Search Strategy and Selection Criteria
A comprehensive literature search was conducted on PubMed, Embase, and the Cochrane Library databases to identify relevant studies published between January 2008 and December 2023.
The search strategy included a combination of relevant keywords and Medical Subject Headings (MeSH) terms, such as "potassium," "dietary intake," "cardiovascular disease," "coronary heart disease," "stroke," "cohort study," and "prospective study." The complete search strategy for PubMed is provided in Appendix A. Additionally, reference lists of included studies and relevant review articles were manually searched for potentially eligible studies.
Also other included study characteristics (author, publication year, study design, location, follow-up duration), participant characteristics (sample size, age, sex), exposure assessment methods (dietary assessment tool, potassium intake categories), outcome ascertainment, risk estimates with corresponding 95% CIs, and adjustment for potential confounders.
Data Extraction and Quality Assessment
Data from the eligible studies were extracted independently using a standardized data collection guidelines from PRISMA. The extracted information was done by comprehensive literature assessments utilizing the following criteria:
Participants and Populations
Inclusion criteria:
1. Study design: Prospective cohort studies published between January 2008 and December 2023.
2. Participants: Adult populations (aged 18 years and above) from the general population, without any specific restrictions on age, sex, or ethnicity.
3. Exposure: Studies that assessed dietary potassium intake using validated dietary assessment methods, such as food frequency questionnaires (FFQs), 24-hour dietary recalls, or dietary records.
4. Outcomes: Studies reporting relative risks (RRs), hazard ratios (HRs), or odds ratios (ORs) with corresponding 95% confidence intervals (CIs) for the association between potassium intake and at least one of the following cardiovascular disease outcomes: coronary heart disease (CHD), stroke, or overall CVD events.
5. Language: Studies published in English.
Exclusion criteria:
1. Study design: Case-control studies, cross-sectional studies, intervention trials, and other non-prospective cohort study designs.
2. Participants: Studies that focused solely on specific subpopulations, such as individuals with pre-existing cardiovascular disease, chronic kidney disease, or other chronic conditions that may significantly alter potassium intake or metabolism.
3. Exposure: Studies that did not assess dietary potassium intake or used non-validated dietary assessment methods.
4. Outcomes: Studies that did not report relevant risk estimates (RRs, HRs, or ORs) for the association between potassium intake and cardiovascular disease outcomes, or studies that only reported intermediate outcomes (e.g., blood pressure) without data on cardiovascular events.
5. Language: Studies published in languages other than English.
Exposures to be Reviewed
The exposure of interest in this systematic review and meta-analysis was dietary potassium intake.
Inclusion criteria:
1. Studies that assessed dietary potassium intake using validated dietary assessment methods, such as: a. Food frequency questionnaires (FFQs) b. 24-hour dietary recalls c. Dietary records
2. Studies that quantified dietary potassium intake in terms of total daily intake (e.g., milligrams or grams per day) or categorized potassium intake into quantiles, tertiles, or other predefined categories for analysis.
3. Studies that used a combination of dietary assessment methods (e.g., FFQ and 24-hour recalls) to estimate potassium intake.
Exclusion criteria:
1. Studies that did not directly assess dietary potassium intake, such as those that only measured serum or urinary potassium levels without estimating dietary intake.
2. Studies that used non-validated or inadequately described dietary assessment methods, which may lead to unreliable estimates of potassium intake.
3. Studies that only reported potassium intake from specific sources (e.g., supplements) without considering total dietary intake from foods and beverages.
4. Studies that did not report sufficient data on potassium intake categories or quantitative measures of intake, precluding their inclusion in the meta-analysis.
The inclusion of studies using various dietary assessment techniques (FFQs, 24-hour recalls, and dietary records) ensured that the review captured a wide range of evidence from different study populations and settings. The exclusion criteria were designed to minimize the inclusion of studies with potentially unreliable or incomplete data on potassium intake, which could bias the meta-analysis results.
Comparator (s)/ control
In this systematic review and meta-analysis, the comparator or control group consisted of individuals with the lowest dietary potassium intake, serving as the reference category against which higher levels of potassium intake were compared.
Inclusion criteria:
1. Studies that reported risk estimates (relative risks, hazard ratios, or odds ratios) for cardiovascular disease outcomes comparing different categories or levels of dietary potassium intake.
2. Studies that defined the reference category as the lowest level of potassium intake, such as the bottom quantile, tertile, or a predefined cut-off point (e.g., <1,500 mg/day).
3. Studies that provided sufficient data to allow for the comparison of cardiovascular disease risk between the highest and lowest categories of potassium intake, or across multiple categories of intake.
Exclusion criteria:
1. Studies that did not report risk estimates for cardiovascular disease outcomes in relation to different levels of potassium intake.
2. Studies that used a reference category other than the lowest level of potassium intake, making it difficult to compare results across studies consistently.
3. Studies that only reported continuous risk estimates (e.g., per 1 g/day increase in potassium intake) without providing data on specific intake categories or levels.
4. Studies that did not provide sufficient data to allow for the comparison of cardiovascular disease risk between different categories of potassium intake.
Study Eligibility Criteria
Studies were eligible for inclusion in this systematic review and meta-analysis if they met the following criteria:
1. Study design: Prospective cohort studies, including nested case-control studies within prospective cohorts. Other study designs within the broader category of prospective cohort studies were also considered.
2. Population: Studies involving adult participants (aged 18 years or older) from the general population, without any specific restrictions on age, sex, or ethnicity. Studies focusing solely on specific subpopulations, such as individuals with pre-existing cardiovascular disease or chronic kidney disease, were excluded.
3. Exposure assessment: Studies that assessed dietary potassium intake using validated dietary assessment methods, such as food frequency questionnaires (FFQs), 24-hour dietary recalls, or dietary records. Studies that measured potassium intake through urinary excretion or biomarkers were also considered.
4. Outcome measures: Studies reporting relative risks (RRs), hazard ratios (HRs), or odds ratios (ORs) with corresponding 95% confidence intervals (CIs) for the association between potassium intake and at least one of the following cardiovascular disease outcomes:
· Coronary heart disease (CHD)
· Stroke
· Overall cardiovascular disease (CVD) events, including myocardial infarction, heart failure, and cardiovascular mortality
5. Language: Studies published in the English language.
6. Publication date: Studies published between January 2008 and December 2023 were considered to ensure the inclusion of the most recent and relevant evidence.
Exclusion criteria:
1. Randomized controlled trials (RCTs): Although RCTs are generally considered the gold standard for assessing the effectiveness of interventions, they were not included in this systematic review. RCTs typically have shorter follow-up periods and may not be feasible or ethical for studying the long-term effects of dietary potassium intake on cardiovascular disease risk.
2. Case-control studies: These studies compare the dietary potassium intake of individuals with cardiovascular disease (cases) to that of healthy individuals (controls). Case-control studies were excluded because they are prone to recall bias and may not accurately reflect the temporal relationship between potassium intake and disease risk.
3. Cross-sectional studies: These studies assess dietary potassium intake and cardiovascular disease prevalence at a single point in time. Cross-sectional studies were excluded because they cannot establish a temporal relationship between potassium intake and disease risk and may be subject to reverse causation bias.
4. Ecological studies: These studies compare dietary potassium intake and cardiovascular disease rates at the population level, rather than at the individual level. Ecological studies were excluded because they are prone to ecological fallacy and may not accurately reflect the association between potassium intake and disease risk at the individual level.
5. Case reports, case series, and other observational study designs that do not meet the inclusion criteria for prospective cohort studies.
Publication Bias and Sensitivity Analyses
Publication bias was assessed visually using funnel plots and quantitatively using Egger's regression test [17]. In the presence of potential publication bias, trim-and-fill analyses were performed to estimate the impact of missing studies on the pooled effect estimates and adjust for potential bias.
Assessment of Certainty of Evidence
The certainty of evidence for each outcome was assessed using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach [28]. The GRADE approach evaluates evidence based on five domains:
1. Risk of Bias:
· We assessed the risk of bias in individual studies using established criteria, such as selection bias, performance bias, detection bias, attrition bias, and reporting bias.
· Studies with high risk of bias were downgraded by one or two levels.
2. Inconsistency:
· We evaluated inconsistency by examining the variability in results across studies.
· Significant heterogeneity (I² > 50%) led to downgrading the certainty of evidence.
3. Indirectness:
· Indirectness was assessed by considering the population, intervention, comparator, and outcomes (PICO) framework.
· Studies that did not directly address our research question were downgraded.
4. Imprecision:
· Imprecision was evaluated by examining the width of confidence intervals and the total number of events.
· Wide confidence intervals or a small number of events led to downgrading the certainty of evidence.
5. Publication Bias:
· We assessed publication bias through visual inspection of funnel plots and statistical tests such as Egger's test.
· Evidence of publication bias led to downgrading the certainty of evidence.
Study Quality Assessment
The methodological quality of the included studies was assessed using the Newcastle-Ottawa Scale (NOS) for cohort studies [13]. The NOS evaluates studies based on three main categories:
1. Selection (4 points):
· Representativeness of the exposed cohort
· Selection of the non-exposed cohort
· Ascertainment of exposure
· Demonstration that the outcome of interest was not present at the start of the study
2. Comparability (2 points):
· Comparability of cohorts on the basis of the design or analysis, with a maximum of two points awarded for controlling for important confounding factors
3. Outcome (3 points):
· Assessment of outcome
· Was follow-up long enough for outcomes to occur?
· Adequacy of follow-up of cohorts
Studies can receive a maximum of 9 points, with higher scores indicating a lower risk of bias and higher quality. Studies with a NOS score ≥ 7 were considered high quality.
Risk of Bias Assessment The risk of bias in the included studies was evaluated using the Risk of Bias Assessment Tool for Nonrandomized Studies (RoBANS) [14]. This tool assesses the risk of bias across six domains: selection of participants, confounding variables, measurement of exposure, blinding of outcome assessments, incomplete outcome data, and selective outcome reporting.
In addition to the NOS and RoBANS, the following characteristics of the included studies were assessed:
· Study design (e.g., prospective cohort, nested case-control)
· Method of dietary potassium intake assessment (e.g., food frequency questionnaire, 24-hour recall, urinary potassium excretion)
· Validity and reliability of the dietary assessment method
· Method of ascertaining cardiovascular disease outcomes (e.g., medical records, self-report, death certificates)
· Validity and reliability of the outcome assessment method
· Adjustment for important confounding factors (e.g., age, sex, body mass index, smoking status, physical activity, other dietary factors)
· Reporting of statistical methods and effect estimates with 95% confidence intervals
· Potential sources of bias (e.g., selection bias, information bias, confounding)
Sensitivity Analyses
Sensitivity analyses were performed to assess the robustness of the findings by excluding studies with a high risk of bias or low methodological quality, as determined by the risk of bias assessment and quality assessment tools. These analyses included:
1. Excluding studies with a high risk of bias or low quality based on the NOS assessment.
2. Using alternative effect estimates (e.g., ORs instead of RRs or HRs) if reported by the studies.
3. Excluding studies that used urinary potassium excretion as a proxy for dietary potassium intake.
4. Using fixed-effect models instead of random-effects models.
All statistical analyses will be performed using the "metafor" package in R (version 4.0.5 or higher). A two-tailed p-value < 0.05 will be considered statistically significant for all analyses.
The meta-analysis was conducted using random-effects models to account for potential heterogeneity across studies.
Statistical Analysis
Risk estimates (RRs, HRs, or ORs) with corresponding 95% confidence intervals (CIs) were extracted from the included studies for the comparison of the highest versus the lowest category of potassium intake. When multiple risk estimates were reported, the most adjusted model was prioritized to account for potential confounding factors.
To synthesize the study results, a meta-analysis was conducted using random-effects models to calculate the pooled risk estimates and 95% CIs for the association between potassium intake and CVD outcomes. The random-effects model was chosen to account for potential heterogeneity across studies.
Heterogeneity across the included studies was quantified using the Cochran's Q statistic and the I² statistic. The I² statistic represents the proportion of total variation across studies due to heterogeneity rather than chance, with values of 25%, 50%, and 75% indicating low, moderate, and high heterogeneity, respectively [16].
Subgroup and Dose-Response Analyses
Subgroup analyses were conducted to explore potential sources of heterogeneity and to investigate whether the relationship between dietary potassium intake and cardiovascular disease risk differs according to certain study or participant characteristics. The following subgroups will be investigated:
1. Sex: Studies were stratified by sex (male vs. female) to assess whether the association between dietary potassium intake and cardiovascular disease risk differs between men and women.
2. Age: Studies were grouped by the mean or median age of the participants at baseline (e.g., <50 years vs. ≥50 years) to investigate whether the association varies by age.
3. Study location: Studies were categorized by geographical region (e.g., North America, Europe, Asia) to assess whether the association differs across populations with potentially different dietary patterns and cardiovascular disease risk profiles.
4. Method of assessing dietary potassium intake: Studies were grouped according to the method used to assess dietary potassium intake (e.g., food frequency questionnaire, 24-hour recall, urinary potassium excretion) to evaluate whether the association varies depending on the assessment method.
5. Follow-up duration: Studies were categorized by the median or mean follow-up duration (e.g., <10 years vs. ≥10 years) to assess whether the association differs by the length of follow-up.
6. Adjustment for confounding factors: Studies were grouped according to whether they adjusted for important confounding factors (e.g., age, sex, BMI, smoking status, physical activity, other dietary factors) to investigate whether the association varies depending on the level of adjustment.
For each subgroup analysis, a random-effects meta-analysis was performed to pool the effect estimates within each subgroup category. The heterogeneity within each subgroup will be assessed using the Cochran's Q test and the I² statistic, as described in the "Strategy for data synthesis" section.
To compare the effect estimates between subgroups, a test for subgroup differences will be conducted using a chi-squared test. A significant test for subgroup differences (p < 0.05) will indicate that the association between dietary potassium intake and cardiovascular disease risk differs significantly between the subgroups.
If there are insufficient studies or data to conduct meaningful subgroup analyses for any of the pre-specified factors, the reasons for not conducting the analyses will be reported.
Additionally, a dose-response analysis was conducted to investigate the potential non-linear relationship between potassium intake and CVD risk. Study-specific risk estimates were extracted across different potassium intake levels, and a two-stage random-effects dose-response meta-analysis was performed using restricted cubic splines [18]. This analysis aimed to determine the shape of the dose-response relationship and identify potential thresholds or ranges of potassium intake associated with the lowest risk of CVD events.