BACKGROUND AND AIMS: Advice to drink plenty of fluid is common in respiratory infections. We assessed whether low fluid intake (dehydration) altered outcomes in adults with pneumonia. METHODS: We systematically reviewed trials increasing fluid intake and well-adjusted, well-powered observational studies assessing associations between markers of low-intake dehydration (fluid intake, serum osmolality, urea or blood urea nitrogen, urinary output, signs of dehydration) and mortality in adult pneumonia patients (with any type of pneumonia, including community acquired, health-care acquired, aspiration, COVID-19 and mixed types). Medline, Embase, CENTRAL, references of reviews and included studies were searched to 30/10/2020. Studies were assessed for inclusion, risk of bias and data extracted independently in duplicate. We employed random-effects meta-analysis, sensitivity analyses, subgrouping and GRADE assessment. Prospero registration: CRD42020182599. RESULTS: We identified one trial, 20 well-adjusted cohort studies and one case-control study. None suggested that more fluid (hydration) was associated with harm. Ten of 13 well-powered observational studies found statistically significant positive associations in adjusted analyses between dehydration and medium-term mortality. The other three studies found no significant effect. Meta-analysis suggested doubled odds of medium-term mortality in dehydrated (compared to hydrated) pneumonia patients (GRADE moderate-quality evidence, OR 2.3, 95% CI 1.8 to 2.8, 8619 deaths in 128,319 participants). Heterogeneity was explained by a dose effect (greater dehydration increased risk of mortality further), and the effect was consistent across types of pneumonia (including community-acquired, hospital-acquired, aspiration, nursing and health-care associated, and mixed pneumonia), age and setting (community or hospital). The single trial found that educating pneumonia patients to drink ≥1.5 L fluid/d alongside lifestyle advice increased fluid intake and reduced subsequent healthcare use. No studies in COVID-19 pneumonia met the inclusion criteria, but 70% of those hospitalised with COVID-19 have pneumonia. Smaller COVID-19 studies suggested that hydration is as important in COVID-19 pneumonia mortality as in other pneumonias. CONCLUSIONS: We found consistent moderate-quality evidence mainly from observational studies that improving hydration reduces the risk of medium-term mortality in all types of pneumonia. It is remarkable that while many studies included dehydration as a potential confounder, and major pneumonia risk scores include measures of hydration, optimal fluid volume and the effect of supporting hydration have not been assessed in randomised controlled trials of people with pneumonia. Such trials, are needed as potential benefits may be large, rapid and implemented at low cost. Supporting hydration and reversing dehydration has the potential to have rapid positive impacts on pneumonia outcomes, and perhaps also COVID-19 pneumonia outcomes, in older adults.