Data on gender contrasts in the risk of incident myocardial infarction by age. The Tromsø Study 1979–2012

The data presented in this article relate to the research article entitled “Risk of incident myocardial infarction by gender: Interactions with serum lipids, blood pressure and smoking. The Tromsø Study 1979–2012” (Albrektsen et al., 2017) [1]. Data quantify the gender differences in the risk of myocardial infarction (MI) in terms of incidence rate ratios (IRR), in subgroups defined by serum lipids, blood pressure and smoking among persons aged 35–54 years, 55–74 years and 75–94 years, respectively. Data also describe the age- and gender-specific linear associations with the coronary heart disease (CHD) risk factors. IRRs for combined categories of age, gender and a CHD risk factor, with each category compared to the same reference group, are also shown. IRRs were calculated as estimates of relative risk in Poisson regression analyses of person-years at risk. Among 33,859 individuals at risk, a total of 622, 1308 and 816 were diagnosed with MI at ages 35–54, 55–74 and 75–94 years, respectively.

Data can be used for evaluation of gender heterogeneity in the association with established CHD risk factors.
Data can be used for comparing risk of incident MI between any subgroups defined by age, gender and a CHD risk factor.
Data can be utilized for exploring issues that can improve knowledge on biological mechanisms underlying the gender contrast in the risk of CHD.
At the community level, data can be utilized for development of gender-specific CHD risk preventive guidelines.

Data
Data displayed in Table 1 are age-and gender-specific incidence rate ratios (IRR) of myocardial infarction (MI) with 95% confidence intervals (CI) for the linear associations with total cholesterol, high-density lipoprotein cholesterol (HDL-C), and HDL-C in percent of total cholesterol. Data shown in Table 2 are the age-specific IRR for gender (men vs. women) within categories defined by the lipid components. Tables 3 and 4 show corresponding numbers for the interaction between gender and systolic and diastolic blood pressure, and Table 5 shows IRR for the interaction between gender and daily smoking. The number of MI-diagnoses within each subgroup defined by age, gender and a CHD risk factor, is also shown. Fig. 1A-F show IRR values from analyses of combined categories of age (35-54, 55-74 and 75-94 years), gender and each CHD risk factor, with each subgroup compared to the same reference group. The data are original and have not been published elsewhere.

Population at risk
Data displayed are calculated on the basis of information from the Tromsø Study in Norway [1,2]. Individual risk factor levels were obtained through questionnaires, blood samples and physical examinations in five repeated surveys in the calendar period 1979-2008. CHD risk factor levels in men and women at start of follow-up are given elsewhere [3]. Dates of MI-diagnoses, emigration and deaths in the period 1979-2012 were obtained from local and national registers. Among 33,859 individuals at risk (51% women), a total of 622 (80 women), 1308 (349 women) and 816 (425 women) had an MI at ages 35-54, 55-74 and 75-94 years, respectively.

Statistical analysis
The data are obtained from Poisson regression analyses of person-years at risk, with IRR of incident MI calculated as estimates of relative risk [4,5]. Information from all repeated surveys was utilized [1]. The data shown in Tables 1-5 are calculated on the basis of two-way interaction models (between gender and a CHD risk factor) in separate analyses of persons aged 35-54, 55-74 and 75-94 years (corresponding to a three-way interaction model). Within each broad age group, IRRs were adjusted for age in 1-year categories. The data displayed in Tables 1-5 quantify subgroup-specific associations with each single risk factor, but provide no information on whether a high-risk group among young people encloses, or possibly crosses the risk level of any older subgroup.
Data displayed in Fig. 1A-F are obtained from analyses of a single variable representing combined categories of age (35-54, 55-74 and 75-94 years), gender and a CHD risk factor (a unique value assigned to each possible value combination). A subgroup in middle-aged women with sufficient number of MI cases, as close as possible to normal-range or unexposed for the CHD risk factor considered, was used as common reference group when calculating the IRRs, and the risk estimates for all subgroups can be compared directly. The internal order of the IRRs will also reflect the rank of absolute risks. To ensure that persons in one particular broad age group were compared with persons exactly 20 years older or younger, additional indicator variables for age were included in the model (1-year categories, original age variable recoded 1-20 within each broad age group). The data presented in Fig. 1A-F provide information on whether a high-risk group in young people encloses the risk level of any older subgroup, but do not quantify the association with each single risk factor.
The regression models used for generating the data are an extension of the two-way interaction models applied in the original research paper [1], to three-way-interaction models used for evaluation of homogeneity across age-groups.

Transparency document. Supporting information
Transparency data associated with this article can be found in the online version at http://dx.doi. org/10.1016/j.dib.2017.07.001.