P773 - LOW-INTAKE DEHYDRATION IS COMMON IN HOSPITALISED OLDER ADULTS: SYSTEMATIC REVIEW AND META-ANALYSIS

P773

LOW-INTAKE DEHYDRATION IS COMMON IN HOSPITALISED OLDER ADULTS: SYSTEMATIC REVIEW AND META-ANALYSIS

L. Hooper^1,*, E. Parkinson^2,3, A. Abdelhamid^1,4, K. Deng¹, O. Edozie¹, C. Fenner¹, L. Frost¹, M. Ganeshamoorthy¹, S. Mohandas⁵, J. Robson¹, A. Sharif¹, C. Wolmarans¹, D. Bunn³

¹Norwich Medical School, University of East Anglia, Norwich, ²Health Innovation East, Cambridge, ³School of Health Sciences, University of East Anglia, ⁴East of England Primary Care Deanery, Norwich, United Kingdom, ⁵International Initiative for Impact Evaluation, 3ie, New Delhi, India

Rationale: Low-intake dehydration (drinking too little) is associated with higher mortality and poor health outcomes in older people. We conducted a high-quality prevalence systematic review to estimate prevalence amongst hospitalised older adults.

Methods: We included studies of hospitalised participants aged ≥65 years assessing dehydration. MEDLINE, Embase, CENTRAL, and CINAHL were searched until July 2024.  Inclusion, data extraction and risk of bias were assessed independently in duplicate.  Data were described and meta-analysed, double-checked using sensitivity analyses. Synthesis without meta-analysis (SWiM) and subgrouping assessed associations with patient and environmental characteristics. Protocol pre-registered: PROSPERO CRD42021293763. 

Results: Forty studies were included, of which 19 (14,398 participants of whom 35% were dehydrated) provided reliable measures of low-intake dehydration (directly measured osmolality >300mOsm/kg or calculated osmolarity using Khajuria Krahn equation >300mmol/L). Meta-analysis suggested 23% (95% CI 17-30%, GRADE moderate-quality evidence) were dehydrated, but heterogeneity was high (I² 98%) and sensitivity analyses suggested higher prevalences. Likely prevalence was a quarter to a third of hospitalised older adults.  Higher prevalence may be associated with impaired cognition, diabetes, renal impairment, and perhaps higher dependency.  Limited evidence suggested that dehydration prevalence reduced only slightly during hospitalisation. 

Conclusion: A quarter to a third of hospitalised older adults have low-intake dehydration on admission, varying according to patient characteristics. Supporting all older adults to drink well in hospital is the appropriate response as dehydration prevalence is high and health consequences severe. Trials are needed to assess effectiveness of in-hospital interventions to support drinking and improve health outcomes. 

Disclosure of Interest: None declared