Table 2.
Statin therapy and rVTE.
| Number of events/Number of patients | Analysis | Crude subhazard ratio (95%-CI) | p-value | Adjusted subhazard ratio (95%-CI) | p-value |
|---|---|---|---|---|---|
| Traditional statistical analysis approach | |||||
| 120/980 | Overall | 0.62 (0.39 to 1.00) | 0.049 | 0.72 (0.44 to 1.19)a | 0.197 |
| 31/972 | Periods with AC | 1.16 (0.54 to 2.49) | 0.707 | 1.34 (0.54 to 3.35)a | 0.525 |
| 89/595 | Periods without AC | 0.52 (0.28 to 0.95) | 0.034 | 0.50 (0.27 to 0.92)a | 0.027 |
| Sensitivity analysis including same variables as the propensity score approach | 0.60 (0.36 to 1.01)b | 0.056 | |||
| Including markers of inflammation | |||||
| 120/980 | Fibrinogen | 0.75 (0.45 to 1.23)c | 0.250 | ||
| 120/980 | ultrasensitive CRP | 0.72 (0.43 to 1.19)c | 0.197 | ||
| Propensity score weighted approach d | |||||
| 99/792 | Overall | 0.60 (0.34 to 1.03) | 0.064 | 0.42 (0.21 to 0.81)e | 0.010 |
| 28/788 | Periods with AC | 1.34 (0.60 to 3.01) | 0.481 | 1.17 (0.47 to 2.90)e | 0.740 |
| 71/477 | Periods without AC | 0.38 (0.18 to 0.84) | 0.016 | 0.20 (0.08 to 0.49)e | <0.001 |
| Including markers of inflammation | |||||
| 99/792 | Fibrinogen | 0.43 (0.22 to 0.84)c | 0.014 | ||
| 99/792 | ultrasensitive CRP | 0.41 (0.21 to 0.80)c | 0.009 |
AC = anticoagulation, CRP = C-Reactive Protein.
aAdjusted for age, gender, symptomatic pulmonary embolism, prior venous thromboembolism, provoked venous thromboembolism, cardiovascular disease (i.e. coronary heart, peripheral arterial or cerebrovascular [stroke, transient ischaemic attack] disease), active cancer, and periods of anticoagulation as a time-varying covariate.
bSensitivity analysis: further adjustment for the same variables as used in the propensity score approach. Adjusted to include age, gender, symptomatic pulmonary embolism, prior venous thromboembolism, provoked venous thromboembolism, cardiovascular disease (i.e. coronary heart, peripheral arterial or cerebrovascular [stroke, transient ischaemic attack] disease), active cancer, periods of anticoagulation as a time-varying covariate, and additionally hypertension, polypharmacy, smoking (never/past/current), body-mass index (≥25).
cTwo separate additional adjustment variables: log-Fibrinogen and log-ultrasensitive C-Reactive Protein (as potential explanatory variables of the association between statins and rVTE). Multiple imputation for missing values of fibrinogen and ultrasensitive C-Reactive Protein was performed.
dVariables used to calculate the propensity score: age, gender, symptomatic pulmonary embolism, prior venous thromboembolism, provoked venous thromboembolism, cardiovascular disease (i.e. coronary heart, peripheral arterial or cerebrovascular [stroke, transient ischaemic attack] disease), active cancer, hypertension, polypharmacy, smoking (never/past/current), and body-mass index (≥25).
eThe adjusted model was weighted according to inverse probability of treatment weights (IPTW) and additionally adjusted for periods of anticoagulation as a time-varying covariate.