THE “EDGE EFFECT” AFTER IMPLANTATION OF β -EMITTING ( 55 CO) STENTS WITH HIGH INITIAL ACTIVITY

Patients with single, de-novo lesion, with length<15mm, without important side branch angiographic presence of calcification at the spot of stenosis, and objective evidence of ischemia were eligible. We analysed neointimal hyperplasia and vascular remodelling in 10 patients who had completed a 6-month angiographic follow-up with intravascular ultrasound study. Summary: The aim of this study was to evaluate the incidence and the cause of “edge restenosis” after implantation of high activity 41.1 µ Ci ± 1.2 µ Ci=1520 kBq ± 44 kBq, β -emitting ( 55 Co) stents. Proton bombarding in cyclotron has brought the radioactivity. Intravascular ultrasound (IVUS) investigation has been completed in 10 patients. The angiographies performed at 6 month revealed restenosis >50 % in 5 cases (50 %). The analysis of edges (5 mm distally and proximally to the last stent struts) showed no significant changes in TVV (187.3 ± 62.60 mm 3 and 176.9 ± 53.5 mm 3 ) but PMV increase significantly (i.e. neointimal proliferation) from 61.9 ± 31.2 mm 3 to 82.2 ± 43.4 mm 3 (p<0.04) and was the major contributor (from 66 %) to lumen volume loss (125.4 ± 40.7 mm 3 and 94.7 ± 22.2 mm 3 , p<0.02). In conclusion, neither statistically significant positive nor negative remodelling at the “stent edges“, were present. Statistically significant increase in plaque+media volume (i.e. neointimal hyperplasia) and reduction in lumen volume were found. The cause of “edge restenosis” was especially (from 66 %) due to increase in plaque+media volume (i.e. neointimal hyperplasia). Probably, main reason for “edge effect”/neointimal hyperplasia was in this trial sharp fall-off in radiation at the edges of the stents.


Implantation technique
The irradiated bare stent was mounted on balloon and then implanted with low pressure (8 atmospheres) without predilatation, thereafter, high pressure post dilatation has been performed with shorter balloon to ensure that the edges of the balloon did not extend beyond the limits of the stent. Intravascular ultrasound was used to ensure optimal stent deployment. One cardiologist has done all stent implantations.
Medication The revascularisation has been performed on standard patient's medication with anteplatelet pre-treatment (ticlopidine) 3 days before procedure. Patients received 10 000 international units of heparin at the initiation of the procedure and activated clotting time was maintained at >300 seconds. All patients received aspirin 100 mg daily indefinitely and ticlopidine 500 mg daily for 6 months to avoid possible late occlusion.

Radioactive stents
The radioactivity has been brought by cyclotron at the Institute of nuclear physics of Czech academy of science in Řež. The main isotope of this radioactive stent is beta-emitter 55 Co with half-life time of 18 hours. Other radioisotops produced in the cyclotron except 55 Co were 56 Co, 57 96 Tc. The dose, calculated for 0.1 mm depth of tissue, was 50-60 Gy and 75 % of this dose was delivered within first 72 hours. The initial activity of the stents was measured and when the activity had decreased to about 40 µCi=1,5 MBq, the stent had been implanted. The use of these radioactive was previously described only in animal study (4). The feasibility and safety of using 55 Co radioactive stents in human has been published recently (16).

Intravascular ultrasound image acquisition analysis
After the final balloon inflation and administration of intracoronary nitrates, IVUS has been performed with a mechanical IVUS system (Clear View, Cardio Vascular Imaging System, CVIS, Boston Scientific Corp, San Jose, CA) working with a sheath-based IVUS catheter incorporating a 30 MHz single-element transducer rotating at 1800 rpm. The IVUS transducer was withdrawn through the stationer imaging sheath by automatic motorised pullback device at fixed speed 0.5mm/second to ensure a constant interval between slices allowing for accurate volumetric analysis. Ultrasound images were recorded on half-inch, high-resolution s-VHS videotape for off-line analysis. This was repeated at the 6-month follow-up.

Quantitative intravascular ultrasound analysis
The IVUS analysis was performed by one cardiologist with experience in IVUS analysis with intraindividual vari-ability from last 100 consecutive IVUS analysis 1.3±2.7 % for LV, 1.9±3.1 % for total vessel volume TVV and 2.2±3.7 % for PMV. The investigated segment was not only stent, but also the adjacent coronary segment 5 mm distal and proximal to the stent. So, the length of analysed area was 28 mm. At the stent edges, the area encompassed by the lumenintima and media-adventitia boundaries defined the luminal and the total vessel volumes, respectively. The difference between luminal and total vessel volumes defined the plaque plus volume. Within the boundaries of the stent total vessel volume, stent volume, neointimal hyperplasia, and lumen volumes were obtained. The neointimal hyperplasia presented was a value measured at follow-up (stent volumelumen volume). In our study the delineation of the total vessel volume boundary was possible in all IVUS analysed patients. All volumetric dates were calculated using the Simpson's formula: V= Σ n i=1 A i. H, where V=volume, A= area of external elastic membrane (EEM), lumen, stent or plaque in a given cross-sectional ultrasound image, H=thickness of the coronary artery slice, that was reported by this digitised cross-sectional IVUS image, and n=number of digitised cross sectional images encompassing the volume to be measured. Validity of this method has been proved previously (9).
The vessel remodelling was described as a change of total vessel volume (∆TVV) during follow-up (TVV at 6 month follow-up minus TVV after procedure), similarly, ∆LV for lumen changes (LV at 6 months minus LV after procedure), ∆PMV for plaque and media volume (PMV at 6 months minus PMV after procedure), and ∆SV for eventual stent volume changes (late recoil) (SV at 6 months minus SV after procedure). NIHV is presented as a mean of neintimal hyperplasia volumes within the stents at 6month follow-up.

Definition and segments of analysis
Stent edges were defined as those volumes axially 5 mm proximal and distal to the final stent strut, stent extremities as volumes axially 5 mm at the both edges of stents and stent body as 8 mm long middle part of stents. Restenosis was defined as an agiographic restenosis >50 % at 6-month follow-up located either at stent edge or stent itself.

Statistical analysis
Quantitative data are presented as a mean ± standard deviation. Volumetric date derived from the IVUS investigations were compared immediately after treatment and at follow-up using the two-tailed paired Student's t-test. A value of p<0.05 was considered statistically significant.

Results
Baseline clinical and procedural characteristics are described in Table 1. Table 2 describes quantitative coronary angiography data pre-and post-intervention and at 6-month follow-up. At   deaths, or stent thrombosis was seen. The angiography revealed restenosis >50 % in 5 cases (50 %). Distal edge restenosis developed in two cases, proximal edge restenosis also in two cases and in one case a "candy wrapper" restenosis was present (Fig. 1). Target lesion revascularisation was performed in three patients (30 %) with objective evidence of ischemia: Two patients with angiographically significant stenosis (>50 %) were not revascularized. One patient refused repeat procedure and second had no evidence of ischemia in spite of angiographic restenosis >50 %. One patient was revascularized due to progression of coronary artery disease with stenosis >50 % at the non-treated segment at 6-month of follow-up. So, the target vessel revascularisation was done in four patients (40 %). IVUS analysis ( Fig. 2  It means reduction of stent lumen volume by 15 % at 6-month. This was due to presence of NIH at both extremities (more but no statistically significant at distal part) of implanted stents. The ingrowth of NIH was inhibited at the body (8 mm long segment in the middle part of the stent) of stents compared to extremities (5mm long segment at each end) of the stents (7.3±5.9 mm 3 versus 28.4±26.2 mm 3 , p<0.05). Also, no chronic recoil of the implanted stents was seen in this group (181.9±80.2 and 177.0±56.2 mm 3 ). The analysis of edges (5mm distally and proximally to the last stent struts) showed no changes in TVV (187.3±62.6 mm 3 and 169.9±53.5 mm 3 ) but PMV increase significantly from 61.9±31.2 mm 3 to 82.2±43.4 mm 3 (p<0.04) and LV decreased from 125.4±40.7 mm 3 to 94.7±22.0 mm 3 (p<0.02) (Fig. 3). This late lumen volume loss was mainly (from 66 %) due to increase in PMV.

Discussion
The results of this study indicate that using single, 18 mm-long, 55 Co radioactive ß-emitting BX Velocity tm stents with high initial activity 41.1 µCi±1,2 µCi=1520 kBq±44 kBq has no beneficial effect in prevention of restenosis in spite of meticulous alertness not to injured the vessel segment proximal and distal to the stent. This no beneficial effect was mainly due to edge restenosis, which was found in five out of ten patients (50%) and due to exuberating NIH at both stent extremities, predominantly localised at the distal edges.
Edge restenosis was mainly due to plaque accommodation. In these high radioactive stents, the neointimal formation was inhibited only in the body of the stents.

Mechanism of edge restenosis
In our study, the edge restenosis with significant late lumen loss was mainly due to an increase in plaque and less due to vessel remodelling. This is with agreement of results published by Kay et al. (6), describing that for stents with activity 6-12 µCi=222-444 kBq is plaque accommodation (neointimal proliferation) the major contributor to lumen loss. On the opposite, Albiero et al. (2) concluded in their study that edge restenosis was mainly due to shrinkage of the vessel for stent with initial activity 12-21 µCi=444-777 kBq.
Since all stents in our study were implanted employing a "nonaggressive strategy" (low pressure without predilatation and high pressure postdilatation has been done with a shorter balloon) we supposed that "geographical miss" was minimize (but "some" balloon injury at the stent ends is always) and so we strongly believe that the edge effect in this trial was mainly due to fall-off in radiation at the edges of the stents.
In agreement with other studies (6), no statistically significant chronic recoil of the stent was found.

Future directions
Two different modalities have been proposed to solve the problem of edge restenosis: the hot-ends or the coldends stents. The hot-ends stents involve literally concentrating the greatest activity at the stent edges. The background for this approach was to extend the area of irradiation beyond the balloon-injured area outside the stent, thereby decreasing the chance of geographical miss (13). However, it has been proved, that subtherapeutic levels of radiation can stimulate proliferation or remodelling in uninjured vessel segments in animal model (10) and so, it might be that increasing the activity at the stent ends will only postpone the restenosis further from the stent. The cold ends stent is another modality. If the edge restenosis is result of negative remodelling (induced by low-dose radiation in an injured area), then the lengthening of the stent could be solution to prevent this negative remodelling. But, this concept was denied by Rotterdam group by publishing their results con-cerning use of cold-end radioactive stents (7). They found increased neointimal hyperplasia in in-stent non-radioactive segments (p<0.017).
According to our study or other latest data and with agreement of others (20,11) we can postulate that use of radioactive stents is safe and feasible but, at present, the problem of edge restenosis remains unsolved and so, these stents should not be clinically used.
However, further therapeutic options are coming on the horizon with promising preliminary data such as drug-eluting stents with rapamycin or paclitaxel (15,5) or biodegradable stents (17).

Conclusion
Single 55 Co radioactive β-emitting stents with high initial activity 41.1 µCi±1,2 µCi=1520 kBq±44 kBq are effective in reducing of neointimal hyperplasia only within the stent body, as measured by IVUS, and they do not solve the problem of restenosis at the stent extremities as well as at the stent edges. Edge restenosis in this high radioactive stents was mainly (from 66%) due to neointimal proliferation.