Hydration status following stroke and the relationship between hydration and functional status at discharge

INTRODUCTION: Dehydration after stroke may influence health outcomes. We assessed prevalence of water-loss dehydration after stroke, associated outcomes, and diagnostic accuracy of bioelectrical impedance analysis (BIA). METHODS: People with acute stroke were prospectively recruited. Hydration (by serum osmolality), stroke severity, functional status, and multi-frequency BIA (Maltron BioScan 920-2) were recorded within 48 hours of admission. Relationships between dehydration and functional status on discharge and hospital mortality were examined. Diagnostic accuracy of total body water (TBW%), intracellular water (ICW%), and extracellular water (ECW%) each as percentage of body weight measured using BIA were assessed against water-loss dehydration, and ROC plots created. RESULTS: Of the 45 patients recruited (60% men, mean age 72.8, SD 11.1 years) with Ischaemic stroke, 8 (18%) had water-loss dehydration (serum osmolality >300mOsm/kg) and a further 16 (36%) had impending dehydration (295-300 mOsm/kg). There were no cut-offs for TBW%, ECW%, ICW% or ECW:ICW ratio for the 27 participants with valid data that provided sensitivity and specificity of_60%, for either impending or current dehydration, although sensitivity analysis using equations developed for older people suggested limited diagnostic accuracy for TBW% with a 46% or 47% cut-off for current and impending dehydration. There were no clear relationships between hydration status after admission and functional status (as assessed by discharge destination) or mortality. CONCLUSION: BIA appears ineffective at diagnosing water-loss dehydration after stroke. Water-loss dehydration is common early post-stroke but we found no evidence that it predicts functional status or mortality at hospital discharge. Research is needed to clarify longer term outcomes.