Abstract: FR-PO414
A Scoring System for Predicting Individual Effects of Statin Treatment in Type 2 Diabetes Mellitus Patients on Hemodialysis
Session Information
- Hemodialysis and Frequent Dialysis - III
November 08, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- 701 Dialysis: Hemodialysis and Frequent Dialysis
Authors
- Wanner, Christoph, University Hospital, Wuerzburg, Germany
- Genser, Bernd, University of Heidelberg, Heidelberg, Germany
- März, Winfried, Synlab Holding Deutschland GmbH, Mannheim, Germany
Background
Randomized clinical trials did not show a benefit of statins in hemodialysis (HD) patients. However, post-hoc analyses of the German Diabetes Dialyse (4D) study indicated that there are subgroups defined by theragnostic markers showing heterogenous treatment effects. We combined the information of multiple markers to a score predicting individual treatment effects.
Methods
We used data from the 4D study, a randomized trial including 1,255 HD patients with type 2 diabetes, randomized to atorvastatin or placebo. We calculated two scores, score 1 (23 predictive markers) and score 2 (17 clinically available markers) and classified patients in groups based on score cut-points indicating changes in effect. In each group we calculated effect estimates with respect to a composite cardiovascular endpoint and all cause death using both trial follow-up (FU) (median: 4 yrs) and long-term FU data (median: 11.5 yrs).
Results
The groups based on score 1 showed completely differential treatment effects: G1) (score < 26, 458 (36%) pts) showed harm: HR=1.54 (95%CI: 1.16-2.03) [Fig. 1a]; G2) (score 26-31, 331 pts (26%)) showed no effect: HR=1.03 (95%CI: 0.72-1.48) [Fig. 1b] and G3) (score>31, 466 pts (38%) showed benefit: HR=0.43 (95% CI: 0.30-0.60) [Fig. 1c]. In G3 statins also reduced all-cause mortality: HR=0.63 (95% CI: 0.48-0.83). Results for score 2 were similar with a smaller group G3 (N=360 pts). For long-term FU the effects were less heterogenous among groups.
Conclusion
The effect of statins in patients on HD is heterogenous and can be predicted by markers that relate to plausible effect modifying mechanisms including cholesterol metabolism, atherosclerosis, protein energy wasting or competing risks. The score will be useful in clinical practice not only to select patients that benefit from statins but also to identify those where treatment is harmful.