Lack of a placebo control group; Design introduces the possibility of biased investigator determined assessment and unbalanced dosage adjustment; Randomization of sequential envelopes may be biased.