2018 Volume 82 Issue 7 Pages 1908-1916
Background: Recent randomized trials have shown the treatment benefits of use of a drug-coated balloon (DCB) over conventional percutaneous transluminal angioplasty (PTA) in patients with femoropopliteal disease. However, the effectiveness and safety of DCB for dialysis patients remain unclear.
Methods and Results: Consecutive dialysis patients, who underwent PTA or DCB for femoropopliteal disease, were assessed retrospectively via 2:1 propensity score matching. Effectiveness and safety endpoints, including binary restenosis, clinically driven target lesion revascularization (CD-TLR), amputations, major adverse cardiac events (MACE), and deaths, were compared between groups. A total of 278 dialysis patients with 339 limbs were eligible for matching: 84 limbs from 77 patients treated with PTA and 46 limbs from 37 patients treated with DCB were compared after matching. Baseline patient and lesion characteristics were not different between groups. Patients treated with DCB had significantly higher rates of freedom from binary restenosis (52.4% vs. 18.6%, P<0.001) and CD-TLR (56.4% vs. 25.9%, P=0.001) at 2 years compared with patients treated with PTA. Both groups had similar outcomes for amputation, MACE, and death. Cox proportional analysis showed that treatment with DCB was independently associated with a reduction of binary restenosis (hazard ratio [HR] 0.368, P=0.001) and CD-TLR (HR 0.390, P=0.004).
Conclusions: This study suggested superior 2-year outcomes using DCB compared with PTA and similar safety profiles in dialysis patients with femoropopliteal disease.
Atherosclerotic lower extremity artery disease (LEAD) is a common manifestation in patients with end-stage renal disease undergoing dialysis and is associated with significant morbidity and mortality.1–3 As compared with the general population, dialysis patients have an incredibly high risk of non-traumatic lower limb amputation from all causes,4 and the survival rates for those patients are even more miserable after lower limb amputation.4,5 Although surgical revascularization is a reasonable treatment for limb salvage in dialysis patients with symptomatic LEAD, especially for critical limb ischemia (CLI), the bypass surgery is technically challenging because of the presence of heavily calcified vessels.1 In addition, dialysis patients have a significantly higher rate of postoperative complications, and lower graft patency, limb salvage and patient survival rates as compared with patients without dialysis.6,7 These observations have led to different recommendations for surgical revascularization in dialysis patients with symptomatic LEAD.8
Remarkable technical advances, together with patient preferences, have changed the revascularization strategies for LEAD from bypass surgery to less invasive endovascular therapy (EVT) in the past decade.9,10 Although the feasibility and effectiveness of EVT in dialysis patients with LEAD has been demonstrated in small series,11–13 the primary patency, limb salvage and survival rates remain poor in dialysis as compared with non-dialysis patients, especially for patients with Trans-Atlantic Inter-Society Consensus (TASC II) C and D lesions.12,13 Proof-of-concept evidence has demonstrated the usefulness of different drug-coated balloon (DCB) technologies in reducing both restenoses and the need for reintervention in comparison with conventional percutaneous transluminal angioplasty (PTA).14–17 However, most of the trials exclude dialysis patients and information regarding the effectiveness of DCB in this group is scarce. We aimed to compare the efficacy and clinical outcomes of DCB vs. PTA with provisional stent implantation in dialysis patients with femoropopliteal disease. Propensity score matching was used to balance the observed covariates between subjects from the treatment groups in order to mimic a randomized study.
The subjects for this study were derived from the Tzuchi Registry of ENDovascular Intervention for Peripheral Artery Disease (TRENDPAD), an ongoing, prospective, physician-initiated, single-center observational registry of patients who underwent EVT for lower limb ischemia, which was started in July 2005. This database was interrogated to identify dialysis patients (either receiving hemodialysis or peritoneal dialysis) with symptomatic femoropopliteal disease treated between January 2009 and June 2016. Patients were categorized according to treatment with a conventional balloon or DCB angioplasty followed by provisional nitinol stent implantation. The local ethics committee and institutional review board approved this study (IRB 06-X17-067), and all study procedures were based on good clinical practices and the applicable laws of various governing bodies. Informed patient consent was waived because of the retrospective study in design.
Study PopulationsDialysis patients presenting with severe claudication or CLI (Rutherford-Becker classification 3–6) were screened. Patients who had de novo, restenosis or in-stent restenotic lesions of the superficial femoral and popliteal arteries were eligible for enrollment. Concomitant interventions for the iliac artery or below-the-knee (BTK) arteries were allowed. Patients with a life-threatening infection, follow-up duration <3 months in surviving patients, or refusal to participate were not eligible. Other exclusion criteria included failure to cross the target lesion with a guidewire, prior use of a drug-eluting stent, an aneurysm in the target vessel, and acute or subacute thrombus in the target limb. Propensity score matching was applied to minimize bias from confounding factors (see Statistical Analysis). The study flowchart is depicted in Figure 1.
Flowchart of study participants. DCB, drug-coated balloon; PTA, percutaneous transluminal angioplasty.
The EVT strategy was left to the discretion of the treating physician. All patients received 100 mg aspirin and 300 mg clopidogrel before EVT. Unfractionated heparin (5,000–10,000 units) was administered during the procedure to maintain an activated coagulation time around 250 s. In the PTA group, dilatation of the target lesion was performed using an uncoated balloon with a vessel/balloon ratio 1:1 after successful crossing of the guidewire. In the DCB group, predilatation with an undersized, shorter, uncoated balloon (0.5–1.0 mm smaller than the reference vessel diameter (RVD)) was performed after crossing the lesion with a guidewire, followed by insertion of a DCB (42 limbs treated with IN.PACT Admiral, Medtronic Ireland, Galway, Ireland; 3 with Freeway, Eurocor GmbH, Bonn, Germany; 1 with Lutonix, Bard, Wexford, Ireland) of appropriate size and length (balloon/vessel diameter ratio 1:1). In cases of lesion length >15 cm, ≥2 DCBs were used with a minimum 5-mm balloon overlap at the edges. The inflation time of the DCBs was 180 s using nominal pressure to allow full drug elution. The details of the interventional procedure and quantitative vascular angiography have been described previously.18,19 For both groups, bailout bare-metal nitinol stents were implanted in cases of suboptimal angiographic result or flow-limiting dissection, determined by residual diameter stenosis (DS) >50%, and translesional pressure gradient ≥10 mmHg.
Angiographic EvaluationAngiograms were acquired in at least 2 orthogonal views at baseline and after the intervention. A radiopaque ruler was used for calibration of the angiographic measurements, including the length and minimum luminal diameter (MLD) of the target lesion and the mean proximal and distal RVD. The percent DS was calculated as [(1−MLD/RVD)×100] at baseline and after the intervention. In addition, the distal runoff vessels were inspected on the completion angiogram for evidence of distal embolization.18
Following EVT, aspirin was continued indefinitely in all patients. Clopidogrel was used for a minimum duration of 3 months in patients receiving nitinol stent implantation or DCB treatment. In patients who could not tolerate the side effects of aspirin treatment, clopidogrel was used indefinitely.
Follow-up ProtocolAfter discharge, all patients were followed at an outpatient clinic. Patients with tissue loss underwent wound debridement, free flap transplant, and hyperbaric oxygen therapy by a plastic surgeon or orthopedist until their wounds healed. Clinical and duplex ultrasound examinations were conducted at 1 week, 1 month, and 3 months after the index procedure, and every 3 months thereafter. Repeat interventions were performed if recurrent symptoms, significant vessel stenosis (≥70%) with dampened Doppler waveform patterns by duplex ultrasound, and an ankle-brachial index (ABI) decrease ≥0.15 were detected. Major events (death, limb amputation, or loss of primary patency) were documented at the time of hospital discharge or during the 3-month follow-up visits. In the event that an office follow-up was not feasible, data were derived from telephone interviews, medical records, local electronic medical databases, and referring physicians.
Definitions and EndpointsThe primary efficacy endpoint was freedom from binary restenosis, defined as diameter restenosis >50% by angiography or peak systolic velocity ratio ≥2.4 by duplex ultrasound. CD-TLR was defined as any reintervention performed within 5 mm of the target lesion after documentation of recurrent clinical symptoms following the index procedure.
Safety endpoints were major amputation (defined as above-ankle amputation) and major adverse cardiac events (MACE: composite of myocardial infarction, stroke, any cause of death or major amputation of the target limb) during the 2-year follow-up.
Procedural success was defined as no procedural complications (death, major target limb amputation, thrombosis of the target lesion, or TLR) before discharge.
Statistical AnalysisIn this retrospective, nonrandomized study comparing conventional PTA and DCB treatment in dialysis patients with femoropopliteal disease, propensity scores were calculated for each subject based on a logistic regression model with a 2-to-1 nearest neighbor method. Baseline characteristics (age, sex, body mass index, history of atrial fibrillation, hypertension, diabetes mellitus, coronary artery disease, stroke, hypercholesterolemia, smoking, clinical presentation) and angiographic characteristics (lesion length, present of in-stent restenosis, severe calcification, chronic total occlusion, BTK runoff vessels) were used as matching criteria. All categorical variables are presented as numbers and percentages and were compared using the chi-square test, whereas continuous data are expressed as a mean±standard deviation and were analyzed using independent t-tests. The C-reactive protein level is presented as median and interquartile range and was logarithmically transformed before statistical analysis because of the strongly skewed distribution. The rates of freedom from binary restenosis and CD-TLR for both groups were assessed using Kaplan-Meier curves and compared with the log-rank test. To identify predictors of binary restenosis and CD-TLR, a univariate Cox proportional hazards model of baseline characteristics, laboratory data, lesion characteristics and angiographic findings was performed first. Each variable with a significant association (P<0.25) was entered into the multivariate proportional hazards model to identify the factors as independent predictors of each time-to-event outcome. Propensity score matching was performed with R version 2.14.0 (R Foundation for Statistical Computing, Vienna, Austria). All other statistical analyses were conducted with the SPSS statistical package for Windows version 21.0 (SPSS, Chicago, IL, USA). P<0.05 was considered statistically significant.
In the designed timeframe, a total of 278 dialysis patients with 339 limbs (291 undergoing conventional PTA; 48 treated with DCB) were eligible for matching. After propensity score matching, the PTA group consisted of 77 patients with 84 limbs, and the DCB group contained 37 patients with 46 limbs. Table 1 summarizes the baseline patient characteristics. There were no differences between the 2 matched groups with regards to age, sex, underlying comorbidities, laboratory data, dialysis duration, antiplatelet medications and clinical presentation of the affected legs. Notably, each group had high proportions of diabetes mellitus (90%), hypertension (80%), and coronary artery disease (80%), and 70% of both groups presented with CLI. The target ABI was also similar between groups after excluding patients with falsely elevated ABIs (≥1.3).
| PTA group (n=84 limbs) |
DCB group (n=46 limbs) |
P value | |
|---|---|---|---|
| Age, years | 67.1±11.2 | 67.9±10.7 | 0.706 |
| Female sex | 41 (48.8) | 25 (54.3) | 0.337 |
| Underlying medical comorbidity | |||
| Diabetes mellitus | 77 (91.7) | 43 (93.5) | 0.501 |
| Hypertension | 69 (82.1) | 38 (82.6) | 0.575 |
| Hypercholesterolemia | 37 (44.0) | 23 (50.0) | 0.320 |
| CAD | 67 (79.8) | 37 (80.4) | 0.560 |
| Stroke | 12 (14.3) | 7 (15.2) | 0.539 |
| Smoking | 31 (36.9) | 15 (32.6) | 0.385 |
| Atrial fibrillation | 9 (10.7) | 3 (6.5) | 0.327 |
| BMI (kg/m2) | 24.3±3.2 | 23.7±2.7 | 0.293 |
| LDL-C (mg/dL) | 92.8±28.5 | 93.4±29.6 | 0.913 |
| Glycohemoglobin (%) | 7.0±1.3 | 7.5±1.5 | 0.062 |
| Hematocrit (%) | 32.4±5.0 | 32.2±6.7 | 0.845 |
| CRP (mg/dL) | 1.33 (0.74–5.75) | 1.74 (0.60–7.25) | 0.575 |
| Albumin (g/dL) | 2.99±0.68 | 2.80±0.53 | 0.109 |
| Dialysis duration (months) | 77.6±70.9 | 83.6±87.5 | 0.684 |
| Dialysis duration | 0.425 | ||
| <1 year | 7 (8.3) | 6 (13.0) | |
| 1–3 years | 16 (19.0) | 8 (17.4) | |
| 3–5 years | 15 (17.9) | 5 (10.9) | |
| 5–10 years | 26 (31.0) | 11 (23.9) | |
| >10 years | 17 (20.2) | 11 (23.9) | |
| Unknown | 3 (3.6) | 5 (10.9) | |
| Clinical presentation | 0.789 | ||
| Claudication | 25 (29.8) | 14 (30.4) | |
| Resting pain | 12 (14.3) | 4 (8.7) | |
| Ulcer | 35 (41.7) | 22 (47.8) | |
| Gangrene | 12 (14.3) | 6 (13.0) | |
| Medications | |||
| Aspirin | 28 (33.3) | 20 (43.5) | 0.169 |
| Thienopyridine | 72 (85.7) | 42 (91.3) | 0.329 |
| Cilostazol | 61 (72.6) | 36 (78.3) | 0.313 |
| Statin | 17 (20.2) | 15 (32.6) | 0.089 |
| Target limb ABI (all patients) | 0.74±0.49 | 0.74±0.43 | 0.990 |
| Target limb ABI (≥1.3)* | 0.54±0.18 | 0.55±0.17 | 0.802 |
Values are mean±SD, n (%) or median (interquartile range). *Calculated by excluding ABI ≥1.3. ABI, ankl-brachial index; BMI, body mass index; CAD, coronary artery disease; CRP, C-reactive protein; DCB, drug-coated balloon; LDL-C, low-density lipoprotein-cholesterol; PTA, percutaneous transluminal angioplasty.
Table 2 summarizes the baseline lesion characteristics and procedural details. The ratios of lesion type, patency of BTK vessels, vessel occlusion, severe calcification, TASC II classification and lesion location were similar between groups. The mean lesion length in the PTA and DCB groups was 173.9±109.2 mm and 198.1±99.5 mm, (P=0.214), respectively. After the intervention, the final minimal luminal diameter (4.32±0.89 vs. 4.18±0.84 mm, P=0.386) and postprocedural DS (18.2±7.9% vs. 17.7±6.6%, P=0.749) were similar between groups. The rates of concomitant intervention for the iliac artery or BTK artery did not differ between groups. Compared with the DCB group, the PTA group had a higher rate of provisional stenting (58.3% vs. 41.3%, P=0.047) and longer length of the implanted stent (183.9±98.5 mm vs. 97.9±63.3 mm, P=0.001). The mean number and length of DCB balloons during the EVT were 1.59±0.86 and 188.5±108.4 mm, respectively. The procedural success and complication rates were not different in the 2 groups.
| PTA group (n=84 limbs) |
DCB group (n=46 limbs) |
P value | |
|---|---|---|---|
| Lesion type | 0.292 | ||
| De novo lesion | 59 (70.2) | 30 (65.2) | |
| Restenotic lesion | 7 (8.3) | 8 (17.4) | |
| In-stent restenosis | 18 (21.4) | 8 (17.4) | |
| Patency of BTK vessels | 0.581 | ||
| ≤1 vessel runoff | 62 (73.8) | 34 (73.9) | |
| ≥2 vessel runoff | 22 (26.2) | 12 (26.1) | |
| Total occlusions | 24 (28.6) | 13 (28.3) | 0.569 |
| Severe calcification | 44 (52.4) | 23 (50.0) | 0.469 |
| TASC classification | 0.549 | ||
| A | 3 (3.6) | 0 (0) | |
| B | 34 (40.5) | 21 (45.7) | |
| C | 33 (39.3) | 19 (41.3) | |
| D | 14 (16.7) | 6 (13.0) | |
| Lesion location | 0.382 | ||
| SFA only | 42 (50.0) | 18 (39.1) | |
| SFA and popliteal artery | 35 (41.7) | 25 (54.3) | |
| Popliteal artery only | 7 (8.3) | 3 (6.5) | |
| Lesion length (mm) | 173.9±109.2 | 198.1±99.5 | 0.214 |
| Preprocedural measurement | |||
| RVD (mm) | 5.07±0.88 | 4.87±0.79 | 0.098 |
| MLD (mm) | 0.81±0.77 | 0.91±0.68 | 0.471 |
| DS (%) | 84.2±14.2 | 80.8±13.7 | 0.196 |
| Postprocedural measurement | |||
| RVD (mm) | 5.28±0.81 | 5.11±0.74 | 0.214 |
| MLD (mm) | 4.32±0.89 | 4.18±0.84 | 0.386 |
| DS (%) | 18.2±7.9 | 17.7±6.6 | 0.749 |
| Concomitant iliac intervention | 3 (3.6) | 1 (2.2) | 0.556 |
| Concomitant BTK intervention | 64 (76.2) | 40 (87.0) | 0.106 |
| Provisional stenting | 49 (58.3) | 19 (41.3) | 0.047 |
| Average stent length (mm) | 183.9±98.5 | 97.9±63.3 | 0.001 |
| Average no. of DCB | 1.59±0.86 | ||
| Mean DCB length | 188.5±108.4 | ||
| Dissection | 0.580 | ||
| 0 | 25 (29.8) | 14 (30.4) | |
| A–C | 54 (64.3) | 27 (58.7) | |
| D–F | 5 (6.0) | 5 (10.9) | |
| Distal embolization | 2 (2.4) | 0 (0) | 0.416 |
| Procedural success | 83 (98.8) | 46 (100) | 0.646 |
Values are mean±SD or n (%). BTK, below-the-knee; DCB, drug-coated balloon; DS, diameter stenosis; MLD, minimal luminal diameter; PTA, percutaneous transluminal angioplasty; RVD, reference vessel diameter; SFA, superficial femoral artery; TASC, Trans-Atlantic Inter-Society Consensus.
The 2-year incidences of the study results in the matched group are shown in Figure 2 and Figure 3. The rates of freedom from binary restenosis and CD-TLR were significantly higher in the DCB group compared with the PTA group (18.6% vs. 52.4%, P<0.001 and 25.9% vs. 56.4%, P=0.001, respectively) (Figure 2A,B). No differences were observed in the rates of freedom from MACE (68.3% vs. 67.8%, P=0.716) (Figure 3A), minor amputation (82.1% vs. 78.6%, P=0.558), major amputation (98.8% vs. 100%, P=0.309), and all-cause death (72.0% vs. 83.7%, P=0.292) (Figure 3B) between the groups at 2 years.
Kaplan-Meier curves of freedom from binary restenosis and CD-TLR. (A) Compared with PTA, the DCB group had a higher binary restenosis-free rate at 2 years (52.4% vs. 18.6%, P<0.001). (B) Compared with PTA, the DCB group had a higher CD-TLR-free rate at 2 years (56.4% vs. 25.9%, P=0.001). CD-TLR, clinically driven target vessel revascularization; DCB, drug-coated balloon(s); PTA, percutaneous transluminal angioplasty.
Kaplan-Meier curves of MACE and all-cause death. The rates of freedom from (A) MACE (67.8% vs. 68.3%, P=0.716) and (B) all-cause death (83.7% vs. 72.0%, P=0.292) were not different between the DCB and PTA groups. DCB, drug-coated balloon(s); MACE, major adverse cardiac event(s); PTA, percutaneous transluminal angioplasty.
The univariate proportional hazards model for factors associated with binary restenosis and CD-TLR is shown in Table S1. The multivariate proportional hazards model for the predictors of binary restenosis and CD-TLR is shown in Table 3. DCB treatment (hazard ratio (HR) 0.368; 95% confidence interval (CI): 0.201–0.674, P=0.001), use of statins (HR 0.378; 95% CI: 0.175–0.818, P=0.014) and cilostazol (HR 0.430; 95% CI: 0.245–0.757, P=0.003) were negative independent factors of 2-year binary restenosis, and DCB treatment (HR 0.390; 95% CI: 0.206–0.740, P=0.004) was the only predictor that was negatively associated with CD-TLR.
| Binary restenosis | CD-TLR | |||
|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | |
| DCB group | 0.368 (0.201–0.674) | 0.001 | 0.390 (0.206–0.704) | 0.004 |
| Total occlusion | 1.380 (0.818–2.328) | 0.227 | 1.725 (0.996–2.988) | 0.052 |
| Postprocedural DS | 1.036 (0.999–1.075) | 0.060 | 1.026 (0.987–1.067) | 0.191 |
| BTK intervention | 0.776 (0.433–1.389) | 0.393 | 0.743 (0.416–1.327) | 0.315 |
| Aspirin | 0.590 (0.326–1.069) | 0.082 | 0.814 (0.416–1.590) | 0.546 |
| Thienopyridine | 0.646 (0.279–1.496) | 0.307 | 0.916 (0.331–2.535) | 0.866 |
| Statin | 0.378 (0.175–0.818) | 0.014 | 0.490 (0.209–1.149) | 0.101 |
| Cilostazol | 0.430 (0.245–0.757) | 0.003 | ||
| CLI | 1.742 (0.989–3.066) | 0.055 | ||
| Female sex | 1.293 (0.754–2.218) | 0.350 | ||
| Diabetes mellitus | 2.813 (0.615–12.87) | 0.182 | ||
| Stroke | 0.751 (0.286–1.970) | 0.561 | ||
CD-TLR, clinically driven target lesion revascularization; CI, confidence interval; CLI, critical limb ischemia; BTK, below-the-knee; DCB, drug-coated balloon; DS, diameter stenosis; HR, hazard ratio.
This is the first study comparing conventional PTA and DCB in dialysis patients with symptomatic femoropopliteal disease. After propensity score matching to minimize the confounding covariates, our results showed a significantly lower rate of binary restenosis and CD-TLR at 2 years following DCB treatment. The rates of minor amputation and death were similar, with rare cases of major amputations in both groups.
Although previous studies have demonstrated the effectiveness and safety of EVT for dialysis patients with LEAD, the clinical outcomes remain poor as compared with non-dialysis patients, with higher rates of restenosis, amputation and all-cause death.12,13 Dialysis patients also have an increased risk of restenosis or loss of patency after nitinol stent implantation in the superficial femoral artery.20 The lesion complexities in dialysis patients with LEAD, including vascular calcification, diffuse vessel disease and a higher proportion of TASC II C&D lesions, contribute to the poor clinical outcomes after EVT.1,12,13
In the past decade, DCB has been included in the armamentarium of endovascular specialists for treatment of femoropopliteal disease. Safety and favorable outcomes, attributed to the reduction of both restenosis and the need for reintervention, are reported in comparison with conventional PTA.14–17,21 However, the efficacy of DCB diminishes in calcified vessels22 and long, complex lesions,18,23 which are usually excluded from randomized trials and commonly observed in dialysis patients.
A direct comparison between DCB and conventional PTA treatment is essential to study the effectiveness of DCB in dialysis patients. In this study, we used propensity score matching to minimize the confounding covariates between groups. The high proportions of severe vessel calcification, TASC II C&D lesions and long lesion length in both groups suggested greater lesion complexity in these patients. The use of DCB significantly reduced the need for provisional and longer stent implantation. The anti-restenotic effect of DCB remained effective in dialysis patients at 2 years, and multivariate analysis showed treatment with DCB was the only independent variable to predict freedom from both binary restenosis and CD-TLR. Regarding the safety profile, the results for the dialysis patients treated with DCB were similar to those of non-dialysis patients, with a lower rate of procedural complications.15
To assess the risk of limb loss in CLI patients, the wound, ischemia, and foot infection (WIfI) classification24 was first introduced in 2014 and was recently reported to predict the wound healing and amputation risk in selected hemodialysis patients with CLI after treatment with EVT.25 Although detailed information on the WIfI classification was not available for study patients enrolled before 2014, the minor and major amputation rates were comparably lower in both groups. However, use of DCB did not improve the rate of MACE or all-cause death in the dialysis patients, which was in line with the results of a previous meta-analysis.17
Schmidt et al reported their DCB outcomes in a real-world registry, which showed the 2-year primary patency and CD-TLR-free rates for complex femoropopliteal lesions (mean lesion length, 24 cm) were 53.7% and 68.6%, respectively.23 They suggested that DCB was effective in delaying rather than preventing restenosis in facing complex lesions. In line with Schmidt’s study, our data showed that dialysis patients undergoing DCB treatment for complex femoropopliteal disease had a reduction in the primary patency and CD-TLR-free rate at 2 years (for primary patency, 75.4% and 52.4% at 1 and 2 years; for CD-TLR-free rate, 74.5% and 56.4% at 1 and 2 years).
Notably, the use of statins and cilostazol significantly reduced the 2-year rate of binary restenosis in the multivariate analysis. The current guideline26 and recent study27 recommend the use of statins to mitigate the risk of cardiovascular events in patients with clinical LEAD, as well as those undergoing open and endovascular interventions.26,27 Statin use also improved the rates of primary and secondary patency in patients with CLI after EVT.28 However, randomized controlled trials show that use of statins in dialysis patients confers no benefit to the reduction of cardiovascular events.29,30 Our study results were in favor of statin use to reduce restenosis in dialysis patients who had undergone EVT for LEAD. However, only a minority of patients in this study underwent statin treatment. A large-scale study with longer follow-up is warranted before routine use of a statin becomes a standard treatment in dialysis patients with LEAD.
Cilostazol, a phosphodiesterase type 3 inhibitor, has antiplatelet aggregation and vasodilatory effects and inhibits smooth muscle cell proliferation.31–33 In patients with femoropopliteal disease who have undergone EVT, treatment with cilostazol is associated with a significant reduction of CD-TLR and in-stent restenosis,34,35 as well as an improvement of long-term patency in dialysis patients after EVT.36 Our study demonstrated a beneficial effect of cilostazol in reducing binary restenosis in dialysis patients after EVT, possibly because of the vasodilatory, anti-inflammatory and antiproliferative effects on vascular smooth muscle cells. However, cilostazol administration did not appear to improve the CD-TLR-free rate in this study.
Study LimitationsThis was a retrospective single-center study and thus lacks randomization. However, using a propensity score-based analysis aimed at excluding significant bias for measured covariates, this study was closer to a real-world scenario than even a prospective randomized trial design with its high patient selection bias. Despite lacking core laboratory analysis to adjudicate the angiographic results, we minimized the interobserver variability by having at least 2 independent operators. Second, the sample size was relatively small. Nonetheless, a literature review suggests that this study may be the largest to report the efficacy and safety of DCB treatment in dialysis patients with femoropopliteal disease. In addition, more than 90% of the patients in the DCB group received IN.PACT Admiral DCB therapy. It is unclear whether the study results can be explored with other DCB technology. Moreover, vessel calcification was assessed via fluoroscopy only, and thus the clinical effect of circumferential calcification on the efficacy of DCB cannot be demonstrated. As well, detailed information on toe–brachial index was not available because it was only measured in patients with non-compressible vessels. Finally, routine follow-up angiography was not performed, and quantitative measurements of late lumen loss were not available.
This matched study suggested that treatment with DCB significantly reduced the rate of binary restenosis and the need of CD-TLR at 2-year follow-up, with a similar safety profile compared with conventional PTA. These results support the clinical benefits of local delivery of paclitaxel in treating complex femoropopliteal diseases in dialysis patients.
The authors thank the cardiac catheterization laboratory medical staff and clinical research coordinators for participating in this study.
Nothing to declare.
Supplementary File 1
Table S1. Cox regression univariate analysis for binary restenosis and CD-TLR
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-18-0077
