Comparison of Framingham 10-Year Cardiovascular Event Risks in Native- and Foreign-born Primary Healthcare Populations in Sweden

The prevalence of cardiovascular (CVD) around the varies by and of Immigrants in may have a higher prevalence of CVD than native-born Swedes, but little is known about their actual risk of cardiovascular (CVE). The Framingham Risk uses cardiovascular risk factors to estimate the 10-year The aim of this study to examine the relationship in Sweden between an elevated 10-year CVE risk and both birthplace and other risk factors. in other countries (including the Middle East). These results contradict reports of higher rates of CVD in Middle-Eastern countries than in Sweden. A cardiovascular risk scoring system modied for region of birth or ethnicity may be needed in Sweden.


Introduction
The primary aim of this study was to examine the relationship between having an elevated 10-year risk of CVE risk and birthplace (Sweden or foreign), in primary healthcare patients living in Sweden. The secondary aim was to determine whether there were any signi cant associations in these populations between an elevated 10-year CVE risk and other potential CVD risk factors that are not part of the FRS, such as age, sex, level of education, anthropometry (waist circumference), physical activity, alcohol consumption, and hs-CRP levels.

Study population
This cross-sectional study was based on data obtained from the '4D Diabetes Project: Screening and Treatment of Prediabetes and Diabetes in Primary Care -A Pilot Study', a Programme 4D sub-project conducted by the Karolinska Institutet and the Stockholm County Council. Participants were recruited for the study from two primary healthcare centres (PHCC) in Stockholm. Participants were offered inclusion in the study if they were between 18 and 74 years old and had no previous history of diabetes or prediabetes. The Jakobsberg and Flemingsberg PHCCs were selected because large portions of their patient populations were born outside of Sweden, and particularly outside of Europe.

Data acquisition
Data were collected from participants at the two PHCCs from 2013 through 2015. At the rst visit, each participant was interviewed to obtain information about demography, country of birth, level of education, and self-reported health and lifestyle habits (physical exercise, smoking, and alcohol consumption). During that visit, BP and waist circumference were measured, and a capillary haemoglobin A1c (HbA1c) test was performed. When participants returned for a second visit, fasting blood samples were obtained, and these were analysed for serum levels of total cholesterol, HDL, calculated LDL-C, triglycerides, and hs-CRP. Also, another BP measurement was taken. For the purposes of our study, the mean values of the two BP readings were used. Also at that visit, each participant with an HbA1c ≥ 39 mmol/mol at the rst visit was offered an oral glucose tolerance test (OGTT).

Participant characteristics
For participant characteristics that represented non-Framingham CVD risk factors, we converted some of the raw data into categorical variables. Level of education categories were 0 to 9 years, 10 to 12 years, and more than 12 years. Physical exercise initially had ve response alternatives, which we collapsed into three categories: no physical activity at all, more than 0 minutes and up to 60 minutes per week, and more than 60 minutes per week. Alcohol consumption categories were less than 4 glasses a week and 4 glasses or more per week. Smoking initially had seven response options, which we collapsed into two categories: quit or never smoked and any active smoking. Diabetes had two categories: yes and no. The participant characteristics of waist circumference and hs-CRP level were employed as continuous variables in the regression analyses.

Framingham risk factors
We calculated a Framingham Risk Score (FRS) for each participant, based on six risk factors, including age, LDL-C, HDL, BP, diabetes, and smoking. Using the system described by Viera and Sheridan, points were either awarded or subtracted for each risk factor based on being above or below cut-offs, which included the following: (a) age of 35 to 39 years in men, 40-44 years in women; (b) LDL-C of 2.59 mmol/L to 4.13 mmol/L for both men and women; (c) HDL of 1.16 mmol/L to 1.54 mmol/L in men and 1.29 mmol/L to 1.54 mmol/L in women; (d) BP of 120/80 to 129/84 in men and 120/80 to 139/89 in women; (e) presence of diabetes; and f) smoking (23). A 10-year risk of CVE was then determined using total points and expressed as a percentage.

Outcome characteristics
For the purposes of this study, we de ned an elevated risk for CVE within 10 years as 10% or higher (23,25). Using this de nition, the two primary (dependent) outcome variables for the study were 10-year CVE risk of 10% or higher and 10-year CVE risk lower than 10%.

Birthplace groups
We divided the participants in this study into two groups based on birthplace: (1) Sweden-born, comprised of those who were born in Sweden, with both parents also born in Sweden; and (2) Foreignborn, comprised of those who were born outside of Sweden or had at least one parent born outside of Sweden. This was based on the o cial de nition of foreign-born used in Sweden (32).

Statistical methods
Descriptive statistics involving the prevalence of demographic, clinical, and outcome characteristics are presented as frequencies and proportions, and those involving Framingham risk factor characteristics are presented as means and standard deviations (SD). Comparisons were performed using Chi-square tests for categorical variables, T-tests for continuous variables following a normal distribution, and Wilcoxon-Rank sum tests for continuous variables not following a normal distribution. For continuous variables, 95% Con dence intervals (CI) were calculated when means were compared. Logistic regression analysis was performed to generate odds ratios (OR) and 95% CI for the outcome of an elevated 10-year CVE risk, and to generate adjusted odds ratios (AOR) for the outcome (adjusted for age, sex, education, waist circumference, physical activity, alcohol consumption, and hs-CRP levels).
A P value of 0.05 or less was considered as signi cant. Calculations were performed using STATA statistical software version 14.0 (StataCorp, College Station, Texas, USA).

Results
A total of 830 participants were included in the study. These participants were born in 69 different countries, with the largest portion born in the Middle East, including 86 in Turkey, 52 in Iran, and 48 in Iraq.

Discussion
In this cross-sectional study, we examined the relationship in participants living in Sweden between elevated 10-year CVE risk and birthplace. We also studied the relationships between elevated 10-year CVE risk and speci c cardiovascular risk factors, including age, sex, level of education, anthropometry (waist circumference), physical activity, alcohol consumption, and hs-CRP levels. The main nding was that, after controlling for other variables, participants born in Sweden were almost twice as likely to have an elevated 10-year CVE risk as those who were foreign-born. Additional ndings were that older age, male sex, lower level of education, larger waist circumference, and abnormal hs-CRP levels were also signi cantly associated with an elevated 10-year CVE risk, regardless of birthplace.
The nding in our study of a higher 10-year CVE risk in the Sweden-born than in the predominately Middle-Eastern foreign-born population was unexpected. This was inconsistent with published studies that have reported higher rates of CVD in countries in the Middle-East than in countries in Western Europe, as well as Sweden (15-19, 30, 31). In our study, foreign-born participants had signi cantly lower systolic and diastolic BP levels than Sweden-born participants, which is similar to ndings in other studies (30,31). In addition, foreign-born participants also had signi cantly lower LDL-C levels, and they were on average 10 years younger than the Sweden-born participants. All of these ndings likely contributed to the fact that a lower proportion of those in the foreign-born group than the Sweden-born group had an elevated 10-year CVE risk, despite the fact that the foreign-born group had a signi cantly lower mean HDL level and a higher prevalence of smoking.
We also observed that participants with an elevated hs-CRP level had an over three-fold higher risk of having an elevated 10-year risk of CVE, compared to those with lower hs-CRP levels. This nding was expected and has been reported by others, who have discussed the potential use of hs-CRP in assessing cardiovascular risk and have noted the lack of consensus regarding its optimal use (10). Although CRP is involved in the immunologic process that triggers vascular remodelling and plaque deposition and has been associated with increased CVE risk, de nitive evidence for its role as a causative factor of CVD is still lacking (10). Nevertheless, it makes sense that high hs-CRP levels were associated with an elevated 10-year CVE risk in our study, given that the effect of hs-CRP is mediated through in ammation, a process that may also in uence some of the Framingham risk factors, such as BP and cholesterol levels.
Our study included participants living in Sweden who had been born in a broad array of regions, which included the Middle East, Asia, and Africa. Previous studies in Sweden on cardiovascular health in foreign-born citizens have largely focused on single countries of birth, such as Iraq (19,30,31). We thought that inclusion of participants with a wider variety of ethnicities might make our results more applicable to immigrant populations throughout Sweden. However, we also acknowledge that differences in the cardiovascular health and risk pro les of individual foreign-born participants, even between those from different countries within each region, may limit the generalizability of our results to all foreign-born people in Sweden.
The issue of cardiovascular risk in immigrant populations in Sweden has received very little attention, despite a number of studies suggesting that the prevalence of CVD in these populations may be substantial (15-19, 30, 31). Studies comparing Middle Eastern and Western countries (including Sweden) have consistently found a higher prevalence of CVD in the Middle East (16,17). Other studies have reported a higher prevalence of CVD for immigrants living in Sweden who are from the Middle East (15,(19)(20)30), though a lower prevalence of CVD when those immigrants are compared to populations in their native countries (21). However, a more recent report of Swedish residents suggested that CVDrelated mortality risk was lower for non-Western-born immigrants than native Swedes (18). Our study showed that participants born in Sweden were almost twice as likely to have an elevated 10-year CVE risk as those who were foreign-born. Although these studies all looked at subtly different outcomes, these outcomes were all closely linked to cardiovascular health. As such, the contradictory results of these studies, including ours, call into question whether the cardiovascular risk scoring systems used for native populations in Western countries are applicable to immigrant populations from the Middle East.
The initial Framingham Heart Study looked at a predominantly white population of European descent when it was done in 1948, but subsequent studies in 1994 and 2003 enrolled more ethnically diverse cohorts (33). The Framingham criteria used for estimating 10-year CVE risk in our study was based on this more diverse data, but they still lack any adjustments for different countries, regions of birth, or ethnicities (23,24). The contradiction between our study results and the known high prevalence of CVD in

Limitations
The data used in our study was dependent on the population speci cally recruited for the 4D Diabetes Project in Sweden. This population involved two cluster samples, comprising all willing participants from two similar primary healthcare locations. Ultimately, the Sweden-born and foreign-born study population characteristics differed signi cantly, particularly with regard to the mean ages of the participants. Given that age is particularly impactful on cardiovascular risk, we adjusted for age in the logistic regression analyses. However, it is still possible that two randomized and more similarly matched study populations would have rendered different results. In addition, no drop-out analysis was conducted of those who chose not to accept recruitment into the project. Without information about the characteristics of those who chose not to participate, the possibility of other selection biases cannot be excluded. It is possible, for example, that Sweden-born and foreign-born patients had different reasons and thresholds for when they sought care at a PHCC. On a related note, another potential limitation of this study was that the data used was gathered from patients who were actively seeking care at a PHCC. These patients were more likely than the general public to be sick and/or have other underlying medical conditions, and this could affect the generalizability of our results to a broader population. On the other hand, cardiovascular risk is most often assessed in those who are at higher risk, because either they are older or have other comorbidities, so it may be that this population was actually optimal to study. Lastly, as this was a crosssectional study, we were unable to ascertain causality in any of the associations. Therefore, the results and conclusions of our study can at most be used to deepen the knowledge base and generate new hypotheses for future studies.

Conclusions
In a study of CVE risk in participants living in Sweden using the Framingham Risk Score, those born in Sweden were almost twice as likely to have an elevated 10-year CVE risk as those born in other countries, a large portion of whom had been born in the Middle East. Older age, male sex, and abnormal hs-CRP levels were associated with an elevated 10-year CVE risk. These results contradict previous reports of higher rates of CVD in residents of Middle-Eastern countries and in Middle-Eastern immigrants living in Sweden. A cardiovascular risk scoring system modi ed for region of birth or ethnicity may be needed in An ethical permit was granted by the Regional Ethical Review Board in Stockholm, review number 2013/2303-31/3. Data were collected after written informed consent was obtained from participants. Data were coded in order to maintain participant privacy. Both the study information sheets and the consent forms were translated from Swedish into the three most-spoken languages among the participants: Turkish, Farsi, and Arabic. Participants with abnormal blood test results were scheduled at the PHCC for further diagnosis and treatment, if indicated. Those subsequently found to have diabetes were booked at the PHCC for regular follow-up.

Consent for publication
Not applicable Availability of data and materials Sharing of the data with other researchers was not included in written informed consent and therefore neither data nor materials are publically available.

Competing interests
The authors declare that they have no competing interests.