RESULTS
Literature search: We identified 1327 articles in our initial
search. After exclusion of duplicate papers and those that did not meet
the inclusion criteria, we identified 11 full-text articles of
potentially relevant studies. After full-text review, we excluded an
additional 4 articles for the following reasons: studies that enrolled
patients with chronic renal diseases or hemodialysis patients, papers
that were conducted on critically ill patients, one document that was
conducted on children. Finally, seven cohort studies were included in
the current systematic review and meta-analysis[5, 6,
14-16, 19, 20]. Four studies had reported effect sizes for all-cause
mortality[6, 15, 19, 20] and 6 studies for CVD
mortality[5, 6, 14-16, 20]. Of these publications,
seven had reported effect sizes for total water[5, 6,
14-16, 19, 20] and four for drinking water[5, 16,
19, 20]. A flow diagram of study selection is shown in Figure
1 .
Characteristics of included studies: Characteristics of
included prospective cohort studies are presented in Tables 1 .
All included studies published between 2002 and 2018. Participants in
these studies ranged from 1055 to 35362 people, with an age range
between 18 and over 70 years. The duration of the follow up of the
included cohort was between 6 and 19.1 years. In total 116816
participants were enrolled in the studies. Among the included studies, 3
studies were done in the USA[5, 6, 19], 2 in
Australia[15, 20], and the rest in
Japan[14] and
Netherland[16]. In total 14754 patients died due
to all-cause in the studies and 7611 deaths were due to CVD. One study
was performed on women[20] and 4
studies[5, 14, 16, 19] were reported effect size
for men and women separately. Out of the remaining studies,
two[6, 15] reported effect sizes for both genders
in combination. To assay fluid/water intake, all studies had used a food
frequency questionnaire and 24h recall. All included studies applied
record linkage for assessment of mortality as the outcome. All studies
adjusted the associations for age. Most cohorts controlled for some
conventional risk factors, including body mass index (n=5), smoking
(n=5), and energy intake (n=3). Some others had also adjusted for
alcohol consumption (n=2) and other dietary variables (n=2). Based on
the ROBINS-E tool, two studies had a low-risk of bias in all components
(Supplemental Table 2 ).