Professional Documents
Culture Documents
org
ABSTRACT
Background Despite increasing incidence of CKD, no evidence-based lifestyle recommendations for CKD
primary prevention apparently exist.
Methods To evaluate the consistency of evidence associating modifiable lifestyle factors and CKD inci-
dence, we searched MEDLINE, Embase, CINAHL, and references from eligible studies from database in-
ception through June 2019. We included cohort studies of adults without CKD at baseline that reported
lifestyle exposures (diet, physical activity, alcohol consumption, and tobacco smoking). The primary out-
come was incident CKD (eGFR,60 ml/min per 1.73 m2). Secondary outcomes included other CKD surro-
gate measures (RRT, GFR decline, and albuminuria).
Results We identified 104 studies of 2,755,719 participants with generally a low risk of bias. Higher dietary
potassium intake associated with significantly decreased odds of CKD (odds ratio [OR], 0.78; 95% confidence
interval [95% CI], 0.65 to 0.94), as did higher vegetable intake (OR, 0.79; 95% CI, 0.70 to 0.90); higher salt intake
associated with significantly increased odds of CKD (OR, 1.21; 95% CI, 1.06 to 1.38). Being physically active
versus sedentary associated with lower odds of CKD (OR, 0.82; 95% CI, 0.69 to 0.98). Current and former
smokers had significantly increased odds of CKD compared with never smokers (OR, 1.18; 95% CI, 1.10 to
1.27). Compared with no consumption, moderate consumption of alcohol associated with reduced risk of CKD
(relative risk, 0.86; 95% CI, 0.79 to 0.93). These associations were consistent, but evidence was predominantly
of low to very low certainty. Results for secondary outcomes were consistent with the primary finding.
Conclusions These findings identify modifiable lifestyle factors that consistently predict the incidence of
CKD in the community and may inform both public health recommendations and clinical practice.
META-ANALYSIS
CKD affects 10% of the world’s population1 and ranks important because the management of patients with
in the top ten noncommunicable diseases contrib- established CKD may require dietary adaptations
uting to disease and disability.2 Its incidence is in-
creasing worldwide,1 and mortality owing to CKD
rose between 2005 and 2017 from 0.9 million to 1.2 Received March 31, 2020. Accepted July 20, 2020.
million deaths annually.2 CKD imposes a significant Published online ahead of print. Publication date available at
economic burden on patients and society; many de- www.jasn.org.
veloped countries spend between 2% and 3% of their Correspondence: Dr. Jaimon T. Kelly, Menzies Health Institute
annual health care expenditure on RRTs (that is, Queensland, Faculty of Medicine, Griffith University, G40 Griffith
Health Centre, Level 8.86, Gold Coast Campus Griffith University,
chronic dialysis or kidney transplantation) for 0.2%
Gold Coast, QLD 4222, Australia, or Dr. Juan-Jesus Carrero,
of the total population.3 Department of Medical Epidemiology and Biostatistics, Kar-
Because prevention is more effective than cure, olinska Institutet, Nobels väg 12A, 171 65 Solna, Sweden. Email:
jaimon.kelly@griffith.edu.au or Juan.jesus.carrero@ki.se
avoiding exposures to hazards that may cause CKD in
the community is a key priority. This is particularly Copyright © 2021 by the American Society of Nephrology
that diametrically differ from those that are needed for primary
Significance Statement
prevention.4 However, no evidence-based lifestyle recommen-
dations for the primary prevention of CKD are apparent. In- Although CKD incidence is increasing, no evidence-based lifestyle
stead, current advice extrapolates from recommendations and recommendations for CKD primary prevention apparently exist. To
evaluate evidence associating modifiable lifestyle factors and in-
inferences drawn from literature concerning cardiovascular
cidence of CKD, the authors undertook a systematic review and
disease, hypertension, and diabetes. 5–7 Furthermore, the meta-analysis. Their analysis, which included 104 observational
World Health Organization discusses CKD as a complication studies of 2,755,719 participants, demonstrated consistency of
of cardiovascular disease and diabetes, which further compli- evidence for a number of measures associated with preventing
cates effective public health messages.8 To address this critical CKD onset, including increasing dietary intake of vegetables and
potassium (21% reduced odds and 22% reduced odds, re-
knowledge gap, we hereby evaluated the consistency of evi-
spectively), increasing physical activity levels (18% reduced odds),
dence associating modifiable lifestyle factors (specifically moderating alcohol consumption (15% reduced risk), lowering
diet, physical activity, alcohol consumption, and tobacco sodium intake (21% increased odds), and stopping tobacco
smoking) and the incidence of CKD. smoking (18% increased risk). In the absence of clinical trial evidence,
these findings can help inform public health recommendations and
patient-centered discussions in clinical practice about lifestyle
measures to prevent CKD.
METHODS
This systematic review and meta-analysis was conducted created in Microsoft Excel, which were checked by a second
according to the Meta-analysis of Observational Studies in author before analysis. Supplemental Table 2 shows the data
Epidemiology checklist for observational studies,9 and it was extracted for analysis. Where we identified more than one publi-
prospectively registered in PROSPERO (CRD42019137328). cation for a single cohort study reporting on the same lifestyle
factor, we used the dataset from the published study contributing
Data Sources and Searches the highest number of participants. Within each meta-analysis, we
We searched MEDLINE, Embase, and CINAHL (database in- conducted a sensitivity analysis examining the effects of substitut-
ception through June 2019) without date or language restric- ing these alternative sources of data on the same cohort. Where
tion (Supplemental Table 1). We used EndNote to deduplicate questions arose concerning a study’s eligibility, missing data, or
and manage citations. For publications in languages other further information on results, we contacted corresponding au-
than English, we used translations by native speakers and thors to request this information. We used author responses to
where not possible, Google Translate. Author pairs completed make decisions on a study’s eligibility and analysis approach.
the title and abstract screening and full-text screening, and We assessed risk of bias for each included study using the
they manually searched the reference lists of eligible studies Newcastle–Ottawa Quality Assessment Scale,13 which as-
and reviews to identify other potentially relevant studies by sesses the methodological quality of the studies across three
snowballing. We included a third investigator (J.T.K. or domains: selection (or representativeness) of cohorts, com-
J.-J.C.) and used consensus to resolve disagreements. parability (of cohorts due to design or analysis), and out-
comes (assessment and follow-up) (Supplemental Table 2).
Study Selection We applied the Grading of Recommendations Assessment,
Inclusion criteria were adult participants without established Development and Evaluation methodology (GRADE) crite-
CKD (i.e., GFR.60 ml/min2) at baseline,10 including those ria to rate the quality of evidence (low, moderate, or high) for
that reported people without CKD as a subgroup of a larger each outcome. In analyses that included five or more studies,
study; cohort studies, prospective or retrospective; and outcome when the primary outcome showed low or moderate hetero-
data (primary or secondary) reported on the basis of levels of geneity, we constructed funnel plots to examine for effects
exposure to a lifestyle factor, including diet (foods or nutrients), that may represent publication, selection, or reporting bias.
physical activity, alcohol consumption, and tobacco smoking.
We excluded studies of participants with albuminuria at base- Data Synthesis and Analyses
line, whenever this information was available. We did not con- We used RevMan 5.3 and Stata version 16.0 (StataCorp, Col-
sider diet pattern exposures in our review, as these have recently lege Station, TX) to meta-analyze data when a particular life-
been systematically evaluated.11,12 style factor was reported for the same outcome from three
The primary outcome was incident CKD, defined as the separate cohorts; otherwise, associations were tabulated and
development of GFR,60 ml/min per 1.73 m 2 during reported using frequencies. Given the observational design
follow-up.10 Secondary outcomes were other surrogate mea- and highly varied sample sizes of the primary studies, we
sures of CKD: RRT, GFR decline, or albuminuria. used the random effects model (DerSimonian and Laird)
for all meta-analyses. Robustness of the association esti-
Data Extraction and Quality Assessment mate was evaluated by comparing if any statistically sig-
One author from each team extracted the data from the in- nificant association identified through random effect
cluded studies using standardized data extraction forms models became nonsignificant under a fixed effects model.
Table 1. Meta-analyses performed in the review showing the association of alcohol consumption, diet, physical activity, and
smoking exposure with the risk of incident CKD
Exposure Studies Participants Association Ratio [95% CI] Heterogeneity I2, % Evidence Qualitya
Diet factors
Fish 3 21,226 OR, 0.94 [0.86 to 1.02] 5 Low
Fruit 4 28,779 OR, 0.91 [0.79 to 1.06] 60 Very low
Vegetables 5 32,054 OR, 0.79 [0.70 to 0.90] 57 Low
Sugar-sweetened beverages 4 22,760 OR, 1.45 [0.97 to 2.15] 68 Very low
Carbohydrates 3 20,238 OR, 1.08 [0.85 to 1.36] 63 Very low
Protein 3 19,835 OR, 1.08 [0.91 to 1.28] 47 Very low
Phosphorus 3 23,466 OR, 1.00 [0.75 to 1.32] 72 Very low
Potassium 7 32,647 OR, 0.78 [0.65 to 0.94] 48 Low
Sodium 6 43,772 RR, 1.21 [1.06 to 1.38] 59 Moderate
Physical activity
High versus low levels 9 70,828 RR, 0.82 [0.69 to 0.98] 83 Very low
Alcohol consumption
High versus low intakes 13 216,291 RR, 0.87 [0.79 to 0.95] 43 Low
Moderate versus low intakes 7 165,415 RR, 0.86 [0.79 to 0.93] 50 Moderate
Tobacco smoking
Never versus former 6 944,931 OR, 1.09 [1.01 to 1.17] 90 Very low
Current or former versus never 12 985,086 OR, 1.18 [1.10 to 1.27] 81 Very low
Summary of findings, which details the reasons for downgrading evidence quality, is reported in Supplemental Table 7.
a
The overall association estimates for all binary outcomes are ex- included incident CKD in n551 studies, RRT in n517 studies,
pressed as odds ratios (ORs) or relative risks (RRs), depending on GFR decline in n532 studies, and albuminuria in n520 studies.
which ratio was predominantly used in the original data, together
with 95% confidence intervals (95% CIs). For studies reporting Lifestyle Hazards and Incident CKD
changes in mean values at the end of their observation follow-up The association between lifestyle factors and incident CKD
periods, we extracted and tabulated the end of the study values was reported in 51 studies of 1,221,018.14–72 The main re-
with their associated variance. If the association compared a life- sults of the meta-analysis are summarized in Supplemental
style factor with a reference exposure that was not homogenous Figure 2.
with the other included lifestyle exposures (e.g., high adherence
versus low adherence), then the ratio methods were inverted and Diet
combined into the meta-analysis. We used chi-squared testing at The association between 57 different dietary factors with
an a of 0.05 and with the I2 statistic, rated low (,25%), moderate incident CKD was described in 31 studies of 176,625
(25%–75%), or high (.75%) to assess for heterogeneity. Supple- participants. 14,20,24,26,27,30,38,39,41,46,47,49,51,57,60,61,66,69,70
mental Table 2 gives details of planned subgroup and sensitivity Meta-analysis was possible for nine dietary factors (Table 1),
analyses to explore reasons for heterogeneity. In addition to of which higher vegetable intake (OR, 0.79; 95% CI, 0.70 to
these linear analyses for alcohol consumption, we also conduc- 0.90; I2557%; evidence quality: low) and higher potassium
ted a dose-response random effects metaregression analysis to intake (OR, 0.78; 95% CI, 0.65 to 0.94; I2548%; evidence
account for potential changes in the risk estimate at different quality: low) were consistently associated with lower odds of
dosages of alcohol consumption (methods are reported in CKD. Higher sodium intake was associated with higher odds
Supplemental Table 2). of CKD (OR, 1.21; 95% CI, 1.06 to 1.38; I2559%; evidence
quality: moderate) (Figure 1, Table 1). Meta-analysis of
studies evaluating carbohydrate intake, fish, fruit, phospho-
RESULTS rus, protein, and sugar-sweetened beverages consumption
showed no clear association (Table 1).
The electronic search retrieved 45,111 citations, of which a total We were unable to pool data on 32 additional dietary
of 104 studies encompassing 2,755,719 participants met the in- factors reported across 21 studies (Supplemental
clusion criteria (Supplemental Figure 1). The characteristics of Table 4).15,16,18,19,23,25,27,30–32,37,45–47,57,59–62,71,73 We identified
included studies are reported in Supplemental Table 3. Lifestyle 19 dietary factors for which one or both available studies
exposures and associations to kidney outcomes varied across the showed an association in the direction indicating decreased
studies, with n556 reporting diet exposures, n518 reporting risk: cereal fiber, coffee, dairy, fiber, folate, legumes, magnesium,
physical activity, n527 reporting alcohol consumption, and nitrate, nuts, nuts and legumes, plant protein, omega-3, DHA,
n530 reporting tobacco smoking. The outcomes evaluated EPA, vitamin B12, vitamin C, vitamin D, vitamin E, and zinc.
Figure 1. There was a reduced odds of CKD in people who were exposed to a higher vegetable intake and potassium intake and an
increased odds of CKD in people who were exposed to a higher sodium intake. Association of (A) vegetable intake, (B) potassium
intake, (C) sodium intake, and incident CKD. Note that the association estimate for each lifestyle factor is presented on the ratio (OR or
RR) that was predominantly used in the included studies. IV, inverse variance.
Figure 2. There is a significant reduced odds of incident CKD from increased levels of physical activity.
We identified two dietary factors for which one or both available smokers (ever smokers) showed higher odds of CKD (OR, 1.18;
studies indicated a harmful relationship: a higher sodium- 95% CI, 1.10 to 1.27; I2581%; evidence quality: very low)
potassium ratio and red and processed meat consumption (Figure 3, Table 1). Compared with people who had never
(Supplemental Table 4). smoked, former smokers had increased odds of CKD (OR,
1.09; 95% CI, 1.01 to 1.17; I2590%; evidence quality: very
Physical Activity low) in six studies of 944,931 participants (Table 1). There was
The association of different levels of physical activity with in- no consistent association with CKD in two studies comparing
cident CKD was described in ten studies of 78,301 participants current with former smokers (Supplemental Table 4).
(Figure 2, Table 1).28,34,42,44,53,54,58,59,63,72 Meta-analysis of
nine studies showed lower odds of CKD in people who were Lifestyle Risk Factors and Secondary Outcomes (RRT,
more physically active compared with those who were less GFR Decline, and Albuminuria)
physically active (OR, 0.82; 95% CI, 0.69 to 0.98; I2583%; The consistency of associations between the lifestyle factors
evidence quality: very low). identified in our primary analysis and other markers of kidney
damage is shown in Figure 4.
Alcohol Intake The association between lifestyle factors and RRT was de-
The association between alcohol intake and incident scribed in 17 studies of 990,723 participants.33,35,37,75–88 The as-
CKD was described in 14 studies of 211,072 partici- sociation between lifestyle factors and GFR decline was described
pants.22,23,28,40,43,50,52,55,63–65,67,68,72 Compared with lower in 32 studies of 108,936 participants,14,20,23,34,46,49,73,86,89–105 and
intakes, meta-analysis showed that higher consumption of the association between lifestyle factors and albuminuria was de-
alcohol (RR, 0.87; 95% CI, 0.79 to 0.95; I2543%; evidence scribed in 20 studies of 512,403 participants.23,39,41,88,95,100,101,105–116
quality: low) and moderate alcohol consumption (RR, 0.86; Table 2 and Supplemental Material have the details. The risk of
95% CI, 0.79 to 0.93; I2550%; evidence quality: moderate) RRT was higher among current and former smokers (RR, 1.59;
were associated with lower risk of CKD (Figure 3, Table 1). 95% CI, 1.30 to 1.94; I2568%; evidence quality: moderate)
In a dose-response metaregression analysis, we modeled the compared with never smokers (Supplemental Figure 4). The
relationship between incident CKD and alcohol intake using risk of GFR decline was lower in persons with higher potassium
restricted cubic splines (Supplemental Figure 3, Supplemental intake (RR, 0.49; 95% CI, 0.31 to 0.79; I2590%; evidence qual-
Table 5) and found that higher alcohol intake was associated ity: very low) and those who were physically active (OR, 0.76;
with lower risk of incident CKD. This association was generally 95% CI, 0.59 to 0.97; I2575%; evidence quality: very low)
observed throughout the whole range of alcohol consumption (Supplemental Figure 5). The risk of albuminuria was lower
considered with a significant effect size (P values for nonline- among physically active persons (OR, 0.88; 95% CI, 0.81 to
arity 50.03). There was no apparent association when people 0.96; I250%; evidence quality: low) and higher among tobacco
who had never consumed alcohol were compared with those smokers (OR, 1.67; 95% CI, 1.23 to 2.26; I2588%; evidence
who had consumed it in the past (former intake) (Table 1). quality: very low) (Supplemental Figure 6).
Other lifestyle hazards did not have sufficient studies to
Smoking allow meta-analysis, but their associations suggested that
The association between smoking and incident CKD was described they may be potentially protective, be potentially harmful,
in 12 studies of 985,086 participants.28,29,35,43,50,52,53,59,63,65,68,74 or have no association to secondary study outcomes on the
Compared with people who never smoked, current and former basis of the direction of the majority ($50%) of studies.
2 2 2
Heterogeneity: Tau = 0.01; Chi = 67.63, df = 13 (P < 0.00001); I = 81% 0.2 0.5 1 2 5
Test for overall effect: Z = 4.58 (P < 0.00001) Favours [Lower exposure] Favours [Higher exposure]
Figure 3. There was a reduced risk of CKD in people who has a higher alcohol consumption and an increased odds of CKD in people
who were exposed to higher tobacco smoking. Association of (A) alcohol consumption, (B) tobacco smoking, and incident CKD. Note
that the association estimate for each lifestyle factor is presented on the ratio (OR or RR) that was predominantly used in the included
studies. IV, inverse variance.
Results are shown in Supplemental Figure 7 and fully detailed same cohort) did not meaningfully change results (Supple-
in Supplemental Tables 3 and 4. In brief, coffee consumption mental Material). We also conducted a post hoc sensitivity
was associated with lower risk of incident CKD and RRT in analysis on the way that the frequency of alcohol consump-
two of two and one of two studies, respectively. Dairy intake tion was reported: higher “daily” consumption of alcohol had
was associated with lower risk of incident CKD and albumin- no significant association with incident CKD (seven studies;
uria in two of four and one of two studies, respectively. Red RR, 0.98; 95% CI, 0.82 to 1.18; I2543%), whereas higher
and processed meats were associated with lower risk of inci- weekly alcohol consumption was associated with a lower
dent CKD, RRT, and albuminuria in one of two, one of one, risk of incident CKD (six studies; RR, 0.82; 95% CI, 0.75 to
and one of one studies, respectively. 0.90; I2532%) (Supplemental Table 6). Subgroup analysis
was conducted to compare studies with participants with
Subgroup and Sensitivity Analyses baseline GFR $89.9–70 ml/min per 1.73 to participants
Subgroup analyses did not show consistent relationships with with GFR$90 ml/min per 1.73 and showed significant asso-
duration of exposure or geographic region. Replacing indi- ciations between study exposures (except for physical activity)
vidual datasets (substituting alternative reports from the and the risk of incident CKD in studies with participants
RR 0.82 [0.69, 0.98] N = 2/3 (67%) OR 0.77 [0.63, 0.93] RR 0.61 [0.42, 0.90]
Physical activity
9 studies protective 5 studies 4 studies
RR 0.85 [0.77, 0.93] N = 1/2 (50%) OR 0.88 [0.72, 1.07] RR 1.03 [0.88, 1.20]
Alcohol consumption
13 studies protective 5 studies 7 studies
RR 1.21 [1.06, 0.38] N = 1/1 (100%) N = 3/4 (75%) OR 1.01 [0.89, 1.14]
Diet: Sodium
6 studies harmful harmful 3 studies
OR 1.18 [1.10, 1.29] RR 1.59 [1.30, 0.94] N = 1/4 (25%) OR 1.67 [1.23, 2.26]
Tobacco smoking
12 studies 8 studies protective 7 studies
Protective Harmful
>50% studies suggest protective No association
>50% studies suggest harmful No evidence
Figure 4. The association of vegetable intake, potassium intake, physical activity, alcohol consumption, sodium intake, and tobacco
smoking was consistent across incident CKD and KRT, however had varying association to GFR decline and albuminuria. NA, not
applicable. aThe four studies of sodium intake and GFR decline were not meta-analyzable because two of the studies reported means
and variance but not ratio data.
baseline GFR $89.9–70 ml/min per 1.73. The association of where 23% were unclear because data were not analyzed using
vegetable intake and incident CKD was robust in both base- ratio statistics to allow statistically data pooling.
line GFR subgroups. Whereas in studies reporting participant
baseline GFR $90 ml/min per 1.73 (which were 37 of 104; 12
studies were able to be meta-analyzed), these associations did DISCUSSION
not attain statistical significance. However, the direction and
magnitude of observed risks in both strata were very similar This study evaluated the association between modifiable life-
(Supplemental Table 6). style factors and incident CKD in the community. In predom-
inately low- to very low–certainty evidence, we found higher
Quality Assessment intake of potassium and vegetables, lower intake of sodium,
The certainty (GRADE) of the evidence for each outcome physical activity, moderate alcohol consumption, and avoidance
meta-analyzed is detailed in Supplemental Table 8. The overall of tobacco smoking to be consistently associated with lower risk
risk of bias across the studies was low (Supplemental Figures 8 of CKD. We identified consistency with these main results when
and 9). Eighty-eight percent of studies had a low risk of bias for we examined their association with other surrogates of kidney
sample, whereas 10% were rated high due to being conducted damage: RRT, GFR decline, and albuminuria.
in populations with established diseases (such as type 2 dia- Higher vegetable intake is associated with higher potassium
betes or hypertension) that are known to increase the risk of intake: both dietary factors were consistently associated with
incident CKD. With respect to follow-up, 82% of studies had fewer kidney outcomes. Low vegetable consumption is a
low risk of bias, and bias was judged unclear in 13% of studies strong predictor of mortality in the general popula-
that reported ,4 years of follow-up; three studies (5%) had a tion.117,118 The mechanism is likely, at least in part, through
high risk of bias due to very short follow-up periods the intermediary diseases that are themselves risks for CKD:
of ,2 years. The risk of bias in the "exposure" domain was cardiovascular disease, hypertension, diabetes, obesity, and
rated high in 92% of studies because lifestyle behaviors were metabolic syndrome. 119 Whether there is additionally a
self-reported; objective measures were used in eight studies more direct relationship between vegetable or potassium in-
(8%). Outcomes were captured using appropriate tools (and take and kidney health requires further study.
rated low risk) in 88% of studies, and rated high risk in 6% of In our analysis, higher potassium intake was associated with
the studies. Potential analysis bias was low in 73% of studies, reduced incident CKD and GFR decline, adding to the growing
Table 2. Meta-analyses performed in the review showing the association of alcohol consumption, diet, physical activity, and
smoking exposure with secondary outcomes (RRT, GFR decline, and albuminuria)
Exposure Studies Participants Association Ratio [95% CI] Heterogeneity I2, % Evidence Qualitya
b
RRT
Tobacco smoking
Former versus current 7 1,126,913 RR, 1.25 [1.13 to 1.39] 39 Moderate
Current or former versus never 8 1,230,390 RR, 1.59 [1.30 to 1.94] 68 Moderate
GFR declinec
Diet factors
Potassium 4 10,729 RR, 0.49 [0.31 to 0.79] 81 Very low
Protein 5 18,507 OR, 1.07 [0.96 to 1.19] 42 Low
Physical activity
High versus low levels 5 15,161 OR, 0.77 [0.63 to 0.93] 75 Very low
Alcohol consumption
High versus low intakes 5 16,580 OR, 0.88 [0.72 to 1.07] 15 Very low
Albuminuria
Diet factors
Sodium 3 11,751 OR, 1.01 [0.89 to 1.14] 0 Very low
Physical activity
High versus low levels 4 110,154 RR, 0.88 [0.81 to 0.96] 0 Low
Alcohol consumption
High versus low intakes 7 220,479 RR, 1.03 [0.88 to 1.20] 59 Low
Tobacco smoking
Current or former versus never 7 184,302 OR, 1.67 [1.23 to 2.26] 88 Very low
a
Summary of findings, which details the reasons for downgrading evidence quality, is reported in Supplemental Table 7.
b
Meta-analysis was not possible for diet factors, physical activity, or alcohol consumption.
c
Meta-analysis was not possible for tobacco smoking.
evidence base for the protective association of potassium intake higher sodium intakes are detrimental, and public health ef-
in other chronic conditions. In the Prospective Urban Rural forts should be prioritized to reduce population-wide sodium
Epidemiology study, higher potassium excretion was associ- consumption. More work is needed on the optimal level of
ated with a lower risk of death and cardiovascular events across sodium restriction; our analysis focused on high versus low
.18 countries worldwide.120 Higher intakes of potassium have intake, but in the original studies, the highest exposures of
also been demonstrated to lower the risk of stroke121 and lower salt defining high intake varied, ranging from $9.88 to
BP in adults with and without hypertension.122,123 In patients 16.27 g/d (sodium $172–283 mmol/d).
in the ONTARGET, a high-vascular risk cohort (most of whom A healthy diet pattern, characterized by higher intake of
did not have CKD at baseline), potassium intake was associated fruit, vegetables, low-fat dairy, fiber, and whole grains and
with a less rapid GFR decline or need for RRT.124 In ONTAR- lower intake of sodium, red meat, and sugar, has been associ-
GET, the relationship between potassium intake and the out- ated with a 30% reduced odds of kidney damage.11,12 Many of
come did not seem to differ by baseline level of albuminuria, the individual diet factors we identified are characteristic of a
but there was a statistically significant interaction suggesting higher diet quality and overall healthier diet pattern. For ex-
that the protective association with potassium might not be ob- ample, taking all of the data into account, diet factors possibly
served in patients with GFR,45 ml/min (advanced CKD). In protective against incident CKD in our study included fruit,
these patients, potassium and with it, fruits and vegetables are dairy, vegetables, fiber, legumes, nuts, potassium, and unsat-
often restricted. When discussing dietary advice, it is critical to urated fatty acids. In contrast, diet factors possibly harmful to
distinguish whether we are concerned with prevention of in- incident CKD in our study included excessive carbohydrate
cident CKD, prevention of progression (and at what level of intake, high energy intake, red and processed meats, saturated
GFR), or management of hyperkalemia and management of fat, sodium, sodium-potassium ratio, and sugar-sweetened
the risk of hyperkalemia.125 beverages. The protective factors, characteristics of a healthy
Similarly, higher sodium intake showed a consistent asso- diet, are also strongly associated with reduced risk of over-
ciation with increased risk of incident CKD, RRT, and GFR weight and obesity,128 which are linked with increased risk
decline. This supports the growing evidence from clinical of incident CKD.129 These associations fundamentally align
trials demonstrating the effect of sodium intake on BP, pro- with guidelines for general healthy eating and cardiovascular
teinuria, and extracellular volume in CKD.126 Despite existing disease prevention, which is helpful and convenient in guiding
controversy surrounding reverse causality at extremely low public health policy for CKD prevention that aligns with
sodium intakes,127 our review supports the message that prevention of other chronic diseases.130
We found that higher levels of physical activity were con- smoking on risk of albuminuria observed in our study, consis-
sistently associated with a lower risk of all study outcomes tent with previous observations.147 The harmful effects of
evaluated. This aligns with the majority of existing studies, smoking may also be mediated by endothelial cell dysfunction,
which suggest that physical activity is associated with reduced inducing advanced glycation end products due to glycotoxins
risk of kidney damage through attenuating cardiovascular dis- present in cigarettes, increased vascular permeability and path-
ease risk,131,132 mortality,133 and rapid declines in kidney ologic vascular changes of kidney disease, and insulin resistance
function.134 Whether a relationship independent of cardiovas- in diabetic and nondiabetic subjects.145 Taken together, these
cular and vascular damage truly exists is not known on the risk factors for kidney function decline clearly support a public
basis of current data. Possible mechanisms include reduced health message to avoid tobacco smoking to prevent kidney
BP, improved glycemic control, angiogenesis, and vascular disease as well as cardiovascular disease.
regeneration by the upregulation of endothelial nitric oxide This study has strengths as the first evidence synthesis of
production and other antioxidant enzymes.135 Although the lifestyle risk and protective factors and primary prevention of
measurement of physical activity was heterogeneous across CKD: it is broad in scope, and we have taken a rigorous ap-
the included studies, the most common categorization of proach to meta-analysis. However, this study has limitations.
higher levels of activity was defined as at least 30 minutes a First, there was scarce evidence for some exposures, particu-
day.44,59 These findings suggest that public health messages larly for diet factors, precluding meta-analysis; for these
should emulate those of primary prevention of cardiovascular factors, our output is simply descriptive. Second, there was
disease: to achieve and maintain a moderate level of physical variation in the definition of the different exposures across
activity of at least 30 minutes a day and avoid extended periods the different lifestyle factors, in the degree of adjustment for
of being sedentary.136,137 potential confounders, and in definition and reporting for
We found an inverse association between alcohol con- study outcomes. For example, there were no consistent adjust-
sumption and incident CKD, RRT, and GFR decline. Although ments for activity, age, education, and sex (as reported in
the relationship between alcohol consumption and kidney Supplemental Table 3) in all studies included in the analyses.
function decline has not been consistent in the current body Furthermore, ten different definitions of incident CKD were
of literature, many clinical studies have indeed shown that used in the included studies, with similar heterogeneity in the
moderate alcohol consumption is associated with lower oc- definitions of the secondary outcomes. This leads to hetero-
currence of CKD.40,55,64,68 In contrast, existing literature has geneity in the analysis and precludes clear recommendations
shown that chronic alcoholism is associated with CKD,56,65 in terms of possible targets, goals, or thresholds. Third, al-
leading to the hypothesis that excessive alcohol consumption though we were able to perform meta-analysis to account
directly damages the kidney, independent of liver dam- for the known nonlinear associations of alcohol consumption
age.138,139 Similarly, some reports included in our review and kidney outcomes, we were unaware of any such assumptions
found that heavy drinking, in contrast with moderate drink- for the other lifestyle factors, and the original studies did not
ing, was associated with increased risk of albuminuria and commonly report their associations in this way. The original
CKD.56,105,140 This distinction (that moderate, but not high, papers usually made a linear assumption and analyzed data
alcohol consumption is associated with lower prevalence with linear models. It is possible that nonlinear effects or
of albuminuria and CKD compared with abstinence or with threshold effects may be missed in this way. Fourth, because
lower intakes) is consistent with a number of other stud- we evaluated cohort studies and not randomized trials, the
ies.139,141,142 There is also the possibility that unadjusted con- findings do not imply causality. However, there were no trials
founders may be at play: for example, social integration as a of this nature in our search, and it is unlikely that studies of
product of moderate alcohol consumption and overall well- sufficient rigor, size, and duration will be conducted, given the
being, which is good for health.143 It would seem, despite resource-intensive nature of such trials. We combined RR and
minor international variations144 and the heterogenous cate- OR; for rare events, these are numerically similar, although
gorization for the higher alcohol consumption (ranging in calculated differently. Fifth, we used the NOS to assess bias
grams per day [range, 20–48 g/d], grams per week [range, rather than the more comprehensive ROBINS-I tool148 be-
210 g/wk], total drinks per day [range, .1–4 drinks per cause our resources to complete the study were finite. Finally,
day], and drinks per week [range, 5–.15 drinks per week]), in order to conduct meta-analysis, each putative risk and pro-
that moderate alcohol consumption in line with public health tective factor was considered separately, and we do not know
guidelines seems safe and unlikely to cause kidney damage in whether there are quantitatively important interactions be-
the general population. tween the different factors (e.g., does the effect of carbohy-
Current and former smokers had higher risk of incident drate intake differ at different levels of physical activity), as no
CKD compared with those who never smoked. Tobacco smok- study has evaluated such hypotheses.
ing has been previously identified as a risk factor for incident Allowing for the limitations of inferences from observa-
CKD and RRT in the general population.145 Although the mech- tional studies, our work suggests that the lifestyle factors
anism is uncertain, smoking may lead to insulin resistance,146 that we have evaluated are probably those that patients have
which might explain the large effect size (67% increase in risk) of the most control over. However, these should not be the only
ones that public health authorities target. For example, other G. Su is supported by the European Renal Association-European Dialysis
potentially modifiable hazards known to increase the risk of and Transplantation Association Young Fellowship Programme and by
Guangdong Provincial Hospital of Traditional Chinese Medicine grant
kidney damage include environmental pollution,149,150 exces- YN2018QL08, outside the submitted work. H. Xu is supported by Loo and
sive body weight related to nutrition,129 inappropriate use of Hans Osterman’s Foundation, outside the submitted work. All remaining
medications (including analgesics, antibiotics, antiretrovirals, authors have nothing to disclose.
or proton-pump inhibitors),151 and heavy metal exposure and
intake (in particular, cadmium).152 Finally, society has a fur-
FUNDING
ther role to play in facilitating more personal choices through
food labeling, alcohol and tobacco labeling, health policy
None.
around exercise, policy around the ease of walking and cycling
in the built environment, and walkability.
We recognize that these findings are on the basis of non-
ACKNOWLEDGMENTS
randomized data with risk of residual and unknown con-
founding, particularly in the setting of studies with low to
Prof. Juan J. Carrero and Dr. Jaimon Kelly contributed to the study
very low certainty of the evidence. However, the generalizabil-
conception; Ms. Ailema González-Ortiz, Dr. Jaimon Kelly, Dr. Skye
ity and overall importance of these findings should not be
Marshall, Dr. Xindong Qin, Dr. Guobin Su, Dr. Hong Xu, and Dr. La
undermined, as they are well in line with public health rec-
Zhang conducted the literature search, extracted data, and appraised
ommendations for preventing related chronic conditions and
risk of bias; Dr. Jaimon Kelly conducted the data analysis; Dr. Jaimon
general healthy lifestyle principles. Therefore, the results can
Kelly was responsible for writing the first draft of the manuscript;
represent important indications for public policy and
Prof. Juan J. Carrero reviewed the first draft of the manuscript;
advocacy and be hypothesis generating for future randomized
Dr. Katrina Campbell and Dr. Catherine M. Clase critically reviewed
trials. It is not trivial to ask patients to change their lifestyle.
the manuscript; and all authors read and approved the final version of
Diet, in particular, is a highly social and cultural issue; further-
the manuscript.
more, many patients are constrained by external factors, in-
This work is endorsed and performed within the European
cluding economic factors, in their decisions. For these reasons,
Renal Nutrition working group, an initiative of and supported by
we believe that lifestyle intervention studies to address unan-
the European Renal Association-European Dialysis and Trans-
swered questions examining the whole range of clinically im-
plantation Association.
portant outcomes, including the incidence and progression of
CKD, continue to be justified.
Increased vegetable and potassium intake, physical activity, SUPPLEMENTAL MATERIAL
and moderate alcohol consumption were consistently associ-
ated with reduced risk of CKD. Increased dietary salt intake This article contains the following supplemental material online at
and tobacco smoking were consistently associated with in- http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/
creased risk of CKD. In the absence of trial evidence, these ASN.2020030384/-/DCSupplemental.
findings can inform both public health recommendations and Supplemental Figure 1. Study flow diagram.
patient-centered discussions in clinical practice. Supplemental Figure 2. Summary of the associations between
modifiable lifestyle risk and protective factors and risk of incident
CKD.
DISCLOSURES Supplemental Figure 3. Dose-response relationship between al-
cohol intake (grams per day) and incident CKD estimated with a
K. Campbell has received consultation, advisory board membership, or random effect metaregression-restricted cubic spline model.
research funding from the Dietitians Association of Australia, Nestle Health
Supplemental Figure 4. Association of tobacco smoking and
Sciences, and Queensland Health, all outside the submitted work. J.J. Carrero
reports grant funding from Astellas, AstraZeneca, Vetenskapsrådet (2019- ESKD.
01059), and ViforPharma; consulting for AstraZeneca and Baxter; speaker Supplemental Figure 5. Summary of the associations between
fees for Abbott, AstraZeneca, Fresenius, Nutricia, and ViforPharma; and sup- modifiable lifestyle risk and protective factors and risk of GFR
port from the Swedish Heart and Lung Foundation, all outside the submitted decline.
work. C.M. Clase has received consultation, advisory board membership, or
Supplemental Figure 6. Summary of the associations between
research funding from Amgen, Astellas, AstraZeneca, Baxter, Boehringer-
Ingelheim, Fresenius Medical Care, Janssen, Johnson & Johnson, Leo Pharma, modifiable lifestyle risk and protective factors and risk of
the Ontario Ministry of Health, Pfizer, Relyspa, Sanofi, and Vifor, all outside albuminuria.
the submitted work. J.T. Kelly has received consultancy fees from Amgen for Supplemental Figure 7. Consistency of associations in lifestyle
travel and lecture presentations and consultancy fees from HealthCert for factors that could not be statistically pooled across the markers of
topics unrelated to the submitted work. J.T. Kelly is supported through a
kidney function decline.
Griffith University Postdoctoral Research Fellowship, outside the submitted
work. A. González-Ortiz is supported by School of Medicine, Programa de Supplemental Figure 8. Risk of bias across the included studies.
Doctorado en Ciencias Médicas, Odontológicas y de la Salud, Consejo Nacio- Supplemental Figure 9. Individual assessment of risk of bias across
nal de Ciencia y Tecnología grant CVU 373297, outside the submitted work. the included studies.
Supplemental Figure 10. Funnel plots for lifestyle hazards and 12. van Westing AC, Küpers LK, Geleijnse JM: Diet and kidney function: A
incident CKD. literature review. Curr Hypertens Rep 22: 14, 2020
13. Wells G, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al.:
Supplemental Figure 11. Funnel plots for lifestyle hazards and
Newcastle-Ottawa Quality Assessment Scale Cohort Studies, Ot-
secondary outcomes. tawa, Ontario, Canada, Ottawa Hospital Research Institute, 2014
Supplemental Material. Data synthesis for secondary outcomes. 14. Araki S, Haneda M, Koya D, Kondo K, Tanaka S, Arima H, et al.: Urinary
Supplemental Table 1. Search terms used across the electronic potassium excretion and renal and cardiovascular complications in
databases. patients with type 2 diabetes and normal renal function. Clin J Am Soc
Supplemental Table 2. Summary of the methods relating to data Nephrol 10: 2152–2158, 2015
15. Asghari G, Yuzbashian E, Mirmiran P, Azizi F: The association between
extraction, risk of bias, metaregression, and planned subgroup and
Dietary Approaches to Stop Hypertension and incidence of chronic
sensitivity analyses. kidney disease in adults: The Tehran Lipid and Glucose Study.
Supplemental Table 3. Characteristics of the included studies. Nephrol Dial Transplant 32[Suppl 2]: ii224-ii230, 2017
Supplemental Table 4. Lifestyle hazards and kidney disease out- 16. Asghari G, Yuzbashian E, Shahemi S, Gaeini Z, Mirmiran P, Azizi F:
comes from individual studies that could not be statistically pooled Dietary total antioxidant capacity and incidence of chronic kidney
disease in subjects with dysglycemia: Tehran Lipid and Glucose Study.
into meta-analysis.
Eur J Nutr 57: 2377–2385, 2018
Supplemental Table 5. Summary of the studies between alcohol 17. Baggio B, Budakovic A, Perissinotto E, Maggi S, Cantaro S, Enzi G,
intake and incident CKD. et al; ILSA Working Group: Atherosclerotic risk factors and renal
Supplemental Table 6. Subgroup analysis for incident CKD. function in the elderly: The role of hyperfibrinogenaemia and smok-
Supplemental Table 7. Results from sensitivity analysis. ing. Results from the Italian Longitudinal Study on Ageing (ILSA).
Supplemental Table 8. GRADE table summarizing the quality of Nephrol Dial Transplant 20: 114–123, 2005
18. Bahadoran Z, Mirmiran P, Ghasemi A, Carlström M, Azizi F, Hadaegh F:
the evidence for each outcome.
Association between dietary intakes of nitrate and nitrite and the risk of
hypertension and chronic kidney disease: Tehran Lipid and Glucose
Study. Nutrients 8: 811, 2016
REFERENCES 19. Bahadoran Z, Mirmiran P, Golzarand M, Davudabadi-Farahani R, Azizi
F: Dietary animal-derived L-arginine intakes and risk of chronic kidney
disease: A 6-year follow-up of Tehran Lipid and Glucose Study. Iran
1. Bello AK, Levin A, Tonelli M, Okpechi IG, Feehally J, Harris D, et al.:
J Kidney Dis 11: 352–359, 2017
Assessment of global kidney health care status. JAMA 317:
20. Bahadoran Z, Mirmiran P, Momenan AA, Azizi F: Allium vegetable
1864–1881, 2017
intakes and the incidence of cardiovascular disease, hypertension,
2. GBD Chronic Kidney Disease Collaboration: Global, regional, and
chronic kidney disease, and type 2 diabetes in adults: A longitudinal
national burden of chronic kidney disease, 1990-2017: A systematic
follow-up study. J Hypertens 35: 1909–1916, 2017
analysis for the Global Burden of Disease Study 2017. Lancet 395:
21. Bomback AS, Derebail VK, Shoham DA, Anderson CA, Steffen LM,
709–733, 2020
Rosamond WD, et al.: Sugar-sweetened soda consumption, hyper-
3. Bello A, Levin A, Tonelli M, Okpechi IG, Feehally J, Harris D, et al;
uricemia, and kidney disease. Kidney Int 77: 609–616, 2010
Global Kidney Health Atlas: A Report by the International Society of
22. Buja A, Scafato E, Baggio B, Sergi G, Maggi S, Rausa G, et al.: Renal
Nephrology on the Current State of Organization and Structures for
Kidney Care across the Globe, Brussels, Belgium, International Soci- impairment and moderate alcohol consumption in the elderly. Results
ety of Nephrology, 2017 from the Italian Longitudinal Study on Aging (ILSA). Public Health Nutr
4. Kelly JT, Campbell KL, Carrero JJ: Primary versus secondary pre- 14: 1907–1918, 2011
vention of chronic kidney disease: The case of dietary protein. J Ren 23. Dunkler D, Dehghan M, Teo KK, Heinze G, Gao P, Kohl M, et al;
Nutr 28: 225–228, 2018 ONTARGET Investigators: Diet and kidney disease in high-risk in-
5. Åkesson A, Weismayer C, Newby PK, Wolk A: Combined effect of low- dividuals with type 2 diabetes mellitus. JAMA Intern Med 173:
risk dietary and lifestyle behaviors in primary prevention of myocardial 1682–1692, 2013
infarction in women. Arch Intern Med 167: 2122–2127, 2007 24. Dunkler D, Kohl M, Teo KK, Heinze G, Dehghan M, Clase CM, et al.:
6. Chiuve SE, Rexrode KM, Spiegelman D, Logroscino G, Manson JE, Dietary risk factors for incidence or progression of chronic kidney
Rimm EB: Primary prevention of stroke by healthy lifestyle. Circulation disease in individuals with type 2 diabetes in the European Union.
118: 947–954, 2008 Nephrol Dial Transplant 30[Suppl 4]: iv76-iv85, 2015
7. Chiuve SE, Fung TT, Rexrode KM, Spiegelman D, Manson JE, 25. Ejtahed HS, Angoorani P, Asghari G, Mirmiran P, Azizi F: Dietary ad-
Stampfer MJ, et al.: Adherence to a low-risk, healthy lifestyle and risk vanced glycation end products and risk of chronic kidney disease.
of sudden cardiac death among women. JAMA 306: 62–69, 2011 J Ren Nutr 26: 308–314, 2016
8. World Health Organization: Global Status Report on Non- 26. Farhadnejad H, Asghari G, Emamat H, Mirmiran P, Azizi F: Low-
communicable Diseases 2014, Geneva, Switzerland, World Health carbohydrate high-protein diet is associated with increased risk of
Organization, 2014 incident chronic kidney diseases among Tehranian adults. J Ren Nutr
9. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, 29: 343–349, 2019
et al.: Meta-analysis of observational studies in epidemiology: A pro- 27. Farhadnejad H, Asghari G, Mirmiran P, Yuzbashian E, Azizi F: Micro-
posal for reporting. Meta-analysis Of Observational Studies in Epi- nutrient intakes and incidence of chronic kidney disease in adults:
demiology (MOOSE) group. JAMA 283: 2008–2012, 2000 Tehran Lipid and Glucose Study. Nutrients 8: 217, 2016
10. Kidney Disease: Improving Global Outcomes (KDIGO): CKD evalua- 28. Foster MC, HwangS-J, Massaro JM, Jacques PF, Fox CS, Chu AY:
tion and management. 2012. Available at: https://kdigo.org/guide- Lifestyle factors and indices of kidney function in the Framingham
lines/ckd-evaluation-and-management/. Accessed February 19, 2020 Heart Study. Am J Nephrol 41: 267–274, 2015
11. Bach KE, Kelly JT, Palmer SC, Khalesi S, Strippoli GFM, Campbell KL: 29. Fox CS, Larson MG, Leip EP, Culleton B, Wilson PW, Levy D: Predictors
Healthy dietary patterns and incidence of CKD: A meta-analysis of of new-onset kidney disease in a community-based population. JAMA
cohort studies. Clin J Am Soc Nephrol 14: 1441–1449, 2019 291: 844–850, 2004
30. Gopinath B, Harris DC, Flood VM, Burlutsky G, Brand-Miller J, Mitchell 48. Miyatake N, Moriyasu H, Sakano N, Tada S, Suzue T, Hirao T: Influence
P: Carbohydrate nutrition is associated with the 5-year incidence of of cigarette smoking on estimated glomerular filtration rate (eGFR) in
chronic kidney disease. J Nutr 141: 433–439, 2011 Japanese male workers. Acta Med Okayama 64: 385–390, 2010
31. Gopinath B, Harris DC, Flood VM, Burlutsky G, Mitchell P: Associations 49. Mun KH, Yu GI, Choi BY, Kim MK, Shin MH, Shin DH: Association of
between dairy food consumption and chronic kidney disease in older dietary potassium intake with the development of chronic kidney
adults. Sci Rep 6: 39532, 2016 disease and renal function in patients with mildly decreased kidney
32. Haring B, Selvin E, Liang M, Coresh J, Grams ME, Petruski-Ivleva N, function: The Korean Multi-Rural Communities cohort study. Med Sci
et al.: Dietary protein sources and risk for incident chronic kidney Monit 25: 1061–1070, 2019
disease: Results from the Atherosclerosis Risk in Communities (ARIC) 50. Nakanishi N, Fukui M, Tanaka M, Toda H, Imai S, Yamazaki M, et al.:
Study. J Ren Nutr 27: 233–242, 2017 Low urine pH Is a predictor of chronic kidney disease. Kidney Blood
33. Haroun MK, Jaar BG, Hoffman SC, Comstock GW, Klag MJ, Coresh J: Press Res 35: 77–81, 2012
Risk factors for chronic kidney disease: A prospective study of 23,534 51. Nam KH, An SY, Joo YS, Lee S, Yun HR, Jhee JH, et al.: Carbohydrate-
men and women in Washington County, Maryland. J Am Soc Nephrol rich diet is associated with increased risk of incident chronic kidney
14: 2934–2941, 2003 disease in non-diabetic subjects. J Clin Med 8: 793, 2019
34. Hawkins M, Newman AB, Madero M, Patel KV, Shlipak MG, Cooper J, 52. Noborisaka Y, Ishizaki M, Yamada Y, Honda R, Yokoyama H, Miyao M,
et al.: TV watching, but not physical activity, is associated with change et al.: The effects of continuing and discontinuing smoking on the
in kidney function in older adults. J Phys Act Health 12: 561–568, 2015 development of chronic kidney disease (CKD) in the healthy middle-
35. Hippisley-Cox J, Coupland C: Predicting the risk of chronic kidney aged working population in Japan. Environ Health Prev Med 18:
disease in men and women in England and Wales: Prospective deri- 24–32, 2013
vation and external validation of the QKidney scores. BMC Fam Pract 53. Obermayr RP, Temml C, Knechtelsdorfer M, Gutjahr G, Kletzmayr J,
11: 49, 2010 Heiss S, et al.: Predictors of new-onset decline in kidney function in a
36. Hu EA, Lazo M, Rosenberg SD, Grams ME, Steffen LM, Coresh J, et al.:
general middle-European population. Nephrol Dial Transplant 23:
Alcohol consumption and incident kidney disease: Results from the ath-
1265–1273, 2008
erosclerosis risk in communities study. J Ren Nutr 30: 22–30, 2020
54. Ogunmoroti O, Allen NB, Cushman M, Michos ED, Rundek T, Rana JS,
37. Hu EA, Selvin E, Grams ME, Steffen LM, Coresh J, Rebholz CM: Coffee
et al.: Association between life’s simple 7 and noncardiovascular dis-
consumption and incident kidney disease: Results from the Atheroscle-
ease: The Multi-Ethnic Study of Atherosclerosis. J Am Heart Assoc 5:
rosis Risk in Communities (ARIC) study. Am J Kidney Dis 72: 214–222,
e003954, 2016
2018
55. Okada Y, Uehara S, Shibata M, Koh H, Oue K, Kambe H, et al.: Habitual
38. Jhee JH, Kee YK, Park JT, Chang TI, Kang EW, Yoo TH, et al.: A diet
alcohol intake modifies relationship of uric acid to incident chronic
rich in vegetables and fruit and incident CKD: A community-based
kidney disease. Am J Nephrol 50: 55–62, 2019
prospective cohort study. Am J Kidney Dis 74: 491–500, 2019
56. Pan CS, Ju TR, Lee CC, Chen YP, Hsu CY, Hung DZ, et al.: Alcohol use
39. Kieneker LM, Bakker SJ, de Boer RA, Navis GJ, Gansevoort RT,
disorder tied to development of chronic kidney disease: A nationwide
Joosten MM: Low potassium excretion but not high sodium excretion
database analysis. PLoS One 13: e0203410, 2018
is associated with increased risk of developing chronic kidney disease.
57. Park I, Xun P, Tsinovoi CL, Klemmer P, Liu K, He K: Intakes of long-
Kidney Int 90: 888–896, 2016
chain omega-3 polyunsaturated fatty acids and non-fried fish in
40. Koning SH, Gansevoort RT, Mukamal KJ, Rimm EB, Bakker SJ, Joosten
relation to incidence of chronic kidney disease in young adults: A
MM; PREVEND Study Group: Alcohol consumption is inversely asso-
25-year follow-up. Eur J Nutr 59: 399–407, 2020
ciated with the risk of developing chronic kidney disease. Kidney Int
58. Ramezankhani A, Tohidi M, Azizi F, Hadaegh F: Application of survival
87: 1009–1016, 2015
tree analysis for exploration of potential interactions between pre-
41. Lee CC, Howard BV, Mete M, Wang H, Jolly S, Adler AI: Association
between fish consumption and nephropathy in American Indians—the dictors of incident chronic kidney disease: A 15-year follow-up study.
Strong Heart study. J Ren Nutr 22: 221–227, 2012 J Transl Med 15: 240, 2017
42. LinH-C, PengC-H, ChiouJ-Y, HuangC-N: Physical activity is associated 59. Rebholz CM, Anderson CA, Grams ME, Bazzano LA, Crews DC, Chang
with decreased incidence of chronic kidney disease in type 2 diabetes AR, et al.: Relationship of the American Heart Association’s impact
patients: A retrospective cohort study in Taiwan. Prim Care Diabetes goals (life’s simple 7) with risk of chronic kidney disease: Results from
8: 315–321, 2014 the Atherosclerosis Risk in Communities (ARIC) cohort study. J Am
43. Michishita R, Matsuda T, Kawakami S, Tanaka S, Kiyonaga A, Tanaka Heart Assoc 5: e003192, 2016
H, et al.: The association between changes in lifestyle behaviors and 60. Rebholz CM, Coresh J, Grams ME, Steffen LM, Anderson CA, Appel
the incidence of chronic kidney disease (CKD) in middle-aged and LJ, et al.: Dietary acid load and incident chronic kidney disease: Re-
older men. J Epidemiol 27: 389–397, 2017 sults from the ARIC study. Am J Nephrol 42: 427–435, 2015
44. Michishita R, Matsuda T, Kawakami S, Tanaka S, Kiyonaga A, Tanaka 61. Rebholz CM, Crews DC, Grams ME, Steffen LM, Levey AS, Miller ER
H, et al.: The joint impact of habitual exercise and glycemic control on 3rd, et al.: DASH (Dietary Approaches to Stop Hypertension) diet and
the incidence of chronic kidney disease (CKD) in middle-aged and risk of subsequent kidney disease. Am J Kidney Dis 68: 853–861, 2016
older males. Environ Health Prev Med 22: 76, 2017 62. Rebholz CM, Young BA, Katz R, Tucker KL, Carithers TC, Norwood AF,
45. Mirmiran P, Bahadoran Z, Golzarand M, Asghari G, Azizi F: Con- et al.: Patterns of beverages consumed and risk of incident kidney
sumption of nitrate containing vegetables and the risk of chronic disease. Clin J Am Soc Nephrol 14: 49–56, 2019
kidney disease: Tehran Lipid and Glucose Study. Ren Fail 38: 63. RyooJ-H, ChoiJ-M, OhC-M, KimM-G: The association between uric
937–944, 2016 acid and chronic kidney disease in Korean men: A 4-year follow-up
46. Mirmiran P, Nazeri P, Bahadoran Z, Khalili-Moghadam S, Azizi F: Di- study. J Korean Med Sci 28: 855–860, 2013
etary sodium to potassium ratio and the incidence of chronic kidney 64. Schaeffner ES, Kurth T, de Jong PE, Glynn RJ, Buring JE, Gaziano JM:
disease in adults: A longitudinal follow-up study. Prev Nutr Food Sci Alcohol consumption and the risk of renal dysfunction in apparently
23: 87–93, 2018 healthy men. Arch Intern Med 165: 1048–1053, 2005
47. Mirmiran P, Yuzbashian E, Asghari G, Sarverzadeh S, Azizi F: Dietary 65. Shankar A, Klein R, Klein BE: The association among smoking, heavy
fibre intake in relation to the risk of incident chronic kidney disease. Br drinking, and chronic kidney disease. Am J Epidemiol 164: 263–271,
J Nutr 119: 479–485, 2018 2006
66. Sugiura T, Takase H, Ohte N, Dohi Y: Dietary salt intake is a significant 86. Van Noordenne ND, Olde Engberink RHG, Van Den Hoek TC, Van
determinant of impaired kidney function in the general population. Den Born BJH, Vogt L: Associations of 15-year average potassium
Kidney Blood Press Res 43: 1245–1254, 2018 intake with long term renal outcome in the outpatient clinical setting.
67. Weiner DE, Tighiouart H, Elsayed EF, Griffith JL, Salem DN, Levey AS: Nephrol Dial Transplant 31: i11, 2016
Uric acid and incident kidney disease in the community. J Am Soc 87. Jin A, KohW-P, Chow KY, YuanJ-M, Jafar TH: Smoking and risk of
Nephrol 19: 1204–1211, 2008 kidney failure in the Singapore Chinese health study. PLoS One 8:
68. Yamagata K, Ishida K, Sairenchi T, Takahashi H, Ohba S, Shiigai T, e62962, 2013
et al.: Risk factors for chronic kidney disease in a community-based 88. Wadén J, Tikkanen HK, Forsblom C, Harjutsalo V, Thorn LM,
population: A 10-year follow-up study. Kidney Int 71: 159–166, 2007 Saraheimo M, et al; FinnDiane Study Group: Leisure-time physical
69. Yoon CY, Noh J, Lee J, Kee YK, Seo C, Lee M, et al.: High and low activity and development and progression of diabetic nephropathy in
sodium intakes are associated with incident chronic kidney disease in type 1 diabetes: The FinnDiane study. Diabetologia 58: 929–936,
patients with normal renal function and hypertension. Kidney Int 93: 2015
921–931, 2018 89. Cirillo M, Cavallo P, Bilancio G, Lombardi C, Terradura Vagnarelli
70. Yoon CY, Park JT, Jhee JH, Noh J, Kee YK, Seo C, et al.: High dietary O, Laurenzi M: Low protein intake in the population: Low risk of
phosphorus density is a risk factor for incident chronic kidney disease kidney function decline but high risk of mortality. J Ren Nutr 28:
development in diabetic subjects: A community-based prospective 235–244, 2018
cohort study. Am J Clin Nutr 106: 311–321, 2017 90. Deriaz D, Guessous I, Vollenweider P, Devuyst O, Burnier M, Bochud
71. Yuzbashian E, Asghari G, Mirmiran P, Zadeh-Vakili A, Azizi F: Sugar- M, et al.: Estimated 24-h urinary sodium and sodium-to-potassium
sweetened beverage consumption and risk of incident chronic kidney ratio are predictors of kidney function decline in a population-based
disease: Tehran Lipid and Glucose Study. Nephrology (Carlton) 21: study. J Hypertens 37: 1853–1860, 2019
608–616, 2016 91. Esmeijer K, Geleijnse J, De Fijter J, Kromhout D, Hoogeveen E: High
72. Kanda E, Muneyuki T, Suwa K, Nakajima K: Alcohol and exercise affect protein intake accelerates kidney function decline in post-myocardial
declining kidney function in healthy males regardless of obesity: A infarction patients: The Alpha Omega cohort study. Nephrol Dial
prospective cohort study. PLoS One 10: e0134937, 2015 Transplant 33[Suppl 1]: i469, 2018
73. Rebholz CM, Tin A, Liu Y, Kuczmarski MF, Evans MK, Zonderman AB, 92. Halbesma N, Bakker SJ, Jansen DF, Stolk RP, De Zeeuw D, De Jong
et al.: Dietary magnesium and kidney function decline: The healthy
PE, et al; PREVEND Study Group: High protein intake associates with
aging in neighborhoods of diversity across the Life Span Study. Am
cardiovascular events but not with loss of renal function. J Am Soc
J Nephrol 44: 381–387, 2016
Nephrol 20: 1797–1804, 2009
74. Tohidi M, Hasheminia M, Mohebi R, Khalili D, Hosseinpanah F,
93. Herber-Gast GC, van Essen H, Verschuren WM, Stehouwer CD,
Yazdani B, et al.: Incidence of chronic kidney disease and its risk fac-
Gansevoort RT, Bakker SJ, et al.: Coffee and tea consumption in re-
tors, results of over 10 year follow up in an Iranian cohort. PLoS One 7:
lation to estimated glomerular filtration rate: Results from the
e45304, 2012
population-based longitudinal Doetinchem Cohort study. Am J Clin
75. Bash LD, Astor BC, Coresh J: Risk of incident ESRD: A comprehensive look
Nutr 103: 1370–1377, 2016
at cardiovascular risk factors and 17 years of follow-up in the Atheroscle-
94. Herber-Gast GM, Biesbroek S, Verschuren WMM, Stehouwer CD,
rosis Risk in Communities (ARIC) study. Am J Kidney Dis 55: 31–41, 2010
Gansevoort RT, Bakker SJ, et al.: Association of dietary protein and
76. Hallan SI, Orth SR: Smoking is a risk factor in the progression to kidney
dairy intakes and change in renal function: Results from the
failure. Kidney Int 80: 516–523, 2011
population-based longitudinal doetinchem cohort study. Am J Clin
77. Jafar TH, Jin A, Koh WP, Yuan JM, Chow KY: Physical activity and risk
Nutr 104: 1712–1719, 2016
of end-stage kidney disease in the Singapore Chinese Health Study.
95. Herber-Gast GM, Boersma M, Verschuren WMM, Stehouwer CDA,
Nephrology (Carlton) 20: 61–67, 2015
Gansevoort RT, Bakker SJL, et al.: Consumption of whole grains, fruit
78. Lew QJ, Jafar TH, Jin A, Yuan JM, Koh WP: Consumption of coffee but
and vegetables is not associated with indices of renal function in the
not of other caffeine-containing beverages reduces the risk of end-
stage renal disease in the Singapore Chinese Health study. J Nutr 148: population-based longitudinal Doetinchem study. Br J Nutr 118:
1315–1322, 2018 375–382, 2017
79. Lew QJ, Jafar TH, Koh HW, Jin A, Chow KY, Yuan JM, et al.: Red 96. Hirahatake KM, Jacobs DR, Gross MD, Bibbins-Domingo KB, Shlipak
meat intake and risk of ESRD. J Am Soc Nephrol 28: 304–312, 2017 MG, Mattix-Kramer H, et al.: The association of serum carotenoids,
80. Lipworth L, Mumma MT, Cavanaugh KL, Edwards TL, Ikizler TA, tocopherols, and ascorbic acid with rapid kidney function decline: The
Tarone RE, et al.: Incidence and predictors of end stage renal disease Coronary Artery Risk Development in Young Adults (CARDIA) study.
among low-income blacks and whites. PLoS One 7: e48407, 2012 J Ren Nutr 29: 65–73, 2019
81. Pscheidt C, Nagel G, Zitt E, Kramar R, Concin H, Lhotta K: Sex- and 97. Jhee JH, Kee YK, Park S, Kim H, Park JT, Han SH, et al.: High-protein
time-dependent patterns in risk factors of end-stage renal disease: A diet with renal hyperfiltration is associated with rapid decline rate of
large Austrian cohort with up to 20 years of follow-up. PLoS One 10: renal function: A community-based prospective cohort study. Neph-
e0135052, 2015 rol Dial Transplant 35: 98–106, 2020
82. Rebholz CM, Grams ME, Steffen LM, Crews DC, Anderson CA, 98. Knight EL, Stampfer MJ, Hankinson SE, Spiegelman D, Curhan GC:
Bazzano LA, et al.: Diet soda consumption and risk of incident end The impact of protein intake on renal function decline in women with
stage renal disease. Clin J Am Soc Nephrol 12: 79–86, 2017 normal renal function or mild renal insufficiency. Ann Intern Med 138:
83. Reynolds K, Gu D, Chen J, Tang X, Yau CL, Yu L, et al.: Alcohol con- 460–467, 2003
sumption and the risk of end-stage renal disease among Chinese men. 99. Krop JS, Coresh J, Chambless LE, Shahar E, Watson RL, Szklo M, et al.:
Kidney Int 73: 870–876, 2008 A community-based study of explanatory factors for the excess risk for
84. Smyth A, Griffin M, Yusuf S, Mann JF, Reddan D, Canavan M, et al.: early renal function decline in blacks vs whites with diabetes: The
Diet and major renal outcomes: A prospective cohort study. The NIH- Atherosclerosis Risk in Communities study. Arch Intern Med 159:
AARP Diet and Health study. J Ren Nutr 26: 288–298, 2016 1777–1783, 1999
85. Stengel B, Tarver-Carr ME, Powe NR, Eberhardt MS, Brancati FL: 100. Lin J, Curhan GC: Associations of sugar and artificially sweetened
Lifestyle factors, obesity and the risk of chronic kidney disease. soda with albuminuria and kidney function decline in women. Clin
Epidemiology 14: 479–487, 2003 J Am Soc Nephrol 6: 160–166, 2011
101. Lin J, Hu FB, Curhan GC: Associations of diet with albuminuria and Prospective Investigation into Cancer and Nutrition (EPIC)-Heart
kidney function decline. Clin J Am Soc Nephrol 5: 836–843, 2010 study. Eur Heart J 32: 1235–1243, 2011
102. Malhotra R, Lipworth L, Cavanaugh KL, Young BA, Tucker KL, 119. Chauveau P, Koppe L, Combe C, Lasseur C, Trolonge S, Aparicio M:
Carithers TC, et al.: Protein intake and long-term change in glomerular Vegetarian diets and chronic kidney disease. Nephrol Dial Transplant
filtration rate in the Jackson Heart Study. J Ren Nutr 28: 245–250, 34: 199–207, 2019
2018 120. O’Donnell M, Mente A, Rangarajan S, McQueen MJ, O’Leary N, Yin
103. Menon V, Katz R, Mukamal K, Kestenbaum B, de Boer IH, Siscovick DS, L, et al; PURE Investigators: Joint association of urinary sodium and
et al.: Alcohol consumption and kidney function decline in the elderly: potassium excretion with cardiovascular events and mortality:
Alcohol and kidney disease. Nephrol Dial Transplant 25: 3301–3307, Prospective cohort study. BMJ 364: l772, 2019
2010 121. Vinceti M, Filippini T, Crippa A, de Sesmaisons A, Wise LA, Orsini N:
104. Ohta Y, Tsuchihashi T, Kiyohara K, Oniki H: High salt intake promotes a Meta-analysis of potassium intake and the risk of stroke. J Am Heart
decline in renal function in hypertensive patients: A 10-year obser- Assoc 5: e004210, 2016
vational study. Hypertens Res 36: 172–176, 2013 122. Whelton PK, He J, Cutler JA, Brancati FL, Appel LJ, Follmann D, et al.:
105. White SL, Polkinghorne KR, Cass A, Shaw JE, Atkins RC, Chadban SJ: Effects of oral potassium on blood pressure. Meta-analysis of ran-
Alcohol consumption and 5-year onset of chronic kidney disease: The domized controlled clinical trials. JAMA 277: 1624–1632, 1997
AusDiab study. Nephrol Dial Transplant 24: 2464–2472, 2009 123. Geleijnse JM, Kok FJ, Grobbee DE: Blood pressure response to
106. Barbato A, D’Elia L, Perna L, Molisso A, Iacone R, Strazzullo P, et al.: changes in sodium and potassium intake: A metaregression analysis of
Increased microalbuminuria risk in male cigarette smokers: Results randomised trials. J Hum Hypertens 17: 471–480, 2003
from the “Olivetti Heart Study” after 8 years follow-up. Kidney Blood 124. Smyth A, Dunkler D, Gao P, Teo KK, Yusuf S, O’Donnell MJ, et al;
Press Res 44: 33–42, 2019 ONTARGET and TRANSCEND investigators: The relationship be-
107. Chang A, Van Horn L, Jacobs DR Jr., Liu K, Muntner P, Newsome B, tween estimated sodium and potassium excretion and subsequent
et al.: Lifestyle-related factors, obesity, and incident micro- renal outcomes. Kidney Int 86: 1205–1212, 2014
albuminuria: The CARDIA (Coronary Artery Risk Development in 125. Clase CM, CarreroJ-J, Ellison DH, Grams ME, Hemmelgarn BR,
Young Adults) study. Am J Kidney Dis 62: 267–275, 2013 Jardine MJ, et al; Conference Participants: Potassium homeostasis
108. Forman JP, Scheven L, de Jong PE, Bakker SJ, Curhan GC, and management of dyskalemia in kidney diseases: Conclusions from
Gansevoort RT: Association between sodium intake and change in a Kidney Disease: Improving Global Outcomes (KDIGO) Controver-
uric acid, urine albumin excretion, and the risk of developing hyper- sies Conference. Kidney Int 97: 42–61, 2020
tension. Circulation 125: 3108–3116, 2012 126. McMahon EJ, Campbell KL, Bauer JD, Mudge DW: Altered dietary
109. Jee SH, Boulware LE, Guallar E, Suh I, Appel LJ, Miller ER 3rd: Direct, salt intake for people with chronic kidney disease. Cochrane Database
progressive association of cardiovascular risk factors with incident pro- Syst Rev 2: CD010070, 2015
teinuria: Results from the Korea Medical Insurance Corporation (KMIC) 127. Stolarz-Skrzypek K, Staessen JA: Reducing salt intake for prevention of
study. Arch Intern Med 165: 2299–2304, 2005 cardiovascular disease--times are changing. Adv Chronic Kidney Dis
110. Kimura Y, Yamamoto R, Shinzawa M, Isaka Y, Iseki K, Yamagata K, 22: 108–115, 2015
et al.: Alcohol consumption and incidence of proteinuria: A retro- 128. Mu M, XuL-F, Hu D, Wu J, BaiM-J: Dietary patterns and overweight/
spective cohort study. Clin Exp Nephrol 22: 1133–1142, 2018 obesity: A review article. Iran J Public Health 46: 869–876, 2017
111. Maeda I, Hayashi T, Sato KK, Koh H, Harita N, Nakamura Y, et al.: 129. Garofalo C, Borrelli S, Minutolo R, Chiodini P, De Nicola L, Conte G: A
Cigarette smoking and the association with glomerular hyperfiltration systematic review and meta-analysis suggests obesity predicts onset
and proteinuria in healthy middle-aged men. Clin J Am Soc Nephrol of chronic kidney disease in the general population. Kidney Int 91:
6: 2462–2469, 2011 1224–1235, 2017
112. O’Seaghdha CM, HwangS-J, Upadhyay A, Meigs JB, Fox CS: Pre- 130. Mozaffarian D: Dietary and policy priorities for cardiovascular disease,
dictors of incident albuminuria in the Framingham Offspring cohort. diabetes, and obesity: A comprehensive review. Circulation 133:
Am J Kidney Dis 56: 852–860, 2010 187–225, 2016
113. Uehara S, Hayashi T, Kogawa Sato K, Kinuhata S, Shibata M, Oue K, 131. Kubota Y, Iso H, Yamagishi K, Sawada N, Tsugane S; JPHC Study
et al.: Relationship between alcohol drinking pattern and risk of pro- Group (Japan Public Health Center): Daily total physical activity and
teinuria: The Kansai healthcare study. J Epidemiol 26: 464–470, 2016 incident cardiovascular disease in Japanese men and women: Japan
114. Wakasugi M, Kazama J, Narita I, Iseki K, Fujimoto S, Moriyama T, et al.: Public Health Center-Based Prospective Study. Circulation 135:
Association between overall lifestyle changes and the incidence of 1471–1473, 2017
proteinuria: A population-based, cohort study. Intern Med 56: 132. Chomistek AK, Cook NR, Rimm EB, Ridker PM, Buring JE, Lee IM:
1475–1484, 2017 Physical activity and incident cardiovascular disease in women: Is the
115. Wakasugi M, Kazama JJ, Yamamoto S, Kawamura K, Narita I: A relation modified by level of global cardiovascular risk? J Am Heart
combination of healthy lifestyle factors is associated with a decreased Assoc 7: e008234, 2018
incidence of chronic kidney disease: A population-based cohort 133. Lear SA, Hu W, Rangarajan S, Gasevic D, Leong D, Iqbal R, et al.: The
study. Hypertens Res 36: 328–333, 2013 effect of physical activity on mortality and cardiovascular disease in
116. Wen J, Hao J, Zhang Y, Liang Y, Li S, Wang F, et al.: Fresh fruit con- 130 000 people from 17 high-income, middle-income, and low-
sumption and risk of incident albuminuria among rural Chinese adults: income countries: The PURE study. Lancet 390: 2643–2654, 2017
A village-based prospective cohort study. PLoS One 13: e0197917, 134. Robinson-Cohen C, Katz R, Mozaffarian D, Dalrymple LS, de Boer I,
2018 Sarnak M, et al.: Physical activity and rapid decline in kidney function
117. Wang X, Ouyang Y, Liu J, Zhu M, Zhao G, Bao W, et al.: Fruit and among older adults. Arch Intern Med 169: 2116–2123, 2009
vegetable consumption and mortality from all causes, cardiovascular 135. Zelle DM, Klaassen G, van Adrichem E, Bakker SJ, Corpeleijn E, Navis
disease, and cancer: Systematic review and dose-response meta- G: Physical inactivity: A risk factor and target for intervention in renal
analysis of prospective cohort studies. BMJ 349: g4490, 2014 care. Nat Rev Nephrol 13: 152–168, 2017
118. Crowe FL, Roddam AW, Key TJ, Appleby PN, Overvad K, Jakobsen 136. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD,
MU, et al; European Prospective Investigation into Cancer and Nu- Hahn EJ, et al.: 2019 ACC/AHA guideline on the primary prevention of
trition (EPIC)-Heart Study Collaborators: Fruit and vegetable intake cardiovascular disease: A report of the American College of cardiol-
and mortality from ischaemic heart disease: Results from the European ogy/American heart association task force on clinical practice
guidelines [published correction appears in J Am Coll Cardiol 74: 145. Xia J, Wang L, Ma Z, Zhong L, Wang Y, Gao Y, et al.: Cigarette smoking
1429–1430, 2019]. J Am Coll Cardiol 74: e177–e232, 2019 and chronic kidney disease in the general population: A systematic
137. Milam RH: Exercise guidelines for chronic kidney disease patients. review and meta-analysis of prospective cohort studies. Nephrol Dial
J Ren Nutr 26: e23–e25, 2016 Transplant 32: 475–487, 2017
138. Latchoumycandane C, Nagy LE, McIntyre TM: Chronic ethanol in- 146. Thomas G, Sehgal AR, Kashyap SR, Srinivas TR, Kirwan JP,
gestion induces oxidative kidney injury through taurine-inhibitable Navaneethan SD: Metabolic syndrome and kidney disease: A sys-
inflammation. Free Radic Biol Med 69: 403–416, 2014 tematic review and meta-analysis. Clin J Am Soc Nephrol 6:
139. Fan Z, Yun J, Yu S, Yang Q, Song L: Alcohol consumption can be a 2364–2373, 2011
“double-edged sword” for chronic kidney disease patients. Med Sci 147. Janssen WM, Hillege H, Pinto-Sietsma SJ, Bak AA, De Zeeuw D, de Jong
Monit 25: 7059–7072, 2019 PE; PREVEND Study Group. Prevention of Renal and Vascular End-stage
140. Lin M, Su Q, Huang H, Zheng Y, Wen J, Yao J, et al.: Alcohol con- Disease: Low levels of urinary albumin excretion are associated with car-
sumption and the risk for renal hyperfiltration in the general Chinese diovascular risk factors in the general population. Clin Chem Lab Med 38:
population. Eur J Clin Nutr 71: 500–505, 2017 1107–1110, 2000
141. Roerecke M, Rehm J: Alcohol consumption, drinking patterns, and 148. Sterne JA, Hernán MA, Reeves BC, Savovic J, Berkman ND,
ischemic heart disease: A narrative review of meta-analyses and a Viswanathan M, et al.: ROBINS-I: A tool for assessing risk of bias in
systematic review and meta-analysis of the impact of heavy drinking non-randomised studies of interventions. BMJ 355: i4919, 2016
occasions on risk for moderate drinkers. BMC Med 12: 182, 2014 149. Wei Y, Wang Y, Di Q, Choirat C, Wang Y, Koutrakis P, et al.: Short term
142. Jespersen T, Kruse N, Mehta T, Kuwabara M, Noureddine L, Jalal D: exposure to fine particulate matter and hospital admission risks and
Light wine consumption is associated with a lower odd for cardio- costs in the medicare population: Time stratified, case crossover study.
vascular disease in chronic kidney disease. Nutr Metab Cardiovasc Dis BMJ 367: l6258, 2019
28: 1133–1139, 2018 150. Blum MF, Surapaneni A, Stewart JD, Liao D, Yanosky JD, Whitsel EA,
143. Dunkler D, Kohl M, Heinze G, Teo KK, Rosengren A, Pogue J, et al; et al.: Particulate matter and albuminuria, glomerular filtration rate, and
ONTARGET Investigators: Modifiable lifestyle and social factors affect incident CKD. Clin J Am Soc Nephrol 15: 311–319, 2020
chronic kidney disease in high-risk individuals with type 2 diabetes 151. Whittaker CF, Miklich MA, Patel RS, Fink JC: Medication safety prin-
mellitus. Kidney Int 87: 784–791, 2015 ciples and practice in CKD. Clin J Am Soc Nephrol 13: 1738–1746, 2018
144. Baranchuk A, Haseeb S, Alexander B: Alcohol consumption guide- 152. Johri N, Jacquillet G, Unwin R: Heavy metal poisoning: The effects of
lines: International discrepancies and variations. BMJ 361: k1630, 2018 cadmium on the kidney. Biometals 23: 783–792, 2010
AFFILIATIONS
1
Menzies Health Institute Queensland, Griffith University, Gold Coast, Queensland, Australia
2
Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden
3
Department of Nephrology, Guangdong Provincial Hospital of Chinese Medicine, The Second Affiliated Hospital, Guangzhou University of
Chinese Medicine, Guangzhou City, Guangdong Province, China
4
Bond University Nutrition and Dietetics Research Group, Faculty of Health Science and Medicine, Bond University, Gold Coast, Queensland,
Australia
5
Nutrition Research Australia, Sydney, New South Wales, Australia
6
Nephrology and Mineral Metabolism Department, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico City, Mexico
7
Department of Medicine, McMaster University, Hamilton, Ontario, Canada
8
Department of Health Research Methods, Evidence and Impact, McMaster University, Hamilton, Ontario, Canada
9
Division of Clinical Geriatrics, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden
albuminuria)
Therapy Diet behaviours: Six studies involving 701,627 participants reported associations of 13
(KRT) different dietary factors to KRT. No meta-analysis was possible for any factor due to having
less than 2 exposures for each diet factor, however 2 studies (100%) showed potassium intake
and 1 study (50%) showed coffee intake to be significantly associated with decreased risks of
KRT. In contrast, four diet factors were found to be potentially harmful to KRT, including
protein, high red meat intake, sodium and diet beverages, with over 50% of studies reporting
Physical activity: Three studies involving 70,058 reported associations of physical activity to
KRT. One of these studies reported a significant reduced risk of KRT, and another study
reported a significant increase risk of a composite outcome of treatment of KRT due to any
cause, or death related to CKD in people with higher levels of physical activity compared to
lower levels. However, these results could not be pooled in meta-analysis and are therefore
consumption to KRT. One of the two studies (50%) reported significant reduced risk of KRT,
however the results could not be pooled in meta-analysis and are therefore reported in
Supplemental Table 4.
smoking to KRT. Meta-analysis of all studies reporting a smoking exposure (ever smoked)
compared to never showed a significant increased risk of KRT (RR 1.59 [95% CI: 1.30, 1.94];
I2=68%; Evidence quality: Moderate) (see Supplemental Figure 1). Meta-analysis also showed
that former (not current) smokers compared to current was significantly associated with
increased risks of KRT (RR 1.25 [95% CI: 1.13, 1.39]; I2=39%; Evidence quality: Moderate)
GFR decline Diet behaviours: A total of 22 individual studies involving 77,274 participants reported
associations of 24 different dietary factors to eGFR decline. Only potassium and protein intake
exposures were able to be pooled into meta-analysis. Higher potassium intake was associated
with a significant decreased risk of eGFR decline (RR 0.49 [95% CI: 0.31, 0.79]; I2=90%;
Evidence quality: Very low), whereas higher protein intake was not (Table 2; Supplemental
Figure 2).
Twenty-two other dietary factors and their associations to eGFR decline were reported across
17 studies, which did not have enough data to pool into meta-analysis and are therefore
magnesium and vitamin A intake were potentially protective against eGFR decline, with over
sodium, saturated fat, carbohydrate intake, sodium-potassium ratio and diet beverages had at
Physical activity: Seven studies involving 27,075 participants reported association of physical
activity levels to GFR decline demonstrated a higher level of physical activity significantly
reduced the odds of eGFR decline (5 studies; 15,161 participants; OR 0.77 [95% CI: 0.63,
0.93]; 75%; Evidence quality: Very low) (Table 2; Supplemental Figure 2). The results of the
other two studies which could not be meta-analyzed concurred with this analysis, with 50% of
Five studies reported smoking and associations to eGFR decline, with only 1 showing
different dietary factors to incident albuminuria. Only 1 meta-analysis was possible on sodium
Eighteen other dietary factors and their associations to incident albuminuria were reported
across 9 studies, which did not have enough data to pool into meta-analysis and are therefore
descriptively reported in Supplemental Table 4. Two diet factors, including fruit and dairy had
at least 50% of studies demonstrating these to protective against incident albuminuria, whereas
5 diet factors, including, sugar-sweetened beverages, saturated fat, cholesterol, red and
processed meat and energy intake had at least 50% of studies demonstrating a potentially
Physical activity: Four studies involving 110,154 participants reported a significant reduced
odds of incident albuminuria (OR 0.88 [95% CI: 0.81, 0.96]; 0%; Evidence quality: Low) in
people participating in higher levels of physical activity compared to lower levels (Table 2;
smoking to incident albuminuria. Meta-analysis of all studies of populations who had ever
smoked compared to those who had never smoked showed a significant increased odds of
incident albuminuria (OR 1.67 [95% CI: 1.23, 2.26]; I2=88%; Evidence quality: Very low).
who had never smoked, or former smokers compared to current smokers (Table 2;
Subgroup A subgroup analysis by duration of exposure showed that physical activity and sodium intake
analysis exposures with durations less than 10 years to be more likely associated with lower risk of
incident CKD compared with exposures greater than 10 years. In contrast, alcohol
consumption and potassium intake exposures appeared to be associated with incident CKD
when exposures were greater than 10 years, compared with exposures of less than 10 years.
Sodium intake exposures were also more likely to be associated with lower risk of incident
CKD in countries outside the US, whereas alcohol consumption was more likely to be
associated with incident CKD in the US compared with other countries. Physical activity was
not associated with incident CKD in the US or other countries. Studies reporting associations
with incident CKD were also more likely to be significant if they reported associations using
a risk ratio (RR) compared with an odds ratio (OR), including potassium intake, alcohol
consumption and tobacco smoking. Other subgroup analyses either were consistent with the
primary finding or were not possible because at least one subgroup had few than 3 studies
Sensitivity In a sensitivity analysis, repeating the analysis by rotating each study reporting associations to
analysis incident CKD from the same cohort of data one by one, there was no change to the primary
finding for any lifestyle exposure, except fish intake, which became statistically significant
when data from Haring et al (1) were substituted for those from Rebholz et al (2) (Supplemental
Table 6).
Table 3. One study in the tobacco smoking and incident CKD (3) analysis and one for
tobacco smoking and albuminuria (4) reported unadjusted associations. One study in the
physical activity incident CKD (5) analysis and one for GFR decline (6) reported unadjusted
associations. We have completed a sensitivity analyses, removing these studies to test the
robustness of analyses, where no change in the primary or secondary findings was observed.
In a sensitivity analysis removing each of these studies from their respective analyses, the
eligible studies
Database Terms
DIET
Line 1
"Diet"[Mesh] OR "Food"[Mesh] OR nutrition*[Title/Abstract] OR
vegetable*[Title/Abstract] OR meat*[Title/Abstract] OR dairy[Title/Abstract]
OR beverage*[Title/Abstract] OR fruit[Title/Abstract] OR
fibre[Title/Abstract] OR fiber[Title/Abstract] OR fish[Title/Abstract] OR
salt[Title/Abstract] OR sodium[Title/Abstract]
Line 2
"Kidney Failure, Chronic"[Mesh] OR "Renal Insufficiency, Chronic"[Mesh]
OR "chronic kidney disease" OR "renal failure*" OR "kidney failure" OR
"renal impairment" OR "kidney impairment" OR "kidney dysfunction" OR
"renal dysfunction" OR "reduced renal function" OR "CKD" OR "progressive
kidney" OR "Glomerular Filtration Rate"[Mesh] OR "GFR" OR "eGFR" OR
"Proteinuria"[Mesh] OR "Proteinuri*" OR "Albuminuria"[Mesh] OR
"Albuminuria" OR "microalbuminuria" OR "End Stage renal disease" OR
"ESRD" OR "End-stage Kidney Disease" OR "ESKD" OR “dialysis” OR
"Renal Replacement Therapy"[Mesh]
Line 3
cohort[Title/Abstract] OR incidence[Title/Abstract] OR
development[Title/Abstract] OR progression[Title/Abstract] OR
association[Title/Abstract]
MEDLINE
strategy
PHYSICAL ACTIVITY
Line 1
“Exercises”[Mesh] OR “Exercise” OR “physical activity”[Mesh] OR
“physical activity” OR “physical activities” OR “Walking”[Mesh] OR
“Walking” OR “Sedentary Behavior”[Mesh] OR “Sedentary” OR “Sitting”
OR “Screen Time” OR “Computer Games” OR “Video Games” OR
“Television”
Line 2
"Kidney Failure, Chronic"[Mesh] OR "Renal Insufficiency, Chronic"[Mesh]
OR "chronic kidney disease" OR "renal failure*" OR "kidney failure" OR
"renal impairment" OR "kidney impairment" OR "kidney dysfunction" OR
"renal dysfunction" OR "reduced renal function" OR "CKD" OR "progressive
kidney" OR "Glomerular Filtration Rate"[Mesh] OR "GFR" OR "eGFR" OR
"Proteinuria"[Mesh] OR "Proteinuri*" OR "Albuminuria"[Mesh] OR
"Albuminuria" OR "microalbuminuria" OR "End Stage renal disease" OR
"ESRD" OR "End-stage Kidney Disease" OR "ESKD" OR “dialysis” OR
"Renal Replacement Therapy"[Mesh]
Line 3
cohort[Title/Abstract] OR incidence[Title/Abstract] OR
development[Title/Abstract] OR progression[Title/Abstract] OR
association[Title/Abstract]
SMOKING AND ALCOHOL
Line 1
Smoke [Mesh Terms] OR Smoke [All Fields] OR Smoke* [All Fields] OR
Smok* [All Fields] OR Smoking. [All Fields] OR Smoker*[All Fields] OR
cigar* [All Fields] OR cigarrete* [All Fields] or Tobacco* [All Fields] or
smoking use [All Fields]
exp Alcohol [Mesh Terms] OR Alcohol [All Fields] OR Alcohol* [All Fields]
OR Alcoholism[ Mesh Terms] OR Alcoholism [All Fields] OR Alcohol
Drinking [Mesh Terms] OR Alcohol Drinking [All Fields] AND (dependen*
or disorder* or drink* or misuse or abuse* or consumption)) [All Fields])
Line 2
"Kidney Failure, Chronic"[Mesh] OR "Renal Insufficiency, Chronic"[Mesh]
OR "chronic kidney disease" OR "renal failure*" OR "kidney failure" OR
"renal impairment" OR "kidney impairment" OR "kidney dysfunction" OR
"renal dysfunction" OR "reduced renal function" OR "CKD" OR "progressive
kidney" OR "Glomerular Filtration Rate"[Mesh] OR "GFR" OR "eGFR" OR
"Proteinuria"[Mesh] OR "Proteinuri*" OR "Albuminuria"[Mesh] OR
"Albuminuria" OR "microalbuminuria" OR "End Stage renal disease" OR
"ESRD" OR "End-stage Kidney Disease" OR "ESKD" OR “dialysis” OR
"Renal Replacement Therapy"[Mesh]
Line 3
cohort[Title/Abstract] OR incidence[Title/Abstract] OR
development[Title/Abstract] OR progression[Title/Abstract] OR
association[Title/Abstract]
Supplemental Table 2. Summary of the methods relating to data extraction, risk of bias, meta-
Data extraction Age, sex, sample size, country/region of study, date of publication,
comorbidity (cardiac disease, diabetes, hypertension), exposure of
interest (s), reference exposure, study risks of bias, date of publication,
baseline kidney function, duration of follow up, statistical analysis
approaches (including adjustment for confounding variables), and all
data relating to our primary and secondary outcomes. Binary outcomes
(incident chronic kidney disease, end-stage kidney disease, eGFR decline
and albuminuria) were extracted as the most adjusted relative risks (RR),
odds ratios (OR), or hazard ratios (HR). If data was reported in figures
and could not be located in text, Supplemental material or through author
contact, then we extracted the data directly from figures using Webplot
designer (https://automeris.io/WebPlotDigitizer/).
Risk of bias These were further subdivided into five key question criteria; Selection
(representativeness of exposed cohort, non-exposed cohort,
ascertainment of exposure, and demonstration that outcome of interest
was not present at the start of the study); Comparability of cohorts on
basis of design or analysis; Outcome (assessment, follow-up length, and
adequacy of follow-up); Selective reporting of outcomes; and other bias
Dose-response A dose-response random-effects meta-regression analysis was performed
meta- based on the natural log of HRs for different categories of alcohol intake
regression (7, 8). There are four studies included in the meta-regression with at least
analysis three quantitative exposure categories of the estimate HRs and their
variance. Because each study used different units of alcohol intake
(grams or number of drinks per day or week), we converted alcohol intake
in gram/day, by the formula alcohol (gram) = 1 drink*13g/drink, and the
median level of each category was assigned to each corresponding HR
(9). If the upper bound in the highest category was not provided, we
assumed that it had the same value.
Subgroup and To explore the sources of heterogeneity, a priori subgroup analyses were
sensitivity conducted exploring the type of ratio reported (risk ratio; odds ratio;
analyses hazard ratio). We also conducted subgroup analyses by country of study
origin, other health conditions, dose of alcohol consumption, baseline
GFR, and length of follow-up. For outcomes where there existed multiple
reports of the same outcome from the same cohort dataset, we conducted
a study-by-study sensitivity analysis substituting included studies which
reported associations from the same cohort dataset, where one study
cohort is removed and another study reporting an association from the
same cohort added in, with the process repeated until all studies from the
same cohort dataset have been rotated through the meta-analysis once.
Subgroup analyses are only reported for analyses where at least 3 studies
are in each subgroup.
Supplemental Table 3. Characteristics of the included studies
Lifestyle exposure Follow-up Adjusted covariates
Citation Population Country Outcomes
/ prognostic factor in statistical models
Japan Diet - Urinary 11 years 1. Incident CKD stage Age, sex, BMI, HbA1c,
potassium 4 total cholesterol, log
excretion 2. eGFR decline triglyceride, log HDL-
24-hour urine cholesterol, LDL-
cholesterol, systolic BP,
collection
Shiga Prospective renin-angiotensin
system inhibitor,
Observational Follow-up Study
hypertension, log
T2DM urinary albumin
N=623 excretion rate, eGFR,
59 ±10 years current smoking, and
57.8% male urinary sodium
Araki 2015 89 ± 19 mL/min/1.73 excretion
Iran Diet - fruit; 6.1 years 1. Incident CKD (<60 Age, sex, smoking, total
vegetable; whole mL/min/1.73 m2) energy intake, BMI,
Tehran Lipid and Glucose grains; low-fat eGFR, triglycerides,
Study (TLGS). Follow-up data dairy; nuts & physical activity,
hypertension and
from 2009-2011 legumes; red &
diabetes
Healthy general population processed meats;
N=1179 sweetened
43 years beverages; sodium
49.5% male Captured using
Asghari 2017 Baseline GFR NP self-reported FFQ
Tehran Lipid and Glucose Iran Diet – antioxidant 3 years 1. Incident CKD (<60 Age, sex, smoking, total
Study (TLGS). Follow-up data capacity (vitamin mL/min/1.73 m2) energy intake, BMI,
from 2009-2011 C; vitamin E; beta- eGFR, triglycerides,
Healthy general population carotene) physical activity,
hypertension and
N=1630 Captured using
diabetes
49.6 ± 10.3 years self-reported FFQ
48.8% male
Asghari 2018 73.3 ± 8.6 mL/min/1.73
Italian Longitudinal Study of Italy Current smokers 4 years 1. Incident CKD Unadjusted
Ageing (ILSA) >20 cigarettes/day (>26.5 mmol/l of
Healthy general population Captured in clinical SCr)
N=678 interview
73.5±6.0 years
Gender NP
Baggio 2005 Baseline GFR NP
Tehran Lipid and Glucose Iran Diet - 5.8 years 1. Incident CKD (<60 Age, sex, diabetes,
Study (TLGS). Follow-up data Nitrite/Nitrate mL/min/1.73 m2) hypertension, eGFR,
from 2006/08 Captured using and smoking, protein
Healthy general population self-reported FFQ intake, fat intake,
potassium intake and
N=2,799
sodium intake
31 years
Bahadoran 42.9% male
2016 80.7 mL/min/1.73
Iran Diet - L-Arginine 6.3 years 1. Incident CKD (<60 Sex and age, BMI,
Tehran Lipid and Glucose Captured using mL/min/1.73 m2) smoking, serum
self-reported FFQ creatinine, diabetes
Study (TLGS)
mellitus, hypertension,
Healthy general population medications use, CVD,
N=1780 daily energy intake,
68 years protein intake, fat
Bahadoran 40.9% male intake, and sodium-
2017 74.8 mL/min/1.73 potassium ratio.
Tehran Lipid and Glucose Iran Diet - Allium 6 years 1. Incident CKD (<60 Sex, age, BMI,
Study (TLGS) vegetables mL/min/1.73 m2) smoking, type 2
Healthy general population Captured using 2. GFR decline diabetes, TG to HDL-C
N=3052 self-reported FFQ (change at follow- ratio, physical activity,
dietary pattern scores
40.3± 14.3 years up)
Bahadoran 44.2% male
2017 77.9 mL/min/1.73
Italy Current smoker 7.8 years 1. Microalbuminuria Unadjusted
vs never smokers (30mg/g to 300mg/g
(independently of and greater that
Olivetti Health Study (OHS) the number or 300mg/g; 24-hour
Healthy general population smoked cigarettes collection)
N=637 per day)
51 years Captured on self-
Gender NP reported
Barbato 2019 90 mL/min/1.73 questionnaire
Atherosclerosis Risk in US Current, former 16 years 1. Incident ESRD or Age, sex, race, gender,
Communities (ARIC) Study and never Death (ESRD (ICD diabetes, BP, CVD,
Healthy general population smokers 9 and 10 kidney LDL, center
N=15,324 Captured on self- transplant or
54 ± 5.8 years reported dialysis); eGFR <
45% male questionnaire 15mL/min/1.73m2)
Bash 2010 93 ± 21 mL/min/1.73
US Diet - Sugar 9 years 1. Incident CKD (<60 Age, sex, body mass
Bomback Atherosclerosis Risk in sweetened mL/min/1.73 m2) index, sodium intake,
2010 Communities (ARIC) Study beverages caloric intake,
Healthy general population Captured using hypertension, diabetes,
N=15,745 self-reported FFQ current tobacco and
53 years alcohol use, education,
49.6% male center, and race
93.3 mL/min/1.73
Italy Alcohol 3.5 years 1. Incident CKD (<60 Age, education,
consumption – mL/min/1.73 m2) antihypertensive,
Categories non- diabetes, blood,
The Italian Longitudinal Study drinker vs, fibrinogen, TC,
smoking, isolated
on Aging (ILSA) study <12g/day, 12 to 24
systolic, HTN, lipid-
Non-CKD elderly population g/Day, 24 to 47 lowering drugs
N=1,542 g/day >48g/day
65 – 84 years Captured on self-
92.7% male reported
Buja 2011 Baseline GFR NP questionnaire
US Diet - Fruit; 15 years 1. Microalbuminuria Age, sex, race, family
vegetables; (sex-race adjusted history of kidney
wholegrains; nuts urine ACR ≥25mg/g disease, education level,
and legumes; low- at 2 separate times) baseline ACD and
energy intake
fat dairy; sugar-
sweetened
beverages; red and
processed meat;
sodium; energy
intake
Captured using
The Coronary Artery Risk self-reported FFQ
Development in Young Adults
(CARDIA) study Currently
Healthy general population smoking or not
N=2,354 currently smoking
18 - 30 years Captured on self-
47.4% male reported
Chang 2013 101.5 mL/min/1.73 questionnaire
Italy Diet – protein 15.9 years 1. GFR decline Sex, age, reduced
Gubbio Study (urine urea (change from kidney function,
nitrogen) baseline to follow- obesity, underweight,
General population (5.2%
up equal to or lower hypertension,
diabetes; 34.5% HTN) hypercholesterolemia,
N=2,845 than Z-score of -1
smoking, ex-smoking,
50 years (eGFR change < diabetes, previous
43.9% male mean -1 standard CVD, urinary
Cirillo 2018 97 mL/min/1.73 deviation of eGFR sodium/potassium ratio,
change in the study habitual physical
cohort)) activity, and habitual
intakes of alcohol, milk
or yogurt, and caffeine-
containing beverages,
baseline eGFR
Switzerland Diet - Urinary 5.4 years 1. GFR decline Age, sex, physical
The Cohort Lausanne study sodium, potassium (change at follow- activity, smoking status,
Healthy general population and sodium-to- up) diabetes,
(100% white) potassium ratio antihypertensive
medication, SBP, BMI,
N=4141 Captured using spot
triglycerides, uric acid,
51.5 years urine collection CRP, 25-
45.7% male using Kawasaki hydroxyvitamin D
Deriaz 2019 88 ± 15 mL/min/1.73 formulae
US Alcohol 5.5 years 1. Albuminuria Age, duration of
consumption – 2. GFR decline diabetes, status of
measured by the albuminuria, sex,
number of ONTARGET
randomization arms,
drinks/week, with 1
GFR, and Δ-UACR to
drink equalling 1.5 progression, BMI, mean
ounces of hard arterial BP, glucose and
liquor or 1 glass of previous use of
beer or wine. angiotensin-converting
The ONgoing Telmisartan enzyme inhibitors or
Alone and in combination with Diet - animal angiotensin-receptor
Ramipril Global Endpoint Trial protein; plant blockers, tobacco use,
(ONTARGET) study protein; CHO physical activity, and
T2DM intake; vegetable; educational level to
study the nutrition-
N=6,213 sodium; potassium;
specific effects of a
67 years fruit healthier lifestyle and
Gender NP Captured using socioeconomic status.
Dunkler 2013 71.5 mL/min/1.73 self-reported FFQ
US Diet - Fruit, 5.5 years 1. Composite outcome Age, gender, duration of
The ONgoing Telmisartan vegetable, fiber, (defined as new diabetes, ONTARGET
Alone and in combination with protein, micro‐, or macro‐ randomization arms,
Ramipril Global Endpoint Trial carbohydrates, albuminuria, a albuminuria status, GFR
and δ-UACR and
(ONTARGET) study trans fat, urine decline in the eGFR
UACR at baseline on
T2DM sodium, urine of more than 5% per the log scale,
N=3,088 potassium Captured year or end‐ stage respectively
65 years using FFQ (except kidney disease)
66.6% male sodium and
Dunkler 2015 72.5 mL/min/1.73 potassium)
Tehran Lipid and Glucose Iran Diet - Glycation 3 years 1. Incident CKD (<60 Age, sex, energy
Study End Products mL/min/1.73 m2) intakes, smoking,
General population (10.7% Captured using physical activity, body
diabetes, 16.2% HTN) self-reported FFQ mass index, and
sodium, diabetes and
N=1,692
HTN
43.4 ± 11.4 years
43% male
Ejtahed 2016 70.4 ± 10.8 mL/min/1.73
Netherlands Diet – Protein 3.4 years 1. GFR decline Age, sex, total energy
Captured using (change at follow- intake, education,
Alpha Omega Cohort
self-reported FFQ up) alcohol, smoking,
Post-MI patients physical activity, RAS
N=2255 blocking drugs, fat
69 years intake, dietary sodium,
80% male diabetes and systolic
Esmeijer 2018 82 ± 20 mL/min/1.73 blood pressure
Iran Diet - 3.6 years 1. Incident CKD (<60 Age, sex, energy intake,
Micronutrient mL/min/1.73 m2) serum triglycerides,
intakes (B1, B2, serum cholesterol, BMI,
B3, B6, B12, hypertension, diabetes,
physical activity, and
Tehran Lipid and Glucose folate, Vitamin A,
smoking
Study (TLGS) follow up data C, D, E, sodium,
from 2006-2008 potassium, calcium,
Healthy general population magnesium,
N=1,692 phosphorous,
43.3 ± 11.4 years selenium, zinc)
Farhadnejad 49.2% male Captured using
2016 Baseline GFR NP self-reported FFQ
Iran Diet - Low CHO, 6.1 years 1. Incident CKD (<60 Age, sex, energy intake,
Tehran Lipid and Glucose high protein (CHO mL/min/1.73 m2) serum triglycerides,
Study (TLGS) 51% compared to serum cholesterol, BMI,
Healthy general population 64.1%; protein hypertension, diabetes,
physical activity, and
N=1,797 15.8% compared to
smoking
37.7 ± 12.2 years 12.9%)
Farhadnejad 48% male Captured using
2018 75.9 mL/min/1.73 self-reported FFQ
Prevention of Renal and Netherlands Diet - Urinary 6.4 years 1. Albuminuria (value Age, BMI, sex, alcohol
Vascular End-stage Disease sodium excretion, at follow-up) intake, smoking status,
(PREVEND) 24hr collection systolic and diastolic
Healthy general population BP, GFR, plasma levels
of glucose and
N=4146
cholesterol, and urinary
43.3 years levels of potassium,
Forman 2012 47.6% male
87 mL/min/1.73 calcium, and creatinine.
Models were also
mutually adjusted for
baseline levels of serum
uric acid and urinary
albumin
Framingham offspring study US Alcohol 6.6 years 1. Incident CKD (<60 Age, sex, eGFR, BMI,
Healthy general population consumption - mL/min/1.73 m2) hypertension, diabetes
N= 1,802 None (0 and dipstick proteinuria,
59.9 ± 9 years drinks/week); low- dietary guidelines
adherence index
45.2% male to-moderate (1–7
Baseline GFR NP drinks/ week in
women and 1–14
drinks/week in
men); high intake
(>7 drinks/week in
women and >14
drinks/week in
men)
Physical activity -
Physical Activity
Index
Smoking - Current
≥1 cigarette/day in
the past year
Captured on self-
reported
Foster 2015 questionnaire
US Smoking - yes vs 18.5 years 1. Incident CKD (<60 Age and sex
Framingham Heart Study no mL/min/1.73 m2)
Healthy general population Captured on self-
N=2585 reported
43 years questionnaire
47.3% male
Fox 2004 112 ± 57 mL/min/1.73
Blue Mountains Eye Study Australia Diet - 5 years 1. Incident CKD (<60 Age and sex, serum
Healthy general population Carbohydrate mL/min/1.73 m2) total homocysteine,
N=2600 intake haemoglobin, and
>50 years Captured using haematocrit
43.2% male self-reported FFQ
Gopinath 2011 Baseline GFR NP
Blue Mountains Eye Study Australia Diet – Dairy and 10 years 1. Incident CKD (<60 Age, receipt of pension,
General population (6.8% DM, calcium mL/min/1.73 m2) BMI, smoking, serum
41.6% HTN) Captured using triglycerides, HTN, and
N=1185 self-reported FFQ history of diagnosed
diabetes
64.1 ± 8.7 years
44.9% male
Gopinath 2016 Baseline GFR NP
Prevention of Renal and US Diet – protein (24- 6.4 years 1. GFR decline Unadjusted
Vascular End-stage Disease h urinary urea (Change at follow-
(PREVEND) excretion (Maroni up)
Healthy general population formula)
N=8,461 Captured using
50 years self-reported FFQ
Halbesma Gender NP
2009 80.6 mL/min/1.73
The Nord-Trøndelag Health Norway Smoking - Never, 10.3 years 1. End-stage kidney Age, education physical
(HUNT) II cohort Former, current disease activity, diabetes, CVD,
General population smoke lipids, waist
N=65,589 Captured on self- circumference, eGFR,
ACR, antihypertensive
47.8 years reported
medication
45.8% male questionnaire
Hallan 2011 98.1 mL/min/1.73
US Diet – Protein 23 years 1. Incident CKD (<60 Age, race, center, sex,
Captured using mL/min/1.73 m2) education level, and
self-reported FFQ total caloric intake,
HDL cholesterol, LDL
cholesterol, TG, TC,
lipid-lowering
medication use, systolic
BP, antihypertensive
Atherosclerosis Risk in medication use, alcohol
intake, current smoker,
Communities (ARIC) Study
physical activity index,
Healthy general population leisure-related physical
N=11,952 activity, total
53.8 years carbohydrate intake,
43.7% male BMI, and waist-to-hip
Haring 2017 103.1 mL/min/1.73 ratio
Campaign Against Cancer and US Smoking - ever vs 20 years 1. Incident CKD (<60 Age, sex, treated
Stroke (CLUE) study never and current mL/min/1.73 m2) diabetes, BP
Healthy general population vs former 2. End-stage kidney
N=23,534 disease
Haroun 2003 41 years
40.8% male Captured on self-
Baseline GFR NP reported
questionnaire
The Health, Aging and Body US Physical activity - 10 years 1. Incident CKD (<60 Age, baseline GFR, sex,
Composition (Health ABC) (0–48 vs >93 mL/min/1.73 m2) race, smoking status,
kcal/kg per week) 2. GFR decline (annual study site, HTN
study
Captured on self- loss of medication use, CVD,
Healthy general population heart failure, diabetes
N=2435 reported >3ml/min/1.73m2
status, pulse pressure,
73.6 ± 2.8 years questionnaire per year eGFR) BMI, HDL, TG, TC,
46.4%-49.8% male CRP, television
Hawkins 2015 79.8-82.5 mL/min/1.73 watching
Netherlands Diet - Coffee and 15 years 1. GFR decline Age, sex, highest
tea (Change at follow- attained level of
Captured using up) education, time-
self-reported FFQ dependent physical
activity, BMI, smoking,
alcohol consumption,
daily energy intake,
energy-adjusted intakes
of fiber, vitamin C, total
protein, fat, and
Doetinchem Cohort Study saturated fat, intake of
General population (28.7% tea (for coffee analysis)
or coffee (for tea
HTN, 1.3% DM, 21.2%
analysis),
hypercholesterolemia) hypercholesterolemia,
N=4722 HTN, and diabetes,
45.5 ± 9.8 years energy-adjusted intakes
Herber-Gast 48% male of magnesium,
2016 108 ± 14.7 mL/min/1.73 potassium, and caffeine
Netherlands Diet – Dairy 15 years 1. GFR decline Age, sex, highest
Captured using (Change at follow- attained level of
self-reported FFQ up) education, time-
dependent physical
Doetinchem Cohort Study activity, BMI, smoking,
General population (1.2% DM, alcohol consumption,
28.4% HTN, 22% daily energy intake,
energy-adjusted intakes
hypercholesterolemia, 8.8%
of fiber, vitamin C, total
obesity) protein, fat, and
N=3798 saturated fat, intake of
45.2 ± 9.7 years tea (for coffee analysis)
Herber-Gast 52.2% male or coffee (for tea
2016 108.6 ± 14.2 mL/min/1.73 analysis),
hypercholesterolemia,
HTN, and diabetes,
energy-adjusted intakes
of magnesium,
potassium, and caffeine
Netherlands Diet - Whole 15 years 1. GFR decline Age, sex, highest
grains, fruit, (change at follow- attained levels of
vegetables up) education, time-
Captured using 2. Albuminuria (ACR; dependent physical
activity, BMI, smoking,
self-reported FFQ change at follow-up)
alcohol consumption,
daily energy intake, and
Doetinchem Cohort Study presence of diabetes,
General population (28.2% HTN, and
hypercholesterolemia,
HTN, 1.2% DM, 22.1
energy-adjusted intake
hypercholesterolemia) of monounsaturated fat,
N=6,113 polyunsaturated fat,
45 years phosphorus,
Herber-Gast 48% male magnesium, vitamin D,
2017 104.6 mL/min/1.73 and calcium
England Smoking - non- 5 years 1. Incident CKD (<60 Age, BMI, systolic BP
and Wales smoker, ex-smoker, mL/min/1.73 m2)
light smoker <10 2. End-stage kidney
cigarettes day, disease
moderate smoker
10-19 cigarettes per
Qresearch database day, heavy smoker
Healthy general population 20 or more
N=788,320 cigarettes per day
35-74 years Captured on self-
Hippisley-Cox 50.6% male reported
2010 86.2± 15.9 mL/min/1.73 questionnaire
The Coronary Artery Risk US Diet - Serum 5 years 1. GFR decline Age, race, sex,
Development in Young Adults Carotenoids, (change at follow- CARDIA center,
(CARDIA) study Tocopherols, and up) education, smoking
General population (58.7% Ascorbic Acid status, alcohol intake,
physical activity, BMI,
white, 3.8% diabetes, 14.2%
and lipids, incident
HTN, 3.2% albuminuria) diabetes, HTN,
N=2,152 albuminuria, and CRP
40 years
Hirahatake 50.9% male
2019 109.3 mL/min/1.73
US Diet – Coffee 24 years 1. Incident CKD (<60 Age, sex, race-center,
Atherosclerosis Risk in consumption mL/min/1.73 m2) education, total energy
Communities (ARIC) Study Captured using 2. End-stage kidney intake, physical activity,
self-reported FFQ disease smoking, alcohol status,
Healthy general population
DASH diet score,
(76% white) diabetes status, BMI,
N=14,209 systolic BP,
54.1 years antihypertensive
44.6% male medication use, baseline
Hu 2018 103 mL/min/1.73 GFR
US Alcohol 24 years 1. Incident CKD (<60 Total energy intake,
consumption - mL/min/1.73 m2) age, ser, race-center,
never drinkers, income, education level,
former drinkers, ≤1 health insurance,
smoking, physical
drink per week, 2
activity, diabetes status,
Atherosclerosis Risk in to 7 drink/week, 8 HTN, BMI, eGFR
Communities (ARIC) Study to 14 drinks/week
Healthy general population and ≥ 15 drinks per
N=12692 week
52.5% male Captured using
Hu 2019 >100 mL/min/1.73 self-reported FFQ
Jafar 2015 Singapore Physical activity - 15.3 years 1. End-stage kidney Age, sex, interview
Never VS the disease year, body mass index,
active group dialect (Hokkein,
defined as Cantonese), education
level, self-reported
individuals who
history of physician
engaged in any diagnosed hypertension,
moderate activity diabetes, heart disease
for 2 hours or more or stroke, alcohol use,
per week, or any smoking, intake of
strenuous activity ginseng, protein intake
on a weekly basis
for at least 30
minutes
Captured using the
The Singapore Chinese Health European
Study Prospective
Healthy general population Investigation in
N=59,552 Cancer (EPIC)
56.1 years study physical
Gender NP activity
Baseline GFR NP questionnaire
Korea Medical Insurance Korea Smoking - current 10 years 1. Proteinuria Age, diabetes, BMI,
Corporation (KMIC) Study smokers, ex- (albuminuria - cholesterol, BP
General population (with HTN smokers, or non- dipstick finding of
and diabetes) smokers 1+ or greater)
N= 157377 Captured on self-
35-59 reported
66.4% male questionnaire
Jee 2005 Baseline GFR NP
Korea Diet – Protein 11.5 years 1. GFR decline Age, sex, eGFR and
Captured using (defined as an daily intake of total
self-reported FFQ annual eGFR decline energy, carbohydrate
Korean Genome and rate 3mL/min/1.73 intake, fat and sodium,
Epidemiology Study database smoking status, alcohol
m2/year)
General population (14.4% status education and
income levels and
HTN, 6.4% DM, 2.4
physical activity, BMI,
dyslipidaemia, 1.6% CVD) systolic BP, history of
N=9,229 HTN and diabetes,
52.0 ± 8.8 years fasting plasma glucose,
48.1% male serum albumin, TC and
Jhee 2019 93.9 ± 14.2 mL/min/1.73 haemoglobin
Korea Diet – Vegetables 8.2 years 1. Incident CKD (<60 Age, sex, total energy
and fruit mL/min/1.73 m2) intake, BMI, systolic
Captured using BP, education level,
self-reported FFQ smoking status, alcohol
Korean Genome and status, physical activity,
Epidemiology Study database history of HTN and
General population (14.4% diabetes, and red or
non-red meat, fish,
HTN, 6.4% DM, 2.4
dairy, egg, legume, nut,
dyslipidaemia, 1.6% CVD) and grain intake, serum
N=9,229 albumin, LDL
52.0 ± 8.8 years cholesterol,
48.1% male haemoglobin, and
Jhee 2019 93.9 ± 14.2 mL/min/1.73 proteinuria levels
Singapore Smoking - "never- 13.3 years 1. Kidney failure (At Age, BMI, dialect,
smokers, former, least one of: 1) education level, history
Singapore Chinese Health Study current serum creatinine of HTN, DM, known
General population (with Captured using level of more than or heart disease or stroke,
alcohol use and intake
diabetes and heart disease) self-reported equal to 500µmol/l
of gingseng
N=63,257 questionnaire or 2) GFR <
55.6 years 15ml/min/1.73m2 or
44.2% male 3) undergoing
Jin 2013 Baseline GFR NP dialysis, or 4) had
undergone kidney
transplant)
Japan Alcohol 3 years 1. Incident CKD (<60 Age, BMI, and eGFR
consumption – mL/min/1.73 m2 or were treated as time-
Measured by >25% reduction in dependent variables
glasses of Sake and GFR at follow-up)
calculated as 2. GFR decline (loss
follows: an amount per year)
of alcohol
consumed in a day
(g of alcohol/day)
Physical activity –
More than 30-
minute exercise
Saitama Cardiometabolic with sweating. The
Disease and Organ Impairment categories of
Study exercise frequency
Healthy general population were as follows: 0=
N=7473 twice a month or
38.8±10.5 years less; 1= once a
74.6% male week; 2= >twice a
Kanda 2015 78.1±15.2 mL/min/1.73 week.
Prevention of Renal and US Diet – Potassium 10.3 years 1. Incident CKD (<60 Age, sex, height,
Vascular End-stage Disease and sodium mL/min/1.73 m2) weight, smoking status,
(PREVEND) Captured using 2. Albuminuria (>30 alcohol consumption,
General population (1.4% DM, self-reported FFQ mg/24 hour) parental history of
CKD, race, diabetes,
11.8% antihypertensive
and urinary potassium,
medicated patients) calcium, urea, and
N=5,315 creatinine excretion,
48.3 years baseline eGFR and
47.5% male UAE
Kieneker 2016 98.2 mL/min/1.73
Japan Alcohol 1.8 years 1. Proteinuria Age, current smokers,
Retrospective cohort – cohort consumption - rare (albuminuria; BMI, mean arterial
name NP drinkers, occasional dipstick urinary pressure, HDL-
Healthy general population drinkers, and daily protein ≥ 1 +) cholesterol,
haemoglobin A1c,
N=177,572 drinker; with
eGFR, current treatment
40-75 years ethanol intake of HTN, dyslipidaemia,
49.9% male ≤19,20-39,40-59 DM and history of CVD
Kimura 2018 77.5 mL/min/1.73 and ≥60g/day
Captured using
self-reported
questionnaire
US Alcohol 11 years 1. GFR decline (eGFR Age, BMI, protein
consumption - ≥20%, ≥25% and intake,
daily alcohol was ≥30%) hypercholesterolaemia,
none, 0.1-4.9d/day, diabetes, HTN and
smoking status
5-14.9g/day, and
15-59.9g/day
Captured using
Nurses’ health study self-reported
Healthy general population questionnaire
N=1,658
30-55 years Diet – protein
100% women Captured using
Knight 2003 >89 mL/min/1.73 self-reported FFQ
US Alcohol 10.2 years 1. Incident CKD Age, sex, height,
consumption - no (eGFR < 60ml/min weight, smoking status,
alcohol, occasional and/or UA >30mg parental history of
<10g/week, light mean of 2 CVD, history of CVD,
education level and
10-69.9 g/week, measurements)
Prevention of Renal and potential mediators like
moderate 70- homeostatic model
Vascular End-stage Disease 210g/week, assessment-insulin
(PREVEND) heavier>210g/week resistance, use of
Healthy general population glucose-lowering drugs,
N=5476 TC to HDL cholesterol
48.4 ± 11.7 years ratio, use of lipid-
47.4% male lowering drugs, BP
Koning 2015 97.3 ± 14.8 mL/min/1.73 drugs
Norway Alcohol 6 years 1. eGFR decline Baseline eGFR
consumption – (change at follow-
Self-reported as (1) up)
alcohol abstention;
(2), ≤ 3 units of
alcohol/week; (3)
Tromsø Study >3 and ≤ 6 units of
General population (56.9% alcohol/week (4) >
HTN, 2.3% DM, 11.8% CVD) 6 units of
N=4441 alcohol/week.
59 years
Kronborg 50.6% male Physical activity –
2008 93 mL/min/1.73 Self-reported active
(>1-hour hard
activity/week;
and/or >3-h light
activity; or inactive
Smoking - Self-
reported as current
smokers; former
smokers; or never
smoked
The Atherosclerosis Risk in US Physical activity - 3 years 1. GFR decline Unadjusted
Communities (ARIC) Study assessed by (change at follow-
Diabetes mellitus (Black & interview based on up)
White ethnicities) a modification of
N=1,434 the questionnaire
56 years developed by
43.1% male Baecke and
Krop 1999 97.2 mL/min/1.73 colleagues
US Diet – Fish intake 5.4 years 1. Incident CKD (<60 Age, sex, center,
(Indians) Captured using mL/min/1.73 m2) diabetes status, TG,
The Strong Heart Study self-reported FFQ 2. Albuminuria systolic BP, WHR,
Healthy general population (Urinary ACR=30 to baseline GFR, smoking,
alcohol consumption,
N=2,261 299 mg/g),
energy intake, protein
39 years macroalbuminuria intake, sodium intake
39.6% male (urinary ACR>300
Lee 2012 102.4 mL/min/1.73 mg/g))
Singapore Diet – Caffeine 16.8 years 1. End-stage kidney Age, sex, dialect group,
Captured using disease year of recruitment, and
self-reported FFQ education, BMI,
smoking status, physical
activity, alcohol
consumption, red meat
intake, total protein
intake, self-reported
Singapore Chinese Health Study
history of diabetes,
Healthy general population HTN, stroke, and CVD,
N=63,257 consumption of other
56.5 years caffeine-containing
44.3% male beverages (coffee, black
Lew 2018 Baseline GFR NP tea, green tea, and soda)
Singapore Diet – Red meat, 15.5 years 1. End-stage kidney Age, gender, dialect,
Singapore Chinese Health Study red & processed disease educational level, year
Healthy general population meats, protein, of interview, BMI,
Lew 2017 N=63,257 physical activity,
56.5 years poultry, fish & smoking status, alcohol
44.3% male shellfish, eggs, use, history of disease
Baseline GFR NP dairy, soy & (diabetes, HTN, stroke,
legumes) heart attack) at baseline,
and total energy intake
Captured using
self-reported FFQ
Nurses’ Health Study US Diet - Sugar 11 years 1. Albuminuria (ACR Age, caloric intake,
General population (54% HTN, sweetened of 25 to 355 ug/mg) HTN, BMI, diabetes,
23% diabetes, 97% Caucasian) beverages 2. GFR decline cigarette smoking,
N=3,318 Captured using (decline >30%) activity, and CVD
67 years self-reported FFQ
100% women
Lin 2011 84 mL/min/1.73
US Diet - Animal 14 years 1. Albuminuria (ACR Age, hypertension,
protein; vegetable of 25 to 355 ug/mg) BMI, diabetes, cigarette
protein; low fat 2. GFR decline smoking, activity, CVD,
dairy; saturated fat; (decline >30%) eGFR and angiotensin-
converting enzyme
monounsaturated
inhibitor/angiotensin-
Nurses’ Health Study fat; animal fat; type 2 receptor blocker
General population (54% HTN, vegetable fat; medication use
23% diabetes, 97% Caucasian) cholesterol;
N=3,348 sodium; vitamin E;
67 years vitamin C; folate
100% women Captured using
Lin 2010 76 mL/min/1.73 self-reported FFQ
Taiwan Physical activity – 2 years 1. Incident CKD (<60 Age, sex, urbanization,
self-reported mL/min/1.73 m2) BMI, education, Marry,
A retrospective cohort study – questionnaire Household income,
drinking, smoking, CCI
cohort name NP
index, dyslipidaemia
T2DM drug, hypertension
N=559 drug, family history of
59±13 years diabetes
48.7% male /hyperlipidaemia
Lin 2014 Baseline GFR NP /HTN/heart disease
Southern Community Cohort US Smoking - ever 4.1 years 1. End-stage kidney age, recruitment, sex,
Study (SCCS) and never disease race, education, annual
General population (with HTN, Captured using household Income,
diabetes high cholesterol) self-reported HTN, diabetes,
MI/CABG, stroke,
N=79,943 questionnaire
hypercholesterolemia
40-79 years
47% male
Lipworth 2012 Baseline GFR NP
Japan Smoking – non- 6 years 1. Hyperfiltration or eGFR, age, BP,
smokers, past- proteinuria antihypertensive
smokers and (hyperfiltration was medication, diabetes
Kansay Healthcare study current-smokers. 117ml/mil per alcohol consumption,
physical activity
General population (with HTN Current 1 to 20, 1.73m2 or higher.
and diabetes) and ≥21 cigarettes Proteinuria 1+ or
N=10,118 per day) higher (30mg/dl or
48.1 ± 4.2 Captured using higher) for dipstick
100% male self-reported examination at
Maeda 2011 84.3 ± 15.4 mL/min/1.73 questionnaire follow-up))
US Diet – Protein 8 years 1. GFR decline Age, diabetes, energy
Captured using (change at follow- from protein intake, sex,
self-reported FFQ up) smoking, BMI, alcohol
use, systolic BP,
diastolic BP, years
Jackson Heart Study between creatinine
measurements, total
General population (57% HTN,
energy intake and
19% DM) percent energy from
N=3,165 saturated fat,
54.4 years polyunsaturated fat,
36% male trans fat, and
Malhotra 2018 97.4 mL/min/1.73 carbohydrate
US Alcohol 5.6 years 1. GFR decline Age, gender, race,
consumption - no (decline >3 ml/min) smoking, diabetes,
alcohol, former, systolic blood pressure,
Cardiovascular health study <1drink, 1-6drinks, diastolic BP, anti-HTN
medications, LDL
General population (elderly) 7-13 drinks and ≥
cholesterol, HDL,
N=4,343 14 drinks prevalent CVD,
72 ± 5 years Captured using prevalent heart failure,
48% male self-reported CRP and fibrinogen
Menon 2010 79 ± 26 mL/min/1.73 questionnaire
Cohort name NP Japan Smoking – yes/ no 5 years 1. Incident CKD Unadjusted
Healthy general population Captured using (eGFR<60ml/min,
N=336 self-reported proteinuria positive
52.5 ± 6.7 years questionnaire (1+ or greater) or
Michishita 100% male both)
2017 81.2 ±6.1 mL/min/1.73
Iran Diet - Nitrate 3 years 1. Incident CKD (<60 Age, sex, BMI,
Tehran Lipid and Glucose smoking, education,
containing mL/min/1.73 m2)
Study – study data from physical activity,
vegetables
2006/08-2009/11 diabetes, and HTN,
Captured using
Mirmiran Healthy general population energy intake, fiber, and
self-reported FFQ
2016 N=1,546 potassium
38.0 ± 12.0 years
43% male
80 mL/min/1.73
Tehran Lipid and Glucose Iran Diet - Sodium, 6.3 years 1. Incident CKD (<60 Sex, age, BMI,
Study Potassium, Sodium mL/min/1.73 m2) smoking, serum
General population (11.7% to Potassium Ratio 2. GFR decline creatinine, diabetes,
CVD, 19% HTN) Captured using (change at follow- HTN, medications,
CVD, history of kidney
N=1,780 self-reported FFQ up)
disease, daily energy
34 years intake, dietary intake of
Mirmiran 40.8% male protein and total fat
2018 74.8 mL/min/1.73
Tehran Lipid and Glucose Iran Diet – Fiber 6.1 years 1. Incident CKD (<60 Age, sex, smoking, total
Study Captured using mL/min/1.73 m2) energy intake, physical
General population (8.2% self-reported FFQ activity, diabetes and
diabetes, 12.5% HTN) using angiotensin-
converting-enzyme
N=1,630
inhibitor
42.8 ± 11.2 years
Mirmiran 49.5% male
2018 73.7 mL/min/1.73
Cohort name NP Japan Smoking – yes v 5 years 1. Incident CKD Unadjusted
Healthy general population no; non-current vs (defined as an eGFR
N=286 non-smokers <60ml/min/1.73m2)
46.4 ± 9.2 years Captured using
Miyatake 100% male self-reported
2010 86.3 ± 15.1 mL/min/1.73 questionnaire
The Korean Multi-Rural Korea Diet – Potassium 4 years 1. Incident CKD (<60 Age, sex, gender, BMI,
Communities Cohort Study Captured using mL/min/1.73 m2) smoking, alcohol and
General population (Mix of self-reported FFQ 2. GFR decline exercise, diabetes, total
normotension and hypertension) (change at follow- cholesterol, triglyceride
and high-density
N=5,064 up)
lipoprotein, calorie,
61.3 years protein, and sodium
37.6% male intake
NTN 75.8 mL/min/1.73; HTN
Mun 2019 73.8 mL/min/1.73
Sakasaki Health cohort Japan Smoking – current 7.7 years 1. Incident CKD (<60 Age, sex, BMI, systolic
Healthy general population vs non-current mL/min/1.73 m2) BP, glucose,
N=1811 Captured using cholesterol, uric acid,
45.5 years self-reported leukocyte count, fasting
urine protein, pH
Nakanishi 61.1% male questionnaire
2012 75.8 ± 11.1 mL/min/1.73
Korean Genome Epidemiology Korea Diet - 11.7 years 1. Incident CKD (<60 Age, sex, baseline
Study Carbohydrate mL/min/1.73 m2 eGFR, WHR, education
General population (42.6% intake and/or the status, marital status,
HTN, 3.6% CVD, 3% Captured using development of smoking status,
exercise, hypertension,
dyslipidaemia) self-reported FFQ proteinuria)
and CVD, haemoglobin,
N=7804 HOMA-IR, albumin,
51.3 ± 8.6 years in non- HDL cholesterol, and
diabetics; 55.4 ± 8.8 years in TG
diabetics
49.7% male
Nam 2019 Baseline GFR NP
Japan Smoking –– non- 6 years 1. Incident CKD (<60 Sex, age, BMI, BP,
smokers, ex- mL/min/1.73 m2 IGR, TC, HDL, TG,
smokers, current and/or the GFR
Cohort name NP smokers limit 1 development of
Healthy general population pack/day or >1pack proteinuria)
N=6998 day
41.7 years Captured using
Noborisaka 58.9% male self-reported
2012 83.8 mL/min/1.73 questionnaire
US Smoking – current 9.5 years 1. Albuminuria Age, sex, systolic BP,
vs non-current (UACR ≥17 mg/g diastolic BP, mean
Framingham Offspring Cohort
Captured using men or ≥ 25mg/g arterial pressure, pulse
Healthy general population pressure, diabetes,
N=1916 self-reported women)
glucose, LDL and HDL,
56 years questionnaire
TG, BMI, waist
O´Seaghdha 52.4% male circumference, WHR,
2010 89.5 mL/min/1.73 UAR, GFR
Vienna Health Screening Vienna Smoking – current 7.1 years 1. Incident CKD (<60 Age, sex, CKD, BMI,
Project smokers, non- mL/min/1.73) Sports=no, uric acid,
Healthy general population smokers, ex- HDL, BP, HTN,
N=17375 smoker glucose
20-89 years Captured using
Obeymayr 53.6 self-reported
2008 93.8 mL/min/1.73 questionnaire
The Multi-Ethnic Study of US Physical activity - 10.2 years 1. Incident CKD (<60 Age, sex, race/ethnicity,
Atherosclerosis (MESA) ideal, intermediate mL/min/1.73) education, and income
Healthy general population and poor
N=6506 Captured using
62.0 ± 10.2 years self-reported
Ogunmoroti 47.2% male questionnaire
2016 Baseline GFR NP
Cohort name NP Japan Diet - 24h home 10.5 years 1. GFR decline (mean Unadjusted
HTN sodium urine change at visit 2
N=133 collection minus GFR at visit
60±9 years 1)
39.8% male
Ohta 2013 71.7±14.6 mL/min/1.73
Japan Alcohol 10.4 years 1. Incident CKD (<60 GFR, age, BMI, systolic
consumption - mL/min/1.73 m2 BP, diastolic BP, fasting
non-drinker, light and/or the plasma glucose,
drinker, moderate development of smoking habits and
regular leisure-time
Kansay Healthcare study drinker, heavy proteinuria)
physical activity
Healthy general population drinkers
N=9116 Captured using
48.2 ± 4.2 self-reported
100% male questionnaire
Okada 2019 85.0 ± 14.1 mL/min/1.73
Taiwans National Health Taiwan Alcohol use 6.47 y and 1. Incident CKD Age, sex, comorbidities
Insurance cohort disorder ICD-09- 7.23 y (eGFR <60ml/min and nonsteroidal anti-
Healthy general population CM codes 303;305 control or urinary albumin inflammatory drugs
N=11639 Captured using group excretion of 30
>20 years verified health mg/day for more
77.59% male records than 3 months)
Pan 2018 Baseline GFR NP
US Diet - Omega-3 25 years 1. Incident CKD (<60 Age, sex, race, study
polyunsaturated mL/min/1.73) center, BMI, education,
fatty acids current smoker, alcohol
Captured using 30- consumption, physical
activity, and total
day food recall
energy, personal kidney
problems, dietary
intakes of magnesium,
calcium, sodium,
potassium, and
Coronary Artery Risk phosphorous, creatinine
Development in Young Adults and glucose, toenail
measurements of
(CARDIA) study
mercury, and cadmium,
Healthy general population toenail selenium. Fried
N=4,133 fish intake was adjusted
25.0 ± 3.6 only in models when
46.7% male the exposure was non-
Park 2019 123.5 ± 15.5 mL/min/1.73 fried fish intake
Vorarlberg Health Monitoring Austria Smoking - ever 17.5 years 1. End-stage kidney Age, sex, BMI, glucose,
and Prevention Program smoke yes vs no disease hypertension,
(VHM&PP) cohort Captured using triglycerides,
Healthy general population self-reported cholesterol, Gamma-GT
N=185341 questionnaire
38.9 years
46.1% male
Pscheidt 2015 Baseline GFR NP
China Smoking - current 7.13 years 1. GFR decline (25% Sex, age, obesity,
Cohort name NP smoke ≥ 10 packs or greater drop in alcohol, total
Healthy general population in the past year eGFR or a sustained cholesterol, triglyceride,
N=2518 Captured using decline in eGFR of HDL, self-reported
health status, education,
50.2 ± 5.9 years self-reported more than 5
physical activity
53.1% male questionnaire ml/min/1.73m2/year
Qin 2015 108.3 ± 13.4 mL/min/1.73 )
Iran Physical activity - 12.4 years 1. Incident CKD (<60 Adjustment details NP
Doing exercise or mL/min/1.73)
labor less than
three times a week
or performing
activities achieving
Tehran Lipid and Glucose < 600 metabolic
Study equivalent of task
Healthy general population (MET)] VS
N=8238 physical inactive
39.3 ± 13.3 years Captured using
Ramezankhani 46.1% male self-reported
2017 75.6±10.5 mL/min/1.73 questionnaire
US Diet - protein, 21 years 1. Incident CKD (<60 Age, sex, race-center,
potassium, calcium, mL/min/1.73) total caloric intake,
Atherosclerosis Risk in phosphorus, diabetes status, HTN
Communities (ARIC) Study magnesium, status,
General population (26% overweight/obese
vegetables
status, smoking status,
African American; 11.4% Captured using education level,
diabetes; 34.2% HTN) self-reported FFQ physical activity, and
N=15,055 baseline eGFR
54.2 years (modelled as linear
44.1% male spline terms with a knot
Rebholz 2015 103.1 mL/min/1.73 at 90 mL/min/1.73 m2)
Atherosclerosis Risk in US Diet - components 23 years 1. Incident CKD (<60 Total caloric intake,
Communities (ARIC) Study of healthy eating mL/min/1.73) age, sex, race, baseline
Rebholz 2016 General population (25% score (fruit & eGFR (linear spline
African American; 11.5% DM; vegetables, sodium, terms with a knot at 90
34% HTN) sugar-sweetened mL/min per 1.73 m2),
N=14,832 beverages, fiber, BMI, physical activity,
54.2 years sodium, fish) diabetes, and HTN
44.1% male Captured using
103.1 mL/min/1.73 self-reported FFQ
Physical activity –
Ideal levels of
physical activity:
≥150 minutes/week
of physical activity;
Poor levels of
physical activity:
None
Captured in an
Interview with
health professional
Smoking – ideal
(Never smoker or
former smoker and
quit>12 months
ago); intermediate
(Former smoker
and quit≤12 months
ago); poor (current
smoker)
US Diet - components 22 years 1. Incident CKD (<60 Age, sex, race, center,
of healthy eating mL/min/1.73) education level,
score (fruit & smoking status, physical
vegetables, sodium, activity, total caloric
intake, and all other
sugar-sweetened
factors in the DASH
beverages, red & diet score, eGFR (linear
Atherosclerosis Risk in
processed meat, spline terms with one
Communities (ARIC) Study
nuts & legumes, knot at 90 mL/min/1.73
General population (36%
wholegrains m2); overweight/obese
African American, 11.6%
sodium) status, diabetes, HTN,
diabetes, 32.5% HTN) systolic BP, use of
Captured using
N=14,882 angiotensin-converting
self-reported FFQ
54 years enzyme inhibitors or
45% male angiotensin receptor
Rebholz 2016 103.5 mL/min/1.73 blockers.
US Diet – Magnesium 5 years 1. GFR decline (3% Age, sex, race,
Healthy Aging in Captured using 2 eGFR decline per education level, health
Neighbourhoods of Diversity 24-hour dietary year) insurance status,
across the Life Span study recalls poverty status, smoking
General population (57% status, total energy
African American, 16% HTN, intake, baseline eGFR,
16% DM. Diet quality score in HTN status, diabetes
Tertile 1 was 35.2 (7.4), Diet status, BMI,
haemoglobin A1c, diet
quality score in Tertile 3 was
quality score (Healthy
52.8 (12.1) (p<0.001)) Eating Index-2010), and
N=1,252 dietary intake of fiber,
47 years sodium, calcium,
41% male potassium, and
Rebholz 2016 96.8 mL/min/1.73 phosphate
US Diet - Beverages: 23 years 1. End-stage kidney Age, sex, race-center,
Diet beverages & disease education level,
Sugar-sweetened smoking status, physical
beverage activity, total caloric
intake, baseline eGFR
Captured using
(linear spline terms with
self-reported FFQ one knot at 90 ml/min
per 1.73 m2), BMI,
Jackson Heart Study diabetes, systolic BP,
General population (13.4% serum uric acid, dietary
diabetes, 29.6% obese, 34.9% acid load, diet quality
HTN, 27% black, 44.5 average (modified Alternative
Healthy Eating Index
diet quality score across the
2010), dietary sodium,
cohort) dietary fructose,
N=15,368 frequency of
54 years consumption of sugar-
45% male sweetened beverages,
Rebholz 2017 98 ± 18mL/min/1.73 dietary phosphorus
US Diet - Beverages: 8 years 1. Incident CKD (<60 Total energy intake,
Alcohol, Fruit and mL/min/1.73) age, sex, education,
vegetable juice, BMI, smoking status,
Diet beverages & physical activity index,
HTN, diabetes, HDL
Jackson Heart Study Sugar-sweetened
cholesterol, LDL
Healthy general population beverage cholesterol, history of
N=3,003 Captured using CVD, and baseline
54 ± 12 years self-reported FFQ eGFR, healthy dietary
36% male pattern and a Southern
Rebholz 2019 98 ± 18 mL/min/1.73 dietary pattern
China Alcohol 8 years 1. End-stage kidney Age, geographic region,
consumption - disease (renal urbanization, education,
non-drinkers, <21 replacement therapy BMI, physical activity,
CHEFS cohort drinks per week or death from renal current cigarette
smoking, systolic BP,
Healthy general population and ≥ 21 drinks per failure)
history of diabetes and
N=65601 week CVD
>40 years Captured using
Reynolds 100% male self-reported
2008 Baseline GFR NP questionnaire
Japan Physical activity - 6 years 1. Incident CKD Age, BMI, eGFR, the
Exercise habits no (composite of eGFR use of anti-hypertensive
exercise habit + no or proteinuria) drugs and anti-
Cohort name NP hyperglycaemia vs hyperlipidemic agents,
and smoking and
Healthy general population exercise habit + no
drinking habits at
N=303 hyperglycaemia baseline
52.2 ± 6.7 years Captured using
100% male self-reported
Ryoma 2017 77.0 ± 10.3 ml/min/1.73 m2 questionnaire
Korea Alcohol 4 years 1. Incident CKD (<60 Age, eGFR, SBP,
consumption - mL/min/1.73) HOMA-IR, triglyceride,
alcohol amount BMI, exercise, diabetes
consumed on a
daily basis ≥
20g/day
Captured using
self-reported
methods
Smoking – Non-
smoker (<100
Census of the population of cigarettes in
Beaver Dam lifetime); former
Healthy general population (stopped smoking 1
N= 3392 year prior); current
62.3 years (current smoker at
Gender NP time of
Shankar 2006 98 mL/min/1.73 examination).
U.S. National Institutes of US Diet - sodium and 14.3 years 1. End-stage kidney Age, gender, BMI,
Health–American Association potassium disease (self- smoking, education,
of Retired Persons Diet and Captured using reported dialysis) ethnicity, physical
Health Study self-reported FFQ activity, diabetes, heart
disease, and stroke
General population (9.2% DM;
43.5% HTN, 14% heart disease,
2.1% stroke)
N=544,635
62.2 ± 5.4 years
59.1% male
Smyth 2016 Baseline GFR NP
Stengel 2003 US Alcohol 13.2 years 1. End-stage kidney Physical activity, BMI,
consumption – disease (Treatment age, gender, race,
NHANES II never; less than ESRD due to any diabetes, CVD, HTN,
Healthy general population once per week; cause or death Systolic BP, TC, eGFR
N=9082 weekly; 1 to 6 related to CKD)
49.13 ± 13.3 years times per week;
47% male daily; one or more
88.1 ± 2.6 mL/min/1.73 per day
Captured using
self-reported
methods
Japan Diet - 24-h urinary 4.8 years 1. Incident CKD (<60 Age, sex, BMI, heart
sodium excretion mL/min/1.73) rate, serum creatinine,
uric acid, TG,
haemoglobin, salt
Cohort name NP intake, proteinuria,
alcohol consumption,
General population (24.8%
smoker status, HTN,
HTN, 7% DM) diabetes, dyslipidaemia,
N=12,126 systolic BP, diastolic
52.1 ± 11.9 BP, fasting plasma
61.3% male glucose, eGFR, LDL-C,
Sugiura 2018 80.8 ± 12.9 mL/min/1.73 HDL-C
Tehran Lipid and Glucose Iran Smoking - never, 9.9 years 1. Incident CKD (<60 eGFR, Impaired fast
Study (TLGS) past former mL/min/1.73) glucose, diabetes, HTN,
Healthy general population smokers CVD, BMI,
N=5082 dyslipidaemia, TG, lipid
medication, obesity
37.3 years
45.1% male
Tohidi 2012 73.8 mL/min/1.73
Japan Alcohol 73,159 1. Proteinuria Age, BMI, smoking
consumption – person- (albuminuria; habits, regular leisure-
non-drinkers; 0.1 to years dipstick urinalysis time physical activity,
23 g ethanol/day; score of >1+) HTN, fasting plasma
glucose, eGFR
23.1-46g
Kansay Healthcare study ethanol/day; 46.1 to
Healthy general population 69g ethanol/day;
N=9154 ≥69.1 g ethanol/day
48.2 ± 4.2 Captured using
100% male self-reported
Uehara 2015 84.7 ± 14 mL/min/1.73 questionnaire
Netherlands Diet - 24-h urinary 12.7 years 1. GFR decline (60% Age, sex, race, kidney
potassium eGFR decline) function, diabetes,
excretion 2. End-stage kidney smoking, CVD history,
Cohort name NP
disease (initiation of 24h creatinine
Healthy general population excretion, HTN,
N=901 RRT
number of
Van 48 ± 14 antihypertensive drugs,
Noordenne Gender NP use of RAAS inhibitors,
2016 101 ± 27 mL/min/1.73 24h sodium excretion
Finland Physical activity - 6.4 years 1. Albuminuria Sex, duration of
Leisure-time (microalbuminuria, diabetes and current
Nationwide multicentre Finnish physical activity ≥20 and < 200 μ g / smoking, HbA1c, mean
Diabetic Nephropathy (LTPA) low vs. minor ≥ 3 0 a n d < arterial pressure, TG,
BMI
(FinnDiane) Study moderate and high 300 mg / 24h; and
T1DM intensity LTPA macroalbuminuria,
N=1,424 Captured using a ≥200 μg/min or
37.0±12.4 years validated self- ≥300 mg/24 h)
48.5% male reported 2. End-stage kidney
Waden 2015 Baseline GFR NP questionnaire disease
Japan Alcohol 1 year 1. Proteinuria Age, sex, BMI, smoke,
Specific Health Check-ups and consumption – (albuminuria; exercise, eating pattern,
Guidance System (SHC) alcohol <20g vs dipstick urinalysis HTN, diabetes,
Healthy general population >20g score of >1+) hypercholesterolemia
N=4902 ethanol/day
66.7 ± 8.4 years Captured using
Wakasugi 41.1% male self-reported
2013 79 ± 13 mL/min/1.73 questionnaire
Japan Physical activity - 1 year 1. Proteinuria Age, HTN, diabetes,
more than 30 (albuminuria; hypercholesterolemia,
minutes at a time; 2 dipstick urinalysis smoking status, BMI,
times weekly; daily score of >1+) alcohol intake, and
healthy eating habits
Specific Health Check-ups and walking equivalent
Guidance System (SHC) amount of physical
Healthy general population activity more than
N=99,404 one hour a day
40-74 years Captured using
Wakasugi 36.9% male self-reported
2017 Baseline GFR NP questionnaire
Atherosclerosis Risks in US Alcohol 8.5 years 1. Incident CKD Serum uric acid, age,
Communities (ARIC) and the consumption – (eGFR decrease of gender, race, History of
Cardiovascular Health Study current use or not ≥15ml/min/1.73 in CVD, diabetes and
(CHS) cohorts alcohol and serum creatinine HTN, education, TC,
GFR, albumin,
Healthy general population Captured using increase ≥0.4mg/dl)
haematocrit
N=13338 self-reported
58.4 years questionnaire
43.4% male
Weiner 2009 90.4 ± 19.5 mL/min/1.73
Handan Eye Study China Diet – Fruit 5.6 years 1. Albuminuria Age, sex, smoking,
Healthy general population Captured using (ACR<30 mg/g) alcohol intake, regular
N=3,574 diet-related self- consumption of fresh
Wen 2018 50 ± 11 years vegetables, BMI, waist
45.4% male reported circumference, physical
96.3 ± 14.3 mL/min/1.73 questionnaire activity, educational
level, diabetes, HTN,
CVD, antihypertensive
drugs use, SBP, TC,
HDL-C, TG, eGFR,
ACR
Australia Alcohol 5 years 1. GFR decline (eGFR Age, sex, log ACR, BP
consumption – ≥10% and final medication, diabetes,
AusDiab study <10 g/d, ≥10g to 30 eGFR >60ml/min) HbA1C, smoking
Healthy general population g/d and ≥30 g/d 2. Albuminuria (ACR status, physical activity,
WHR
N=6259 ethanol/day ≥2.5) in males and
>25 years Captured using ≥3.5mg/mmol in
44.9% male self-reported females)
White 2009 Baseline GFR NP questionnaire
Japan Alcohol 10 years 1. Incident CKD Age, eGFR, proteinuria,
consumption – (<60ml/min or haematuria, BP,
Non-alcohol dipstick proteinuria impaired glucose
consumption, >1 +) tolerance, serum lipids,
obesity
occasional, less
than 20g/day and
more than 20g/day
Captured using
self-reported
Cohort name NP questionnaire
Healthy general population
N= 123764 Smoking - non-
60 years smoker; previous
Yamagata 33.1% male smoker; current
2007 80.9 mL/min/1.73 smoker
Korean Genome and Korea Diet – Sodium 10.2 years 1. Incident CKD Age, sex, GFR, protein
Epidemiology Study Captured using (<60ml/min) intake, fat intake,
General population (14.6% DM, self-reported FFQ education, income,
3% CVD, analysis split by diabetes, TG, exercise,
and serum albumin
hypertensive status)
N=4,871
51 years
48.6% male
Yoon 2018 94 mL/min/1.73
Korea Diet - Dietary 9 years 1. Incident CKD Age, sex, WHR,
phosphorus density (<60ml/min) average protein intake,
Korean Genome and Captured using education, income,
Yoon 2017 Epidemiology Study self-reported FFQ marital status, smoking
General population (20.3% status, history of HTN,
HTN, 3.4% CVD, analysis split eGFR, fasting glucose,
by diabetes status) serum albumin, and
N=6719 HDL cholesterol
53.5 years
48.9% male
93 mL/min/1.73
Tehran Lipid and Glucose Iran Diet - Sugar 3.3 years 1. Incident CKD Age, sex, energy
Study sweetened (<60ml/min) intakes, smoking,
General population (16.2% beverages physical activity, BMI,
HTN, 10.7% DM) Captured using sodium, diabetes and
HTN
N=2,382 self-reported FFQ
45 ± 12.4 years
Yuzbashian 45.5% male
2016 70.4 ± 10.8 mL/min/1.73
Abbreviations: ACD: Alveolar capillary dysplasia, BMI: body mass index, CABG: Coronary artery bypass grafting, CKD: chronic kidney disease, CRP: c-reactive protein, CVD:
cardiovascular disease, FFQ: food frequency questionnaire, GFR: glomerular filtration rate, HbA1c: glycated haemoglobin, HOMA-IR: Homeostatic Model Assessment of Insulin
Resistance, HTN: hypertension, IGR: Impaired glucose regulation, MI: myocardial infarction, NP: not published, NTN: normotension, RAS: Renal artery stenosis, RRT: renal
replacement therapy, SCr: serum creatinine, T2DM: type 2 diabetes, TC: total cholesterol, TG: triglycerides, UAE: urine albumin excretion, WHR: waist-hip-ratio
Supplemental Table 4. Lifestyle hazards and kidney disease outcomes from individual
Incident CKD
Alcohol consumption exposure 3 cohorts; 7 ≤2 studies from N=1 (14%)
associations different cohorts harmful
with eligible data
to pool; no meta-
analysis
Buja 2011 Abstainers alcohol vs 1,430 RR 0.2 [95% CI:
(men) former men 0.05, 0.87]
Buja 2011 Abstainers alcohol vs 112 RR 1.25 [95% CI:
(women) former women 0.76, 1.88]
Hu 2018 ≤1 Drink per week vs 12,692 HR 1.11 [95% CI:
former 1.00, 1.23]
Kanda 2015 Non-obese male: 7,473 Adjusted OR
No alcohol consumed 1.226 [95% CI
vs >140g/week unable to
determine]
Kanda 2015 Non-obese female: Adjusted OR 1.92
No alcohol consumed [95% CI unable to
vs >140g/week determine]
Kanda 2015 Obese male: Adjusted OR
No alcohol consumed 1.398 [95% CI
vs >140g/week unable to
determine]
Kanda 2015 Obese female: Adjusted OR
No alcohol consumed 1.371 [95% CI
vs >140g/week unable to
determine]
Diet exposures
Advanced Glycation 1 cohort ≤2 studies; no N=0 (0%)
End Products meta-analysis
Ejtahed 2016 Total Advanced 1,692 OR 1.45 [95% CI:
Glycation End Products 0.90, 2.35]
(>9,908 compared to
<6,218KU/d)
Amino acid (L- 1 cohort ≤2 studies; no N=0 (0%)
arginine) meta-analysis
Bahadoran Total L-Arginine (1.33 1,780 RR 1.30 [95% CI:
2017 compared to 2.92) 0.79, 2.00]
Animal protein 3 cohorts ≤2 studies from N=1 (34%)
different cohorts; harmful
no meta-analysis
Dunkler 2015 Animal protein (Median 3,088 OR 0.847 [95%
(composite 0.26 compared to CI: 0.727, 0.978]
outcome) 0.71g/kg/day)
Haring 2017 Animal protein (<3.36 11,952 OR 0.91 [95% CI:
compared to 0.78, 1.06]
>69.6g/day)
Rebholz 2015 Animal protein (Q1 15,055 HR 1.06 [95% CI:
compared to Q4) 0.91, 1.23]
Bread & Cereal 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Gopinath 2011 Bread & Cereal intake 2,600 OR 0.52 [95% CI:
(328 compared to 82g) 0.25, 1.08]
B-group (1-6) 1 cohort ≤2 studies; no N=0 (0%)
Vitamins meta-analysis
Farhadnejad Thiamine (B1) - (2.99 1,692 OR 1.11 [95% CI:
2018 compared to 1.11mg) 0.57, 2.16]
Riboflavin (B2) - (3.29 OR 1.70 [95% CI:
compared to 1.04mg) 0.92, 3.14]
Niacin (B3) - (34.41 OR 1.56 [95% CI:
compared to 12.52mg) 0.80, 3.07]
Pyridoxine (B6) - (3.08 OR 1.60 [95% CI:
compared to 1.11mg) 0.83, 3.09]
Calcium 3 cohorts ≤2 studies with N=1 (34%)
eligible data to protective
pool; no meta-
analysis
Gopinath 2016 Calcium (>1226 2,600 Association not
compared to <543.1mg) reported – p>0.05
Farhadnejad Calcium (1660.2 1,692 OR 0.79 [95% CI:
2018 compared to 619.9mg) 0.39, 1.57]
Rebholz 2015 Calcium (Q1 compared 15,055 HR 0.80 [95% CI:
to Q4) 0.69, 0.92]
Cereal fiber 2 cohorts ≤2 studies; no N=1 (50%)
meta-analysis protective
Gopinath 2011 Cereal fiber (13.3 2,600 OR 0.50 [95% CI:
compared to 2.9g) 0.24, 1.03]
Mirmiran 2018 Cereal fiber (33.5 1,780 OR 0.68 [95% CI:
compared to 13.6g/day) 0.47, 0.98]
Coffee 1 cohort ≤2 studies; no N=2 (100%)
meta-analysis protective
Gopinath 2016 Coffee (>cups 1,185 OR 0.83 [95% CI:
compared to never) 0.72, 0.97]
Hu 2018 Coffee (>3 cups/day 14,209 HR 0.84 [95% CI:
compared to none) 0.75, 0.94]
Dairy 3 cohorts; 4 ≤2 studies from N=2 (50%)
associations different cohorts; protective
no meta-analysis
Asghari 2017 Low-fat dairy (Q1 1,630 HR 0.64 [95% CI:
compared to Q5) 0.38, 1.08]
Haring 2017 High fat dairy (0.13 11,952 HR 0.93 [95% CI:
compared to 1.61 0.81, 1.06]
servings/day)
Haring 2017 Low fat dairy (0 HR 0.75 [95% CI:
compared to 2.04 0.65, 0.85]
servings/day)
Rebholz 2016 Low-fat dairy (<0.1 14,882 HR 0.84 [95% CI:
compared to 1.4-10.8 0.75, 0.95]
servings/day)
Diet beverages 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Rebholz 2019 Diet beverages (T1 3,003 OR 0.80 [95% CI:
compared to T3) 0.51, 1.25]
Fiber 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Mirmiran 2018 Total fiber (33.5 1,780 OR 0.47 [95% CI:
compared to 13.6g/day) 0.28, 0.76]
Folate 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Farhadnejad Vitamin B9 (folate) - 1,692 OR 0.44 [95% CI:
2018 (628.2 compared to 0.24, 0.80]
245.8ug)
Fruit & Vegetable 1 cohort ≤2 studies; no N=0 (0%)
(combined intake) meta-analysis
Rebholz 2015 Fruit and vegetable 14,832 HR 0.97 [95%CI:
intake (<4.5 cups 0.89, 1.07]
compared to >4.5
cups/day)
Fruit juice 2 cohorts ≤2 studies; no N=0 (0%)
meta-analysis
Rebholz 2019 Fruit and vegetable 3,003 OR 1.19 [95% CI:
juice (T1 compared to 0.75, 1.90]
T3)
Yuzbashian Fruit juice (never 2,382 OR 1.04 [95% CI:
2016 compared to >2/week) 0.55, 1.96]
Legumes 2 cohorts ≤2 studies; no N=2 (50%)
meta-analysis protective
Haring 2017 Legumes (0.07 11,952 OR 0.83 [95% CI:
compared to 0.68 0.72, 0.95]
servings/day)
Mirmiran 2018 Legume fiber (0.3 1,780 OR 0.75 [95% CI:
compared to 2.34g/day) 0.50, 1.11]
Magnesium 2 cohorts ≤2 studies; no N=2 (100%)
meta-analysis protective
Farhadnejad Magnesium (581.3 1,692 OR 0.41 [95% CI:
2018 compared to 224.9mg) 0.22, 0.76]
Rebholz 2016 Magnesium (Q1 15,055 HR 0.72 [95% CI:
compared to Q4) 0.60, 0.85]
Nitrates 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Bahadoran Nitrate (<7.69 2,799 OR 0.50 [95% CI:
2016 compared to 0.24, 0.89]
>10.7mg/day)
Nitrites 1 cohort; 2 ≤2 studies; no N=0 (0%)
associations meta-analysis
Mirmiran 2016 Total nitrate containing 1,546 OR 0.93 [95% CI:
vegetables (<203 0.43, 2.02]
compared to
>332g/day)
Bahadoran Nitrite (<355 compared 2,799 OR 0.76 [95% CI:
2016 to >511mg/day) 0.43, 1.24]
Nuts 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Haring 2017 Nuts (0.03 compared to 11,952 OR 0.81 [95% CI:
0.86 servings/day) 0.72, 0.92]
Nuts & legumes 2 cohorts ≤2 studies; no N=1 (50%)
meta-analysis protective
Asghari 2017 Nuts & legumes (Q1 1,630 HR 0.52 [95% CI:
compared to Q5) 0.30, 0.90]
Rebholz 2016 Nuts & legumes (<0.4 14,882 HR 0.91 [95% CI:
compared to 1.4-10.6 0.81, 1.03]
servings/day)
Plant protein 3 cohorts ≤2 studies from N=2 (67%)
different cohorts; protective
no meta-analysis
Dunkler 2015 Plant protein (Median 3,088 OR 0.903 [95%
(composite 0.26 compared to CI: 0.769, 1.060]
outcome) 0.71g/kg/day)
Haring 2017 Vegetable protein 11,952 OR 0.76 [95% CI:
(<12.2 compared to 0.64, 0.91]
>24.5g/day)
Rebholz 2015 Vegetable protein (Q1 15,055 HR 0.72 [95% CI:
compared to Q4) 0.61, 0.85]
Red & processed 2 cohorts ≤2 studies; no N=1 (50%)
meats meta-analysis harmful
Asghari 2017 Red and processed meat 1,630 HR 1.162 [95%
(Q1 compared to Q5) CI: 0.689, 1.960]
Haring 2017 Red and processed meat 11,952 OR 1.23 [95% CI:
(0.3 compared to 1.93 1.06, 1.42]
servings/day)
Selenium 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Farhadnejad Selenium (175.1 1,692 OR 1.13 [95% CI:
2018 compared to 60ug) 0.62, 2.04]
Sodium-potassium 1 cohort ≤2 studies; no N=1 (100%)
ratio meta-analysis harmful
Mirmiran 2018 Sodium to potassium 1,780 OR 1.52 [95% CI:
ratio (>1.29 compared 1.01, 2.30]
to <0.80)
Sugar 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Gopinath 2011 Sugar intake (169 2,600 OR 2.07 [95% CI:
compared to 86.2g) 0.93, 4.59]
Trans-saturated fat 1 cohort ≤2 studies; no N=0 (0%)
intake meta-analysis
Dunkler 2015 Trans-fat (yes v no) 3,088 OR 1.004 [95%
(composite CI: 0.845, 1.193]
outcome)
Unsaturated fatty 1 cohort ≤2 studies; no N=3 (100%)
acids meta-analysis protective
Park 2019 Omega 3 FA (every SD 4,133 HR 0.72 [95% CI:
= 0.19 g/day increase) 0.58, 0.90]
DHA (every SD = 0.09 HR 0.68 [95% CI:
g/day increase) 0.56, 0.84]
EPA (every SD = 0.08 HR 0.77 [95% CI:
g/day increase) 0.61, 0.95]
Vitamin A 2 cohorts ≤2 studies; no N=0 (0%)
meta-analysis
Asgari 2017 Beta-carotene 1,179 OR 0.91 [95% CI:
(1156±129 compared to 0.59, 1.42]
2185±141ug/1000kcal)
Farhadnejad Vitamin A (859.9 1,692 OR 1.15 [95% CI:
2018 compared to 197.4ug) 0.61, 2.16]
Vitamin B12 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Farhadnejad Vitamin B12 - (7.4 1,692 OR 0.57 [95% CI:
2018 compared to 2.2ug) 0.34, 0.93]
Vitamin C 2 cohorts ≤2 studies; no N=2 (100%)
meta-analysis protective
Asgari 2017 Vitamin C (57.8±31.9 1,179 OR 0.60 [95% CI:
compared to 0.38, 0.93]
94.3±46.8mg/1000kcal)
Farhadnejad Vitamin C (268.2 1,692 OR 0.38 [95% CI:
2018 compared to 52.2mg) 0.21, 0.69]
Vitamin D 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Farhadnejad Vitamin D (4.09 1,692 OR 0.39 [95% CI:
2018 compared to 0.61ug) 0.21, 0.70]
Vitamin E 2 cohorts ≤2 studies; no N=1 (50%)
meta-analysis protective
Asgari 2017 Vitamin E (5.0±1.7 1,179 OR 0.79 [95% CI:
compared to 0.48, 1.32]
5.0±1.5mg/1000kcal)
Farhadnejad Vitamin E (17.61 1,692 OR 0.45 [95% CI:
2018 compared to 6.06mg) 0.22, 0.92]
Zinc 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Farhadnejad Zinc (17.4 compared to 1,692 OR 1.29 [95% CI:
2018 6.5mg) 0.65, 2.58]
1 cohort; 4 ≤2 studies from N=2 of 4
associations different cohorts; associations
Physical activity
no meta-analysis (50%)
harmful
Kanda 2015 Non-obese male: 7,473 Adjusted OR
<2 times/month vs >2 1.417 [95% CI
times/week unable to
determine]
Kanda 2015 Non-obese female: Adjusted OR
<2 times/month vs >2 1.494 [95% CI
times/week unable to
determine]
Kanda 2015 Obese male: Adjusted OR
<2 times/month vs >2 1.842 [95% CI
times/week unable to
determine]
Kanda 2015 Obese female: Adjusted OR
<2 times/month vs >2 1.576 [95% CI
times/week unable to
determine]
2 cohorts; 3 ≤2 studies from N=1 (34%)
Smoking exposure associations different cohorts; protective
no meta-analysis
Noborisaka Ex smoker vs 6,998 OR 1.17 [95% CI:
2012 continuous - male 0.51, 2.69]
Noborisaka Ex smoker vs OR 0.61 [95% CI:
2012 continuous - female 0.44, 0.84]
Rebholz 2016 Former smoker and quit 14,832 OR 1.03 [95% CI:
≤ 12 months ago 0.76, 1.40]
compared to current
GFR decline
Alcohol consumption 2 cohorts; 6 ≤2 studies from N=3 of 6
associations different cohorts; associations
no meta-analysis (50%)
harmful
Kanda 2015 Non-obese male: 7,473 Parameter
ln(alcohol), natural estimate = 0.0925
logarithm values of an (SEM 0.0377)
amount of alcohol p = 0.014
consumed and decrease
in GFR
Kanda 2015 Non-obese female: Parameter
ln(alcohol), natural estimate = 0.0897
logarithm values of an (SEM 0.0795)
amount of alcohol p = 0.260
consumed and decrease
in GFR
Kanda 2015 Obese male: Parameter
ln(alcohol), natural estimate = 0.204
logarithm values of an (SEM 0.061)
amount of alcohol p = 0.0009
consumed and decrease
in GFR
Kanda 2015 Obese female: Parameter
ln(alcohol), natural estimate = 0.203
logarithm values of an (SEM 0.250)
amount of alcohol p = 0.42
consumed and decrease
in GFR
Kronborg 2008 Males: >6 units of 4,441 β 0.40 (95% CI:
alcohol/week vs 0.05 to 0.76)
abstinence
Kronborg 2008 Females: >6 units of β −0.13 (95% CI:
alcohol/week vs −0.78 to 0.53)
abstinence
Diet exposures
Animal protein 3 cohorts ≤2 studies with N=0 (0%)
associations; no
meta-analysis
Dunkler 2013 Animal protein (0.81 6,213 OR 1.093 [95%
compared to CI: 0.990, 1.244]
0.27/kg/day)
Esmeijer 2018 Animal protein (Per 0.1 2,255 β annual change
g/kg actual body in eGFR: -0.12
weight) [95% CI: -0.27,
0.11]
Lin 2010 Animal protein (61.2 3,348 OR 0.91 [95% CI:
compared to 41.9g/day) 0.65, 1.27]
Antioxidants 2 cohorts ≤2 studies; no N=2 (29%)
meta-analysis protective
Hirahatake Total antioxidant (118.1 2,152 OR 0.51 [95% CI:
2017 compared to 30.5) 0.32, 0.80]
Lycopene (44.1 OR 0.84 [95% CI:
compared to 34.5) 0.58, 1.24]
a-Tocopherol (1.48 OR 0.76 [95% CI:
compared to 1.09) 0.48, 1.20]
y-Tocopherol (0.22 OR 0.96 [95% CI:
compared to 0.28) 0.63, 1.47]
Ascorbic acid (10.3 OR 0.74 [95% CI:
compared to 6.77) 0.50, 1.09]
Carbohydrates 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Dunkler 2013 High CHO food (2 6,213 OR 1.246 [95%
compared to 21.46 CI: 1.071, 1.449]
servings/day)
Cholesterol 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2010 Cholesterol (316 3,348 OR 1.12 [95% CI:
compared to 184g/day) 0.81, 1.56]
Coffee 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Herber-Gast Coffee (<1 compared to 6,113 β annual change
2017 >6cups/day) in eGFR: -0.01
[95% CI: -0.25,
0.09]
Dairy 2 cohorts; 3 ≤2 studies with N=1 (34%)
associations associations; no protective
meta-analysis
Herber-Gast Milk and milk products 6,113 β annual change
2017 (610.2 compared to in eGFR: 0.06
159.4g) [95% CI: -0.08,
0.19]
Low-fat dairy (410.5 β annual change
compared to 48.1g) in eGFR: 0.07
[95% CI: -0.05,
0.19]
Lin 2010 Low fat dairy protein 3,348 OR 0.71 [95% CI:
(4.6 compared to 0.52, 0.97]
0.82g/day)
Diet beverages 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Lin 2010 Artificial sweetened 3,348 OR 2.02 [95% CI:
soda (>2/day compared 1.36, 3.01]
to <1/month)
Folate 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2010 Folate (398 compared 3,348 OR 1.22 [95% CI:
to 244ug/day) 0.80, 1.88]
Fruit 2 cohorts ≤2 studies; no N=0 (0%)
meta-analysis
Dunkler 2013 Fruit and fruit juice (4 6,213 OR 0.945 [95%
compared to 17.32 CI: 0.862, 1.035]
serves)
Herber-Gast Fruit (330.1 compared 6,113 β annual change
2017 to 54g) in eGFR: 0.03
[95% CI: -0.06,
0.11]
Magnesium 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Rebholz 2016 Magnesium (46.9-101.4 1,252 OR 0.50 [95% CI:
compared to 131.9- 0.26, 0.95]
468.2mg/1,000 kcal)
Plant protein 3 cohorts ≤2 studies with N=0 (0%)
associations; no
meta-analysis
Dunkler 2013 Plant protein (0.3 6,213 OR 1.136 [95%
compared to CI: 0.997, 1.295]
0.14g/kg/day)
Esmeijer 2018 Plant protein (Per 0.1 2,255 β annual change
g/kg actual body in eGFR: -0.14
weight) [95% CI: -0.37,
0.08]
Lin 2010 Vegetable protein (24.9 3,348 OR 0.93 [95% CI:
compared to 17.5/day) 0.66, 1.30]
Protein 5 cohorts; ≤2 studies with N=2 (19%)
8 associations; no harmful
associations meta-analysis
Dunkler 2013 Animal protein (0.81 6,213 OR 1.093 [95%
compared to CI: 0.990, 1.244]
0.27/kg/day)
Dunkler 2013 Plant protein (0.3 OR 1.136 [95%
compared to CI: 0.997, 1.295]
0.14g/kg/day)
Esmeijer 2018 Total protein (Per 0.1 2,255 β annual change
g/kg actual body in eGFR: -0.12
weight) [95% CI: -0.04, -
0.19]
Halbesma Protein (0.26-0.99 8,461 Mean change: Q1:
2009 compared to 1.38- -0.45 vs Q4: -0.41
3.27g/kg/day)
Hirahatake Total protein 3,798 Reported not
2017 (82.4±10.5 compared to significant (p>0.1;
88.8±9.6g/day) data not reported)
Jhee 2019 Protein (1.7g/kg/day 9,226 OR 1.32 [95% CI:
compared to 1.02, 1.73]
0.6g/kg/day)
Lin 2010 Animal protein (61.2 3,348 OR 0.91 [95% CI:
compared to 41.9g/day) 0.65, 1.27]
Vegetable protein (24.9 OR 0.93 [95% CI:
compared to 17.5/day) 0.66, 1.30]
Low fat dairy protein OR 0.71 [95% CI:
(4.6 compared to 0.52, 0.97]
0.82g/day)
Malhotra 2018 Protein (18.7 compared 3,165 Mean ± SD Q1: -
– non-diabetes to 10.7% total energy 15.9 ±2.8 vs Q4: -
from protein) 20 ±1.7
Malhotra 2018 Protein (18.7 compared Mean ± SD Q1: -
– diabetes to 10.7% total energy 11.1 ±0.8 vs Q4: -
from protein) 11.3 ±1.0
Saturated fat 1 cohort ≤2 studies; no N=2 (100%)
meta-analysis harmful
Lin 2010 Saturated fat (22.9 3,348 OR 1.62 [95% CI:
compared to 14.9g/day) 1.02, 2.59]
Animal fat (37.9 OR 1.49 [95% CI:
compared to 22.3g/day) 1.08, 2.07]
Sodium 4 cohorts ≤2 studies with N=3 (75%)
associations; no harmful
meta-analysis
Deriaz 2019 Sodium (each 1 SD 4,141 β annual change
increase in sodium in eGFR: -0.07
intake) [95% CI: -0.11, -
0.04]
Dunkler 2013 Sodium (24 hour urine) 6,213 OR 1.029 [95%
(3.46 compared to CI: 0.901, 1.177]
6.41g/day)
Lin 2010 Sodium (2.4 compared 3,348 OR 1.53 [95% CI:
to 1.7g/day) 1.11, 2.09]
Ohta 2013 Sodium (<8 compared 133 Mean ± SD
to >8g/day) <8g/day: -0.41
±1.10 vs >8g/day:
-0.83 ±1.19
- correlation to
salt intake R= -
0.19, p=0.03
Sodium to Potassium 2 cohorts ≤2 studies; no N=1 (50%)
ratio meta-analysis harmful
Deriaz 2019 Sodium to Potassium 4,141 β annual change
ratio (each 1 SD in eGFR: -0.05
increase in sodium [95% CI: -0.02, -
intake) 0.08]
Mirmiran 2018 Sodium to Potassium 1,780 β annual change
ratio in eGFR: -0.81
(p=0.12)
Sugar-sweetened 1 cohort ≤2 studies; no N=0 (0%)
beverages meta-analysis
Lin 2010 Sugar soda (>1/day 3,348 OR 1.56 [95% CI:
compared to <1/day) 0.84, 2.91]
Tea 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Herber-Gast Tea (<1 compared to 6,113 β annual change
2017 >6cups/day) in eGFR: -0.01
[95% CI: -0.11,
0.04]
Unsaturated fats 1 cohort ≤2 studies; no N=1 (0%)
meta-analysis
Lin 2010 Monounsaturated fat 3,348 OR 1.06 [95% CI:
(25.1 compared to 0.64, 1.76]
16.9g/day)
Vegetable fat (Q4 OR 1.06 [95% CI:
compared to Q1) 0.78, 1.45]
Vitamin A 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis protective
Lin 2010 Vitamin A (b-carotene) 3,348 OR 0.60 [95% CI:
(5.3 compared to 0.42, 0.85]
2.2mg/day)
Vitamin E 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2010 Vitamin E (292.9 3,348 OR 0.68 [95% CI:
compared to 0.47, 1.00]
4.9mg/day)
Vegetables 4 cohorts ≤2 studies with N=1 (25%)
associations; no protective
meta-analysis
Behadoran Allium vegetable 39g 3,052 Mean (95% CI)
2017 compared to 1g/week change: (-7.8
(95% CI: -6.8, -
8.7) in T1 versus
(-4.6 (95% CI: -
4.1, -5.3) in T
Dunkler 2013 Vegetables (5 compared 6,213 OR 0.956 [95%
to 21 serves) CI: 0.827, 1.106]
Herber-Gast Vegetable (168.7 6,113 β annual change
2017 compared to 72.7g) in eGFR: 0.01
[95% CI: -0.08,
0.11]
Lin 2010 Vegetable protein (24.9 3,348 OR 0.93 [95% CI:
compared to 17.5/day) 0.66, 1.30]
Whole grains 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Herber-Gast Whole grains (139 6,113 β annual change
2017 compared to 0.27g) in eGFR: 0.01
[95% CI: -0.08,
0.09]
Physical activity 2 cohorts; 6 ≤2 studies; no N=3 of 6
associations meta-analysis associations
(50%)
harmful
Kanda 2015 Non-obese male: 7,473 Parameter
Exercise frequency (1 estimate = -0.491
increase in the score) (SEM 0.104)
p = 0.0001
Kanda 2015 Non-obese female: Parameter
Exercise frequency (1 estimate = -0.847
increase in the score) (SEM 0.214)
p = 0.0001
Kanda 2015 Obese male: Parameter
Exercise frequency (1 estimate = -0.326
increase in the score) (SEM 0.170)
p = 0.056
Kanda 2015 Obese female: Parameter
Exercise frequency (1 estimate = -0.949
increase in the score) (SEM 0.519)
p = 0.072
Kronborg 2008 Males: active versus 4,441 β −0.01 (95% CI:
inactive −0.17 to 0.16)
Kronborg 2008 Females: active versus β 0.30 (95% CI:
inactive 0.13 to 0.47)
Smoking exposures
Smokers Current or 4 cohorts ≤2 studies with N=1 (25%)
Former vs Never associations with protective
can be statistically
pooled; no meta-
analysis
Barbato 2019 Smokers vs never 637 Smokers mean
smokers GFR change: -15
(95% CI: -16.5, -
13.5)
Non-smokers
mean GFR
change: -15.8
(95% CI: -17.1, -
14.5)
Foster 2015 Current smoker yes vs 1,803 OR 1.19 [95% CI:
no 0.78, 1.81]
Miyatake 2010 Current smoker yes vs 286 Smokers mean
no GFR change: -1.9
± 12.3
Non-smokers
mean GFR
change: -5.0 ±
12.1
P=0.03
Kronborg 2008 Males: Current smokers 4,441 β 0.16 (95% CI:
versus non-smokers −0.08 to 0.39)
Kronborg 2008 Females: Current β 0.29 (95% CI:
smokers versus non- 0.07 to 0.50)
smokers
Qin 2015 Current smoker yes vs 5,244 OR 0.90 [95% CI:
no 0.61, 1.33]
Former smokers 1 cohort; 2 ≤2 studies; no N=0 (0%)
versus non-smokers associations meta-analysis
Kronborg 2008 Males: Former smokers 4,441 β 0.08 (95% CI:
versus non-smokers −0.14 to 0.30)
Kronborg 2008 Females: Former β 0.04 (95% CI:
smokers versus non- −0.18 to 0.25)
smokers
Incident albuminuria
Diet exposures
Animal protein 2 cohorts ≤2 studies; no N=0 (0%)
meta-analysis
Dunkler 2013 Animal protein (0.27 6,213 OR 0.931 [95%
compared to CI: 0.807, 1.074]
0.81g/kg/day)
Lin 2010 Animal protein (61.2 3,348 OR 1.43 [95% CI:
compared to 41.9g/day) 0.88, 2.31]
Carbohydrates 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Dunkler 2013 High CHO food (2 6,213 OR 1.106 [95%
compared to 21.46 CI: 0.934, 1.310]
servings/day)
Cholesterol 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Lin 2010 Cholesterol (316 3,348 OR 1.64 [95% CI:
compared to 184g/day) 1.06, 2.55]
Dairy 2 cohorts ≤2 studies; no N=1 (50%)
meta-analysis protective
Chang 2013 Low fat dairy 2,354 Median (IQR): 0.8
servings/day (0.2-1.4) in people
with
microalbuminuria;
Median (IQR): 0.7
(0.3-1.8) in people
without
microalbuminuria;
P=0.02
Lin 2010 Low fat dairy protein 3,348 OR 0.91 [95% CI:
(4.6 compared to 0.59, 1.39]
0.82g/day)
Diet beverages 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2011 Artificial sweetened 3,318 OR 0.92 [95% CI:
soda (>2/day compared 0.52, 1.65]
to <1/month)
Energy intake 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Chang 2013 Energy intake 2,354 Median (IQR):
(kcal/day) 3210.1 (±1442.4)
in people with
microalbuminuria;
Median (IQR):
2714.2 (±1171.7)
in people without
microalbuminuria;
P=0.01
Fish 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lee 2015 Fish (>15 compared to 2,261 OR 1.21 [95% CI:
0g per day) 0.77, 1.92]
Fruit 3 cohorts ≤2 studies with N=3 (75%)
associations; no protective
meta-analysis
Chang 2013 Fruit servings/day 2,354 Median (IQR): 0.9
(0.3-2.0) in people
with
microalbuminuria;
Median (IQR): 1.2
(0.6-2.2) in people
without
microalbuminuria;
P=0.05
Dunkler 2013 Fruit and fruit juice (4 6,213 OR 0.904 [95%
compared to 17.32 CI: 0.818, 0.999]
serves)
Herber-Gast Fruit (330.1 compared 6,113 β annual change
2016 to 54g) in eGFR: 0.03
[95% CI: -0.06,
0.11]
Wen 2018 Fruit (never compared 3,574 OR 0.56 [95% CI:
(ACR) to >3/weekly) 0.38, 0.83]
Micronutrients 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2010 Folate (398 compared 3,348 OR 0.78 [95% CI:
to 244ug/day) 0.46, 1.34]
Potassium 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Dunkler 2013 Potassium (24h urine) 6,213 OR 0.862 [95%
(1.7 compared to CI: 0.732, 1.015]
2.71g/day)
Red and processed 1 cohort ≤2 studies; no N=1 (100%)
meat meta-analysis harmful
Chang 2013 Red and processed meat 2,354 Median (IQR): 3.6
(servings/day) (2.4-5.6) in people
with
microalbuminuria;
Median (IQR): 2.4
(1.4-4.0) in people
without
microalbuminuria;
P=<0.001
Saturated fats 1 cohort ≤2 studies; no N=1 (50%)
meta-analysis harmful
Lin 2010 Saturated fat (22.9 3,348 OR 0.16 [95% CI:
compared to 14.9g/day) 0.86, 2.99]
Animal fat (37.9 OR 1.66 [95% CI:
compared to 22.3g/day) 1.08, 2.57]
Sugar-sweetened 2 cohorts ≤2 studies; no N=1 (50%)
beverages meta-analysis harmful
Chang 2013 Sugar-sweetened 2,354 Median (IQR): 1.2
beverages (0.3-2.3) in people
(servings/day) with
microalbuminuria;
Median (IQR): 0.6
(0.1-1.6) in people
without
microalbuminuria;
P=0.004
Lin 2011 Sugar soda (>1/day 3,318 OR 0.79 [95% CI:
compared to <1/day) 0.23, 2.68]
Unsaturated fats 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2010 Monounsaturated fat 3,348 OR 1.40 [95% CI:
(25.1 compared to 0.71, 2.74]
16.9g/day)
Vegetable fat (Q4 OR 0.76 [95% CI:
compared to Q1) 0.49, 1.17]
Vegetables 3 cohorts ≤2 studies with N=0 (0%)
associations; no
meta-analysis
Chang 2013 Vegetable servings/day 2,354 Median (IQR): 3.3
(2.1-4.5) in people
with
microalbuminuria;
Median (IQR): 2.9
(1.8-4.5) in people
without
microalbuminuria
Dunkler 2013 Vegetables (5 compared 6,213 OR 0.974 [95%
to 21 serves) CI: 0.828, 1.147]
Herber-Gast Vegetable (168.7 6,113 β annual change
2016 compared to 72.7g) in eGFR: 0.01
[95% CI: -0.08,
0.11]
Lin 2010 Vegetable protein (24.9 3,348 OR 0.65 [95% CI:
compared to 17.5/day) 0.42, 1.02]
Vegetable protein 2 cohorts ≤2 studies; no N=0 (0%)
meta-analysis
Dunkler 2013 Plant protein (0.14 6,213 OR 1.031 [95%
compared to CI: 0.894, 1.188]
0.3g/kg/day)
Lin 2010 Vegetable protein (24.9 3,348 OR 0.65 [95% CI:
compared to 17.5/day) 0.42, 1.02]
Vitamin A 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2010 Vitamin A (b-carotene) 3,348 OR 1.07 [95% CI:
(5.3 compared to 0.66, 1.72]
2.2mg/day)
Vitamin E 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lin 2010 Vitamin E (292.9 3,348 OR 1.57 [95% CI:
compared to 0.97, 2.57]
4.9mg/day)
Whole grains 2 cohorts ≤2 studies; no N=0 (0%)
meta-analysis
Chang 2013 Whole grains 2,354 Median (IQR): 1.4
servings/day (0.6-2.3) in people
with
microalbuminuria;
Median (IQR): 1.5
(0.7-2.7) in people
without
microalbuminuria
Herber-Gast Whole grains (139 6,113 β annual change
2017 compared to 0.27g) in eGFR: 0.01
[95% CI: -0.08,
0.09]
Smoking exposures
Smoker former v 2 cohorts; 3 ≤2 studies; no N=0 (0%)
current associations meta-analysis
Jee 2005 Non-smoker vs ex- 104,523 RR 0.97 [95% CI:
smoker women 0.46, 2.03]
Jee 2005 Non-smoker vs ex- RR 0.97 [95% CI:
smoker men 0.88, 1.06]
Noborisaka Ex-smoker vs non- 6,998 OR 1.29 [95% CI:
2012 smokers 0.48, 3.42]
End-stage kidney disease
Alcohol consumption exposure
Alcohol - high vs low 2 cohorts ≤2 studies; no N=1 (50%)
intake meta-analysis protective
Stengel 2003 Alcohol never vs daily 65,601 RR 0.90 [95% CI:
0.40, 2.20]
Reynolds 2008 Non-drinkers vs ≥ 21 9,082 RR 0.51 [95% CI:
drinks per wk 0.29, 0.87]
Diet exposures
Caffeine 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lew 2018 Caffeine (>300mg/day 63,257 HR 0.93 [95% CI:
compared to none- 0.73, 1.18]
<100mg/day)
Coffee 1 cohort ≤2 studies; no N=1 (50%)
meta-analysis protective
Hu Coffee (>3 cups/day 14,209 HR 0.83 [95% CI:
compared to none) 0.54, 1.29]
Lew 2018 Coffee (>2 cups/day 63,257 HR 0.82 [95% CI:
compared to none- 0.71, 0.96]
<1cup/day)
Dairy 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lew 2018 Dairy (Q1 lowest 63,257 HR 1.11 [95% CI:
compared to Q4 0.92, 1.35]
highest)
Diet beverages 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Rebholtz 2017 Diet soda consumption HR 1.64 [95% CI:
(<1 compared to >7 per 1.18, 2.28]
week)
Eggs 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lew 2018 Eggs (Q1 lowest 63,257 HR 1.00 [95% CI:
compared to Q4 0.83, 1.30]
highest)
Fish & Shellfish 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lew 2018 Fish & Shellfish (Q1 63,257 HR 1.07 [95% CI:
lowest compared to Q4 0.89, 1.30]
highest)
Potassium 2 cohorts ≤2 studies; no N=2 (100%)
meta-analysis protective
Smyth 2016 Potassium (<2.3 544,635 HR 1.27 [95% CI:
compared to >4.3g/day) 1.02, 1.57]
Van Potassium (>80 901 HR 0.06 [95% CI:
Noordenne compared to 0.007, 0.47]
2016 <60mmol/day)
Protein 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Lew 2018 Total protein (53.1g/day 63,257 HR 1.24 [95% CI:
compared to 1.05, 1.46]
>57.6g/day)
Red meat 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Lew 2018 Red meat (48.8/day 63,257 HR 1.40 [95% CI:
compared to 12.5g/day) 1.15, 1.71]
Sodium 1 cohort ≤2 studies; no N=1 (100%)
meta-analysis harmful
Smyth 2016 Sodium (<1.7 compared 544,635 HR 1.29 [95% CI:
to >3.6g/day) 1.02, 1.62]
Soy & Legumes 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lew 2018 Soy & legumes (Q1 63,257 HR 0.83 [95% CI:
lowest compared to Q4 0.39, 1.01]
highest)
Sugar-sweetened 2 cohorts ≤2 studies; no N=0 (0%)
beverages meta-analysis
Lew 2017 Soda (>1 cup/day 63,257 HR 1.08 [95% CI:
compared to none- 0.75, 1.55]
<monthly)
Tea 1 cohort ≤2 studies; no N=0 (0%)
meta-analysis
Lew 2018 Black tea (>2 cups/day 63,257 HR 0.88 [95% CI:
compared to less than 0.62, 1.26]
weekly)
Green tea (>2 cups/day HR 0.69 [95% CI:
compared to less than 0.28, 1.71]
weekly)
≤2 studies
reporting solely N=2 (67%)
Physical activity exposure 3 cohorts
ESKD; no meta- protective
analysis
Jafar 2015 Never VS strenuous 59,552 HR 0.58 [95% CI:
aerobic activity 0.37, 0.90]
Stengel 2003 Low physical activity 9,082 RR 0.45 [95% CI:
VS. high physical 0.24, 0.83]
activity
Waden 2015 Low VS. moderate and 1,424 HR 0.71 [95% CI:
high intensity LTPA 0.34, 1.49]
Incident hyperuricemia
Diet exposures
Sugar-sweetened 1 cohort ≤2 studies; no N=0 (0%)
beverages meta-analysis
Bomback 2010 Sugar sweetened 15,745 OR 1.17 [95% CI:
beverages (Comparing 0.95, 1.43]
>1 soda to >1 soda)
Supplement Table 5. Summary of the studies between alcohol intake and incident CKD
Alcohol Assigned Type of
intake, Dose, No. No. Adjusted HR study
Author g/day g/day cases Total (95% CI) SE
Foster 0.0 0.0 64 536 1 . Cohort
1.86-19.5 10.68 81 945 0.84(0.78,0.91) .20
19.5 19.5 26 321 0.80(0.74,0.87) .27
Hu 0.0 0.0 1007 3118 1 . Cohort
1.86g 0.93 810 2960 0.98(0.97,0.99) .05
3.72-13g 8.36 683 2592 0.86(0.81,0.92) .06
14.86-26g 20.43 247 1029 0.80(0.73,0.87) .07
27.86g 34.35 193 754 0.77(0.70,0.85) .09
Koning 0.0 0.0 100 1285 1 . Cohort
1.43g 0.72 50 860 0.99(0.98,0.99) .18
1.43-10g 5.72 96 1949 0.90(0.86,0.94) .15
10-30g 20.0 45 1121 0.80(0.73,0.87) .19
30g 45.0 9 261 0.76(0.66,0.86) .36
Okada 0.0 0.0 224 1377 1 . Cohort
0.1-23g 11.55 571 3847 0.83(0.77,0.90) .08
23.1-46g 34.55 317 2871 0.77(0.70,0.85) .09
46.1g 57.55 118 1021 0.73(0.62,0.87) .12
Supplemental Table 6: Subgroup analysis for Incident CKD
Variable (No of studies) Odds Ratio (95% CI)
Vegetables
Population:
Healthy NA
Type 2 diabetes NA
Duration:
<10 years NA
>10 years NA
Country of origin:
USA NA
Other NA
Association estimate:
Odds Ratio NA
Hazard Ratio NA
Baseline GFR:
GFR ≤89.9 mL/min/1.73 0.76 (0.60, 0.97)
GFR ≥90 mL/min/1.73 0.77 (0.70, 0.85)
Potassium intake
Population:
Healthy NA
Type 2 diabetes NA
Duration:
<10 years 0.77 (0.62-0.95)
>10 years 0.72 (0.42-1.25)
Country of origin:
USA 0.88 (0.78-0.99)
Other 0.58 (0.36-0.93)
Association estimate:
Odds Ratio 0.81 (0.63-1.03)
Hazard Ratio 0.73 (0.51-1.05)
Baseline GFR:
GFR ≤89.9 mL/min/1.73 0.68 (0.49, 0.95)
GFR ≥90 mL/min/1.73 0.89 (0.76, 1.04)
Sodium intake
Population:
Healthy NA
Type 2 diabetes NA
Duration:
<10 years 1.32 (1.02-1.70)
>10 years 1.14 (0.98-1.31)
Country of origin:
USA 1.10 (0.98-1.24)
Other 1.28 (1.04-1.58)
Association estimate:
Odds Ratio NA
Hazard Ratio NA
Baseline GFR:
GFR ≤89.9 mL/min/1.73 1.34 (1.08-1.67)
GFR ≥90 mL/min/1.73 1.08 (0.94-1.25)
Physical activity
Population:
Healthy NA
Type 2 diabetes NA
Duration:
<10 years 0.66 (0.45-0.97)
>10 years 0.78 (0.78-1.10)
Country of origin:
USA 0.89 (0.74-1.09)
Other 0.67 (0.44-1.01)
Association estimate:
Odds Ratio NA
Hazard Ratio NA
Baseline GFR:
GFR ≤89.9 mL/min/1.73 1.02 (0.89, 1.16)
GFR ≥90 mL/min/1.73 0.77 (0.56, 1.06)
Alcohol consumption
Population:
Healthy NA
Type 2 diabetes NA
Duration:
<10 years 0.88 (0.77-1.01)
>10 years 0.83 (0.72-0.95)
Country of origin:
USA 0.78 (0.69-0.88)
Other 0.90 (0.81-1.01)
Association estimate:
Odds Ratio 0.88 (0.73-1.06)
Hazard Ratio 0.86 (0.78-0.93)
Baseline GFR:
GFR ≤89.9 mL/min/1.73 0.87 (0.78, 0.98)
GFR ≥90 mL/min/1.73 0.94 (0.55, 1.61)
Tobacco smoking
Population:
Healthy NA
Type 2 diabetes NA
Duration:
<10 years 1.27 (1.09-1.49)
>10 years 1.13 (1.04-1.23)
Country of origin:
USA 1.49 (1.38-1.61)
Other 1.12 (1.06-1.19)
Association estimate:
Odds Ratio 1.22 (0.93-1.59)
Hazard Ratio 1.20 (1.11,1.29)
Baseline GFR:
GFR ≤89.9 mL/min/1.73 1.11 (1.04, 1.18)
GFR ≥90 mL/min/1.73 1.41 (1.21, 1.63)
Supplemental Table 7. Results from sensitivity analysis substituting data from secondary
publications of the same cohort dataset. The shaded rows represent the study citation included in
№ of Relative Certainty
Study design Risk of bias Inconsistency Indirectness Imprecision (95% CI)
studies
I2
Incident CKD
Fish intake and incident CKD
3 observational not serious serious a not serious not serious OR 0.94 ⨁⨁◯◯
studies (0.86 to 1.02) LOW
Fruit intake and incident CKD
4 observational serious b serious c not serious not serious OR 0.91 ⨁◯◯◯
studies (0.79 to 1.06) VERY LOW
Vegetable intake and incident CKD
5 observational not serious not serious serious d not serious OR 0.79 ⨁⨁◯◯
studies (0.70 to 0.90) LOW
Sugar-sweetened beverage consumption and incident CKD
statistically pooled across the markers of kidney function decline. Summary of the number and
Supplemental Table 4.
Supplemental Figure 8. Risk of bias across the included studies
Supplemental Figure 9: Individual assessment of risk of bias across the included studies
Supplemental Figure 10: Funnel plots for outcomes with at least 5 studies reporting
associations to a lifestyle hazard and secondary outcomes (ESKD, GFR decline and albuminuria)
References
1. Haring B, Selvin E, Liang M, Coresh J, Grams ME, Petruski-Ivleva N, et al. Dietary Protein
Sources and Risk for Incident Chronic Kidney Disease: Results From the Atherosclerosis
2. Rebholz CM, Anderson CA, Grams ME, Bazzano LA, Crews DC, Chang AR, et al.
Relationship of the AHA impact goals (life's simple 7) with risk of chronic kidney disease:
Results from the ARIC cohort study. Journal of the American Heart Association.
2016;5(4):e003192.
association between changes in lifestyle behaviors and the incidence of chronic kidney
97.
Microalbuminuria Risk in Male Cigarette Smokers: Results from the “Olivetti Heart
Study” after 8 Years Follow-Up. Kidney and Blood Pressure Research. 2019;44(1):33-42.
6. Krop JS, Coresh J, Chambless LE, Shahar E, Watson RL, Szklo M, et al. A community-
based study of explanatory factors for the excess risk for early renal function decline in
blacks vs whites with diabetes: the Atherosclerosis Risk in Communities study. Archives
7. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-
8. Friberg E, Orsini N, Mantzoros CS, Wolk A. Alcohol intake and endometrial cancer risk:
2009;18(1):355-8.