Introduction Low-intake dehydration is associated with higher mortality and poor health outcomes in older adults. Contributing factors (physiological, physical and cognitive decline, incontinence worries and reduced social drinking) may be exacerbated by illness, unfamiliar environment and routines in hospitalised older adults. We conducted a high-quality systematic review and meta-analysis to estimate dehydration prevalence amongst hospitalised adults (≥65 years) at admission. Methods We included studies of hospitalised participants measuring serum osmolality or osmolarity. MEDLINE, Embase, Cochrane CENTRAL and CINAHL were searched (inception to July 2024). Inclusion, data extraction and risk of bias were assessed independently in duplicate. Data were described and synthesised in random effects meta-analysis (Meta-XL 5.3) and synthesis without meta-analysis (SWiM). Subgrouping assessed associations between dehydration prevalence and renal impairment, cognitive impairment, diabetic status, dependence, ethnicity, age, sex and economic status. Results 19 studies (of 13 097 assessed) reported the most reliable low-intake dehydration measures (directly measured osmolality >300 mOsm/kg, calculated osmolarity using Khajuria Krahn equation >300 mmol/L). 14 398 hospitalised participants were included of whom 5039 (35%) were dehydrated at or near admission. While meta-analysis suggested 23% (95% CI 17% to 30%, Grading of Recommendations, Assessment, Development and Evaluations (GRADE) moderate-quality evidence) were dehydrated, heterogeneity was high (I 2 98%) and different meta-analytic models suggested higher prevalence. Likely prevalence was a quarter to a third of hospitalised older adults. Subgrouping was generally not feasible, but SWiM suggested higher prevalence may be associated with impaired cognition, diabetes, renal impairment, and perhaps higher dependency. Limited evidence suggested dehydration prevalence reduced only slightly during hospitalisation. Conclusions Between a quarter and a third of older adults have low-intake dehydration on hospital admission, varying with patient characteristics. Supporting older adults to drink well in hospital may be the appropriate response given high prevalence and severe health consequences. Trials are needed to assess effectiveness of in-hospital interventions to support drinking and improve health outcomes. PROSPERO registration number CRD42021293763.
Hooper et al. (Wed,) studied this question.