What does body mass index measure in amyotrophic lateral sclerosis and why should we care?

Abstract

The editorial by Dr. Kasarskis1 commented on the fact that “Neither NIV nor PEG use were entered as covariates into the Paganoni et al.2 models, but it appears that the rate of PEG placement in their population was low (0.7%) compared to national surveys”–a reference to our study also published in this journal.2 We aim to clarify that we included only baseline predictors of survival in our analyses and therefore the low frequency of feeding tubes was secondary to the fact that this number referred to baseline feeding tubes only. After reviewing the primary data and including alternative methods of capturing PEG tube placement (including an ALSFRS swallowing score of ≤1 at enrollment), we have increased our estimate of the number of feeding tubes at the time of enrollment to 11 (2.5%). This did not affect our analysis in any way, as baseline feeding tubes did not enter into the proportional hazards models (log-rank test for survival in subjects with PEG at baseline vs. no PEG, P = 0.15). Dr. Kasarskis is correct in citing that a significantly larger number of subjects received feeding tubes during the course of the studies: 19% in the Celecoxib trial; 10% in the Single Center Cohort study; and 22% in the Topiramate study. However, due to the fact that this was a time-varying covariate, we did not include this in our analysis of baseline predictors of survival. We intend to undertake a follow-up study examining weight changes before and after feeding tube placement in ALS clinical trials.