Richard’s Reviews: Diabetes decision aids and lowering LDL
Posted on 10th December 2014 by Robert Kemp
This week we look at a pair of trails examining the evidence of new treatment for lowering LDL (or bad) cholesterol in familial hypercholesterolaemia and one which assesses patient-tailored decision aids in diabetes.
PCSK9 inhibition with evolocumab (AMG 145) in heterozygous familial hypercholesterolaemia (RUTHERFORD-2): a randomised, double-blind, placebo-controlled trial
Does monoclonal antibody directed against PCKS9 reduce LDL cholesterol in people with heterozygous familial hypercholesterolaemia when added to standard care?
Randomised double blind placebo controlled trial.
Familial hypercholesterolaemia is a genetic condition that causes constantly elevated LDL cholesterol. High levels of LDL, or bad, cholesterol is associated with an increased risk of cardiovascular events such as heart attacks. Drugs that lower LDL cholesterol, such as statins, are taken by these patients and much of the general population to lower their cardiovascular risk.
Recently a new target has emerged for lowering cholesterol PCSK9. PCSK9 aids in targeting the LDL receptor for lysosomal degradation in cells. Blocking PCSK9 results in more LDL receptors available to increase catabolism of plasma LDL.
Evolocumab is a monoclonal antibody designed to inhibit PCSK9 and reduce LDL cholesterol. The RUTHERFORD – 2 trial was designed to test if evolocumab,can reduce LDL cholesterol in patients with heterozygous familial hypercholesterolaemia.
331 patients were randomised to receive injections on either placebo or evolocumab every 2 weeks or month for 12 months in addition to their current lipid lowering therapy. At the end of the 12 weeks treatment with evolocumab was associated 2.4mmol reduction in LDL over placebo corresponding to a 59.2% reduction with no additional risk of adverse events.
Take Home Message
This trial provides good evidence that evolocumab is safe in the short term and can be added to current lipid lowering therapies in patients with heterozygous familial hypercholesterolaemia to further lower LDL.
Much of the population, including people with familial hypercholesterolemia, take LDL lowering therapies, in the form of statins, to reduce their risk of cardiovascular events. This trial shows good evidence that evolocumab can reduce LDL but does that mean a reduction in risk of cardiovascular events?
The outcome of LDL cholesterol used in this trial is what is described as a surrogate outcome. A surrogate outcome is a marker for the real world outcome we are interested in but is often easier to measure. In this case our real world outcome is cardiovascular risk but our surrogate outcome, our marker, is LDL cholesterol. Often when designing these trials we hope that the improvement in our surrogate outcome will lead to improvements in our real world outcome; this is the case for statins which have been shown to lower the risk of cardiovascular events. However, this is not always the case and while improvements in surrogate outcomes can give us a hint if a drug works, proof ultimately relies on demonstrating improvement in our real world outcome, this trial gives us a hint evolocumab may work, but not proof.
Effects of a patient-oriented decision aid for prioritising treatment goals in diabetes: pragmatic randomised controlled trial
Does a patient-orientated decision aid aimed at prioritising treatment goals improve patient empowerment in type 2 diabetes patients when compared to usual care?
A major challenge for evidence-based medicine is translating evidence generated from populations, statistics and risk into meaningful tools to enable shared decision making between patients and doctors. This trial tested a patient decision aid against usual care for improving patient empowerment and treatment decisions in type 2 diabetes. While most decision aids are aimed at one treatment decision those with type 2 diabetes face multiple clinical areas (glucose, lipids, blood pressure) which all require treatment decisions.The decision aid let patients choose the outcomes that matter to them and then presented tailored risk information and treatment options for each clinical domain.
In 18 GP practices in the Netherlands 344 people with type 2 diabetes were randomised to receive the decision aid or continue on normal care. Patients were followed up for and assessed outcomes on patient empowerment and treatment decisions. The results of the study was that use with the decision aid produced a non significant improvement of 0.039 (−0.056 to 0.134) on a 5 point scale of patient empowerment when compared to usual care. In addition, there was no significant change in treatment intensification when compared to usual care.
Take Home Message
This patient decision aid did not improve empowerment of patients or treatment decisions when compared to usual care.
On the face of it, this trial suggests the patient decision aid has failed, but you can’t be so sure. This trial has limitations which the authors acknowledge. One is the use of the surrogate outcome, ‘patient empowerment’. It is assumed that improving patient empowerment will lead to improvements in involvement and adherence to treatment resulting in better clinical outcomes. We don’t know if the aid improved adherence to treatment or patients clinical outcomes, the authors don’t present results.
In addition uptake and use of the decision aid varied greatly across the intervention group and the content of patient-doctor consultations was not assessed in the trial. Without controlling for this we can’t properly gauge the aid’s success. It may have been that some clinicians and patients in the intervention group used the aid a lot and received a great deal of benefit, while others may have used it very little and got no benefit. This could mask any benefit when comparing the intervention group to the control group. In an editorial, decision aid designers from the Mayo Clinic argue controlling for the content of the consultations and the extent to which the decision aid was used is not only needed to see how successful the aid is but could also identify interactions that truly benefit patients.
Richard’s blog post can be found here.