Advanced Search

Study Preview



Study Title and Description

Coffee consumption and risk of chronic disease in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Germany study.



Key Questions Addressed
1 For [population], is caffeine intake above [exposure dose], compared to intakes [exposure dose] or less, associated with adverse effects on cardiovascular outcomes?
  • Comments Comments (
    0
    ) |

Primary Publication Information
  • Comments Comments (
    0
    ) |
TitleData
Title Coffee consumption and risk of chronic disease in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Germany study.
Author A Floegel,T Pischon,MM Bergmann,B Teucher,R Kaaks,H Boeing,
Country
Year 2012
Numbers

Secondary Publication Information
There are currently no secondary publications defined for this study.


Extraction Form: Cardiovascular Design
Design Details
Question... Follow Up Answer Follow-up Answer
What outcome is being evaluated in this paper? Cardiovascular
  • Comments Comments (
    0
    ) |
What is the objective of the study (as reported by the authors)? We investigated prospectively the association between coffee consumption and the risk of chronic diseases, including type 2 diabetes (T2D), myocardial infarction (MI), stroke, and cancer.
  • Comments Comments (
    0
    ) |
Provide a general description of the methods as reported by the authors. Information should be extracted based on relevance to the SR (i.e., caffeine related methods) Study population: Adults aged mainly between 35 and 65 y were recruited from the general population in Potsdam, Heidelberg, and surrounding areas with response rates of 22.7% in Potsdam and 38.3% in Heidelberg. A total of 53,088 adults (30,255 women and 22,833 men) consented to participate during the recruitment examination, which took place between 1994 and 1998. A self-administered questionnaire on sociodemographic and lifestyle factors and an FFQ were also completed. After baseline, follow-up questionnaires were administered every 2–3 y to identify incident cases of chronic diseases, including T2D, MI, stroke, and cancer. For the current analysis, we considered the data until the fourth follow-up period. Coffee consumption: The EPIC-Germany participants were inquired about their habitual consumption of "caffeinated coffee" and "decaffeinated coffee" in the self-administered semi-quantitative FFQ at baseline. Ten frequency categories were provided, ranging from "never" to "less than 1/mo" to "5 times/d or more." In addition, the usual portion size was assessed with categories ranging from 0.5 cups to 3 cups (1 cup was defined as 150 mL). From the frequency and portion size, the usual consumption of caffeinated and decaffeinated coffee was estimated. Ascertainment of disease: Incident cases of T2D, MI, stroke, and cancer were identified by participants’ self-report of the disease occurrence, disease-relevant medication, or change in diet or cause of death information. Potential incident diseases were further verified by contacting the treating physician and/or clinic to obtain medical records or via linkage to cancer and hospital registries.
  • Comments Comments (
    0
    ) |
How many outcome-specific endpoints are evaluated? 3
  • Comments Comments (
    0
    ) |
What is the (or one of the) endpoint(s) evaluated? (Each endpoint listed separately) Combined CVD.
  • Comments Comments (
    0
    ) |
List additional health endpoints (separately). 2 Stroke.
  • Comments Comments (
    0
    ) |
List additional health endpoints (separately).3 MI.
  • Comments Comments (
    0
    ) |
List additional health endpoints (separately).4
  • Comments Comments (
    0
    ) |
List additional health endpoints (separately).5
  • Comments Comments (
    0
    ) |
List additional health endpoints (separately).6
  • Comments Comments (
    0
    ) |
Clinical, physiological, other Clinical
  • Comments Comments (
    0
    ) |
What is the study design? Cohort
  • Comments Comments (
    0
    ) |
Randomized or Non-Randomized?
  • Comments Comments (
    0
    ) |
What were the diagnostics or methods used to measure the outcome? Objective
  • Comments Comments (
    0
    ) |
Optional: Name of Method or short description Incident cases of T2D, MI, stroke, and cancer were identified by participants’ self-report of the disease occurrence, disease-relevant medication, or change in diet or cause of death information. Potential incident diseases were further verified by contacting the treating physician and/or clinic to obtain medical records or via linkage to cancer and hospital registries
  • Comments Comments (
    0
    ) |
Caffeine (general)
  • Comments Comments (
    0
    ) |
Coffee, Chocolate, energy drink, gum, medicine/supplement, soda, tea, other?
  • Comments Comments (
    0
    ) |
Measured or self reported? Self-report
  • Comments Comments (
    0
    ) |
Children, adolescents, adults, or pregnant included? Adults
  • Comments Comments (
    0
    ) |
What was the reference, comparison, or control group(s)? (e.g. high vs low consumption, number of cups, etc.) <1 cup/day was the reference group. Habitual caffeinated and decaffeinated coffee consumption were considered as exposure variables. The population was divided into 5 coffee consumption categories (<1 cup/d, 1 to <2 cups/d, 2 to <3 cups/d, 3 to <4 cups/d, and >/=4 cups/d).
  • Comments Comments (
    0
    ) |
What were the listed confounders or modifying factors as stated by the authors? (e.g. multi-variable components of models.  Copy from methods) For the current analysis, we excluded participants with missing information on coffee consumption (n = 17), those with self-reported chronic diseases (T2D, MI, stroke, or cancer) at baseline (n = 5536) or with unknown disease status (n = 2392), those with missing information on lifestyle factors and other covariates (n = 1830), and those who never completed a follow-up questionnaire (n = 654). Descriptive statistics of the covariates were ageand sex-adjusted and reported across coffee-consumption categories as arithmetic means 6 SEs for continuous variables or as percentages for categorical variables. The primary endpoint for this analysis was defined as the first incident event occurring for T2D, MI, stroke, or cancer, whichever came first. Cox proportional hazards analysis was used to estimate multivariableadjusted HRs as a measure of the RR and 95% CIs, with age as the primary time-dependent variable (entry and exit time of each participant defined as the age at recruitment and age at first event of T2D, MI, stroke, cancer, or censoring, respectively). Furthermore, the disease-specific HRs for T2D, MI, stroke, and cancer across categories of coffee consumption were estimated. Coffee consumers of ,1 cup/d were used as the reference group. We calculated 3 different multivariate models for each exposure and outcome. All models were stratified by integers of age at recruitment to be less sensitive to violation of the proportional hazards assumption. Model 1 was the crude model and stratified by age at recruitment and center (Potsdam or Heidelberg) and adjusted for sex. Model 2 was additionally adjusted for classic risk factors, particularly smoking (nonsmokers; smoking intensity: <15, 15–24, or >/=25 cigarettes/d, other smoking; former smokers: gave up smoking <10, 11–20, or >20 y previously; smoking duration: <10, 11–20, 21–30 y, 31–40 y, or >40 y), alcohol consumption (non-consumers; women: .0–6, >6–12, or >12 g/d; men: >0–12, .12–24, or >24 g/d), physical activity (average of cycling and sports during summer and winter; in h/wk), education (none to primary school, technical to secondary school, or higher education including university), employment (yes or no), vitamin and mineral supplement use (during past 4 wk: yes or no), total energy intake (kcal/d), and tea intake (cups/d). Model 3 included the covariates of model 2 with additional adjustment for risk factors that could be considered intermediates, particularly BMI (in kg/m2), waist-to-hip ratio (continuous), and prevalent hypertension (at least one of the following: systolic blood pressure >/=140 mmHg, diastolic blood pressure >/=90 mmHg, blood pressure–lowering medication, or self-reported: yes or no). In addition, in models 2 and 3, caffeinated and decaffeinated coffee consumption (cups/d) was mutually adjusted. We also calculated continuous models because a dose-dependent effect was assumed. Furthermore, the significance of linear trends across the caffeinated coffee consumption categories was tested by assigning each participant the median of the category and modeling this value as a continuous variable. In addition, we conducted a competing risk analysis for Cox regression based on multivariate model 3 and used the likelihood ratio test to compare the disease-specific risk associations. Interactions were tested between coffee consumption (continuous), chronic disease risk, and the covariates age, sex, center (Potsdam or Heidelberg), alcohol intake (categorical), smoking status (categorical), BMI (>/=30, yes or no), and hypertension (yes or no). Models with and without multiplicative interaction terms were compared by using the likelihood ratio test. If an interaction was observed, a stratified analysis was conducted. For sensitivity analysis, we calculated disease-specific HRs, including subjects with any prevalent chronic disease, except the disease under investigation.
  • Comments Comments (
    0
    ) |
What conflicts of interest were reported? No conflicts of interest were reported by the authors.
  • Comments Comments (
    0
    ) |
Refid 22338038
  • Comments Comments (
    0
    ) |
What were the sources of funding? Supported by the Federal Ministry of Science, Germany (grant 01 EA 9401), the European Union (grant SOC 95 201408 05F02), and the German Cancer Aid (grant 70-2201-Bo2).
  • Comments Comments (
    0
    ) |




Results & Comparisons

No Results found.