Soft as statin.

August 24, 2017. A series of major articles has us all pretty confused about the role of statins in cancer care. A retrospective Danish study in JCO last week showed a nearly 20% reduction in prostate cancer mortality with post-diagnosis statin use. An accompanying editorial did a commendable job of summarizing the literature on this topic and prescribing future trial designs, but it stopped short of recommending regular statin use among men with prostate cancer. Then there’s this week’s article in JAMA Onc which sets out to debunk this kind of data altogether by highlighting significant selection bias and immortal time bias. They, too, found a survival benefit with statin use within an independent sample of >17K SEER cancer patients...until adjusting for selection and immortal time biases, which rendered a flat hazard ratio of 1.00 for cancer-specific death. Palm to face. Our take away: this is a poignant reminder that retrospectively obtained big data almost always means big bias. Oh, and prescribing statins to decrease cancer death is a soft call at best.