Endothelial dysfunction and drug-eluting stents

Background

Percutaneous coronary intervention (PCI) with stent implantation is effective treatment in patients with acute coronary syndromes as and stable CAD patients with significant myocardial ischemia.

The major limitation of stent implantation effectiveness has been, until the introduction of drug-eluting stents (DES), in-stent restenosis (ISR). DES use however, has led to a significant reduction of ISR, however it was found that DES can cause delayed endothelial healing and endothelial dysfunction in the stented vessels (1).

This issue has raised safety concerns because delayed endothelialisation of the stent struts can contribute to in-stent thrombosis in certain patients. The impaired vasomotor endothelial function after DES implantation has recently been described, however the clinical significance of this finding needs to be further evaluated. In addition, the degree of endothelial healing and vasomotor dysfunction seems to depend on whether the old or new generation of DES is used (2) (3).

1 - Assessment Methods

In assessing vasomotor function, most investigators have used intracoronary infusion of acetylcholine (Ach) followed by bolus injection of nitrate and angiographic assessment of the artery segments adjacent to the implanted stent. Standard protocol involves withdrawal of antiangial medication for 1-3 days, and infusion of incremental doses of Ach through the coronary catheter or selective infusion catheter (20 - 100 µg) with 2.5 to 5-min intervals between doses. After reaching the maximum dose, intracoronary bolus of nitroglycerin is administered.

Acetylcholine infusion leads to dilatation of normal coronary arteries with preserved endothelial integrity, mainly by nitric oxide release (NO). In the setting of endothelial injury Ach induces paradoxical vasoconstriction (4).
Some protocols mandate the infusion of increasing doses of sodium nitropruside (10 µg/min) followed by a bolus of nitroglycerin (5). Angiographic images are recorded after each dose of Ach, and finally < 2 min after the bolus of nitroglycerin. Eventually the diameter of the target segment in diastole is measured in two orthogonal views and averaged. Usually segments located within 5 mm proximally and distally of the stent edges are assessed by this method (6) (4).

2 - Mechanisms

There are several mechanisms which may contribute to endothelial dysfunction after implantation of DES (2):

Reduced bioavailability of vasorelaxants (nitric oxide, NO)

• Relative increase of vasoconstictive factor
• Inflammatory response
• Impaired maturation of endothelial progenitor cells (EPCs) to endothelial cells (EC)
• Impaired mobilisation of EPCs and endothelialisation of the stent struts
• Direct toxicity of the drug eluted from stent
• Polymer toxicity
• Hypersensitivity
• Increased production of free radicals

Proper reendothelialisation of the stented segment is a prerequisite for the formation of functionally mature endothelium which serves as a source of vasoactive, antiinflammatory and antithrombotic substances. Several studies demonstrated that circulating bone marrow-derived EPCs contribute to the repair of the endothelium after injury, most likely by repopulating the site of stent implantation. Stent expansion causes significant injury to the artery wall.

The endothelial lining is disrupted and the local inflammatory response is activated. The vessel wall injury also leads to the recruitment of circulating monocytes, EPCs and platelet deposition on the site of endothelial disruption. The ultimate goal of this reaction is to restitute the endothelial integrity over the area of dilated segment (reendothelialisation). The time between the initial endothelial injury caused by stent struts and full endothelialisation is significantly shorter after implantation of bare metal stents (approximately 30 days) than DES (> 6 months) (7). The reparatory mechanism is in part dependent on the recruitment of circulating EPCs, with their subsequent adherence to the struts surface as well as areas of arterial wall located between the struts (8). This mechanism is altered after implantation of drug-eluting stents (DES). For example, mobilisation of EPC is reduced after implantation of paclitaxel (PES) and sirolimus-eluting (SES) stents (Wojakowski et al. AHA 2009, Circulation. 2009;120:S967-S968.).

Data from the histopathology studies and in vivo angioscopy show that part of the stent struts are not covered by endothelial cells for as long as 6 months after PCI. The effect is caused by direct inhibitory effects of drug released from the stent. Absence of functional EC on a substantial part of the stent struts could be at least in part responsible for reduction of vasoactive substances secreted normally by the endothelium, such as NO. This effect can be augmented following implantation of long stents or several stents in the same vessel (2). The vasomotor dysfunction can also be attributed to the direct effects of the drug released from the stent.

Sirolimus was shown to reduce the production of angiogenic vascular endothelial growth factor (VEGF) and responsiveness of the endothelial cells to VEGF (5). Obata et al. showed that in patients with acute MI sirolimus 2 weeks after implantation SES can be detected in arterial blood and coronary sinus (5). Importantly, several studies showed that segments proximal to the stent show abnormal endothelial-dependent vasoreactivity (9). The presence of endothelial dysfunction in non-stented segments of the vessel can also be explained by the propagation of drug through vasa vasorum. Therefore the endothelial healing and functional recovery can also be suppressed in segments located proximally and distally to the stent (2). Long term influence of sirolimus and paclitaxel cannot be explained solely based on their release kinetics, because most of sirolimus (80%) is eluted from the stent surface within 28 days after implantation. Additionally, PES platforms show biphasic kinetics of drug elution, with rapid release in the first 48 h after implantation and subsequent slow release over 14 days (10).

3 - Data from animal models

Several experimental studies in animal models shed some light in patomechanism of the endothelial dysfunction after implantation of DES. Studies carried out in pigs have consistently shown that PES and SES cause impaired endothelium-dependent vasodilation in proximal and distal segments of the stented vessel one month after implantation. In all studies the endothelium-independent vasoreactivity was preserved (11). Shiroto et al. showed that paclitaxel activated the expression and activity of Rho-kinase, which is involved in the coronary vasospasm in vascular smooth muscle cells. The use of a Rho-kinase antagonist abolished the endothelium-dependent vasoconstriction after implantation of SES and PES in porcine coronary arteries (12).

Pendyala et al. showed that function of not only the conduit, but also the resistance vessels is negatively affected by implantation of PES. This effect coexists with local inflammatory reaction, increased production of reactive oxygen species and increased sensitivity to vasoconstrictors such as endothelin-1 (11). Murine and rat studies have shown decreased activity of eNOS and increased production of free radicals after implantation of SES (13). A hypersensitivity reaction to the polymer probably could contribute to the endothelial dysfunction (2).

4 - Data from clinical studies

Impaired endothelium-dependent vasomotor function after stent implantation in patients was described 10 years ago for the first time, however the finding that this phenomenon is more pronounced in patients receiving DES was published by Hofma et al. in 2005. They showed that 6 months after implantation of SES Ach infusion led to paradoxical vasoconstriction distally to the stented segment. Endothelium independent vasomotor function was not compromised. Implantation of BMS was not associated with such impairment of vasomotor function (4).  Kim et al. measured the endothelium-dependent vasoreactivity in 75 patients after DES  implantation (SES and PES). Six months after PCI both types of DES produced abnormal vasoconstriction in segments distal to the stent after Ach infusion (9). Vessel occlusion as well as reperfusion injury in acute myocardial infarction (MI) create the milieu which induce profound endothelial dysfunction and abnormal response to vasoactive agents.

DES have become widely used in patients with acute MI. Obata et al recently showed that implantation of SES leads to worsening of endothelial vasomotor dysfunction, as well as a lesser increase in coronary blood flow after infusion of Ach into previously recanalised infarct-related artery. The endothelial dysfunction was shown 2 weeks after primary PCI (5).

5 - Differences between stents

Recent data from clinical trials using several types of DES suggest that significant endothelial dysfunction appears limited to first generation DES. Comparison of studies in which endothelium-dependent vasorelaxation was assessed using intracoronary infusion of Ach showed that implantation of PES and SES led to significant impairment of this reaction. On the other hand newer DES [zotarolimus eluting (ZES), everolimus-eluting (EES) and biolomus-eluting (BES) stents] do not produce the endothelial dysfunction. Such a difference can be attributed to use of biocompatible polymers or polymer free platforms and less toxic lipophilic drugs. Also the kinetics of drug elution and abluminal strut coverage leading to reduced systemic levels of drug might play an important role in the preservation of endothelial function. Several clinical studies have shown that endothelial dysfunction was absent after 6-9 months following the implantation of ZES (2) (14-15). In vitro studies have also shown a higher expression of eNOS in endothelial cells (2) (15). Lastly, the endothelialisation of the stent struts after implantation of EES is more complete in comparison to ZES, PES and SES (16).    

6 - Clinical relevance

Abnormal vasomotor response is a sign of endothelial dysfunction and a prognostic factor for adverse cardiac events (17). The consequences of endothelial dysfunction might go beyond the induction of coronary vasospasm because the endothelium secretes anti-thrombotic substances, such as prostacyclin or tissue plasminogen activators. Impaired endothelial secretory function might produce a thrombogenic milieu. So far no data has clearly shown the association between the DES-induced endothelial dysfunction and clinical outcome including the risk of in-stent thrombosis. This issue clearly needs verification in prospective clinical trials.