Multivariable Cox Regression Models of PRU Quintiles With Clinical Outcomes
| Hazard Ratio | 95% Confidence Interval | p Value | |
|---|---|---|---|
| Stent thrombosis, definite or probable∗ | |||
| PRU Q1, referent | 1.00 | — | — |
| PRU Q2 vs. Q1 | 1.13 | 0.51–2.48 | 0.76 |
| PRU Q3 vs. Q1 | 1.37 | 0.64–2.90 | 0.41 |
| PRU Q4 vs. Q1 | 1.71 | 0.84–3.49 | 0.14 |
| PRU Q5 vs. Q1 | 2.32 | 1.17–4.59 | 0.02 |
| Clinically relevant bleeding† | |||
| PRU Q1, referent | 1.00 | — | — |
| PRU Q2 vs. Q1 | 0.74 | 0.60–0.93 | 0.010 |
| PRU Q3 vs. Q1 | 0.71 | 0.57–0.89 | 0.003 |
| PRU Q4 vs. Q1 | 0.63 | 0.50–0.79 | <0.001 |
| PRU Q5 vs. Q1 | 0.61 | 0.47–0.77 | <0.001 |
| All-cause mortality‡ | |||
| PRU Q1, referent | 1.00 | — | — |
| PRU Q2 vs. Q1 | 0.90 | 0.58–1.40 | 0.64 |
| PRU Q3 vs. Q1 | 1.22 | 0.82–1.84 | 0.33 |
| PRU Q4 vs. Q1 | 1.17 | 0.78–1.73 | 0.45 |
| PRU Q5 vs. Q1 | 1.08 | 0.72–1.60 | 0.72 |
MI = myocardial infarction; PR = platelet reactivity; PRU = P2Y12 reaction units; Q = quintile(s).
↵∗ Model includes PR quintile, clopidogrel usage (time dependent), insulin treatment, previous MI, history of dialysis, ST-segment elevation MI, baseline white blood cell count, baseline platelet count.
↵† Model includes PR quintile, clopidogrel usage (time dependent), age, history of renal insufficiency, history of hypertension, current smoking status, number of vessels treated per patient, baseline hemoglobin, baseline white blood cell count.
↵‡ Model includes PR quintile, clopidogrel usage (time dependent), age, male sex, history of diabetes, previous MI, history of renal insufficiency, current smoking status, ST-segment elevation MI, non–ST-segment elevation MI, baseline hemoglobin, creatinine clearance, white blood cell count.