2180Estimating individual lifetime benefit and bleeding risk of adding rivaroxaban to aspirin for patients with stable cardiovascular disease: results from the COMPASS trial

Background

The Cardiovascular Outcomes for People Using Anticoagulation Strategies (COMPASS) trial has demonstrated that adding low-dose rivaroxaban to aspirin in patients with stable atherosclerotic disease on average reduces recurrence of cardiovascular disease (CVD) events, but increases the risk of major bleeding. For clinical practice, it is important to be able to weigh the absolute benefit from the intervention in terms of lower cardiovascular risk against the absolute increase in risk for major bleeding.

Purpose

The aim of this study was to estimate the individual lifetime benefit and harm of adding low-dose rivaroxaban to aspirin in patients with stable cardiovascular disease by predicting individual months free from CVD events gained and individual months free from major bleeding lost.

Methods

Analyses were based on data of patients with established CVD in the COMPASS trial (n=27,390) and SMART prospective cohort study (n=8,139). The externally validated lifetime SMART-REACH model for recurrent CVD was used to predict life expectancy free of stroke and myocardial infarction, based on the following predictors: sex, current smoking, diabetes mellitus, systolic blood pressure, total cholesterol, creatinine, number of locations of CVD, history of atrial fibrillation, and history of congestive heart failure. A new Fine & Gray competing-risk adjusted Cox proportional hazard model was derived in the COMPASS study population for prediction of life expectancy free from major bleeding, including the same predictors as the SMART-REACH model and additionally ethnicity, geographical region, and history of bleeding requiring transfusion. These lifetime estimates were then combined with hazard ratios from the COMPASS trial to estimate lifetime treatment effects from adding low-dose rivaroxaban to aspirin, expressed in terms of 1) months free from stroke or myocardial infarction gained, and 2) months free from major bleeding lost.

Results

External goodness-of-fit of the SMART-REACH model in the COMPASS study was sufficient. The newly developed major bleeding risk model also showed sufficient external goodness-of-fit in the SMART cohort. The median predicted individual gain in life-expectancy free of stroke or MI from added low-dose rivaroxaban was 16 months (range 1–48 months), while the median predicted individualized lifetime lost in terms of major bleeding was 2 months (range 0–20 months) (Figure 1A). Predicted benefit was higher than predicted harm in more than 90% of the study population. An interactive calculator for use in clinical practice will be made available (example in figure 1B).

Figure 1

Conclusions

There is a wide distribution in lifetime gain and harm from adding low-dose rivaroxaban to aspirin in individual patients with stable CVD. Using these lifetime models, benefits and bleeding risk can be weighed for and with each individual patient, to support treatment decision making in clinical practice.

P4990Heterogeneity of treatment effects from an intensive lifestyle weight loss intervention on cardiovascular events in patients with type 2 diabetes: data from the Look AHEAD trial

Background

The Action for Health in Diabetes trial (Look AHEAD) randomized overweight and obese patients with type 2 diabetes to either an intensive lifestyle intervention (ILI) or diabetes support and education (DSE). The trial was stopped early for futility after a median follow-up of 9.6 years due to a lack of effect on cardiovascular disease outcomes, despite beneficial effects on metabolic control and cardiovascular risk factors. Subgroup analyses identified no subgroups based on baseline characteristics with a significant treatment effect. However, traditional simple subgroup analyses have several disadvantages compared to a multivariable risk-based approach to identify heterogeneity of treatment effects (HTE).

Purpose

To explore the possible presence of HTE of an ILI on the occurrence of major cardiovascular events (4-point MACE: nonfatal myocardial infarction, nonfatal stroke, hospitalization for angina, and death from cardiovascular causes) in overweight or obese patients with type 2 diabetes mellitus, and to identify patient characteristics associated with treatment.

Methods

In 4,901 patients from Look AHEAD, a ridge penalized Cox regression model to predict treatment effect of ILI versus DSE on the risk of MACE was derived including all possible treatment-by-covariate interaction terms. Next, the ability of the model to predict HTE was confirmed by calculating hazard ratios (HR) and absolute risk change in quartiles of predicted treatment effect, thereby leaving randomization intact. Finally, baseline patient characteristics were compared between quartiles of predicted treatment effect.

Results

During a median follow-up of 9.4 years, 799 events occurred (Fig. 1A). The derived risk model showed good internal calibration, with a C-statistic for discrimination of 0.73 (95% confidence interval [95% CI] 0.71–0.73). The median estimated absolute treatment effect on 10 year risk for MACE with ILI was −1.3% and varied substantially, ranging from −39% to +43% (Fig. 1B). In quartile 1, the quartile with the highest benefit, there was a significant treatment benefit of ILI versus DSE (HR 0.64; 95% CI 0.49–0.83), while there was no effect from treatment in quartiles 2 and 3 (HR 0.81, 95% CI 0.58–1.14, and 1.13, 95% CI 0.80–1.60, respectively), and a detrimental effect in quartile 4 (HR 1.37, 95% CI 1.09–1.73) (Fig. 1C). Patient characteristics most notably associated with higher benefit of ILI were higher age, male sex, higher socio-economic status, no history of cardiovascular disease, no use of insulin, higher blood pressure, lower HbA1c, and the presence of micro-albuminuria but absence of macro-albuminuria.

Figure 1

Conclusion

This post-hoc analysis of the Look AHEAD trial shows evidence of considerable HTE of an intensive lifestyle intervention aimed at weight loss for reducing MACE. Future research into ILI for MACE risk reduction should be specifically aimed at subgroups of patients with a high likelihood of treatment benefit.

Acknowledgement/Funding

None