OBJECTIVES: The current study aimed to conduct a systematic review and meta-analysis to explore the efficacy and safety of tube feeding in patients with advanced dementia. DESIGN: Systematic review and meta-analysis. SETTING AND PARTICIPANTS: PubMed, Medline, Embase, and Cochrane Library were searched from inception until March 7, 2020, to obtain relevant studies. INTERVENTION: Feeding with nasogastric tube or percutaneous endoscopic gastrostomy (PEG). MEASURES: We evaluated the associations of tube feeding and the risk of mortality, period of survival days, tube-related complications, and nutritional status. Data from original studies were synthesized by using a random-effects model. Each selected article was assessed for bias using the Newcastle-Ottawa Scale. A narrative synthesis and pooled analyses are reported. RESULTS: Twelve trials were eligible, involving 1805 patients with tube feeding (mean age: 82.8 years; 71.3% female) and 3861 without tube feeding (mean age: 82.7; 68.7% female). For mortality rate, patients with advanced dementia with tube feeding are associated with significantly higher mortality rate [k = 8; odds ratio (OR) 1.79; 95% confidence interval (CI) 1.04-3.07; P = .03]. Initially, no association was found for the risk of pneumonia and pressure sore between groups. However, sensitivity analysis showed patients with advanced dementia with PEG tube feeding have significantly higher risk of pneumonia (OR 3.56; 95% CI 2.32-5.44; P < .001) and pressure sore (OR 2.25; 95% CI 1.92-2.63; P < .001). Finally, no association was found for the survival period and nutritional status between groups. CONCLUSIONS AND IMPLICATIONS: This meta-analysis indicates that tube feeding is associated with increased mortality rate and possible tube-related complications, but not improves with prolonging survival days and nutritional status. Shared decision-making routinely before insertion of a tube between caregivers and physicians is recommended.