Stent-Related Adverse Events Continue Out to 5 Years: Meta-analysis

Ischemic adverse events related to coronary stenting continue to accrue beyond the first year regardless of the type of stent implanted, and there does not appear to be a point at which that risk levels off, according to a large, patient-level meta-analysis.

Even with contemporary DES, such as the second-generation devices that replaced the sirolimus-eluting Cypher stent (Cordis/Johnson & Johnson) and paclitaxel-eluting Taxus stent (Boston Scientific), very-late stent-related events, including ischemia-driven TLR and stent thrombosis, occur at a rate of roughly 2% per year.

“While we can’t extrapolate our 5-year results to 20 years, one would think these risks persist beyond 5, 10, or 15 years for the patients, many of whom are undergoing revascularization in their 50s and 60s,” lead investigator Mahesh Madhavan, MD (NewYork-Presbyterian Hospital/Columbia University Irving Medical Center, New York, NY), told TCTMD. “For the individual patient, these risks are certainly real, and finding strategies to mitigate this risk over time is crucially important.”

Robert Byrne, MD (Mater Private Hospital, Dublin, Ireland), who wasn’t involved in the study, agreed.

“A large proportion of the patients we treat are middle-aged and have a life expectancy that goes out beyond 12 months or the 2 or 3 years we do in clinical trials,” he said. “These 5- and 10-year follow-up data give us a better picture of the entire life cycle of the devices we implant. These studies are important, and we should be doing them as much as possible even though it’s challenging to reach out and maintain contact.”

For the individual patient, these risks are certainly real, and finding strategies to mitigate this risk over time is crucially important. Mahesh Madhavan  

In the new analysis, which was published online ahead of the February 18, 2020, issue of the Journal of the American College of Cardiology, the researchers identified 19 clinical trials, among them RAVEL, HORIZONS-AMI, PLATINUM, and several studies each from the SIRIUS, TAXUS, ENDEAVOR, SPIRIT, COMPARE, and TWENTE series. In addition to Cypher and Taxus, which were classified as first-generation DES (DES1), the second-generation DES (DES2) studied in the trials included everolimus-eluting stents (Xience; Abbott Vascular and Promus/Promus Element; Boston Scientific), fast- and slow-release zotarolimus-eluting stents (Endeavor; Medtronic and Resolute; Medtronic, respectively), and the biolimus-eluting stent with a bioabsorbable-polymer coating (Nobori; Terumo).

Madhavan said it’s well documented that newer DES are more effective and safer than the Cypher and Taxus stents, as well as BMS. However, the trials tend not to focus on the risk to the patient after that first year of revascularization. “It’s an understudied question,” he said. “Potentially, if we could identify predictors [of adverse events] or aspects of care that could be refined in terms of patient management, we could target those so that the rate of late events could be improved.”

Linear Increase Without Plateau

In total, 25,032 patients were included in the meta-analysis, among them 3,718 received BMS, 7,934 DES1, and 13,380 DES2. In the DES2 group, more than 60% were treated with an everolimus-eluting stent, 24.7% with a zotarolimus-eluting stent, and 13.4% with the biolimus-eluting stent. Not surprisingly, the 1-year adjusted risk of MACE (cardiac death, MI, or ischemia-driven TLR) declined progressively after treatment with BMS, DES1, and DES2 (17.9%, 8.2%, and 5.1%, respectively; P < 0.0001).

The median duration of follow-up was 4.1 years. Between 1 and 5 years, the rate of MACE and target lesion failure occurred in 9.4% and 8.2% of patients, respectively. The event rates “increased linearly with all stent types with no plateau evident at 5 years,” the investigators report. Very-late MACE occurred in 11.0% of patients treated with DES1, 9.7% of patients treated with BMS, and 8.3% of those who received DES2 (P < 0.0001). For target lesion failure, the event rate between 1 and 5 years was 9.5% with DES1, 7.7% with BMS, and 7.7% with DES2 (P = 0.006).

“After the first year, all stent types are associated with adverse events,” said Madhavan. “It’s about 2% to 3% per year regardless of the stent type. Interestingly, bare-metal stents and second-generation drug-eluting stents were the most similar in terms of event rates. It was actually the first-generation stents, by a statistical comparison, that fared the worst between 1 and 5 years.” 

Overall, the very-late MACE rate was higher with DES1 compared with DES2 (RR 1.33; 95% CI 1.19-1.47), but the MACE rate was not significantly different between DES1 and BMS. Rates of target lesion failure were higher with DES1 compared with BMS (RR 1.27; 95% CI 1.01-1.56) and DES2 (RR 1.23; 95% CI 1.08-1.40). In the network meta-analysis, which allowed researchers to make broader comparisons, DES2 was associated with a significantly lower risk of very-late MACE compared with DES1, as well as a lower risk of target lesion failure, MI, and stent thrombosis. Very-late rates of adverse events, including MI and stent thrombosis, were lower with BMS than with DES1.

In a multivariate analysis, current smoking and the presence of diabetes were predictors of MACE between 1 and 5 years, as was a previous CABG surgery or PCI. There was no difference in risk among patients who presented with ACS or stable ischemic disease.

To TCTMD, Byrne pointed out that 10-year data from ISAR-TEST 4 and ISAR-TEST 5 have also shown the accrual of stent-related events over time. Even between 5 and 10 years, stent-related adverse events occurred at roughly 1% to 1.5% per year, he said. Long-term data from the SPIRIT trials also showed similar results.

The truth of the matter is that, based on this analysis, we’ve made significant progress within the first year. Sripal Bangalore

“It’s a wake-up call to know that these patients continue to accrue events and we really need to intensify our efforts in terms of general secondary prevention,” said Byrne. “It’s a reminder to keep your foot on the pedal with risk factors, even as you get beyond 5 years or more from the stent procedure.”      

For Sripal Bangalore, MD (NYU Langone Medical Center, New York, NY), who also wasn’t involved in the study, the results from this large meta-analysis reinforce the message that improvements in procedural technique and stent design affect event rates within the first year, but beyond that the risk of MACE is largely the same between BMS, DES1, and DES2.

“Many of us have been saying the stents are fantastic compared with older-generation stents and we’ve made significant progress in reducing hard endpoints, but the truth of the matter is that, based on this analysis, we’ve made significant progress within the first year,” said Bangalore. While the absolute event rates are lower with DES2, “the event curves are parallel” between 1 and 5 years, which raises the question whether more can be done to lower reduce the rate of very-late events.

Both Bangalore and Byrne noted that the present study did not include the newest-generation ultrathin-strut stent with bioabsorbable polymer. Even the Nobori stent with the bioresorbable polymer has thick struts and is not representative of the latest technology, said Byrne.

“It reinforces the need for continued innovation and potentially for looking [again] at the stent platform, such as the thinner ones that people are working on,” added Bangalore.

Outside of the interventional realm, Madhavan said there are several developments in recent years that could help improve clinical outcomes in patients who received a coronary stent. For example, novel cholesterol-lowering drugs like the PCSK9 inhibitors, better antithrombotic strategies, and agents targeting inflammation could help.

“Over time, you could hypothesize that some combination could translate to better outcomes,” he said.

Devices and Pharmacotherapy

In an editorial, Manel Sabaté, MD (Hospital Clínic de Barcelona, Spain), and Michael Mack, MD (Baylor Scott & White Health, Dallas, TX), agree that thinner stent struts could reduce the risk very-late adverse events, as could excellent implantation techniques that factor in patient anatomy—including features like calcified plaques, diffuse disease, thrombotic lesions, and bifurcations—and clinical characteristics.

They point out, however, that only 40% of clinical events between 1 and 5 years in the present analysis are related to the stent. As such, “further device improvements would therefore be expected to have minimal incremental benefit to overall patient outcomes,” write Sabaté and Mack. For the prevention of very-late MACE not associated with the target lesion, they recommended aggressive pharmacotherapy, highlighting several drug classes, including PCSK9 inhibitors, anti-inflammatory agents, sodium glucose co-transporter 2 inhibitors, and antiplatelet therapies, that have been shown to lower the risk in selected patients.

“Therapeutic approaches combining devices and techniques with pharmacological agents may have synergistic benefits for the patient as a whole,” they write. “It is about time to combine efforts among medical devices, implantation techniques, and adjunctive pharmacological agents to ensure the best for our patients.”