Review of the TAXUS-IV trial

A Polymer-Based, Paclitaxel-Eluting Stent in Patients with Coronary Artery Disease

N Engl J Med 2004;350:221-31

Background: For the past year we have been posting reviews of seminal trials in cardiovascular medicine. It is our anticipation that these will ultimately be published in a textbook format that will be indexed by major subject headings and the reviews will be presented in chronological order. However, in curating postings for Substack we have had to jump around in order to maintain some consistency in the topics being presented.

We started this year by reviewing medical therapies for patients with acute coronary syndrome. After that we moved to the management of patients with mostly stable coronary artery disease and have completed reviews on trials involving CABG and percutaneous coronary interventions compared to medical therapy and to each other, in the case of patients with left main and multivessel disease.

In completing that stream of trials, we intentionally skipped trials that have been instrumental in developing those techniques, especially coronary stenting. While perhaps not that important to general readers, medical trainees, especially in the field of cardiovascular medicine need to be familiar with these trials. They are not intended to address questions involving stenting versus medical care but instead, to address the question of “If you’re going to stent, is it better to use product A or B?”

In that vein, we recently reviewed the RAVEL trial that compared sirolimus-eluting stents to bare metal stents. This is an issue of interest because a common problem following PCI is restenosis of the treated area due to the process of neointimal proliferation that involves the migration and proliferation of smooth muscle cells from the injured arterial wall. The idea behind drug coated stents was that the drug coatings would reduce this process locally (at the level of stented arterial wall) by blocking the process of neointimal proliferation and hyperplasia, which is not the same as atherosclerosis.

RAVEL was a small trial showing that a sirolimus-eluting stent improved the surrogate endpoint of in-stent luminal loss at 6 months compared to a bare metal stent. The TAXUS-IV trial was undertaken for similar purposes but on a larger scale and sought to test a more clinically relevant endpoint. It sought to test the hypothesis that a paclitaxel-eluting stent would reduce ischemia-driven target-vessel revascularization compared to a bare metal stent.

*Note to learners: A common parlance for describing stents in the clinical setting is to refer to them based on generation (e.g., first, second, or third generation). First generation stents are bare metal stents. Second and third generation stents are drug eluting stents with newer generations often featuring improved biocompatibility and drug delivery mechanisms. Sirolimus- and paclitaxel-eluting stents are considered second generation stents.

Patients: Patients had to be 18 years of age or older, have stable or unstable angina or provokable ischemia, and were undergoing PCI for a single, previously untreated lesion in a native coronary artery. Angiographic inclusion required a single target lesion with a reference-vessel diameter on visual examination of 2.5 to 3.75 mm and a lesion length of 10 to 28 mm that could be covered by a single study stent.

There were many exclusion criteria that can be summarized as follows: acute MI, complex coronary disease (including left main, ostial target lesion or bifurcating target lesion), complex patient and predisposition to bleeding.

Baseline characteristics: The average age of patients was 62 years and 72% were men. Approximately 30% of patients had diabetes with nearly a quarter requiring insulin. Over 20% were smokers and 30% had a previous MI. The average LV EF was 55%.

The target lesion was located in the LAD in 40%, the circumflex in close to 30% and the right coronary artery in 30%. The reference-vessel diameter was >/=3.0 mm in over 75% of patients. The average lesion length was 13 mm, average reference-vessel diameter was 2.75 mm, average minimal luminal diameter was 0.92 mm, and average % stenosis was 66%.

Procedures: Patients were assigned in equal proportions in a double-blind fashion to treatment with either the paclitaxel-eluting stent or a visually indistinguishable bare-metal stent. Unfractionated heparin was administered according to standard practice, and the use of glycoprotein IIb/IIIa inhibitors was at the operator’s discretion. After mandatory balloon dilation, patients received an appropriately-sized stent. A postprocedural electrocardiogram was obtained, and cardiac enzymes were measured every 8 hours for 24 hours. Patients took 325 mg of aspirin daily indefinitely and 75 mg of clopidogrel daily for 6 months. Clinical follow-up was scheduled at 1, 4 and 9 months and yearly thereafter for 5 years.

Endpoints: The primary end point was the 9 month incidence of ischemia-driven target-vessel revascularization. It was considered to be “ischemia driven” if the stenosis of the target vessel was at least 50% of the luminal diameter on the basis of quantitative analysis with either: 1) ECG changes while the patient was at rest or 2) a functional study indicating ischemia in the distribution of the target vessel. It was also considered “ischemia driven” if there was a 70% stenosis in conjunction with recurrent symptoms alone.

*It should be noted that in this case “ischemia driven” does not necessarily mean symptom driven.

Major adverse cardiac events were defined as death from cardiac causes, MI, or ischemia-driven target-vessel revascularization. Target-vessel failure was defined as death, MI or ischemia-driven target vessel revascularization related to the target vessel.

Analysis was based on the intention-to-treat principle. A total of 1172 patients were needed to detect a 40% relative reduction (6% absolute reduction) in the primary endpoint based on an anticipated event rate of 15% in the bare-metal stent group. This sample size would have 85% power with an alpha level of 0.05 to detect the difference described above while allowing for a drop out rate of 10%.

Results: A total of 1,326 patients were enrolled over a 3 month period from 73 US centers and 1,314 were included in the final analysis with 662 in the paclitaxel-eluting stent group and 652 in the bare-metal stent group. The initial angiographic results were similar in the 2 groups.

At 9 months, paclitaxel-eluting stents reduced the primary endpoint of ischemia-driven target vessel revascularization by 61% (4.7% vs 12.0%; RR 0.39; 95% CI 0.26-0.59). There were no differences in death from cardiac causes (1.4% vs 1.1%), MI (3.5% vs 3.7%), or stent thrombosis (0.6% vs 0.8%).

In a prespecified subset of patients who underwent coronary angiography at 9 months, paclitaxel-eluting stents were associated with better angiographic features compared to bare-metal stents.

Conclusions: In patients with stable and unstable angina (not acute MI), paclitaxel-eluting stents significantly reduced ischemia-driven target vessel revascularization at 9 months of follow-up with a number needed to treat of approximately 14 patients. There were no differences in any hard endpoints.

While some would hail this as a remarkably positive trial we have reservations. Firstly, the primary endpoint is not symptom-driven and it should be regarded as a surrogate endpoint. At the time this trial was undertaken it was routine practice for patients to undergo surveillance testing with ECG’s and functional tests following coronary revascularization and this is likely how the majority of patients came to undergo revascularization (not via a symptom-driven route). Had these elective revascularizations not occurred, it is unknown whether it would have resulted in any deleterious consequences.

Second, patients enrolled in the trial were highly selected and no information is provided in the main manuscript on how many underwent screening. Commonly-occurring angiographic features of coronary lesions, for which patients undergo PCI, were excluded (i.e., ostial and bifurcating lesions, lesions in vessels with reference vessel diameters <2.5 mm, etc.).

Third, there were no differences in hard outcomes like death from cardiac causes, MI or stent thrombosis but the trial was not powered for these events. It cannot be determined from this study whether target-vessel revascularization even improved symptoms or any other proxy measure of quality of life.

In conclusion, the TAXUS-IV trial demonstrated that paclitaxel-eluting stents reduced the surrogate endpoint of target lesion revascularization at 9 months compared to bare-metal stents in highly selected patients with stable coronary disease. Whether this translates into improvements in hard endpoints or even patient-centered endpoints cannot be answered by this study.