To the Editor.
—In their meta-analysis on the reversal of left ventricular hypertrophy in essential hypertension, Dr Schmieder and colleagues1 identified 471 references that reported results on regression of left ventricular mass with antihypertensive drugs. They then applied strict selection criteria and only the results of 39 clinical trials with a doubleblind, randomized, controlled, parallel-group design were retained. The authors are to be congratulated for the important and meticulous work they have done. However, I was somewhat surprised by the subsequent statistical analysis of the data. Whereas the individual studies were designed as comparative studies, the authors used each treatment arm of these selected trials as a separate observation and, therefore, gave up the major advantage of the original design. In my view, a more appropriate analysis would have been not only to calculate the changes within each treatment arm, but also to analyze the differences between the changes
Fagard R. Reversal of Left Ventricular Hypertrophy. JAMA. 1996;276(20):1636-1637. doi:10.1001/jama.1996.03540200022019