Coffee consumption and cardiovascular diseases and mortality in patients with type 2 diabetes: A protocol for a systematic review and dose-response meta-analysis of cohort studies

Coffee consumption and cardiovascular diseases and mortality in patients with type 2 diabetes: A protocol for a systematic review and dose-response meta-analysis of cohort studies Hossein Shahinfar, Ahmad Jayedi, Tauseef Ahmad Khan, Sakineh Shab-Bidar Review question Is there an association between coffee drinking and the incidence of cardiovascular disease (CVD) and mortality in patients with type 2 diabetes (T2D)? Searches The systematic literature search will be conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses: the PRISMA statement. The systematic search will be conducted by using pre-defined search terms in PubMed, Scopus, and Web of Siences. Titles and abstracts will be screened according to the pre-defined inclusion and exclusion criteria to identify eligible studies. Full texts will be retrieved and independently assess for eligibility by two review authors. Any disagreements will be resolved by consensus. Reference lists of all potential relevant articles and reviews will be screened to find further potential relevant studies. The systematic search will be restricted to articles published in English. Search terms: (Coffee OR caffeine OR coffea OR beverage OR beverages) AND (Mortality OR death OR survival OR "cardiovascular disease" OR "cardiovascular diseases" OR "coronary heart disease" OR "coronary artery disease" OR "cardiovascular event" OR "cardiovascular events" OR "ischemic heart disease" OR "coronary event" OR "coronary events" OR stroke OR "myocardial infarction" OR CVD OR CHD OR IHD) AND (“Prospective OR prospectively OR retrospective OR retrospectively OR observational OR longitudinal OR cohort OR cohorts OR nested OR case-control OR "relative risk" OR RR OR "hazard ratio" OR HR OR "odds ratio" OR "follow-up" OR "follow up") AND (Diabetes OR diabetic) Types of study to be included Prospective observational studies. Study Selection Inclusion and exclusion criteria were defined according to the PICOS (population, intervention/exposure, comparator, outcome, and study design) framework. Published prospective cohort studies will be considered eligible for inclusion in the present meta-analysis if they had the following criteria: 1) prospective observational studies that were conducted in patients with T2D aged 18 years or older; 2) reported coffee consumption as exposure 3) considered all-cause and cause-specific mortality and CVD including coronary heart disease (CHD), stroke, and total CVD vents as the outcomes of interest; 4) provided estimates of relative risk, hazard ratio (HR) or rate ratio with corresponding 95% confidence intervals (CIs) for ≥2 quantitative categories of coffee consumption; and 5) reported the number of cases and noncases and person-years in each category of a coffee consumption. Studies that were conducted in the general population and reported the results in patients with T2D in the subgroup analyses or in supplemental files will also be eligible. Review studies, interventional studies, and studies conducted in populations other than patients with T2D will be excluded. Two independent investigators (HSH and AJ) will carry out an initial screening of all titles and abstracts from retrieved papers to identify potential eligible studies that should be included in the analysis. If the same dataset had been published in more than one publication, we will include the one with greater participants. Condition or domain being studied The following outcome will be investigated: All-cause mortality, CVD mortality, CHD mortality, cancer mortality, total CHD, and stroke. Participants/population Inclusion: Patients with T2D aged 18 years and older. Exclusion: General population, healthy adults, patients with type 1 diabetes, children, adolescent, and pregnant or breastfeeding women Intervention(s), exposure(s) Coffee drinking Comparator(s)/control Low intake of coffee or dose-response. Main outcome(s) All-cause mortality Additional outcome(s) Incidence of CVD including CHD and stroke, and mortality from CVD, CHD, and cancer. Data extraction (selection and coding) Two independent researchers (HSH and AJ) will record the following characteristics from the identified studies: first author’s name, date of publication, country, age range, sex, study participants, number of cases, duration of follow-up, method of assessment of coffee consumption, and list of variables that were entered into the multivariable model as potential confounders. Risk of bias (quality) assessment Quality of the original studies included in meta-analysis will be evaluated using a 9-point Newcastle–Ottawa Scale. Accordingly, studies with 1–3, 4–6, and 7–9 points will be rated poor, fair, and high quality, respectively. Two independent investigators (HSH and AJ) will perform the quality assessment to examine the possible risk of bias associated with each of the included studies. The certainty of the evidence will be assessed by using the GRADE tool. This tool grades the evidence as high, moderate, low, or very low quality. Observational studies such as prospective cohort studies start as low-quality evidence that can be downgraded or upgraded on the basis of pre-specified criteria. The criteria used to downgrade the evidence include study limitations, inconsistency, indirectness, imprecision, and publication bias. The criteria used to upgrade the quality of the evidence include a large magnitude of association, a dose–response gradient, and attenuation by plausible confounding. Disagreements will be solved by consulting the principal investigator (SS-B). Strategy for data synthesis We will perform a random-effect meta-analysis to estimate the HR and its 95%CI as the effect size in the present meta-analysis. For studies that reported the effect size as relative risks, we will consider them as equal to HR. Three types of analyses will be carried out in the present meta-analysis. First, we will perform a pairwise meta-analysis to combine the HRs for the highest compared with the lowest category of coffee drinking in each primary study. For studies with sex-specific effect sizes, we will combine sex-specific estimates using a fixed-effects model and will use the combined effect size for the analyses. The Cochran Q and I2 statistics will be used to test for heterogeneity. We will perform a series of subgroup analyses to detect potential sources of heterogeneity, based on gender, geographic location, follow-up duration, number of participants and adjustments for main confounders including energy (yes or no), body mass index (BMI) (yes or no), smoking (yes or no), alcohol drinking (yes or no) and physical activity (yes or no). Publication bias will be examined by visual inspection of funnel plots when at least 10 primary studies were available. Formal statistical assessment of funnel plot asymmetry will also be done with Egger’s regression asymmetry test and Begg’s test. A trim and fill method will be used to detect the effect of probable missing studies on the overall effect size. We will also conduct a sensitivity analysis, in which each prospective cohort study will be excluded in turn to examine the influence of that study on the overall estimate. Second, we will perform a random-effects dose-response meta-analysis to estimate the HRs of CVD and mortality for a 1 cup/d increment in coffee drinking using the method presented by Greenland and colleagues. In this method, distribution of cases and personyears and the reported effect estimates across categories of coffee drinking will be required. We will consider the median of coffee consumption (cup per day) in each category. For studies that reported coffee consumption as a range, we will estimate the midpoint in each category by calculating the midpoint of the lower and upper bounds. When the highest and lowest categories were open-ended, we will assume the same interval as those of the adjacent category. Finally, we will perform a one-stage linear mixed effects meta-analysis to clarify the shape of the dose-response associations when at least three studies were available. This method estimates the study specific slope lines and combines them to obtain an overall average slope in a single stage, and allows to include studies with two categories of exposures in dose-response analysis. Statistical analyses will be conducted using STATA version 15.0. A two-tailed P value of less than 0.05 will be considered significant. Analysis of subgroups or subsets NA. Type and method of review Meta-analysis of prospective cohort studies Anticipated or actual start date 01 November 2020 Anticipated completion date 01 December 2020 Funding sources/sponsors None. Conflicts of interest None known Language English Country Iran

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.