To the Editor.
—Dr Edeki and colleagues1 provide evidence that appears to support their conclusion that an individual's debrisoquin phenotype is an important determinant of β-blockade following timolol eyedrops, and metabolism of timolol is inhibited and β-blockade increased by coadministration of oral quinidine. But they failed to discuss two troublesome issues that may affect the internal and external validity of their study.First, the authors failed to report the methods they used to select the sample size of 13. In general, randomized controlled trials begin with a calculation to predetermine sample size before recruitment is started. The authors are no doubt aware that these methods include selection of a desired power (typically 0.80 or 0.90), a reasonable estimate of treatment effect (drug vs placebo) in units of the main outcome measure, an estimate of pooled variance, and a level of significance, 1—α (the authors accepted α ≤.05). They cited
Kalenak JW. Excessive β-Blockade With Timolol Eye Drops. JAMA. 1996;275(13):985. doi:10.1001/jama.1996.03530370023014