of the subjects’ records.
None.
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-18-0592
