Possible carry over effects: crossover from low sodium dietary advice back to normal diet?
Slagman and colleagues evaluate in a factorial design the impact of
low sodium diet and Valsartan using a crossover randomised controlled
trial, without any washout period. The authors claim that a washout period
is not necessary because of the short half life of Valsartan. I'm not
convinced that a short half life necessary rules out a possible carryover
effect. But, more importantly, surely there must be some sort of carry
over effect between patients crossing over from a low sodium dietary
advice back to normal sodium intake? If dietary advice is given, how long
is this advice retained - surely a very difficult question to answer which
would ultimately rule out a cross-over design, let alone a crossover
design with no wash-out period? The authors do indeed test for carryover
effects, we are told these are not significant, but we are not given
treatment effects by sequence (only P-values) and lack of significance
does not mean lack of effect (with only 52 people in the study there may
have been insufficient power to demonstrate treatment by sequence
interactions).
Furthermore, the information provided on randomisation is limited, and it
is stated that the participants remained blind through the entire study:
how is it possible to blind to dietary advice?
Rapid Response:
Possible carry over effects: crossover from low sodium dietary advice back to normal diet?
Slagman and colleagues evaluate in a factorial design the impact of
low sodium diet and Valsartan using a crossover randomised controlled
trial, without any washout period. The authors claim that a washout period
is not necessary because of the short half life of Valsartan. I'm not
convinced that a short half life necessary rules out a possible carryover
effect. But, more importantly, surely there must be some sort of carry
over effect between patients crossing over from a low sodium dietary
advice back to normal sodium intake? If dietary advice is given, how long
is this advice retained - surely a very difficult question to answer which
would ultimately rule out a cross-over design, let alone a crossover
design with no wash-out period? The authors do indeed test for carryover
effects, we are told these are not significant, but we are not given
treatment effects by sequence (only P-values) and lack of significance
does not mean lack of effect (with only 52 people in the study there may
have been insufficient power to demonstrate treatment by sequence
interactions).
Furthermore, the information provided on randomisation is limited, and it
is stated that the participants remained blind through the entire study:
how is it possible to blind to dietary advice?
Competing interests: No competing interests