Studying interventions to prevent the progression from prehypertension to hypertension: does TROPHY win the prize?

and most individuals with systolic blood pressures in the range of 130–139 mm Hg will develop hypertension. Given the burden of cardiovascular disease and stroke attributable to hypertension, delaying its progression using an intervention in the prehypertension stage would be attractive. Testing whether a temporary treatment alters the progression from prehypertension to hypertension, however, raises specific methodological challenges. The research question is not, “Do antihypertensive medications lower blood pressure?” but rather, “Does temporary treatment alter the subsequent course of an individuals’ blood pressure?” Because the focus is on durable effects after treatment has ended, outcomes must be assessed using only data from the post-treatment period for the intervention and control groups. Using data obtained during the initial active-treatment period to assess outcomes will inevitably give the spurious impression of a sustained difference in blood pressure when this may not exist. The recent Trial of Preventing Hypertension (TROPHY) examined whether treating patients with candesartan for two years resulted in a sustained reduction in blood pressure after candesartan was discontinued. TROPHY reported that treatment with candesartan significantly reduced the risk of incident hypertension over the four-year study interval. However, given the way the study’s primary endpoint was defined, these results are likely invalid. Here we describe how the methods of TROPHY biased the findings in the third and fourth years of the study (when both groups were treated with placebo) so it may have appeared there were persistent differences in incident hypertension and higher blood pressures in the placebo group even if the two groups, in actuality, did not differ. We make recommendations for future trials seeking to demonstrate whether lasting differences in blood pressure can be achieved through short-term interventions.