Sirs, We are grateful to Drs Turner and Hamlin for their comments regarding our study.1, 2
All included patients had a documented diagnosis of ulcerative colitis (UC), and were selected on the basis of their disease activity score. Histology was not available at the time of inclusion. It is certainly disturbing that 19% of patients had a histological index of 0, and this recurrent problem in treatment trials in UC needs to be addressed.
Patients were already under stable treatment with 5-ASA, so the literature concerning rate of response in naïve patients was not used to perform a power calculation. Actual response was higher than expected, however, treatment difference was as expected, and the power is not relevant before a statistically significant result for the primary endpoint. The study was not powered for secondary endpoints and could be underpowered for them. This is standard in clinical trials and is not related to the issue of the placebo response.
The results of the 1 g/day and 2 g/day groups were numerically different but this is to be expected also under the null hypothesis of no difference for effect of random variation. A priori, no difference was anticipated, and the comparison between the combined dose groups vs. placebo was planned as the primary analysis. The study was not powered to compare the doses, therefore, a posteriori, the only judgement can be clinical. When looking at the 95% confidence intervals of the difference between each dose and placebo, the results do not contradict the expectation of no difference, being the shift between the intervals of approximately 5%.
Pancolonoscopy was required only at baseline to document the extension of the disease, if not recently performed. In all other cases only the affected part of the colon was examined.
Finally it is not our intention to propose propionyl-L-carnitine as a treatment for mild-to-moderate UC at this stage, and two phase III confirmatory trials are expected to start in the near future.