In the clinical practice of interventional cardiology, the idea has long been accepted that coronary artery disease (CAD) is chiefly (although not only) a local process, one best treated with local methods. Segmental coronary luminal stenoses, occlusions, or unstable plaques are the most favorable targets for interventions that are aimed at preventing or limiting manifestations of ischemia and myocardial loss, including arrhythmias, congestive heart failure, and death. Coronary interventional treatment has typically consisted of mechanical solutions—surgical bypass and angioplasty. Over the years, however, several systemic pharmaceutical treatments have been introduced to decrease recurrence, stabilize lesions, or cause disease regression at target sites.¹ Here, we review the evolution and current status of local drug delivery in the coronary arteries, with an emphasis on novel catheters that have retractable needles.

In the late 1980s, angioplasty with the use of bare-metal stents (BMSs) was introduced to obviate the most serious limitations of stand-alone balloon angioplasty: acute elastic recoil, plaque ulceration or dissection, subacute negative vessel remodeling, and fibrocellular proliferation. Angioplasty with BMSs enabled physicians to optimize the early results of vascular recanalization while avoiding elastic recoil and preventing dissection and early reclosure.^2,3 Nevertheless, BMS-angioplasty did not solve the biological problems that were created by producing localized coronary trauma with a rigid, metallic device: rare, but catastrophic, postoperative acute stent thrombosis, and, more frequently, delayed in-stent restenosis. Eventually, systemic platelet anti-aggregants, namely clopidogrel and aspirin, proved effective in significantly limiting early stent thrombosis.^2,3 Fibrocellular growth (mostly by means of cell proliferation and smooth-muscle-cell migration from the media level), platelet activation and aggregation, inflammation, and fibrin deposition were recognized as the mechanisms that initiate restenosis.^4,5

Subacute in-stent restenosis was not treated effectively until the advent of drug-eluting stents (DESs), which deploy drugs that inhibit fibrocellular proliferation and migration.^6,7 These stents are covered with porous polymer coatings that absorb the chosen drug, transport it to the target site, and deliver it over a specified time interval. Such stents are quite effective in repressing intimal growth; they lower the rate of restenosis (or late luminal loss) by 50% to 80%.^7–9 Nevertheless, their advent introduced or left unsolved the following limitations:

• The cost of a stent increased approximately 3-fold.^10,11 This is a serious clinical and health-policy problem: multivessel CAD has evolved into a de facto indication for angioplasty, although stent-angioplasty in multiple-vessel disease is not yet approved by the U.S. Food and Drug Administration (FDA).¹²
• The problem of restenosis was not totally resolved,¹³ especially in cases involving complex lesions, which comprise up to 70% of lesions that are treated with DESs.^14–16 (Again, the FDA has never approved the deployment of any stent for the treatment of these lesions.)
• With the use of DESs in complex lesions, the incidence of stent thrombosis increased significantly in complex lesions, both early and late (1 to 4 years) after stent deployment.^17,18
• As a result of the preceding 2 limitations, the probability of acute myocardial infarction or death after coronary-stent interventions did not lessen upon the use of DESs in comparison with BMSs.^15,16 Because DESs were designed to prevent major clinical events, this finding was sobering.⁸

• Any coupling of a drug agent with a given device generates a long, expensive approval process, because drug agents must be specifically licensed for such use. The cost is inevitably transferred to the patient or to the health-maintenance system.
• The stent's coating itself could cause biological interference. The coating may not be covered by a vital endothelium; when the coating eventually breaks down, the probability of inflammatory and thrombotic complications increases, as is evidenced by delayed fibrin deposition.^19,20
• Most DESs deliver the active drug endoluminally. This method of delivery likely inhibits endothelial migration and seeding onto the stent—a process that is thought to be essential in the prevention of stent thrombosis.^17,18
• After DES implantation, restenosis is usually focal,^7,8,13 unlike the diffuse restenosis that is commonly seen after the implantation of a BMS. This is likely caused by the specific conditions at the restenosis site:
  • Restenosis after DES implantation typically occurs at the level of severe (calcific) lesions that were not adequately dilated before stent deployment. Under these conditions, high-pressure balloon postdilation is frequently applied. This treatment may lead to compression in the coating, which could cause drug elution to occur too early and too quickly. During the healing phase after stenting, when the vessel is most vulnerable to restenosis, local drug availability may be inadequate.
  • Proliferative, diffuse CAD may require a largeramount of the inhibiting drug than is routinely carried by a DES, as in patients who have diabetes mellitus, long lesions, diffuse in-stent restenosis, or venous grafts. These patients still experience increased restenosis rates after DES use.^{8,13,15,16,19,20}
  • At the site of DES implantation, side branches are at higher risk of restenosis, probably because of the presence of an unopposed stent and the likelihood of “snow-plowing” of the compressed atherosclerotic plaque.^12,14 In this circumstance, dispensing a higher dosage of the inhibiting drug(s) than is routinely available in a DES may be beneficial.
  • The edges of a DES—particularly the proximal edge—may also be sites of increased fibrocellular proliferation and could routinely require an increased drug concentration or longer duration of administration, especially if the arterial disease is diffuse.^8,15,16
  • Manufacturing defects in the coating of an individual DES may also increase the risk of restenosis.²¹

• The guiding catheter itself (Fig. 1) may be used for acute endoluminal delivery, as in the administration of thrombolytics or IIb/IIIa platelet-receptor inhibitors. Unfortunately, this simple approach has not proved effective for preventing restenosis. Prevention seems to require prolonged administration—possibly for an entire month, as has been suggested by the results of animal experiments.²³

  Fig. 1 Computer-imaged rendition of selective administration of drugs in a coronary artery by means of a guiding catheter. Local retention of the drug is low and not site-specific.
• Different kinds of balloon catheters (Fig. 2) have been used in multiple ways (such as in leaky or porous balloons^24,25) to deliver solution-carrying active substances locally to the target site. However, the balloon-catheter approach imparts quite a low efficiency of drug delivery (the ratio of the drug delivered into the vascular wall vs the total amount delivered), and the risk of intimal or endothelial damage could be a limiting factor, especially in the presence of high-velocity jets. In addition, the treated artery becomes temporarily occluded during the delivery process.

  Fig. 2 Computer-imaged rendition of the leaky-balloon technique, which results in low-efficiency drug delivery.
• Drugs have been applied directly on balloons or BMSs by means of various techniques, in the absence of a coating, with initially promising results in slowing intimal growth in coronary and peripheral arterial disease.^26,27 In an animal experiment, approximately 90% of the drug that was administered via this method was lost or delivered in the bloodstream.²⁷
• Leaky coils²⁸ were used in the Dispatch® catheter (Boston Scientific/SCIMED, Inc.; Fremont, Calif), but delivery was deemed inefficient, and this method was soon abandoned.
• Needle-row balloons (Interventional Technologies, Inc.) (Fig. 3) featured triple metal rows of cone-shaped needles, each 0.25 mm in height. These balloons were studied in phase-2 clinical trials but did not attain clinical usage.²⁹

  Fig. 3 Computer-imaged rendition of the needle-row balloon, which enables deeper drug penetration into the vascular wall (A) but also allows drug leakage and washout after balloon deflation (B).

Fig. 4 Computer-imaged rendition of the Binlab retractable-needle catheter (local injection device), which is designed to permit reliable, deep penetration and drug retention in the vascular wall.

Fig. 5 Photomicrograph shows histologic cross-sections of vessels that were treated with local injection devices in human cadaver specimens. A) Typical remnants of needle penetration, showing the deep imprints in the exact sites of penetration in the (more ...)

Although further studies are needed, the LID approach promises to simply and efficiently solve many limitations of current DES. Chief among them are the following:

• The excessive cost (especially when multiple vessels or lesions are involved).³³
• The otherwise complex and expensive approval process. Because the LID involves neither permanent implantation nor a life-threatening procedure, its developers could submit only a 510(k) premarket notification to the FDA for clearance of the LID—a much less onerous process of approval than the usual Investigational Device Exemptions (IDE) process.
• The presence of a potentially thrombogenic stent-coating.
• Potential toxic or hypersensitivity³⁴ effects of stent-coating on the intimal and endothelial cells, which could delay re-endothelialization (LIDs would be used jointly with BMSs).
• Adaptable delivery of various quantities of different drug combinations, as suggested by the nature of various local intimal substrates (for example, thrombotic vs calcific vs soft/fibrous vs lipid-laden plaques). The use of local injection devices could improve the efficiency and precision of local administration of the desired drug(s), as decided by the operator, according to protocols that still need to be developed. In this regard, it is relevant to emphasize that the lipophilic sirolimus molecule can easily cross the cell membrane and bind to cytosolic receptors intracellularly, enhancing chronic tissue retention.³⁵
• The dependence of DES on long-term antiplatelet inhibitors, which could still be insufficient in a substantial percentage of patients.³⁶

In addition, LID technology might improve the results of even stand-alone balloon angioplasty for patients in whom stents are contraindicated (for example, in clopidogrel nonresponders or in patients who have superficial femoral artery lesions).

By enabling simple local administration of medical therapy or experimental biological substances, LIDs could also be used for the preliminary testing of new drugs and their combinations in animals or human beings, and for cellular or genetic therapy; such testing would be much more cost-effective than the testing of DES models. For instance, the Bavaria LID has been successfully used to establish an interesting experimental model of atherosclerotic plaques in pigs, by the administration of cholesteryl linoleate into the intima.³¹

The most relevant remaining question regarding the LID approach concerns the possibility of toxic effects if a drug is administered in “excessive doses” in procedures that could be difficult to standardize. In particular, the risk of delayed aneurysm formation or rupture could be a concern.³⁷

In providing local drug delivery for vascular interventions, catheter-based retractable-needle devices may offer some unique benefits.³⁸ These devices could also presage novel, cost-effective interventions for the treatment of nonocclusive plaques that are considered vulnerable to rapid progression.^32,39 Eventually, LIDs might enable plaque stabilization (for example, in an unobstructed left main trunk or a venous graft lesion) by reducing the lipid burden or the inflammatory-cell population, or by locally administering direct antithrombin or tissue-factor inhibitors. These are only some of the possible investigational and therapeutic options that might be realized with these devices.

We acknowledge the editorial support of Virginia Fairchild for helping to prepare this article; of Hasitha Samarajeewa for providing the computer renditions of Figures 1 through 4; and of Tomas Klima, MD, for supplying the histologic preparations.