|
What’s new? Despite multiple advances in coronary stent design and implantation techniques, in-stent restenosis remains a significant clinical issue. Recurrence of in-stent restenosis is a rare event, but still little is known about its optimal treatment. Our multicenter retrospective analysis of two most widely applied strategies suggests that implantation of another drug-eluting stent in the restenotic lesion might be preferred over the use of a drug-eluting balloon, as it helps avoid future revascularizations. |
Introduction
Recurrent in-stent restenosis (R-ISR) is defined as a second event of ISR after successful treatment of an initial ISR lesion [1]. Despite advances in stent technology and implantation technique, still up to 12% of patients undergoing percutaneous coronary intervention (PCI) with current drug-eluting stents (DES) experience target lesion failure within 5 years, with the majority of those events attributed to ISR [2]. There are very limited data on the prevalence of R-ISR, which was observed in at least 1.4% of all patients undergoing PCI in a large retrospective cohort [3]. While implantation of another stent or using a drug-eluting balloon (DEB) are both valid strategies with similar effectiveness in the first occurrence of ISR, the preferred approach to R-ISR is debatable [4]. Specific concerns of stenting include adding another layer of metallic scaffolding, which may lead to progression of luminal narrowing and an increased risk of thrombotic events. Conversely, the use of DEB may be associated with mechanical complications, has limited potential of reducing the neointimal tissue burden, and has recently been associated with excess mortality in the context of peripheral interventions [5, 6]. Moreover, according to the current guidelines, R-ISR should prompt consideration of surgical revascularization [7]. Hence, the results of trials evaluating treatment of the first ISR could not be directly extrapolated to R-ISR. Additionally, data on the management of R-ISR come mostly from the BMS and first-generation DES era with limited experience with new-generation thin-strut DES (thin-DES) [8]. Therefore, we aimed to utilize the data from the contemporary DEB-DRAGON registry (NCT04415216) to compare the outcomes of DEB and thin-strut DES in treatment of R-ISR.
Methods
Population
The DEB-DRAGON is a multicenter observational registry conducted in thirteen high-volume catheterization laboratories in Poland, which collected data from patients with coronary ISR treated with PCI between February 2008 and October 2019. Long-term follow-up was obtained from the National Health Fund. The current analysis utilized the DEB DRAGON registry to select patients with R-ISR in previously implanted DES (i.e. with one or more stent layers present within the lesion). Patients with the first incidence of ISR, ISR in a bare-metal stent or in bypass grafts, treated with a thick-strut DES, patients who were treated with both stent and a drug-eluting balloon, as well as those who underwent simultaneous PCI of multiple coronary territories were excluded. Derivation of the final study cohort is shown in Figure 1.
Patients were divided into two groups according to the type of interventional treatment received at the time of R-ISR: a thin-DES (strut thickness <100 µm) or a paclitaxel-eluting DEB. The following thin-strut stents were used: Xience (Abbott Vascular Devices, Santa Clara, CA, US), Resolute (Medtronic CardioVascular, Santa Rosa, CA, US), Promus (Boston Scientific, Natick, MA, US), Synergy (Boston Scientific, Natick, MA, US), Orsiro (Biotronik AG, Bulach, Switzerland), Alex (Balton, Warszawa, Poland). The paclitaxel-DEB types were: Agent (Boston Scientific, Natick, MA, US), Elutax (Aachen Resonance GmbH, Aachen, Germany), Essential (iVascular, Barcelona, Spain), In.Pact (Medtronic Vascular, Santa Clara, CA, US), Pantera Lux (Biotronik AG, Buulach, Switzerland), Restore DEB (Cardionovum GmbH, Bonn, Germany), SeQuent Please Neo (B.Braun Interventional Group, Ltd, Melsulgen, Germany). All PCIs were performed by certified interventional cardiologists in accordance with standard procedures at each catheterization laboratory. No routine angiographic follow-up was recommended. The patients’ data were anonymized in each center, combined into a database, and statistically analyzed as a single cohort. Chronic kidney disease was defined as estimated glomerular filtration rate <60 ml/min/1.73 m2, hyperlipidemia as low-density lipoprotein cholesterol concentration >116 mg/dl or current lipid-lowering treatment, hypertension as blood pressure >140/90 mm Hg or current antihypertensive treatment, peripheral artery disease as prior lower limb or carotid revascularization or current ischemic symptoms with >50% vessel luminal stenosis. The study was approved by the local ethics committees of each participating center. The patients’ data were protected according to the requirements of Polish law, General Data Protection Regulation (GDPR), and hospital Standard Operating Procedures. The study was conducted in accordance with the Declaration of Helsinki.
Endpoints
The primary endpoint was target lesion revascularization (TLR). Secondary endpoints were target vessel revascularization (TVR), myocardial infarction (MI), all-cause death, and device-oriented adverse cardiac events (DOCE, including cardiac death, TLR, and target vessel MI). All endpoints were defined according to the definitions of endpoints for clinical trials [9].
Statistical analysis
Continuous variables were presented as means with standard deviations and compared with Student t-test in case of normal distribution. Variables with non-normal distribution were presented as medians and interquartile ranges and compared with the Mann-Whitney U test. Normality was assessed with the Kolmogorov-Smirnov test. Discrete variables were expressed as counts and percentages and compared with the χ2 test. Crude incidence of adverse events was presented with Kaplan-Meier survival curves and compared with a long-rank test. Propensity-score matching with the nearest neighbor method was used to adjust for baseline differences. The variables selected for matching are listed in the Supplementary material, File S1. The validity of logistic regression was assessed using the Hosmer-Lemeshow goodness-of-fit test. The model was well calibrated (χ2 = 4.13; P = 0.85). The propensity model yielded a concordance index (C-index) of 0.75 (95% confidence interval [CI], 0.70–0.81). The association of treatment and selected variables with TLR were assessed with the Cox proportional hazard model using a backward multivariable procedure in the matched population. Hazard ratios (HR) with the corresponding 95% CI were estimated. A test for non-proportionality of hazards based on Schoenfeld residuals did not reveal violations of the proportionality assumptions. The two-sided P-value <0.05 was considered significant. The statistical analysis was performed using SAS 9.4 (SAS Institute, Cary, NC, US).
Results
The final analysis included 311 patients with DES-ISR, of whom 86 (mean age [standard deviation, SD], 65.3 [10.0] years, 69.7% male) received a thin-DES and 225 (mean age [SD], 67.7 [9.9] years, 67.1% male) were treated with a DEB. There were substantial differences in terms of baseline clinical characteristics (Table 1).
|
|
Unmatched population |
Propensity score-matched population |
||||
|
DEB n = 225 (72.3%) |
Thin-DES n = 86 (27.7%) |
P-value |
DEB n = 57 |
Thin-DES n = 57 |
P-value |
|
|
Demographic data |
||||||
|
Age, years, mean (SD) |
67.7 (9.9) |
65.3 (10.0) |
0.06 |
66.2 (9.7) |
65.9 (11.4) |
0.87 |
|
Male sex |
151 (67.1) |
57 (66.3) |
0.89 |
39 (68.4) |
37 (48.7) |
0.69 |
|
BMI, kg/m2, mean (SD) |
28.5 (4.6) |
27.6 (3.8) |
0.22 |
27.2 (3.9) |
27.4 (3.8) |
0.82 |
|
Discharge diagnosis |
||||||
|
Chronic coronary syndrome |
74 (32.9) |
20 (23.3) |
0.27 |
18 (31.6) |
15 (26.3) |
0.54 |
|
Unstable angina |
97 (43.1) |
39 (45.3) |
28 (49.1) |
23 (40.3) |
||
|
NSTEMI |
49 (21.8) |
23 (26.7) |
10 (17.6) |
16 (28.1) |
||
|
STEMI |
5 (2.2) |
4 (4.6) |
1 (1.7) |
3 (5.3) |
||
|
CAD history |
||||||
|
Previous MI |
169 (75.1) |
64 (74.4) |
0.90 |
37 (64.9) |
41 (71.9) |
0.42 |
|
Previous CABG |
67 (29.8) |
21 (24.4) |
0.35 |
17 (29.8) |
14 (24.6) |
0.53 |
|
Risk factors and comorbidities |
||||||
|
Diabetes mellitus — Requiring insulin |
110 (48.9) 43 (19.1) |
34 (39.5) 4 (4.6) |
0.14 <0.01 |
18 (31.6) 5 (8.8) |
23 (40.3) 4 (7.0) |
0.33 1.00 |
|
Hypertension |
201 (89.3) |
83 (96.5) |
0.04 |
55 (96.5) |
55 (96.5) |
1.00 |
|
Hyperlipidemia |
184 (81.8) |
71 (82.6) |
0.87 |
47 (82.5) |
47 (82.5) |
1.00 |
|
Chronic kidney disease — On dialysis |
70 (31.1) 3 (1.3) |
15 (17.1) 0 (0.0) |
0.02 0.56 |
9 (15.8) 0 (0) |
12 (21.1) 0 (0) |
0.47 NA |
|
Atrial fibrillation |
42 (18.7) |
9 (10.5) |
0.08 |
6 (10.5) |
8 (14.0) |
0.57 |
|
Current smoker |
43 (19.1) |
26 (30.2) |
0.04 |
15 (26.3) |
14 (24.6) |
0.83 |
|
Family history of CAD |
50 (24.9) |
25 (29.1) |
0.46 |
16 (32.0) |
17 (29.8) |
0.81 |
|
Pulmonary disease |
21 (9.3) |
12 (13.9) |
0.24 |
7 (12.3) |
7 (12.3) |
1.00 |
|
Peripheral artery disease |
48 (21.3) |
14 (16.3) |
0.32 |
13 (22.8) |
8 (14.0) |
0.23 |
|
Left ventricular ejection fraction, %, mean (SD) |
48.9 (11.4) |
46.8 (10.8) |
0.14 |
48.4 (11.0) |
48.0 (9.8) |
0.84 |
|
Time to ISR, months, median (IQR) |
18.0 (7.0–36.0) |
18.6 (10.3–40.8) |
0.19 |
18.3 (7.0–41.0) |
18.6 (10.8–32.5) |
0.60 |
|
Current ISR event |
||||||
|
Second |
145 (64.4) |
54 (62.8) |
0.79 |
34 (60.7) |
38 (66.7) |
0.51 |
|
Third or further |
80 (35.6) |
32 (37.2) |
22 (39.3) |
19 (33.3) |
||
|
Type of stent with ISR |
N = 90 |
N = 20 |
|
N = 27 |
N = 14 |
|
|
First-generation DES |
|
|
|
|
|
|
|
SES |
4 (57.1) |
2 (33.3) |
0.592 |
1 (50.0) |
3 (75.0) |
1.0 |
|
PES |
3 (42.9) |
4 (66.7) |
1 (50.0) |
1 (25.0) |
||
|
Second-generation DES |
|
|
|
|
|
|
|
SES |
13 (15.7) |
5 (38.5) |
0.007 |
4 (16.0) |
5 (50.0) |
0.13 |
|
BES |
1 (1.2) |
2 (15.4) |
1 (4.0) |
1 (10.0) |
||
|
ZES |
16 (19.3) |
0 |
2 (8.0) |
0 (0) |
||
|
EES |
53 (63.9) |
6 (46.2) |
18 (72.0) |
4 (40.0) |
||
The prevalence of insulin-dependent diabetes mellitus (4.6 vs. 19.1%; P <0.01) and chronic kidney disease (17.1 vs. 31.1%; P = 0.02) was lower in patients treated with thin-DES compared with DEB, while hypertension (96.5 vs. 89.3%; P = 0.04) and current smoking (30.2 vs. 19.1%; P = 0.04) were more frequent in the thin-DES group compared with the DEB group. Around two-thirds of patients were treated for the second and one-third for the third or further episode of ISR within the same lesion. The angiographic characteristics were similar, except for more frequent final post-PCI TIMI-3 flow (100 vs. 93.8%; P = 0.01) and fewer patients with only one layer of previously implanted stents (20.9 vs. 44.9%; P <0.001) in patients treated with thin-DES compared with DEB (Table 2). Dual antiplatelet treatment was prescribed for a longer period in patients treated with thin-DES.
|
|
Unmatched population |
Propensity score-matched population |
||||
|
DEB n = 225 (72.3%) |
Thin-DES n = 86 (27.7%) |
P-value |
DEB n = 57 |
Thin-DES n = 57 |
P-value |
|
|
Angiography |
||||||
|
1-vessel disease |
135 (60.0) |
46 (53.5) |
0.24 |
33 (57.9) |
33 (57.9) |
0.82 |
|
2-vessel disease |
57 (25.3) |
30 (34.9) |
16 (28.1) |
18 (31.6) |
||
|
3-vessel disease |
33 (14.7) |
10 (11.6) |
8 (14.0) |
6 (10.5) |
||
|
Bifurcation |
30 (13.3) |
8 (9.3) |
0.33 |
9 (15.8) |
7 (12.3) |
0.59 |
|
Thrombus |
2 (0.9) |
3 (3.5) |
0.13 |
1 (1.7) |
0 (0.0) |
1.00 |
|
Severe calcification |
4 (1.8) |
3 (3.5) |
0.40 |
2 (3.5) |
3 (5.3) |
1.00 |
|
Diameter stenosis, %, mean (SD) |
84.0 (11.7) |
84.0 (8.8) |
0.97 |
80.9 (12.5) |
83.5 (8.6) |
0.20 |
|
Target lesion |
||||||
|
Left main |
13 (5.8) |
6 (7.0) |
0.69 |
7 (12.3) |
5 (8.8) |
0.54 |
|
Left anterior descending |
104 (46.2) |
39 (45.3) |
0.89 |
24 (42.1) |
25 (43.9) |
0.85 |
|
Left circumflex |
102 (45.3) |
39 (45.3) |
0.99 |
27 (47.4) |
29 (50.9) |
0.71 |
|
Right coronary artery |
117 (52.0) |
54 (62.8) |
0.09 |
28 (49.1) |
29 (50.9) |
0.85 |
|
Original stent-length, mm, mean (SD) |
24.3 (9.6) |
21.3 (7.8) |
0.10 |
26.1 (10.9) |
20.9 (8.0) |
0.06 |
|
Original stent diameter, mm, mean (SD) |
3.1 (0.5) |
3.0 (0.4) |
0.81 |
2.9 (0.5) |
3.0 (0.4) |
0.32 |
|
Prior stent layers |
|
|
|
|
|
|
|
1 |
101 (44.9) |
18 (20.9) |
<0.001 |
22 (38.6) |
13 (22.8) |
0.16 |
|
2 |
94 (41.8) |
51 (59.3) |
24 (42.1) |
33 (57.9) |
||
|
>2 |
30 (13.3) |
17 (19.8) |
11 (19.3) |
11 (19.3) |
||
|
ISR morphology |
||||||
|
Focal |
107 (50.0) |
42 (48.8) |
0.16 |
26 (46.4) |
29 (50.9) |
0.53 |
|
Diffuse |
79 (36.9) |
28 (32.6) |
23 (41.1) |
18 (31.6) |
||
|
Proliferative |
20 (9.4) |
15 (17.4) |
5 (8.9) |
9 (15.8) |
||
|
Occlusive |
8 (3.7) |
1 (1.2) |
2 (3.6) |
1 (1.7) |
||
|
Balloon pre-dilatation |
||||||
|
Length, mm, mean (SD) |
16.5 (4.0) |
16.2 (4.7) |
0.64 |
15.9 (3.2) |
15.9 (5.0) |
0.98 |
|
Diameter, mm, mean (SD) |
3.0 (0.6) |
2.9 (0.7) |
0.62 |
2.9 (0.5) |
2.9 (0.8) |
0.88 |
|
Device data |
||||||
|
Length, mm, mean (SD) |
22.2 (6.6) |
20.7 (9.1) |
0.17 |
21.5 (7.0) |
21.1 (8.9) |
0.80 |
|
Diameter, mm, mean (SD) |
3.1 (0.5) |
3.1 (0.5) |
0.70 |
3.0 (0.4) |
3.1 (0.5) |
0.22 |
|
Post-procedure |
||||||
|
Residual stenosis |
21 (9.3) |
9 (10.5) |
0.76 |
9 (15.8) |
5 (8.8) |
0.25 |
|
TIMI 3 |
211 (93.8) |
86 (100) |
0.01 |
57 (100) |
57 (100) |
NA |
|
Complications |
||||||
|
Perforation |
1 (0.4) |
0 (0.0) |
1.00 |
0 (0) |
0 (0) |
NA |
|
Dissection |
6 (2.7) |
2 (2.3) |
1.00 |
0 (0) |
2 (3.5) |
0.49 |
|
No-reflow |
2 (0.9) |
0 (0.0) |
1.00 |
0 (0) |
0 (0) |
NA |
|
Intracoronary imaging |
12 (5.3) |
6 (7.0) |
0.59 |
1 (1.7) |
5 (8.8) |
0.21 |
|
Duration of DAPT after PCI |
||||||
|
<3 months |
37 (16.4) |
3 (3.5) |
<0.01 |
5 (8.8) |
3 (5.3) |
0.58 |
|
4–6 months |
31 (13.8) |
12 (13.9) |
6 (10.5) |
9 (15.6) |
||
|
7–12 months |
157 (69.8) |
71 (82.6) |
46 (80.7) |
45 (78.9) |
||
|
Pharmacotherapy at discharge |
||||||
|
Clopidogrel |
194 (86.2) |
73 (84.9) |
0.76 |
46 (80.7) |
47 (82.5) |
0.81 |
|
Ticagrelor |
25 (11.1) |
12 (13.9) |
0.49 |
10 (17.5) |
9 (15.8) |
0.80 |
|
Prasugrel |
3 (1.3) |
1 (1.2) |
0.91 |
0 (0.0) |
1 (1.7) |
1.00 |
|
Oral anticoagulant |
34 (15.1) |
6 (7.0) |
0.06 |
5 (8.8) |
5 (8.8) |
1.00 |
The median follow-up was 31 months (range, 2–121 months). The primary endpoint of TLR occurred in 18 (20.9%) patients who received thin-DES and 61 (27.1%) patients treated with DEB (hazard ratio [HR], 0.57; 95% CI 0.33–0.98; log-rank P = 0.04) (Figure 2).
There was no significant difference between the thin-DES and DEB groups in terms of secondary endpoints, including DOCE (27.9 vs. 31.1%; P = 0.12), TVR (27.9 vs. 31.1%; P = 0.11), MI (20.9 vs. 20.9%; P = 0.29), and all-cause death (11.6 vs. 4.9%; P = 0.25 respectively) (Figure 3A–D).
The rates of each component of DOCE were similar among both groups (Table 3).
|
|
Unmatched population |
Propensity score-matched population |
||||||||||
|
DEB n = 225 (72.3%) |
Thin-DES n = 86 (27.6%) |
36-month event rate (95% CI) DEB |
36-month event rate (95% CI) DES |
HR (95%CI) Reference: DEB |
P-value (log-rank) |
DEB |
Thin-DES |
36-month event rate (95% CI) DEB |
36-month event rate (95% CI) DES |
HR (95%CI) Reference: DEB |
P-value (log-rank) |
|
|
TLR |
61 (27.1) |
18 (20.9) |
30.0 (23.9–37.1) |
19.0 (9.8–28.2) |
0.57 (0.33–0.98) |
0.039 |
19 (33.3) |
10 (17.5) |
33.5 (19.5–47.5) |
14.1 (4.3–23.8) |
0.38 (0.17–0.86) |
0.013 |
|
TVR |
70 (31.1) |
24 (27.9) |
34.4 (27.0–41.7) |
24.8 (14.2–35.4) |
0.69 (0.43–1.11) |
0.12 |
23 (40.3) |
14 (24.6) |
40.4 (26.1–54.7) |
21.5 (10.2–32.9) |
0.47 (0.24–0.94) |
0.026 |
|
MI |
47 (20.9) |
18 (20.9) |
22.0 (15.6–28.4) |
18.5 (9.6–27.4) |
0.78 (0.45–1.37) |
0.39 |
13 (22.8) |
11 (19.3) |
21.1 (9.9–32.3) |
17.8 (7.1–28,6) |
0.67 (0.29–1.56) |
0.33 |
|
TV-MI |
21 (9.3) |
6 (7.0) |
9.6 (5.0–14.1) |
6.2 (0.9–11.6) |
0.62 (0.25–1.54) |
0.29 |
6 (10.5) |
4 (7.0) |
7.4 (0.4–14.4) |
7.7 (0.4–14.9) |
0.17 (0.01–2.88) |
0.41 |
|
CV Death |
8 (3.6) |
7 (8.1) |
5.0 (1.4–8.0) |
7.9 (1.8–14.0) |
1.87 (0.67–5.22) |
0.22 |
0 (0) |
4 (7.0) |
0.0 |
8.4 (0.4–16.3) |
7.98 (0.30–211) |
0.06 |
|
All-cause death |
11 (4.9) |
10 (11.6) |
5.1 (1.7–8.5) |
10.7 (3.6–17.7) |
1.67 (0.69–4.04) |
0.25 |
2 (3.5) |
5 (8.8) |
1.8 (0.0–5.3) |
8.4 (0.4–16.3) |
1.58 (0.31–8.08) |
0.49 |
|
DOCE |
70 (31.1) |
24 (27.9) |
35.1 (27.7–42.6) |
24.7 (14.5–34.9) |
0.69 (0.43–1.10) |
0.12 |
19 (33.3) |
13 (22.8) |
34.0 (19.7–48.2) |
20.0 (6.6–33.5) |
0.52 (0.25–1.09) |
0.07 |
Propensity score matching yielded 57 well-matched pairs of patients who received a thin-DES or DEB. Baseline clinical and angiographic differences were balanced with no significant between-group difference (Tables 1 and 2). The primary endpoint of TLR was less prevalent in patients who received thin-DES compared with those treated with DEB (17.5 vs 33.3% respectively; HR, 0.38; 95% CI, 0.17–0.86; P = 0.01). Similarly, the risk of TVR was lower in patients treated with a thin-DES compared with a DEB (24.6 vs. 40.3% respectively; HR, 0.47; 95% CI, 0.24–0.94; P = 0.03). The difference favoring thin-DES over DEB in terms of DOCE did not meet statistical significance (22.8 vs. 33.3% respectively; HR, 0.52; 95% CI, 0.25–1.09; P = 0.07) and was mainly driven by repeated revascularizations. There was no difference in terms of MI, all-cause- and cardiac mortality between patients who received thin-DES compared with DEB (Table 3, Central illustration).
Multivariable Cox proportional hazard analysis with backward selection of variables performed on the matched sample demonstrated that treatment with thin-DES was an independent predictor of freedom from TLR (HR, 0.33; 95% CI, 0.133–0.841; P = 0.02), along with shorter length of the original stent, hypertension, R-ISR in the non-LMCA location and presentation with NSTEMI (Table 4).
|
Variable |
HR (95% CI) |
P-value |
|
Total length of implanted stents (1 mm increment) |
1.079 (1.034–1.126) |
<0.001 |
|
Hypertension |
0.070 (0.021–0.239) |
<0.001 |
|
Lesion location in left main |
8.383 (2.840–24.745) |
<0.001 |
|
NSTEMI |
0.166 (0.039–0.713) |
0.02 |
|
Thin-DES vs. DEB |
0.334 (0.133–0.841) |
0.02 |
Discussion
To our best knowledge, this is the largest study to date to examine the long-term outcomes after treatment of recurrent ISR with a thin-strut DES or a DEB. The main findings of our analysis are as follows: (1) R-ISR in DES is associated with a high risk of future cardiac events, especially of repeated revascularizations; (2) interventional treatment of recurrent R-ISR in DES with thin-strut DES results in fewer subsequent target lesion revascularizations compared with treatment with a DEB, and (3) both treatment modalities were associated with similar long-term risk of device-oriented composite endpoint, as well as of all-cause and cardiovascular mortality.
The data on long-term outcomes after PCI for R-ISR are not well known. In a study by Kubo et al. [10], the 4-year incidence of TLR after implantation of DES for R-ISR was 27.6%, in line with the frequency of 20.9% observed in our study at almost 3 years. However, in an analysis by Kawamoto et al. [11], the frequency of TLR at 2 years was 27.7% for DES and 38.3% for DEB. Theodoropoulos et al. [3] reported an even higher incidence of TLR of 45% at 2- years after treatment of recurrent ISR. These excessive event rates can be attributed to including only patients with ISR with two preexistent metallic stent layers, which was not a prerequisite in our study. However, the incidence of TLR observed in our study (20.9% and 27.1% for thin-DES and DEB) is still higher than reported after treatment of de novo coronary lesions even with new-generation DES (from 5.0% to 11.9% of TLR at 5 years) prompting close post-procedural follow-up [2, 12]
The results of our analysis suggest the superiority of thin-DES over DEB in the treatment of R-ISR in terms of the need for future target lesion revascularizations. This finding substantially differs from the results of our prior study on patients with the first episode of ISR, in which there was no advantage of either thin-DES or DEB in the propensity-score matched groups regarding any of the long-term outcomes [13]. In the absence of randomized trials, evaluating patients with recurrent ISR, all available evidence comes from retrospective analyses, which so far have yielded largely conflicting conclusions. Similar to our findings, a signal toward the superiority of DES compared with DEB was observed in the previously mentioned study by Kawamoto et al. [11] (27.7 vs. 38.3% of TLR at 2 years respectively), which did not meet statistical significance, possibly due to smaller sample size. On the other hand, in a report by Wang et al. [14] (n = 172), the 1-year incidence of TLR was significantly higher after DES compared with DEB (27.8 vs. 15.1%; P = 0.04 respectively). Interestingly, there were no baseline differences between the study groups, and thus no statistical adjustment was done. Compared with both these studies, our analysis has a larger sample size due to multicenter design and longer follow-up.
Treatment of recurrent ISR with stent implantation remains controversial since adding yet another metallic layer inside the vessel may lead to a potential decrease of lumen area and progressive luminal obstruction. However, this happens mostly in heavily resistant calcified lesions interfering with proper stent expansion [15]. In more compliant lesions, new-generation thin-strut DES may provide the benefit of a larger lumen area while maintaining the radial force. We, therefore, hypothesize that restricting our analysis only to new generation thin-strut DES might be a factor contributing to improved long-term outcomes in comparison with DEB.
Recent meta-analyses raised some important concerns about excess late mortality associated with using paclitaxel-eluting balloons for peripheral interventions [6, 16]. The potential explanation of the dose-dependent association between the drug and mortality was systemic toxicity of high-dose crystalline paclitaxel delivered with larger peripheral angioplasty balloons. This finding was not confirmed for similar devices used for coronary revascularization, which contain significantly less drug [17]. Similarly, our analysis did not show any excess mortality associated with DEB. On the contrary, there were numerically fewer deaths among patients treated with DEB compared with thin-DES (4.9 vs. 11.6%), which was not statistically significant.
Limitations
To obtain adequate sample size, we had to include patients treated with different types of DES, as well as several DEB platforms. Moreover, despite the current recommendations, the use of intravascular imaging in our cohort was infrequent, which precluded precise characterization of underlying ISR mechanism and subsequent tailored therapy [18]. Other available technologies postulated for the management of recurrent ISR, such as intravascular brachytherapy, laser atherectomy, ultra-high pressure balloon dilatations, and intravascular lithotripsy, were not applied. Furthermore, details on using cardiovascular medications, as well as on the control of hypertension and the prevalence of periprocedural myocardial infarction, are not available in the DEB-DRAGON registry. Due to the retrospective study design, even propensity score matching cannot fully exclude the influence of unmeasured confounders on the study results. Finally, the operator’s experience may have influenced the outcomes in this unique group of patients, especially in light of high anatomical complexity of the disease [19]. However, data on the operator’s volume were not available in the registry.
Conclusions
In patients with recurrent DES, ISR treatment with implantation of a thin-strut DES was superior to the use of DEB in terms of long-term risk of target lesion revascularization with similar risk of myocardial infarction, as well as all-cause and cardiovascular mortality.
Supplementary material
Supplementary material is available at https://journals.viamedica.pl/kardiologia_polska.
Article information
Acknowledgment: The DEB-DRAGON registry was initiated on the Scientific Platform of the “Club 30” of the Polish Cardiac Society.
Conflict of interest: None declared.
Funding: None.
Open access: This article is available in open access under Creative Common Attribution-Non-Commercial-No Derivatives 4.0 International (CC BY-NC-ND 4.0) license, allowing to download articles and share them with others as long as they credit the authors and the publisher, but without permission to change them in any way or use them commercially. For commercial use, please contact the journal office at kardiologiapolska@ptkardio.pl.
