Simple model predicts risk of stent thrombosis, but is it necessary?

It is possible to predict the risk that a patient will develop stent thrombosis (ST) after PCI for treatment of acute coronary syndrome using an assessment tool that considers several readily available variables, according to a study published in the November issue of the Journal of the American College of Cardiology: Cardiovascular Interventions.

ST is a serious complication of stent implantation that occurs rarely but with often devastating consequences. George D. Dangas, MD, of Mount Sinai Medical Center in New York and colleagues set out to develop a model that could predict which patients may be at high risk of developing ST. They conducted a patient-level analysis of data from the HORIZONS-AMI (Harmonizing Outcomes With Revascularization and Stents in Acute MI) and ACUITY (Acute Catheterization and Urgent Intervention Triage Trial) studies to identify significant factors that point to increased risk of developing ST.

From a cohort of 6,139 patients undergoing PCI, Dangas et al randomly divided them 2:1 into a development group and a validation group. They used Cox regression methods to identify the following variables that were independent predictors of ST and assigned hazard ratios (HR):

  • Type of ACS (ST-segment elevation myocardial infarction, non-ST segment elevation ACS with ST deviation, or non-ST segment elevation without ST changes) ( HR 1.39);
  • Current smoking (HR 1.64);
  • Insulin-dependent diabetes mellitus (HR 2.41);
  • Prior PCI (HR 1.71);
  • Baseline platelet count (HR 1);
  • Pre-PCI anticoagulant treatment (HR 0.63);
  • Aneurysmal/ulcerated lesion (HR 2.18);
  • Baseline Thrombolysis in Myocardial Infarction (TIMI) flow grade 0/1 (HR 2.36);
  • Final TIMi flow grade 3 (HR 0.49); and
  • Number of treated vessels (HR 1.75).

Using this method, the researchers were able to assign an integer score ranging from 0 to 4 to each of the variables. A total score of 1 through 6 was considered low risk; a score of 7 to 9 represented moderate risk, and patients with scores of 10 to 12 were at high risk of developing ST within a year of stent implantation. Applying the risk scoring protocol to the development group, the researchers found that the score was an accurate predictor of ST.

“We documented highly statistically significant incremental ST rates with increasing risk score integer values in both the development and validation cohorts. These trends paralleled each other in the derivation and development cohorts,” the authors wrote.

The authors pointed out that they did not incorporate all known ST risk factors into the scoring system, but they asserted that those factors they omitted—such as results of intravascular ultrasound imaging, genotyping and thienopyridine resistance or non-compliance—are rarely available or considered prior to stenting.

As study limitations, the authors cited the absence of data on deployment pressure or CYP2C19 alleles in the HORIZONS-AMI and ACUITY studies, and the fact that very few patients in these studies received thrombus extraction before STEMI angioplasty. They suggested that further analysis of the scoring tool’s utility with other study populations is desirable.

But an accompanying editorial by Ron Waksman, MD, and Israel Barbash, MD, from MedStar Washington Hospital Center in Washington, D.C., urged caution. Although Waksman and Barbash congratulated the researchers on the simplicity of the scoring system, they noted that the variables the score considers are known to have an impact on the risk of developing ST, and asked, “Is there a need for another scoring system to predict what is already known?” 

The editorialists questioned whether participants in the HORIZONS-AMI (comprised of high-risk patients) and ACUITY trials (comprised of lower-risk patients) were an appropriate population for the development of this tool, because of their different risk profiles. They also criticized the fact that the score purports to predict ST risk for a year post-PCI.

“Because the pathophysiological development of neointimal formation is time dependent, the validity of a risk score that is formulated on acute, subacute and late ST patients is grossly confounded,” Waksman and Barbash wrote.

Finally, they noted that study participants received first generation drug-eluting stents, and given the improved performance of second-generation drug-eluting stents, the editorialists suggested that the risk score’s relevance may be in question.

Waksman and Barbach conceded that predictive models can be useful, but cautioned that users must be thoroughly cognizant of any given tool’s limitations and must exercise best clinical judgment in accordance with each patient’s individual circumstances.