ENHANCE Study: Basically, the Ultrasound Images Were Crap
A review of the lengthy timeline released by Schering-Plough (SP) on Friday provides further insight into the delay of the ENHANCE* study results, for those who care to plow (pardon the pun) through it. Study designers may well be kicking themselves for not electing to use intravascular ultrasound (IVUS) to assess arterial atherosclerotic plaque way back in 2001, when it was considered, as opposed to going solely with the lower-tech B-mode† ultrasound imaging method.
According to the SP timeline, data-quality reviews of study ultrasound images by SP biostatisticians during the latter half of 2005 identified "biologically implausible" results. This description is defined later in the timeline as "significant fluctuations in IMT‡ [intima-media thickness] (in individual patients) within short periods of time and over time, which would not be expected to occur in nature." In other words, widely different and biologically unexplainable measures of IMT from ultrasound to ultrasound in the same patient were observed, suggesting poor quality and/or validity of the ultrasound measurements in this study. Information from the timeline also indicates that there were concerns that the same segment of artery was not measured in each patient from assessment point to assessment point, possibly accounting for the "implausible" variability of the data.
Given these data problems, it was then decided that a "synchronous" reading of the ultrasound images might account for and accommodate the wide variability in the ultrasound data found within each study participant; however, this process required the installation of new computer equipment, which further delayed the assessment of the ultrasound data in April 2006. The synchronous reading of the ultrasound images nevertheless began in 2006 and continued through the end of the year. But an "[i]nitial statistical review of the final received dataset…indicates that data quality problems still exist." The two quality issues identified were 1) remaining, significant fluctuations in IMT among individual patients, "which would not be expected to occur in nature," and 2) missing data, because some individual ultrasound images were too poor to read or did not assess the correct arterial segment.
Beginning in January 2007, SP hired independent contractors and convened independent experts to provide advice and recommend fixes in an attempt to salvage the suboptimal ultrasound data of the ENHANCE study. The consensus from experts was that, because the common carotid artery (CCA) images (as opposed to other arterial images included in the primary endpoint of the study) provided the most reliable and consistent measurements, the endpoint should be changed to include only measurements of the CCA, and the original primary endpoint would become a secondary endpoint. In other words, the CCA ultrasound images were less crappy than the other carotid ultrasound images. (This decision to change the primary endpoint, on the basis of data quality, accounted for the brouhaha among interested parties.) However, an advisory board convened by SP later recommended that the primary endpoint should not be changed.
Meanwhile, the deadlines for the presentation of the ENHANCE study data at the 2007 annual meetings of the American College of Cardiology and the American Heart Association were missed in March and November, respectively. Then, as everyone knows who cares, a press release was issued by the study sponsors announcing the negative results of the ENHANCE study in January 2008.
The ultimate problem is that, given the poor quality of the ultrasound data in the ENHANCE study, it is still unknown if the combination of ezetimibe and simvastatin is associated with greater atherosclerotic-plaque stability, or even reduction, than simvastatin alone in patients with heterozygous familial hypercholesterolemia.
*ENHANCE = Ezetimibe and Simvastatin in Hypercholesterolemia Enhances Atherosclerosis Regression. The study was an international 2-year, randomized, double-blind, controlled trial comparing the combination of ezetimibe and simvastatin (Vytorin) with simvastatin alone (Zocor) in patients with heterozygous familial hypercholesterolemia. The originally designated primary endpoint was the change in the average far-wall intima media thickness (IMT) of the right and left common carotid arteries (CCA), carotid bulb, and internal carotid arteries on a per-patient basis.
†B = brightness; imaging provides a 1-dimensional graphical display, with brightness corresponding to the amplitude of reflected ultrasound.
‡Surrogate marker of atherosclerotic plaque.
0 TrackBacks
Listed below are links to blogs that reference this entry: ENHANCE Study: Basically, the Ultrasound Images Were Crap.
TrackBack URL for this entry: http://bmartinmd.com/cgi-bin/mt/mt-tb.cgi/25
2 Comments
Leave a comment
Interesting.
How come AZ managed to get a licence extantion with their data from METEOR.
Were they just better at studies?
Thanks Insider for the blog hat tip and question. I'm certainly not the best person to answer, but I can possibly provide some insight, given my read of the studies.
There are a number of differences between the ENHANCE and METEOR studies that likely affected their ultimate outcomes. First, they were conducted in 2 different patient populations: ENHANCE assessed pts with heterozygous familial hypercholesterolemia; METEOR, pts with a low Framingham risk score and mild-moderate subclinical atherosclerosis. Given that the ENHANCE pts had a clear genetic predisposition to atherosclerosis, it may be more difficult to trump that predisposition with treatment than to affect milder disease. Second, ENHANCE compared statin therapy with an add-on blocker of cholesterol absorption, and METEOR compared a statin with placebo. In this context, it's probably much easier to demonstrate a statistically significant treatment difference against baseline placebo therapy than against an active comparator. Third, although we don't have a complete, peer-reviewed description of how intima-media thickness (IMT) was assessed in ENHANCE, there may have been some methodological differences between the primary endpoints of the 2 studies (both assessed with B-mode ultrasound) to account for the presumed greater variability of measurements in the ENHANCE study.