Determinants of attaining and maintaining a low cardiovascular risk profile—the Doetinchem Cohort Study

Abstract

Background: While maintenance of a low cardiovascular risk profile is essential for cardiovascular disease (CVD) prevention, few people maintain a low CVD risk profile throughout their life. We studied the association of demographic, lifestyle, psychological factors and family history of CVD with attainment and maintenance of a low risk profile over three subsequent 5-year periods. Methods: Measurements of 6390 adults aged 26–65 years at baseline were completed from 1993 to 97 and subsequently at 5-year intervals until 2013. At each wave, participants were categorized into low risk profile (ideal levels of blood pressure, cholesterol and body mass index, non-smoking and no diabetes) and medium/high risk profile (all others). Multivariable-adjusted modified Poisson regression analyses were used to examine determinants of attainment and maintenance of low risk; risk ratios (RR) and 95% confidence intervals (95% CI) were obtained. Generalized estimating equations were used to combine multiple 5-year comparisons. Results: Younger age, female gender and high educational level were associated with higher likelihood of both maintaining and attaining low risk profile ( P < 0.05). In addition, likelihood of attaining low risk was 9% higher with each 1-unit increment in Mediterranean diet score (RR: 1.09, 95% CI: 1.02–1.16), twice as high with any physical activity versus none (RR: 2.17, 95% CI: 1.16–4.04) and 35% higher with moderate alcohol consumption versus heavy consumption (RR: 1.35, 95% CI: 1.06–1.73). Conclusion: Healthy lifestyle factors such as adherence to a Mediterranean diet, physical activity and moderate as opposed to heavy alcohol consumption were associated with a higher likelihood of attaining a low risk profile.

Introduction

Maintenance of a low cardiovascular risk profile [i.e. ideal levels of blood pressure, cholesterol and body mass index (BMI), non-smoking and no diabetes] is essential for the effective prevention of cardiovascular disease (CVD). Several studies have demonstrated the benefits of a low risk profile, measured at a single point in time, in relation to the risk of coronary heart disease, stroke and total CVD. ^1–6 Recently, we indicated that adults who maintained a low risk profile over a period of 11 years had a 2.5 times lower risk of CVD when compared with adults who were low risk at baseline but deteriorated over time, emphasizing the importance of adults keeping their low risk status. ⁷

Unfortunately, most adults ‘lose’ their low risk status during young adulthood or middle age, and for those not at low risk, the likelihood of attaining a low risk profile is very low. ^7–9 Little is known about determinants that influence the likelihood of losing and achieving this low risk status. Identification of modifiable factors associated with maintaining and achieving low risk is necessary for the development of effective preventive strategies to increase the proportion of adults with a low risk profile. In addition, it is important to characterize groups that face a higher likelihood of losing their low risk status and who may benefit from earlier and/or more intensive interventions. This study investigated the association of lifestyle, demographic and psychosocial factors, history of CVD and family history of diabetes and myocardial infarction with (i) maintaining a low risk profile versus losing a low risk status and (ii) attaining a low risk profile versus remaining medium/high risk profile.

Methods

Population

The Doetinchem Cohort Study is an ongoing study which involves an age- and sex-stratified random sample of men and women aged 20–59 years in 1987–91, drawn from the civil registries of Doetinchem, the Netherlands. From 1987 to 91 (wave 1), 20 155 men and women were invited to undergo a clinical examination, of whom 62% ( N = 12 405) participated. Of these, a two-third random sample of 7768 participants was re-invited to be examined after a 6 year-interval in 1993–97 (wave 2, N = 6117) and subsequently at 5-year intervals in 1998–2002 (wave 3, N = 4918), 2003–2007 (wave 4, N = 4520) and 2008–2012 (wave 5, N = 4018). Details are described elsewhere. ¹⁰ As of the second wave, extensive information on diet, physical activity and psychosocial factors were gathered. Therefore, data from waves 2–5 were used for this study and wave 2 was considered to be the baseline examination. This resulted in 6390 participants who attended at least one examination between waves 2 and 5.

Measures

Weight, height, diastolic and systolic blood pressure, total cholesterol (TC), high-density lipoprotein (HDL) cholesterol and blood glucose were measured by trained staff according to standardized protocols. Mean diastolic and systolic blood pressure levels measured at wave 4 were unexpectedly higher compared with the blood pressure values in the other waves. No causes could be identified: therefore the blood pressure values at wave 4 were statistically corrected. This is extensively described in Supplementary Methods . Lifestyle factors, demographic characteristics, psychosocial factors and medical history were collected with standardized questionnaires completed by the participants. Details of these measurements have been described elsewhere ¹⁰ and in Supplementary Methods .

Determinants

All determinants were assessed at waves 2–4. Educational status of the participant and his/her partner (if any) were categorized as low (intermediate secondary education or less), intermediate (intermediate vocational or higher secondary education) and high (higher vocational education or university). Occupation was categorized as currently employed, homemaker or unemployed/retired/unfit for work. Marital status was categorized as married/civil union, unmarried and widow/divorced/other. Self-reported weekly alcohol intake was categorized according to recommendations of the European guidelines on CVD prevention as no alcohol consumption, moderate alcohol consumption (1–10 g/day for women and 1–20 g/day for men) or heavy alcohol consumption (>10 g/day for women and >20 g/day for men). ¹¹ Sleep duration was assessed as self-reported usual duration of sleep per 24-hour period and was categorized as short (≤6 h/day), intermediate (7–8 h/day) and long (≥9 h/day). Subjects were categorized into four groups according to the amount of physical activity performed at work and for recreational purposes using the validated index of Wareham et al. ¹² : inactive, low active, intermediate active and high active. A healthy diet was operationalized with an 8-scale modified Mediterranean diet score ¹³ that assigned values of 0 to 1 to eight nutritional components. Details are available in Supplementary Methods . All four psychosocial domains of the Dutch version of the RAND-36 were used to obtain scores for vitality, mental health, social role functioning and emotional role functioning. ^14,15 Scores are summed and transformed to a 0–5 scale, with higher scores indicating better psychosocial wellbeing.

Definition of risk profile

At each wave, participants were categorized into low risk profile and medium/high risk profile (i.e. all others). In accordance with our previous work, ⁷ similar to other studies ^1–6 and recent recommendations, ¹⁶ a low risk profile was defined as untreated systolic/diastolic blood pressure <120/<80 mmHg, untreated TC/HDL-cholesterol ratio < 4.0, BMI < 25 kg/m ² , currently non-smoking and no diabetes. Diabetes was defined based on self-reported history (i.e. response to the question, ‘Do you have diabetes?’ yes/no) and/or a random glucose concentrations of ≥11.1 mmol/l. All other participants, i.e. those with intermediate or high levels of risk factors, were defined as having a ‘medium/high risk profile’. Figure 1 schematically shows how four 5-year risk profiles were constructed: (i) ‘maintained low risk profile’: low risk two waves consecutively; (ii) ‘lost low risk status’: low risk at one wave and medium/high risk at the following wave; (iii) ‘attained low risk profile’: medium/high risk at one wave and low risk at the following wave and (iv) ‘remained medium/high risk profile’: medium/high risk two waves consecutively.

Figure 1

Schematic representation of the four 5-year risk profiles. ^a Low risk profile was defined as including all of the following: systolic blood pressure < 120 mmHg, diastolic blood pressure < 80 mmHg, not taking antihypertensive medication, TC/HDL ratio < 4.0, not taking cholesterol-lowering medication, BMI < 25 kg/m ² , not smoking and no history of diabetes. People who were not low risk at baseline were classified as having a medium/high risk profile. ^b Wave _x : Wave 2, 3 or 4.

Data analyses

Since the exclusion of participants with missing data may lead to biased results and loss of precision, ^17,18 missing values were multiple imputed as 20 datasets, using the ‘multivariate imputation by chained equations’ method in R (version 3.0.0). ¹⁹ Of the total population ( N = 6390), 4%, 23%, 29% and 37% had missing data in waves 2, 3, 4 and 5, respectively. Missing outcome data were multiple imputed since this may be superior over exclusion of the data due to high correlations between the outcome variables at consecutive waves. ^20,21 The transition from low risk to medium/high risk and vice versa was determined over three consecutive 5-year periods, i.e. waves 2–3, waves 3–4 and waves 4–5. If participants did not participate in both waves at each end of a 5-year period, we excluded the multiple-imputed data of that period, excluding 2813 of the 19 170 observations from the analyses.

A Poisson regression model using a sandwich variance estimator was used to obtain risk ratios (RRs) and 95% confidence intervals (CIs) adjusted for clustering for the associations of lifestyle, demographic and psychosocial factors, history of CVD and family history of diabetes and myocardial infarction with maintaining versus losing low risk profile and attaining low risk profile versus remaining medium/high risk profile ( figure 1 ). This method has been developed for longitudinal studies with correlated binary outcomes. ²² To combine the three 5-year periods and take the correlations amongst repeated observations on the same participants into account, generalized estimating equations with exchangeable structure were performed. When analysing changes over a 5-year period, covariates measured at the beginning of that 5-year period were used. Analyses were adjusted for age and sex in model 1 and additionally for occupational status, marital status, attained education, history of CVD, Mediterranean diet score, physical activity, alcohol consumption and sleep duration in model 2. Psychosocial factors were not adjusted for lifestyle factors, which are potential mediators in these associations. SAS software version 9.3 was used to perform all analyses. The analyses on determinants of maintaining a low risk profile were based on 852 participants who had a low risk profile at any of waves 2–4, resulting in 1325 measurements. The analyses on determinants of attaining a low risk profile were based on 6184 participants with 15 032 measurements who had a medium/high risk profile at any of waves 2–4.

As a sensitivity analysis, we categorized participants based on their cholesterol and blood pressure levels irrespective of the use of antihypertensive or cholesterol-lowering medication, because medication does lower cardiovascular risk although not to the extent of lifelong naturally low level. ^23,24 Compared with the main analyses, in sensitivity analyses few participants ( N ≤ 26 each wave) were reclassified from medium/high to low risk profile and results were the same and therefore not shown.

Results

The average age from 1993 to 97 (wave 2) was 46.3 (range: 26–65) years ( table 1 ). Participants with a low risk profile at baseline were more often women, younger, higher educated and had more favourable values for metabolic and lifestyle factors than those with a medium/high risk profile.

Table 1

Baseline characteristics (1993–97) by risk status

	Total population, n = 6368	Low risk profile, an = 601	Medium/high risk profile, bn = 5767
Demographic factors
Age (years), mean (SD)	46.3 (10.1)	41.0 (8.8)	46.9 (10.1)
Sex (women)	3383 (53%)	483 (80%)	2900 (50%)
Education (low)	3597 (56%)	245 (41%)	3352 (58%)
Occupation (employed)	3871 (61%)	405 (67%)	3466 (60%)
Civil status (married)	5219 (82%)	487 (81%)	4732 (82%)
In the risk definition
BMI (kg/m 2 ), mean (SD)	25.9 (3.8)	22.4 (1.6)	26.2 (3.8)
SBP (mmHg), mean (SD)	125 (17)	108 (7)	127 (16)
DBP (mmHg), mean (SD)	80 (11)	70 (6)	81 (11)
TC/HDL, mean (SD)	4.3 (1.5)	3.0 (0.5)	4.5 (1.5)
Currently smoking	1981 (31%)	0 (0%)	1981 (34%)
Diabetes mellitus	106 (1.7%)	0 (0.0%)	106 (1.8%)
Lifestyle factors
Physical activity (inactive)	730 (11%)	29 (5%)	702 (12%)
MDS (scale: 0–8), mean (SD)	4.0 (1.5)	4.2 (1.6)	4.0 (1.5)
Alcohol intake (g/day), median (IQR)	6 (0–16)	3 (0–10)	6 (0–17)
Sleep duration (≤6 h/day)	1011 (16%)	79 (13%)	933 (16%)

SD, standard deviation; IQR, interquartile range; SBP, systolic blood pressure; DBP, diastolic blood pressure; TC/HDL, TC/HDL ratio, MDS, Mediterranean diet score.

a: Low risk profile was defined as including all of the following: systolic blood pressure < 120 mmHg, diastolic blood pressure < 80 mmHg, not taking antihypertensive medication, TC/HDL ratio < 4.0, not taking cholesterol-lowering medication, BMI < 25 kg/m ² , not smoking and no history of diabetes.

b: Persons not low risk at baseline were classified as medium/high risk profile.

Factors associated with maintaining low risk profile

Of those having a low risk profile at any given wave, only 43% maintained that low risk profile, while 57% lost their low risk status by the following wave. Of those who were low risk at baseline, only 18% still had a low risk profile after 15 years of follow-up.

In multivariable adjusted analysis, age, gender and education were the only factors significantly associated with maintaining a low risk profile ( P < 0.05) ( table 2 ); adults with high levels of education were 29% (RR: 1.29, 95% CI: 1.03–1.61) more likely to maintain a low risk profile compared with adults with low educational attainment. Women had a 38% higher likelihood of maintaining a low risk profile compared with men (RR: 1.38, 95% CI: 1.07–1.79) and every 10-year increase in age, the likelihood of maintaining a low risk profile decreased by 26% (RR: 0.74, 95% CI: 0.65–0.84).

Table 2

Determinants of maintaining a low risk profile (maintained low risk profile versus lose low risk profile)

	RR (95% CI)
	Model 1 a	Model 2 b
Demographic factors
Age (per 10 years)	0.74 (0.67–0.83)	0.74 (0.65–0.84)
Sex (women)	1.37 (1.07–1.75)	1.38 (1.07–1.79)
Education attainment
Low	ref (1.00)	ref (1.00)
Intermediate	1.12 (0.91–1.38)	1.12 (0.91–1.39)
High	1.29 (1.04–1.59)	1.29 (1.03–1.61)
Education attainment partner
Low	ref (1.00)	ref (1.00)
Intermediate	1.10 (0.87–1.40)	1.07 (0.83–1.37)
High	1.24 (0.98–1.56)	1.13 (0.86–1.49)
Occupation
Employed	ref (1.00)	ref (1.00)
Homemaker	1.01 (0.80–1.27)	1.01 (0.79–1.27)
Other	1.01 (0.71–1.45)	1.00 (0.70–1.43)
Civil status
Married/civil union	ref (1.00)	ref (1.00)
Unmarried	0.97 (0.76–1.26)	0.97 (0.75–1.25)
Widow/divorced/other	1.01 (0.71–1.44)	0.98 (0.68–1.40)
Lifestyle factors
MDS (per unit increase)	1.01 (0.96–1.07)	1.01 (0.95–1.07)
Physical activity
Inactive	ref (1.00)	ref (1.00)
Low active	0.94 (0.63–1.41)	0.94 (0.62–1.43)
Intermediate active	0.85 (0.57–1.29)	0.86 (0.56–1.32)
High active	0.87 (0.59–1.28)	0.88 (0.59–1.32)
Alcohol intake
None	1.10 (0.86–1.40)	1.12 (0.88–1.43)
Moderate c	1.10 (0.86–1.42)	1.10 (0.86–1.41)
Heavy c	Ref (1.00)	Ref (1.00)
Sleep duration
Short (≤6 h/ day)	Ref (1.00)	Ref (1.00)
Intermediate (7–8 h/ day)	0.92 (0.73–1.15)	0.92 (0.74–1.15)
Long (≥9 h/ day)	0.91 (0.58–1.44)	0.93 (0.58–1.48)
History
CVD history	0.76 (0.16–3.73)	0.77 (0.15–3.95)
Parental history MI	0.94 (0.76–1.15)	0.96 (0.78–1.18)
Parental history DM	0.82 (0.63–1.07)	0.82 (0.63–1.07)
Psychosocial factors (range: 0–5)
Mental health	1.11 (0.98–1.27)	1.11 (0.97–1.26)
Vitality	1.04 (0.93–1.16)	1.04 (0.93–1.16)
Social role functioning	1.00 (0.92–1.09)	1.00 (0.92–1.09)
Emotional role functioning	1.02 (0.96–1.09)	1.02 (0.96–1.09)

DM, diabetes mellitus; MDS, Mediterranean diet score; MI, myocardial infarction. Significant associations (at P < 0.05) are printed in bold.

a: Model 1: adjusted for age and gender.

b: Model 2: adjusted for age, gender educational attainment, occupation, civil status, history of CVD and all lifestyle factors. Psychosocial factors were not adjusted for lifestyle factors since lifestyle factors might be an intermediate between the relation with change in risk profile.

c: Moderate alcohol intake: 1–10 g/day for women and 1–20 g/day for men; heavy alcohol intake: >10 g/day for women and >20 g/day for men.

Factors associated with attaining low risk profile

Of those having medium/high risk profile at any wave, 97% remained at medium/high risk and only 3% attained a low risk profile by the following wave. After 15 years of follow-up, 95% of those who were medium/high risk at baseline still had a medium/high risk profile.

Among those at medium/high risk at any wave, a higher educational level was associated with a higher likelihood of attaining a low risk profile by the subsequent wave ( table 3 ). Independent of individual educational attainment, having a highly educated partner also increased the likelihood of attaining a low risk profile by 71% (RR: 1.71, 95% CI: 1.24–3.35). Each 10-year increase in age was associated with 42% lower likelihood of attaining low risk profile (RR: 0.58, 95% CI: 0.52–0.65) and women were three times more likely to attain a low risk profile than men (RR: 3.06, 95% CI: 2.38–3.93).

Table 3

Determinants of attaining a low risk profile (attained low risk profile versus remained medium/high risk profile)

	RR (95% CI)
	Model 1 a	Model 2 b
Demographic factors
Age (per 10 years)	0.57 (0.52–0.62)	0.58 (0.52–0.65)
Sex (women)	2.66 (2.12–3.33)	3.06 (2.38–3.93)
Education attainment
Low	ref (1.00)	ref (1.00)
Intermediate	1.48 (1.16–1.89)	1.37 (1.07–1.75)
High	2.39 (1.87–3.04)	2.11 (1.63–2.74)
Education attainment partner
Low	ref (1.00)	ref (1.00)
Intermediate	1.39 (1.07–2.80)	1.19 (0.89–1.59)
High	2.41 (1.85–3.12)	1.71 (1.24–3.35)
Occupation
Employed	ref (1.00)	ref (1.00)
Homemaker	0.64 (0.48–0.86)	0.80 (0.59–1.07)
Other	0.89 (0.66–1.21)	1.05 (0.77–1.42)
Civil status
Married/civil union	ref (1.00)	ref (1.00)
Unmarried	1.00 (0.72–1.39)	0.95 (0.68–1.32)
Widow/divorced/other	1.09 (0.76–1.56)	1.12 (0.78–1.60)
Lifestyle factors
MDS (per unit increase)	1.13 (1.06–1.21)	1.09 (1.02–1.16)
Physical activity
Inactive	ref (1.00)	ref (1.00)
Low active	2.40 (1.28–4.47)	2.17 (1.16–4.04)
Intermediate active	3.65 (1.41–4.97)	2.35 (1.25–4.43)
High active	2.32 (1.27–4.26)	2.16 (1.17–3.98)
Alcohol intake
None	0.92 (0.72–1.19)	1.10 (0.85–1.42)
Moderate c	1.31 (1.03–1.68)	1.35 (1.06–1.73)
Heavy c	Ref (1.00)	Ref (1.00)
Sleep duration
Short (≤6 h/ day)	Ref (1.00)	Ref (1.00)
Intermediate (7–8 h/ day)	1.25 (0.94–1.67)	1.18 (0.89–1.58)
Long (≥9 h/ day)	1.02 (0.61–1.73)	1.08 (0.64–1.83)
History
CVD history d	—	—
Parental history MI	0.81 (0.66–1.01)	0.84 (0.67–1.04)
Parental history DM	0.80 (0.62–1.03)	0.83 (0.64–1.07)
Psychosocial factors (range: 0–5)
Mental health	1.16 (1.01–1.34)	1.13 (0.98–1.30)
Vitality	1.10 (0.97–1.24)	1.09 (0.96–1.24)
Social role functioning	1.06 (0.95–1.19)	1.05 (0.94–1.18)
Emotional role functioning	1.06 (0.99–1.13)	1.05 (0.98–1.12)

DM, diabetes mellitus; MDS, Mediterranean diet score; MI, myocardial infarction. Significant associations (at P < 0.05) are printed in bold.

a: Model 1: adjusted for age and gender.

b: Model 2: adjusted for age, gender educational attainment, occupation, civil status, history of CVD and all lifestyle factors. Psychosocial factors were not adjusted for lifestyle factors since lifestyle factors might be an intermediate between the relation with change in risk profile.

c: Moderate alcohol intake: 1–10 g/day for women and 1–20 g/day for men; heavy alcohol intake: >10 g/day for women and >20 g/day for men.

d: Model did not converge since no persons with history of CVD attained low risk profile.

A healthy diet, any amount of physical activity and moderate alcohol consumption were associated with improvements in risk profiles. Each 1-unit increment in the Mediterranean diet score was associated with a 9% higher likelihood of attaining a low risk profile, versus remaining at medium/high risk (RR: 1.09, 95% CI: 1.02–1.16) ( table 3 ). Any physical activity—low, intermediate and high physical activity—was associated with a more than twofold higher likelihood of attaining a low risk profile compared with being physically inactive ( P < 0.05). Participants who consumed a moderate amount of alcohol had a 35% higher likelihood of attaining low risk profile compared with heavy consumers (RR: 1.35, 95% CI: 1.06–1.73).

Role of individual major CVD risk factors in maintaining or attaining low risk

Participants who were not able to maintain a low risk profile, lost their low risk status mainly due to unfavourable changes in blood pressure levels, followed by changes in TC/HDL-cholesterol and BMI ( Supplementary table S1 ). Among those with a medium/high risk status, attainment of a low risk profile was again largely due to improvements in blood pressure levels and to a smaller extent, due to improvements in TC/HDL-cholesterol ratio, BMI and smoking status ( Supplementary table S2 ).

Discussion

For those with an existing low risk profile, the highly educated, women and younger participants were more likely to maintain a low risk profile; however, lifestyle factors did not seem to affect the likelihood of maintaining a low risk profile. Medium/high risk people who attained a low risk profile were also more likely to be highly educated, female and young but in addition to that were also more likely to adhere to the Mediterranean diet, were more often physically active and more often had moderate alcohol consumption rather than heavy alcohol consumption. Changes in blood pressure were the main contributors of both losing and attaining a low risk profile, followed by changes in BMI and TC/HDL-cholesterol ratio.

Although numerous studies have demonstrated the benefits of a low risk profile, low risk remains rare. ^1–7 The proportion of adults with a low risk profile was small in this study population but relatively high compared with the prevalence of low risk observed in most US and European studies, i.e. ranging from 10 to 44% of young adults (age range: 18–39 years) ^1,9 to 3–7% of middle-aged adults (age range: 35–79 years). ^1–4,6 A large proportion of adults lost their favourable risk status over time in this study, similar to the CARDIA participants. ⁹ Moreover, while pharmaceutical and/or lifestyle interventions to prevent CVD are largely directed towards adults with high risk profiles, most CVD cases occur among untreated adults with slightly elevated levels. ²⁵ This underscores the need to identify potential modifiable factors associated with achieving and maintaining low risk profile to facilitate the development of strategies to increase the proportion of adults at low risk and to characterize groups that face higher likelihood of loss of low risk status and exposure to adverse risk profiles who may benefit from earlier/ more intensive interventions.

The observed associations of education, gender and age with maintenance of a low risk profile are consistent with previous studies on individual CVD risk factors. It has been shown that low education, male gender and older age are independent determinants for most major CVD risk factors. ^26,27 Unfavourable changes in blood pressure were the main contributors to loss of low risk status followed by change in cholesterol and BMI. It was therefore unexpected that no lifestyle factors were associated with the maintenance of a low risk profile in this study, since physical activity and diet are important determinants for the individual major CVD risk factors. ^28,29 Although we used repeated measurements, interim changes between measurements in lifestyle factors might have occurred, which may have attenuated the results. In addition, salt intake was not sufficiently measured in our population but might be important since reducing salt intake across the population is an effective way to lower blood pressure. ³⁰

We did find modifiable determinants associated with attaining a low risk profile among those with medium/high risk profile; adherence to a Mediterranean diet, physical activity and moderate alcohol consumption increased the likelihood of attaining a low risk profile. This is consistent with the established relationship of a Mediterranean diet and alcohol intake with overweight/obesity, hypercholesterolemia, hypertension and diabetes. ^28,31–34 A dose-response relationship between physical activity and the major risk factors has often been observed, i.e. higher physical activity is associated with more favourable risk factor levels. ²⁹ We showed that any physical activity compared with none was similarly associated with a higher likelihood of attaining a low risk profile.

Our finding that only 5% of participants with elevated risk attained a low risk profile after 15 years stresses the importance of maintaining a low risk profile from young adulthood onwards. Efforts to increase the proportion of adults with a low risk profile should be especially targeted towards those with low educational levels, as this group was most vulnerable to developing unfavourable risk. The current findings also indicate that men are more susceptible to losing low risk status. For those not at low risk, healthy diets, physical activity and moderate alcohol intake can lead to improvements in risk profile. Our results indicate more specifically that even small amounts of physical activity can improve risk profiles.

The strength of this study is that extensive information about risk factors and lifestyle factors was objectively obtained at four points in time over a 15-year period, by the same group of trained staff. Limitations of this study include reliance on self-reported data on lifestyle factors. However, the physical activity and food-frequency questionnaires used have been shown to be reproducible and valid, ^11,35,36 and self-reported lifestyles were shown to be associated with CVD in the present population. ^37–39 Still, recall and misclassification bias due to socially desirable answers may have occurred, possibly resulting in attenuated associations. We focused on attainment and maintenance of an overall low risk profile since this is associated with the lowest risk of CVD. ⁷ Because of this classification, it was not possible to detect small changes in risk status which would require computation of a continuous risk score. Furthermore, individuals who participate in cohort studies are generally healthier and better educated than non-responders. Participants who were excluded and those who dropped out during follow-up also had slightly less favourable levels of the major risk factors at wave 1 (1987–91) ( Supplementary table S3 ). The underrepresentation of individuals with the worst/unhealthiest levels of the determinants may have mitigated the true difference between the lowest and highest categories, resulting in some underestimation of the observed associations. This potential bias may have been partly addressed by multiple imputation of missing values.

In conclusion, age, gender and educational attainment were the major determinants of attaining and maintaining low risk profile. Participants with lower educational levels and men had lower likelihood of attaining and maintaining low risk status; therefore, these groups may benefit from early, intensive interventions. Lifestyle factors—diet, moderate alcohol intake and physical activity—were associated with a higher likelihood of attaining low risk profile; these should therefore be a fundamental part of CVD prevention programs among adults. The low rate of attaining low risk profile underscores the difficulty of improving overall risk status once adverse risk factors have developed and subsequently the importance of maintaining a low risk status from young adulthood onwards. Finally, more research on modifiable determinants of maintaining a low risk profile is especially needed, to inform the development of effective strategies to promote the achievement and maintenance of low risk profiles.

Funding

The Doetinchem Cohort Study was supported by the Ministry of Health, Welfare and Sport of the Netherlands and the National Institute for Public Health and the Environment (grant number: S/260236/01/LC).

Conflicts of Interest : None declared.

Acknowledgements

The authors thank the field workers of the Municipal Health Services in Doetinchem (C te Boekhorst, I Hengeveld, L de Klerk, I Thus, and ir. C de Rover) for their contribution to the data collection for this study. Project director is Prof. Dr. ir. WMM Verschuren. Logistic management was provided by J Steenbrink and P Vissink and the secretarial support by EP van der Wolf. The data management was provided by ir. A Blokstra, Drs. AWD van Kessel and ir. PE Steinberger. For statistical advice, Prof. Dr. HC Boshuizen (Centre for Nutrition, Prevention and Health Services, National Institute for Public Health and the Environment Bilthoven, The Netherlands) is gratefully acknowledged.

Key points

• Age, gender and educational attainment were the major determinants of attaining and maintaining a low risk profile.

• Men and low educated adults may therefore benefit from early, intensive interventions to attain and maintain favourable risk profiles.

• Adherence to a Mediterranean diet, physical activity and moderate as opposed to heavy alcohol consumption were associated with a higher likelihood of attaining but not with maintaining a low risk profile.

• These lifestyle factors should therefore be a fundamental part of cardiovascular disease prevention programs among middle-aged adults.