Clinical Trials

Clinical Trials: A Practical Guide ■❚❙❘• a change in the levels of disease markers, such as a change in bloodpressure as a predictor of a future occurrence of a stroke or kidney disease• the macroscopic appearance of tissues, such as the endoscopicvisualization of an area of erosion in the stomach that is consideredby gastroenterologists to be the precursor of stomach bleedsAdvantages of surrogate endpointsLike a composite endpoint, use of a surrogate endpoint can reduce the samplesize needed for a study and thereby the duration and cost of performing a clinicaltrial. Surrogate endpoints are particularly useful when conducting Phase II screeningtrials to identify whether a new intervention has an effect, since a change is oftenseen in a surrogate endpoint long before an adverse event occurs. If there issufficient change, it might then be justifiable to proceed to a large definitive trialwith clinically meaningful outcomes.Limitations of surrogate endpointsIt is important to realize that surrogate endpoints are only useful if they are a goodpredictor of a clinical outcome. If this relationship is not clearly defined, surrogateendpoints can be misleading. One classic example is the following case of threeparticular antiarrhythmic treatments.ExampleThree drugs (encainide, flecainide, and moricizine) were found to reducearrhythmias and received FDA approval for use in patients with life-threateningor severely symptomatic arrhythmias. This was based on the assumption thatif irregularities in heart rhythm were reduced, there should be lower numbersof cardiac deaths from disturbances of heart rhythm. More than 200,000 patientsper year took these drugs in the US. However, it was subsequently shown in theCAST (Cardiac Arrhythmia Suppression Trial) study of 2,309 subjects that ratesof death were higher in patients taking these drugs compared to those onplacebo [7–9].Health-economic endpointsEconomic endpoints are becoming increasingly important for new treatments.The Medical Research Council in the UK believes that clinical trial reports shouldalways be accompanied by economic evaluations [10]. The growing number ofeconomic analyses being submitted to medical journals led to the then editorof the BMJ announcing a policy that stated that economic results from clinicaltrials would now always have to be submitted with their respective clinical results,or they would not be accepted [10]. Therefore, assessments of health economics43