Introduction: Multiple factors contribute to increased risk of dehydration in amyotrophic lateral sclerosis (ALS), which contributes to shortened survival independent of nutritional status. The assessment of hydration by doubly labeled water is restricted due to the limited availability of this gold standard technique for clinical use. This prompted us to examine the utility of urine-specific gravity (USG) as a predictor of hydration need in ALS subjects. Material and Methods: Using data from a multicenter study of 80 ALS subjects with 250 visits, we conducted a secondary analysis of the original data set from doubly labeled water experiments. We used a cross-section of the data (one visit per 75 subjects) in the model selection step ("test set"), and a repeated measures analysis in the validation step with data from 63 subjects and 142 follow-up visits. The sensitivity to detect inadequate water turnover rate (a surrogate for water intake) was the goal of the predictive model presented for clinical use. Results and discussion: The final predictive model to estimate water requirement included USG, gender, body mass index, and the ALSFRS gross motor subscale score. We developed a best-fit equation to estimate water intake from USG, determine hydration status, and improve clinical care of real-world ALS subjects.
Keywords: ALS; doubly labeled water; urine specific gravity; water intake; water turnover.