Skip to main content
. Author manuscript; available in PMC: 2017 Feb 1.
Published in final edited form as: Circ Cardiovasc Interv. 2016 Feb;9(2):e002730. doi: 10.1161/CIRCINTERVENTIONS.115.002730

Table 5.

Predictors of stent thrombosis: Generalized Linear mixed Model (GLMM) estimates

Variable OR [95% CI] P-value
Univariate
Age > 63 years 1.02 [0.46-2.24] 0.965
Ankle brachial index > 0.66 1.13 [0.50-2.54] 0.772
Diabetes mellitus 1.22 [0.55-2.70] 0.615
Hypertension 1.85 [0.53-6.47] 0.333
Hyperlipidemia 1.12 [0.33-3.86] 0.856
Smoking 1.37 [0.58-3.22] 0.469
Critical limb ischemia 1.32 [0.54-3.24] 0.536
Coronary artery disease 1.55[0.61-3.96 0.352
Prior myocardial infarction 1.70 [0.76-3.80] 0.191
Congestive heart failure 1.27 [0.50-3.27] 0.611
Chronic kidney disease 0.43 [0.10-1.88] 0.260
Ostial SFA 2.33 [0.95-5.76] 0.065
Distal SFA 0.44 [0.13-1.50] 0.186
Popliteal 1.06 [0.24-4.66] 0.942
Lesion length >133 mm 2.22 [0.99-5.01] 0.054
Chronic total occlusion 4.33 [1.27-14.75] 0.019
In stent restenosis 5.07 [1.91-13.48] 0.001
Stent type
    Bare metal stents 1.00 ---
    Zilver PTX™ drug-coated stents 1.48 [0.41-5.30] 0.545
    Viabahn™ stent graft 3.72 [1.28-10.81] 0.006
    Supera™ mimetic stent 1.94 [0.54-7.02] 0.307
Stent length >180 mm 2.01 [0.91-4.41] 0.082
Multivariate
Prior myocardial infarction 1.85 [0.80-4.27] 0.150
Chronic total occlusion 3.46 [0.98-12.20] 0.054
In stent restenosis 5.30 [1.83-15.32] 0.002
Stent type
    Bare metal stents 1.00 ---
    Zilver PTX™ drug-coated stents 1.15 [0.31-4.23] 0.710
    Viabahn™ stent graft 2.57 [0.82-5.05] 0.105
    Supera™ mimetic stent 2.25 [0.60-8.37] 0.224
Lesion length >130 mm 1.86 [0.80-4.31] 0.149

CI – confidence interval, OR – odds ratios, SFA – superficial femoral artery, OR – Odds Ratios, CI- Confidence Intervals

*For Dual antiplatelet therapy (DAPT) variable, while convergence criterion is met but a solution is questionable because a covariance (G) matrix is not positive.

Stepwise logistic regression used to allow a variable into the model with a significance level of 0.30 and 0.35 for a variable to stay in the model to select GLMM to predict ST. The final selected predictors were prior myocardial infarction, chronic total occlusion, in stent restenosis, stent types, and lesion length >130 mm, simultaneously. The Hosmer and Lemeshow goodness-of-fit test for the final selected model was χ2=8.638 (degree of freedom=7) p=0.2796. The model's fit statistics AIC=216.36.