The effect of drug eluting stent underdeployment on restenosis rates in patients with small coronary arteries

Jennifer E. Taylor, University of Cambridge School of Clinical Medicine, Addenbrooke's Hospital, Road, Cambridge, CB2 0SP
Zefoula Beytoula, Division of Cardiology, State University of New York, Stony Brook, New York, USA
William E. Lawson, Division of Cardiology, State University of New York, Stony Brook, New York, USA

When the use of DES is warranted, options currently available in the United States for treatment of stenotic coronary arteries < 2.5mm in diameter include the placement of an under-inflated 2.5 mm DES or deployment of a 2 or 2.25 mm bare metal stent (BMS). The significant risk of restenosis by the use of BMS makes the latter option less attractive, however clinical outcomes of target-vessel restenosis in patients receiving underdeployed drug eluting stents have not been studied sufficiently.

The medicinal component of DES deployed at sub-nominal pressures would be expected to reduce the incidence of in-stent restenosis. However, underdeployment of DES may lead to non-uniform strut distribution, which may then affect its rheological and pharmacokinetic properties. This study seeks to determine if there is an increased incidence of target-vessel restenosis in patients who received 2.5 mm DES inflated at sub-nominal pressures compared to patients who received 2.5 mm DES inflated at nominal pressure. The incidence of in-stent restenosis of DES deployed at sub-nominal pressures was found to be higher than those deployed at nominal pressures. Recurrent symptoms requiring additional percutaneous intervention occurred significantly earlier in target-vessels of those patients receiving DES at sub-nominal pressures (relative to non-target vessels) compared to patients who received 2.5 mm stents at nominal pressures (relative to non-target vessels).


Current percutaneous revascularization treatment options in the United States for atherosclerosis in patients with small coronary arteries (<2.5mm in diameter) such as diabetics include the underdeployment of 2.5 mm drug eluting stents (DES) or full inflation of 2 or 2.25 mm bare metal stents (BMS). This work investigates the incidence of restenosis that occurs with the use of underdeployed DES compared to fully deployed DES.

In-stent restenosis (ISR) rates have dropped significantly with the advent of DES compared to those observed with the use of BMS [1]. Controlled release of anti-proliferative agents such as Sirolimus and Paclitaxel from the polymeric coatings of stents showed a substantial reduction in the neointimal hyperplasia that causes late lumen loss. Paclitaxel and Sirolimus inhibit vascular smooth muscle cell proliferation through over expression of microtubules and inhibition of inflammatory cell proliferation, respectively.

Stents are typically deployed at nominal pressure or greater to ensure stent apposition, uniform dilation, and maintenance of laminar flow. Drug eluting properties as well as drug interaction with the intimal wall are highly dependent on these properties.

The currently available sizes of DES in the United States are limited to 2.5, 3.0, and 3.5 mm in diameter. Therefore, in cases where arteries are smaller than 2.5 mm in diameter the interventionist must decide whether to use an undersized 2.5 mm DES or a fully deployed BMS (available in 2 and 2.25 mm sizes). Diabetic patients in particular are at increased risk for restenosis [2].

The eluting and distribution properties of the drugs from the polymer coatings of the DES are engineered for an ideal situation where the DES is deployed at nominal pressure so the stent struts fully expand to make a perfect fit with the arterial wall. Mass transport of the drug from stent to endothelium is highly dependent on the properties of both the stenosed arterial wall and polymer film [3].

Sub-nominal pressure deployment is used to match the stent with arterial size when the stent deployed at nominal pressures would be too large for the vessel. Recent studies have shown that the ‘bigger-is-better’ approach to BMS placement in small vessels compared to large leads to increased incidence of late lumen loss [4]. This may be attributed to non-laminar blood flow, which decreases shear stress within the vessel thereby contributing to vascular cell proliferation, inflammation, and adverse remodelling of the vessel wall [5]. Furthermore, exposure of vessel wall sub-intima can lead to platelet adherence and thrombus formation [6]. In some cases, there may be a significant risk of vessel rupture or extensive dissection if nominal pressure stent deployment is achieved and the resulting stent diameter is too large for the vessel to accommodate it.



The incidence of target-vessel restenosis was examined in three hundred fifty three patients receiving 2.5 mm DES. They were divided into two groups: those who received DES deployed at nominal pressure and those who received DES deployed at sub-nominal pressure. Stents deployed at sub-nominal pressure were considered ‘under-inflated’ and corresponded to stenotic arteries and implanted DES <2.5mm in diameter, while vessels with stents deployed at nominal pressure had diameters of 2.5mm. The nominal pressures used were 11 ATM for Cypher and 9 ATM for Taxus stents as defined by the manufacturer packaging. The study included patients who received either a 2.5 mm DES at nominal or sub-nominal pressure within the time period of July 2004 through January 2006. Patients with complex lesions, such as bifurcation, ostial, segmental, chronic total occlusion, and graft lesions, were excluded from this study.

Indications and Stenting Procedure

All percutaneous revascularization treatments were performed at Stony Brook University Hospital, which is the exclusive tertiary care center in Suffolk County, New York, USA. Indications included both salvage and elective procedures. Percutaneous interventions were performed by standard techniques. Post-stent high pressure non compliant balloon dilation was not performed in sub nominal stent deployment.

A standard pharmaceutical therapy was followed in most cases. During the procedures, patients were given 325 mg aspirin and 300 mg clopidogrel. Unfractionated heparin or angiomax (bivalirudin) was routinely given during percutaneous intervention (PCI) and intracoronary nitroglycerin was delivered during the procedure for maximal vessel dilation.

Follow-Up Analysis and Endpoints

Patients were followed up for at least one year post-PCI. The endpoints of the study were characterized as either target-vessel restenosis or non-target vessel coronary disease as found in subsequent catheterization.

Statistical Analysis

Confounding factors contributing to ISR include in the analysis: age, male gender, prior revascularization, stent lengths, and Diabetes Mellitus. The frequency of target lesion repeat revascularization was examined by univariate analysis since confounding factors were found to be insignificantly different between the two groups. Student t-test and chi-square testing were used to evaluate patients with clinically indicated repeat coronary intervention (for signs, symptoms, or objective evidence of myocardial ischemia) with significance <0.05. The acuteness of coronary disease in both target- and non target-vessel restenosis was assessed by measuring the time (in days) between the initial and return PCI.


Out of 353 patients (345 patients received Cypher stents, 8 patients received Taxus stents) in the study, there were 150 patients with 2.5mm drug-eluting stent deployed at nominal pressure (11 ATM Cypher; 9 ATM Taxus) and 203 patients with 2.5 mm drug-eluting stent deployed with less than nominal pressure (8.78±1.26 ATM in Cyper patients). The baseline demographics of those patients who had stents that were deployed with nominal pressure versus sub-nominal pressure were not significantly different in: age (62±12 years vs. 60±11 years) and male gender (60% vs. 64%) as shown in table 1. Differences in confounding factors such as prior revascularization (70% vs. 61%), stent length (15.8±6.0mm vs 17.7±6.2mm), and Diabetes Mellitus (29% vs. 36%) were also found to be insignificant (table 1).

Repeat PCI was needed in 45 patients (22%) in the sub-nominal stent deployment group versus 26 (18%) patients in the nominal group over the observed period, which was not significantly different as shown in table 1. However, in patients who had repeat PCI, the target-vessel restenosis rate of those with underdeployed stents was significantly higher (46.7%) than for those in the nominal group (17.3%) as shown in table 2.

As shown in table 2, the sub-nominal and nominal groups returned after an average of 143±126 and 225+/-62 days, respectively, for target-vessel restenosis. For non target-vessel restenosis, patients returned for further PCI after 250 +/- 175 and 221 +/- 156 days for sub-nominal and nominal groups, respectively. Comparison of the number of days to return for further PCI in target verses non-target lesions of sub-nominal and nominal groups is shown in table 3. Patients with target-vessel restenosis returned significantly more quickly than those with non target-vessel lesions in the sub-nominal group. There was significant difference found comparing the days to return of patients with target-vessel and non-target vessels within the nominal group.


Confounding Factors

Confounding factors were not significantly different between the two groups. Although recent evidence suggests that Paclitaxel eluting stents are less effective at preventing restenosis [7], there was an insignificant number of patients who received Paclitaxel stents over Sirolimus DES and they were almost evenly distributed between sub nominal and nominal groups. Factors contributing to risk of ISR particularly in patients with DM include longer lesion length and smaller baseline reference vessel sizes [8]. Previous work has shown, however, that abnormal glucose tolerance as opposed to smaller vessel size is a more significant confounding factor of ISR [9]. The angiographic ISR rate for a stent 2 mm in diameter (15mm length) is predicted as 47% in non-diabetics and increases to 58% in diabetic patients (Figure 1). Post stent high pressure inflation with a 2.0 or 2.25 mm non compliant balloon could have potentially improved uniformity of expansion and stent apposition [10].

Figure 1 - Predicted in-stent angiographic restenosis rate vs. stent diameter.
Figure 1 - Predicted in-stent angiographic restenosis rate vs. stent diameter.

Clinical Outcomes

This study showed that a higher incidence of restenosis occurred in patients that received 2.5 mm DES stents deployed at sub-nominal pressures. Furthermore, patients within the sub-nominal group required target-vessel revascularization (TVR) procedures at an early time compared to those treated for disease progression in a non target-vessel. This distinction was not observed in the nominal group perhaps because the baseline vessel sizes were larger. A difference may be found after a longer observation period than the time examined in this study. In addition to disparity in baseline vessel diameter, factors related to the physical characteristic of an underdeployed stent may play a role in the overall structure of the stent and the morphology of the polymeric film from which anti-proliferation drugs are released. These factors may affect uniform expansion of the stent, the apposition of the stent to the arterial wall, and its drug release properties.

Factors of Underdeployment

Sub nominal pressure leads to a lesser gain in cross-sectional area. Given the pressures used, the cross-sectional area would be at least 10% less than those achieved by a nominal pressure stent. Previous work has shown that even at nominal pressure, stents typically deploy to as little as 60% of the manufacturer’s intended cross-sectional area [11]. Furthermore, deployment of stents at sub-nominal pressure may lead to poor stent apposition [12]. These events may cause unintended decrease in overall dose delivery. Takashima et al. found that the contact area of linked stents was increased with deployment pressure in relation to stent geometry [13]. Low contact area of stent struts may result from higher radial and longitudinal recoil. Non-linear expansion properties can result in unevenly distributed contact of the stent with the arterial wall and ‘dog-boning’ of the stent within the vessel [14]. Poor stent apposition has been found to increase the risk of thrombosis or in-stent restenosis by under-dosing of the arterial wall due to too few contact points [15]. Research has shown that the number of stent struts normalized to stent cells at the contact point is a predictor of the amount of intimal hyperplasia leading to restenosis [16].

Drug Delivery Mechanism

Previous research has shown that the drug delivery is dependent on the mass transport properties of the drug through arterial wall [17]. Polymeric coatings deployed at nominal pressure will stretch to an optimum thickness having intended delivery properties whereas the polymeric coating of underdeployed stents may be thicker and irregular. Increased thickness of the coating will result in a higher concentration of drug being delivered at fewer contact points, which may result in overdosing in these areas and under-dosing in areas of poor contact. Drug delivery properties were shown to be highly dependent on the thickness of both the arterial wall and polymeric layer [18]. Overdosing by these points of high concentration may give rise to ‘black holes’ [19] in the arterial wall which are hypocellular and contain large areas of proteoglycan as reported by intravascular ultrasound (IVUS) [20].

Limitations of the Study

The findings of this work are limited without the use of post procedural IVUS. This technique could resolve the surface morphology of stents deployed at nominal and sub-nominal pressures. This could be used to determine post stent adequacy of expansion and the degree of malapposition as a function of ‘dog-boning’.


This research finds that there is a higher incidence of restenosis in patients receiving 2.5 mm DES deployed at sub-nominal pressures compared to those patients receiving 2.5 mm DES deployed at nominal pressure. Furthermore, restenosis of target-vessels within this group occurred more rapidly compared to progression of disease in non-target lesions of nominally deployed stents. Altered diffusion properties of the drug due to geometric differences brought about by underdeployment may contribute to these observations. Other possible reasons for this difference include vessel size, non-uniform expansion, stent malapposition, drug absorption and diffusion. In view of the higher restenosis risk seen with this strategy, consideration should be given to high pressure post deployment dilation with a 2 or 2.25 mm non compliant balloon or routine IVUS. Further investigation is warranted.

Table 1-Probability values for confounding factors
Characteristic Sub nominal Nominal Ψ2 /
(n=203) (n=150)
Age, years (+/- SD) 60 +/-11 62+/-12 1.624 0.11
Stent Length, mm (+/- SD) 17.7+/-6.2 15.8+/-6.0 3.903 0.42
Male, (%) 64 60 0.440 0.51
Diabetes Mellitus, (%) 36 29 1.763 0.18
Prior Revascularization, (%) 61 70 2.631 0.10
Return PCI (%) 22 18 0.684 0.41
Table 2 – Probability factors for target-vessel restenosis and days to second percutaneous intervention
Characteristic Sub nominal Nominal Ψ2 /
(n=45) (n=26)
Target-vessel Restenosis (%) 46.7 17.3 4.639 0.03
Days to 2nd PCI, Target-vessel (+/- SD) 143+/-126 225+/-62 1.399 0.17
Days to 2nd PCI, Non-target-vessel (+/- SD) 250+/-175 221+/-156 0.591 0.56
Table 3 – Probability values for days to second percutaneous intervention
Target Vessel Non-target Vessel Probability Values
Type of Deployment mean STD n mean STD n t DF P
Nominal 225 62 5 221 156 22 0.056 25 0.542
Sub-nominal 143 126 21 251 175 24 2.345 43 0.024

1. Eisenberg MJ., et al. 2006. Review of randomized clinical trials of drug-eluting stents for the prevention of in-stent restenosis. American Journal of Cardiology, 98(3), 375-382.

2. West NE., et al. 2004. Clinical and angiographic predictors of restenosis after stent deployment in diabetic patients. Circulation, 109, 867-873.

3. Ai L., et al. 2006. A coupling model for macromolecules transport in a stenosed arterial wall. International Journal of Heat Mass Transfer, 49, 1568–1591.

4. Moussa, I., et al. 1998. Does “the bigger the better” hypothesis after coronary stenting apply in small vessels? Journal of the American College of Cardiology, 31(21001), 141A.

5. Weissman, NJ. 2003. Vascular remodelling. Do we need yet another study? Journal of the American College of Cardiology, 42(5), 811-813.

6. Schwartz, RS. 1998. Pathophysiology of restenosis: interaction of thrombosis, hyperplasia, and/or remodelling. American Journal of Cardiology, 81(7), 14E–17E.

7. Mahilli, J., et al. 2006. Randomized trial of paclitaxel- and sirolimus-eluting stents in small coronary vessels. European Heart Journal, 27(3), 260-266.

8. Radke, PW., et al. 2006. Comparison of coronary restenosis rates in matched patients with versus without diabetes mellitus. American Journal of Cardiology, 98(9), 1218-1222.

9. Otsuka, Y., et al. 2000. Abnormal glucose tolerance, not small vessel diameter, is a determinant of long-term prognosis in patients treated with balloon coronary angioplasty. European Heart Journal, 21(21), 1790-1796.

10. Romagnoli, E., et al. 2008. Drug-Eluting Stenting. The Case for Post-Dilation. Journal of the American College of Cardiology, 1, 22-31.

11. Takano, Y., et al. 2001. Optimizing stent expansion with new stent delivery systems. Journal of the American College of Cardiology, 38, 1622-1627.

12. Singh, H., et al. 2005. Mycotic aneurysm of left anterior descending artery after sirolimus-eluting stent implantation a case report. Catheterization and Cardiovascular Intervention, 65, 282–285.

13. Takashima, K., et al. 2007. Simulation and experimental observation of contact conditions between stents and artery models. Medical Engineering and Physics, 29, 326-335.

14. Migliavacca, F., et al. 2002. Mechanical behavior of coronary stents investigated through the finite element method. Journal of Biomechanics, 35, 803–811.

15. Fujii, K., et al. 2005. Stent underexpansion and residual reference segment stenosis are related to stent thrombosis after sirolimus-eluting stent implantation. Journal of the American College of Cardiology; 45, 995–998.

16. Delfour, MC., et al. 2005. Modelling and design of coated stents to optimize the effect of the dose. SIAM Journal of Applied Mathematics, 65, 858–881.

17. Yang, N., et al. 2006. Modelling of low-density lipoprotein (LDL) transport in the artery – effect of hypertension. International Journal of Heat and Mass Transfer, 49, 850–867.

18. Pontrelli, G., et al. 2007. Mass diffusion through two-layer porous media: an application to the drug-eluting stent. International Journal of Heat and Mass Transfer; 50, 3658-3669.

19. Kay, IP., et al. 2003. The black hole echolucent tissue observed following intracoronary radiation. International Journal of Cardiovascular Intervention, 5, 137–142.

20. Costa, MA., et al. 2006. Intravascular ultrasound characterization of the “black hole” phenomenon after drug-eluting stent implantation. American Journal of Cardiology, 97, 203-206.