107 Increased Hospitalisation Associated with Nasogastric Feeding in Advanced Dementia: Experience of A Geriatric Unit in Singapore

Abstract

Abstract Introduction Tube feeding is not recommended in persons with advanced dementia and dysphagia; it does not improve nutrition, reduce infections or mortality. In Singapore, limited data exists on the prevalence of oral versus tube feeding in persons with advanced dementia, and the alignment between the mode of administering food and fluids and a prior advanced care plan (ACP) discussion. Methods Retrospective review of 120 patients from January to October 2017 was conducted to examine the prevalence of oral and tube feeding in advanced dementia patients with ACP, and the alignment of the chosen mode of intake with their ACP. Outcomes include rates of readmission, pneumonia, ACP revisions and mortality at six months. Results 42 subjects fulfilling criteria were analysed. 81.0% continued oral feeding. There was a statistically significant difference in the decision for tube insertion between oral and tube feeding groups (p=0.02), which was influenced by the ACP discussion. Concordance with ACP discussion in the tube feeding group was mixed, with 50% agreeing for tube insertion and 33.3% who were clinically guided. Surrogates made the decisions in 90.5% of the discussions. Six-month readmission rate was 32.4% for orally-fed patients versus 75% for tube-fed patients (p=0.045). 52.9% of admissions were attributed to pneumonia. 8.8% of orally fed patients revised their ACP to opt for a trial of treatment in their own home instead of readmission to the hospital. There was no difference in 6-month mortality (p=0.123). Conclusions Oral feeding was continued for most advanced dementia patients. Tube feeding does not offer survival benefit or prevent aspiration at 6 months, yet is associated with increased hospitalisations. ACP discussion should be initiated earlier to provide persons with dementia the opportunity to express their preferences for the mode of food and fluid intake.