Forty-year trends in cardiovascular risk factors in Finland

Abstract

Background: Finland has experienced remarkable changes in population levels of coronary heart disease risk factors and mortality over the past decades. The National FINRISK studies have monitored risk factors in major non-communicable diseases from 1972 to 2012. The 40-year changes in those risk factors are presented. Methods: Study population included participants aged 30–59 years in the series on independent random population samples. Data were collected in 5-year intervals in 1972–2012. FINRISK studies so far comprised 53 589 men and women who participated in a health examination, gave a venous blood sample and filled in questionnaires. Serum total cholesterol, systolic and diastolic blood pressure, and body mass index (BMI) were measured using standardized protocol, and smoking status was recorded. Results: Total serum cholesterol decreased remarkably until 2007, but after that has increased. Systolic blood pressure has continued to decline over time since 1972, while decrease in diastolic blood pressure has levelled off during the last 10 years. Smoking prevalence has markedly decreased. BMI has increased in the population, but most significantly in the earlier survey years, not the past 10 years. Conclusions: After three decades of favourable development, the population risk factor levels showed some increase in total cholesterol and diastolic blood pressure. This emphasizes the need for continued efforts towards national disease prevention and health promotion.

Introduction

Finland has experienced a great change in coronary heart disease (CHD) incidence and mortality over the past 35 years. ^1,2 In 1972, the North Karelia Project was implemented to combat the world’s highest CHD incidence rates at the time. ^3,4 The intervention program was highly successful, substantially decreasing serum total cholesterol levels, blood pressure levels and smoking prevalence. In consequence the CHD incidence was considerably lowered. ^1,5–8 Key elements of the intervention were reduced intake of saturated fat and sodium, increased consumption of vegetables and reduced smoking rates. Later, these principles were introduced to the entire population in Finland with support of the National Public Health Institute. Since launching of the North Karelia Project, CHD mortality rate decreased by 80% in the working aged population, which is clear evidence that lifestyle intervention can prevent disease. ^1,2,4

Cardiovascular diseases (CVDs) are still among the leading causes of death in Finland, despite the major changes since the 1970s. The key risk factors include high serum total cholesterol, elevated systolic and diastolic blood pressure, smoking and obesity. To date, development of these risk factors has mostly been towards a more favourable risk factor profile of the population. ^6,7 During 1972–2007, serum total cholesterol levels, systolic and diastolic blood pressure, and smoking dramatically decreased. ⁷ It is important to remember that all these risk factors are also recognized risk factors for a number of non-communicable diseases (NCDs), not just CVDs.

Risk factor monitoring has been very important to the prevention of NCDs in Finland. With 40 years of experience in systematic and standardized health examinations, these studies provide health planners and policymakers with invaluable information to guide health promotion work. The aim of this study was to report 40-year population trends in CHD risk factors in Finland. The length of systematic population-based CVD risk factor monitoring in Finland is unique and not available in any other country.

Methods

The National Institute for Health and Welfare, THL (formerly National Public Health Institute) has carried out risk factor surveys in 5-year intervals since 1972. The first and second surveys in 1972 and 1977 were carried out in eastern Finland in North Karelia and Northern Savo to evaluate the North Karelia Project. In subsequent survey years, new areas were added to improve national representativeness of the monitoring; Turku and Loimaa regions in southwestern Finland in 1982; the capital area including cities of Helsinki and Vantaa in 1992 and provinces of Northern Ostrobothnia and Kainuu (former Oulu province) in northwestern Finland in 1997.

For each survey, an independent random sample was drawn from the national population register. For the surveys in 1972 and 1977, a random sample of 6.6% of the population born between 1913 and 1947 was drawn in both survey areas. Later surveys drew a random sample stratified by sex and 10-year age group from the population aged 25–64 years separately for each survey area. For the 2007 and 2012 surveys, the age group was 25–74 years. As the sampling has differed slightly between the surveys, a common age range for all these surveys, also used in current analyses, was 30–59 years. Supplementary appendix table 1 presents the samples and participation rates by area and sex in detail, comprising altogether 53 589 participants. Participation rates have varied from year to year, showing a downward trend with rates above 90% in the 1970s and above 60% in the most recent surveys. To ensure comparable data on risk factor trends, the methodology of the surveys has been kept as similar as possible. This applies to the study years 1972 and 1977, where surveys were largely comparable and also to the study year 1982 and onwards, where the survey methodology has closely followed the WHO MONICA protocol ⁹ and the later recommendations of the European Health Risk Monitoring project. ¹⁰ Ethical approval has been covered according to commonly required procedures over the study years: confidentiality, anonymity and data protection have been assured. From 1997 onwards a written informed consent has been obtained from each participant. Surveys have obtained permissions from the local Ethics Committees that have varied over time, including the Coordinating Ethics Committee for the Uusimaa Hospital District in years 2007 and 2012.

Table 1

Serum total cholesterol (mmolċl ⁻¹ ) by area, year and sex

Area and year	Men							Women
	n	Mean	SD	≤5.0  mmolċl −1 (%)	5.0–6.49  mmolċl −1 (%)	6.5–7.99  mmolċl −1 (%)	≥8.0  mmolċl −1 (%)	n	Mean	SD	≤5.0  mmolċl −1 (%)	5.0–6.49  mmolċl −1 (%)	6.5–7.99  mmolċl −1 (%)	≥8.0  mmolċl −1 (%)
North Karelia
1972	1742	6.92	1.32	6	34	41	20	1880	6.81	1.37	7	36	38	18
1977	1762	6.52	1.22	11	41	37	11	1817	6.40	1.35	13	44	31	11
1982	1229	6.30	1.18	12	47	32	8	1268	6.12	1.32	20	46	25	9
1987	1138	6.28	1.19	12	49	30	8	1246	6.01	1.25	21	47	25	7
1992	519	5.92	1.11	20	51	26	3	600	5.61	1.10	31	51	15	3
1997	537	5.73	1.09	26	50	22	2	573	5.63	1.12	29	51	16	4
2002	492	5.62	1.09	30	53	14	3	551	5.36	0.96	39	51	9	1
2007	357	5.41	1.01	36	50	12	1	397	5.20	1.00	45	45	9	1
2012	303	5.46	1.10	34	50	13	3	365	5.37	0.99	38	51	10	1
Northern Savo
1972	2513	6.68	1.21	7	40	41	13	2607	6.66	1.29	9	39	37	14
1977	2601	6.58	1.23	8	42	38	11	2731	6.33	1.28	14	45	31	10
1982	1206	6.26	1.21	13	51	28	8	999	6.02	1.28	21	48	24	7
1987	599	6.23	1.25	15	47	29	9	629	5.59	1.18	22	52	21	5
1992	581	5.96	1.09	19	51	25	4	619	5.58	1.04	31	51	17	2
1997	539	5.72	1.01	23	58	17	2	611	5.49	0.99	34	49	16	1
2002	449	5.58	1.14	30	53	16	1	555	5.34	0.93	38	51	10	1
2007	365	5.31	1.02	41	48	9	2	409	5.15	0.96	47	45	7	1
2012	318	5.42	1.06	37	50	11	2	375	5.25	0.90	41	50	10	0
Turku/Loimaa
1982	1233	6.07	1.11	16	52	26	5	1291	5.94	1.19	21	50	23	6
1987	566	6.07	1.20	18	49	27	6	612	5.84	1.22	26	46	22	6
1992	562	5.85	1.08	22	53	22	4	609	5.54	1.10	33	49	16	2
1997	529	5.57	1.03	30	51	17	2	568	5.47	0.97	33	53	13	2
2002	448	5.57	1.04	32	49	18	1	518	5.25	0.97	44	44	10	1
2007	317	5.31	0.97	35	55	9	1	403	5.06	0.93	50	45	5	0
2012	297	5.37	0.98	38	49	11	1	353	5.22	1.03	45	44	10	1
Helsinki/Vantaa
1992	527	5.74	1.09	26	51	21	2	542	5.41	1.05	39	47	12	2
1997	481	5.59	1.12	30	53	15	2	557	5.38	0.98	38	48	13	1
2002	428	5.49	1.12	33	52	13	2	489	5.19	0.97	46	45	8	1
2007	314	5.36	0.89	34	57	9	1	383	5.09	0.90	49	44	7	0
2012	305	5.26	1.04	39	51	9	1	333	5.24	1.05	43	47	7	3
Northern Ostrobothnia/Kainuu
1997	505	5.81	1.03	22	52	24	2	565	5.67	1.04	26	55	17	3
2002	438	5.61	1.05	29	53	17	2	534	5.41	0.94	35	53	11	1
2007	364	5.39	1.00	36	51	13	1	397	5.20	0.90	43	49	7	1
2012	325	5.65	1.05	29	50	19	2	367	5.35	0.95	37	51	11	1
Men	72–77	77–82	82–87	87–92	92–97	97–02	02–07	07–12
Year	<0.001	<0.001	0.299	<0.001	<0.001	<0.001	<0.001	0.019
Area	0.001	0.182	<0.001	0.005	<0.001	0.002	0.277	<0.001
Area × year	<0.001	0.553	0.361	0.377	0.542	0.440	0.646	0.267
Women
Year	<0.001	<0.001	<0.001	<0.001	0.029	<0.001	<0.001	<0.001
Area	<0.001	0.007	<0.001	0.019	<0.001	<0.001	<0.001	0.002
Area × year	0.160	0.57	0.977	0.247	0.276	0.445	0.493	0.963

Note: Numbers in pairwise comparisons for survey years are P values (GLM, Bonferroni correction). Standard deviation, SD.

The surveys included a self-administered questionnaire, physical measurements and blood samples. The questionnaire, together with an invitation to the health examination, was sent by mail to all the subjects. Trained nurses carried out physical measurements and blood sampling in local health centres or other survey sites. Blood pressure was measured by mercury sphygmomanometers in all surveys. In 1972 and 1977, the cuff bladder size was 13 cm by 23 cm; in 1982–1997, 13 cm by 42 cm; and in 2002–2012, 14 cm by 36 cm. Measurements were made from the right arm in a seated position with at least a 5-min rest before the measurement. The first phase of Korotkoff sounds was recorded as systolic blood pressure and the fifth phase as diastolic blood pressure. Height was measured to the nearest 0.5 cm (1972–1997) and 0.1 cm (2002 onwards). Weight was measured in light clothing to the nearest 100 g with a beam balance scale. Body mass index (BMI) was calculated as weight in kilograms divided by squared height in metres (kg m ⁻² ) and further categorized into normal weight, overweight and obesity according to the World Health Organization. ¹¹ Smoking was assessed by structured questions in the self-administered questionnaire, where the participants were classified into three categories: smokers were those who had smoked regularly for more than 1 year and had smoked during the previous month. Ex-smokers had smoked regularly, but quit at least 1 month before the survey. Never-smokers had never smoked regularly.

The venous blood samples were centrifuged at the field survey sites and transferred daily for cholesterol measurements to the laboratory of THL. In 2007 and 2012, the sera were frozen immediately after separation and transferred weekly in dry ice to the laboratory for analyses. In 1972 and 1977, serum total cholesterol was determined from fresh samples using the Liebermann–Burchard method. In the other surveys, the analyses have been carried out using an enzymatic method. As the enzymatic method gave 2.3% lower values than the Liebermann–Burchard method, the serum total cholesterol values in 1972 and 1977 have been corrected by the same amount. We have described methods, instruments and reagents for cholesterol measurement between 1972 and 2002 previously. ¹² In 2007 and 2012, serum total cholesterol was measured by an enzymatic assay (Abbott laboratories, Abbott Park, IL, USA) with Abbott Architect c8000 clinical chemistry analyser. The laboratory in the National Institute for Health and Welfare has taken part in both national and international quality assurance systems first with the WHO centre in Prague and the last three surveys with the Center for Disease Control and Prevention in Atlanta. Based on the quality diagnostics, the systematic errors for each survey year were as follows. 1982: −0.06%, 1987: −0.56%, 1992: −0.84%, 1997: −1.41%, 2002: +2.0%, 2007: +0.7% and 2012: 0.0%. The serum total cholesterol values were corrected respectively in our analyses.

Results are presented as means and proportions of each risk factor by area, year and sex. We used Bonferroni-corrected general linear models for continuous variables and logistic regression analysis for categorical variables to study change in risk factor between each 5-year interval. Only areas where the measurements were carried out in both years were included in analyses. Study year, area and age were covariates in the models, including an interaction term between area and year. Analyses followed the earlier reports on the FINRISK Studies ⁷ and employed SAS software (version 9.3, Cary, NC, USA).

Results

Serum cholesterol

Serum total cholesterol declined steadily from 1972 to 2007 and turned to a rise in 2012 ( table 1 ). The changes from 1972 to 2007 have been substantial in North Karelia and Northern Savo. Total cholesterol levels declined by 21% (from 6.92 mmol l ⁻¹ to 5.41 mmol l ⁻¹ ) in North Karelia men and 23% (from 6.81 mmol l ⁻¹ to 5.20 mmol l ⁻¹ ) in North Karelia women and by 20% (from 6.68 mmol l ⁻¹ to 5.31 mmol l ⁻¹ ) in Northern Savo men and 22% (from 6.66 mmol l ⁻¹ to 5.15 mmol l ⁻¹ ) in Northern Savo women. The pooled 2012 mean for all areas in cholesterol was 5.43 mmol l ⁻¹ in men and 5.29 mmol l ⁻¹ in women. Only men in Helsinki/Vantaa still had lower total cholesterol levels in 2012 when compared with earlier surveys. Women had lower cholesterol levels than men over the years, yet the increase between 2007 and 2012 was more pronounced in women ( F = 24.7, P < 0.001) than in men ( F = 5.5, P = 0.019). We also observed statistically significant differences between areas; urban areas (Turku/Loimaa and Helsinki-Vantaa) had lower cholesterol levels than rural ones (North-Karelia, Northern Savo and Northern Ostrobothnia/Kainuu). Over the 40-year period, the whole distribution of serum cholesterol level in population has shifted to the left.

Blood pressure

Systolic blood pressure levels have declined significantly during the survey years, except for statistically non-significant periods 2002–2007 in men and 1997–2002 in women ( table 2 ). This downward trend has been observed in both sexes and all study areas, except for Helsinki/Vantaa with a slight increase in women’s systolic blood pressure. Some areas had somewhat larger decreases in the systolic blood pressure levels than others, showing statistically significant interactions between years and areas ( table 2 ). In North Karelia, the systolic blood pressure levels declined by 9% in men and 15.3% in women over 40 years’ time. In 2012, the pooled mean for systolic blood pressure was 134.4 mm Hg (95% confidence intervals, CI: 133.6–135.3) in men and 128.1 mm Hg (95% CI: 127.4–128.9) in women. The means for 2007 were 136.6 mm Hg (95% CI: 135.8–137.4) in men and 129.9 mm Hg (95% CI: 129.2–130.6) in women. In the population, the shift to the left in the distribution of systolic blood pressure has taken place steadily over the survey years.

Table 2

Systolic blood pressure (mm Hg) by area, year and sex

Area and year	Men							Women
	n	Mean	SD	≤130 mm Hg (%)	130–139 mm Hg (%)	140–159 mm Hg (%)	≥160 mm Hg (%)	n	Mean	SD	≤130 mm Hg (%)	130–139 mm Hg (%)	140–159 mm Hg (%)	≥160 mm Hg (%)
North Karelia
1972	1744	148.6	21.0	16	19	39	26	1887	152.6	26.0	18	17	29	36
1977	1765	142.9	17.7	21	22	41	16	1834	141.2	21.1	31	19	33	17
1982	1227	144.7	19.2	21	21	36	22	1267	141.1	19.7	31	19	31	19
1987	1139	143.9	18.9	22	21	38	19	1249	138.7	20.6	36	19	28	17
1992	521	141.5	19.2	28	23	31	17	611	135.3	20.9	45	16	25	14
1997	539	139.7	19.6	32	22	30	17	576	132.8	19.1	48	18	23	11
2002	493	137.1	17.8	36	22	30	12	550	132.0	21.1	52	17	19	12
2007	356	138.5	17.5	31	23	34	12	395	134.0	20.3	46	16	27	11
2012	303	135.2	16.5	40	26	26	8	366	129.3	18.0	54	22	17	7
Northern Savo
1972	2520	146.0	21.0	20	22	36	23	2621	147.2	25.5	25	18	29	28
1977	2607	145.7	19.2	19	21	39	22	2747	142.9	22.1	29	19	32	20
1982	1207	146.5	18.3	17	19	42	23	999	143.3	22.3	29	19	32	20
1987	599	144.4	18.5	20	20	39	21	631	138.9	20.5	36	19	29	17
1992	582	140.1	18.2	29	24	30	17	622	135.5	21.7	45	17	25	13
1997	539	138.9	17.4	30	27	32	11	610	133.2	18.0	46	20	26	9
2002	449	137.5	18.0	36	23	29	12	554	132.3	20.3	50	19	21	10
2007	365	140.0	17.8	34	22	32	12	409	130.9	18.4	51	19	22	8
2012	318	136.6	17.1	37	26	27	10	375	129.2	16.7	55	22	18	5
Turku/Loimaa
1982	1231	143.6	19.0	22	23	36	20	1293	135.7	20.0	42	20	24	14
1987	566	139.2	16.7	27	26	34	13	614	135.7	21.1	41	22	25	13
1992	562	139.2	16.7	25	28	35	11	612	133.5	19.4	48	21	21	10
1997	529	136.0	16.4	35	25	32	8	569	129.3	18.9	55	17	20	8
2002	448	136.8	17.4	38	23	29	10	519	131.6	18.9	53	16	21	10
2007	321	135.5	17.4	40	22	29	9	407	125.9	16.1	61	19	17	3
2012	297	131.2	15.9	50	23	22	5	352	125.0	16.1	65	17	14	3
Helsinki/Vantaa
1992	527	137.0	17.4	34	26	29	11	545	132.4	19.1	48	23	20	9
1997	482	136.9	18.7	38	20	29	13	560	129.4	18.1	55	19	20	6
2002	429	134.1	16.7	40	22	30	8	493	127.9	18.3	58	20	15	7
2007	316	134.4	17.0	40	25	27	8	383	126.9	17.7	57	21	17	5
2012	305	132.7	14.1	47	25	24	5	332	127.5	15.9	60	19	17	4
Northern Ostrobothnia/Kainuu
1997	507	138.2	18.7	31	27	29	13	573	131.5	18.8	51	18	21	9
2002	439	138.1	17.7	33	24	32	11	534	132.7	19.0	48	19	24	9
2007	365	137.0	17.3	35	24	32	10	399	131.8	18.7	50	20	22	8
2012	325	136.2	16.7	38	25	28	10	366	129.5	17.6	55	18	21	5
Men	72–77	77–82	82–87	87–92	92–97	97–02	02–07	07–12
Year	<0.001	0.013	<0.001	<0.001	0.004	0.005	0.987	<0.001
Area	0.543	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Area × year	<0.001	0.934	0.009	0.023	0.152	0.115	0.487	0.246
Women
Year	<0.001	0.780	<0.001	<0.001	<0.001	0.610	0.005	<0.001
Area	0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Area × year	<0.001	0.793	0.005	0.686	0.408	0.101	<0.001	0.016

Note. Numbers in pairwise comparisons for survey years are P values (GLM, Bonferroni correction). Standard deviation, SD.

Diastolic blood pressure levels have markedly declined over the years in all areas, but are levelling off or even on the rise in the past 5–10 years ( table 3 ). The pooled means for all areas in 2012 were 85.5 mm Hg (95% CI: 84.9–86.0) in men and 79.6 mm Hg (95% CI: 79.1–80.0) in women. In 2007, the means were 83.1 mm Hg (95% CI: 82.5–83.6) in men and 77.3 mm Hg (95% CI: 76.8–77.7) in women.

Table 3

Diastolic blood pressure (mm Hg) by area, year and sex

Area and year	Men							Women
	n	Mean	SD	≤80 mm Hg (%)	80–89 mm Hg (%)	90–99 mm Hg (%)	≥100 mm Hg (%)	n	Mean	SD	≤80 mm Hg (%)	80–89 mm Hg (%)	90–99 mm Hg (%)	≥100 mm Hg (%)
North Karelia
1972	1742	92.0	12.2	13	27	34	26	1886	92.4	13.5	15	27	29	29
1977	1764	88.6	11.0	19	32	33	17	1834	86.3	10.9	24	36	28	11
1982	1227	86.7	12.7	27	30	26	17	1267	84.5	12.0	33	33	22	11
1987	1139	88.1	11.8	22	31	30	17	1249	83.2	11.5	36	36	19	9
1992	521	84.6	12.4	33	36	21	11	610	79.5	11.3	49	32	14	5
1997	539	84.3	11.7	35	33	22	10	576	80.2	11.3	47	32	15	5
2002	493	83.0	11.6	36	33	22	8	550	77.6	11.6	56	28	12	4
2007	356	83.2	12.1	34	36	21	9	395	77.9	11.7	55	27	15	3
2012	303	84.3	12.1	37	29	23	11	366	78.9	11.4	52	30	15	4
Northern Savo
1972	2520	93.3	11.8	9	27	38	27	2620	91.3	12.1	14	29	34	23
1977	2607	92.6	11.9	11	31	30	28	2747	88.4	11.9	22	33	27	18
1982	1207	88.9	13.3	22	28	29	21	999	84.8	11.9	33	32	12	11
1987	599	89.1	11.1	19	31	31	19	631	83.9	11.3	33	36	22	10
1992	582	83.8	11.9	33	36	20	11	622	79.7	11.9	49	29	15	7
1997	539	86.0	11.8	27	32	27	14	611	80.9	11.0	44	34	17	5
2002	449	81.7	11.6	40	35	18	6	554	76.5	11.4	59	27	12	3
2007	365	84.5	11.6	31	36	22	11	409	78.5	10.9	50	33	14	2
2012	318	87.1	11.3	23	37	29	11	375	81.2	10.5	42	37	16	5
Turku/Loimaa
1982	1231	86.7	12.3	26	32	26	16	1293	81	11.7	45	31	18	7
1987	566	85.9	11.5	27	35	25	12	614	81.9	11.1	41	35	17	8
1992	562	85.1	12.3	30	33	25	12	612	81.2	11.0	43	33	18	6
1997	529	86.0	10.7	25	37	28	10	569	81.0	10.2	42	36	17	4
2002	448	84.0	10.7	33	37	22	7	519	79.3	11.0	51	30	16	3
2007	321	82.1	12.2	38	33	21	8	407	76.0	11.1	60	28	10	3
2012	297	83.7	11.1	33	38	21	8	352	77.9	10.2	55	31	11	3
Helsinki/Vantaa
1992	527	85.3	12.0	31	31	25	13	545	81.3	11.9	44	32	17	7
1997	482	85.0	12.1	29	35	24	12	560	80.4	10.8	47	33	14	6
2002	429	80.2	11.7	49	28	15	7	493	75.5	10.9	63	26	10	2
2007	316	82.7	11.2	36	36	21	7	383	76.4	10.4	61	26	12	1
2012	305	85.3	11.0	29	40	19	12	332	79.2	10.3	52	30	16	3
Northern Ostrobothnia/Kainuu
1997	507	85.0	11.7	30	34	26	10	573	80.3	11.0	46	33	15	6
2002	439	82.7	12.0	36	35	20	9	534	77.6	11.5	55	28	13	3
2007	365	83.0	11.1	36	34	23	7	397	77.7	10.8	55	30	12	3
2012	325	86.9	11.0	23	39	24	13	366	80.5	10.8	48	30	17	5
Men	72–77	77–82	82–87	87–92	92–97	97–02	02–07	07–12
Year	<0.001	<0.001	0.804	<0.001	0.140	<0.001	0.035	<0.001
Area	<0.001	<0.001	<0.001	0.149	0.100	<0.001	0.027	<0.001
Area × year	<0.001	0.045	0.010	<0.001	0.025	0.003	<0.001	0.164
Women
Year	<0.001	<0.001	0.119	<0.001	0.847	<0.001	0.984	<0.001
Area	0.007	<0.001	<0.001	0.260	0.037	<0.001	0.002	<0.001
Area × year	<0.001	<0.001	0.007	<0.001	0.085	0.002	<0.001	0.175

Note: Numbers in pairwise comparisons for survey years are P values (GLM, Bonferroni correction). Standard deviation, SD.

Smoking

In men, the proportion of smokers has declined markedly in all areas over the years, although the statistical significance of the change has somewhat varied between the survey years, showing periods of time with only modest changes in smoking prevalence ( table 4 ). The prevalence of smoking has even halved in North Karelia and Northern Savo during the 40 years. The pooled mean percentage of smokers was 27% in 2012, showing a statistically marginally significant decline ( P = 0.053). The proportion of never-smokers has grown continuously in all areas and survey years.

Table 4

Smoking by area, year and sex

Area and year	Men				Women
	n	Smoker (%)	Ex-smoker (%)	Never smoker (%)	n	Smoker (%)	Ex-smoker (%)	Never smoker (%)
North Karelia
1972	1802	52	20	28	1955	10	2	88
1977	1733	44	27	29	1824	10	5	85
1982	1185	36	30	35	1238	15	8	77
1987	1157	36	28	36	1274	15	12	73
1992	517	32	25	43	610	17	13	70
1997	535	31	26	43	573	16	16	68
2002	523	33	24	42	586	22	14	63
2007	384	31	27	42	428	18	22	60
2012	335	27	25	47	389	23	17	60
Northern Savo
1972	2566	50	22	29	2693	11	3	86
1977	2557	44	27	28	2729	12	6	82
1982	1181	42	24	34	986	15	8	77
1987	606	41	24	35	636	15	11	74
1992	582	37	26	37	620	19	14	67
1997	538	31	30	39	610	17	16	67
2002	499	38	24	38	591	20	18	62
2007	395	30	26	45	438	22	24	54
2012	355	26	27	47	423	14	20	66
Turku/Loimaa
1982	1198	39	28	32	1275	22	11	67
1987	576	38	27	34	625	23	13	64
1992	561	39	28	33	612	23	15	62
1997	528	34	26	41	569	22	16	62
2002	508	34	22	43	569	25	15	60
2007	341	31	25	45	426	19	17	63
2012	322	28	26	46	394	18	19	63
Helsinki/Vantaa
1992	527	36	27	37	543	30	15	55
1997	483	33	28	39	561	26	19	54
2002	473	40	20	40	537	27	21	53
2007	346	31	25	45	415	25	23	52
2012	337	26	25	49	362	18	22	60
Northern Ostrobothnia/Kainuu
1997	499	32	23	45	564	20	18	62
2002	487	34	26	40	567	21	19	61
2007	386	26	29	45	422	21	21	58
2012	363	28	26	47	415	19	25	56
Men	72–77	77–82	82–87	87–92	92–97	97–02	02–07	07–12
Year	<0.001	<0.001	0.661	0.191	0.012	0.007	<0.001	0.053
Area	0.275	0.008	0.001	0.008	0.176	0.297	0.209	0.848
Area × year	0.188	0.013	0.964	0.414	0.568	0.436	0.495	0.744
Women
Year	0.824	<0.001	0.813	0.113	0.149	0.004	0.101	0.045
Area	0.018	0.120	<0.001	<0.001	<0.001	<0.001	0.027	0.390
Area × year	0.466	0.234	0.875	0.428	0.915	0.316	0.259	0.008

Note: Numbers in pairwise comparisons for survey years are P values (Logistic regression analysis with Bonferroni correction).

In women, smoking was rare in the 1970s. Smoking prevalence increased in all areas until the years 2002–2007, then turned to a decrease ( table 4 ). The amount of never-smokers has continuously decreased, particularly in North Karelia and Northern Savo where time trends are available for 40 years. Smoking prevalence has still significantly decreased during the last 5 years, except in North Karelia. The pooled mean for smoking was 19% in 2012. There were statistically significant regional differences in smoking behaviour.

Body mass index

In men, the mean BMI has increased continuously over the years, but has shown one period of levelling off in 1987–1992 and a second in 2002–2012 ( Supplementary appendix table 2 ). The pooled mean BMI was 27.2 kg m ⁻² in 2007 and 27.4 kg m ⁻² in 2012 and this change was no longer statistically significant. The proportion of normal-weight men has decreased and that of obese men has increased in all areas over the survey years. There were also large regional differences in BMI.

In women, mean BMI increased significantly until 1987 and has thereafter continued to increase, but no longer statistically significantly ( Supplementary appendix table 2 ). BMI remained unchanged in women from 2007 to 2012, with a pooled mean 26.5 kg m ⁻² for both years, and potentially showing a levelling off in the upward 40-year trend of BMI. The proportion of obese women has increased over time in all survey years and areas, yet there has been some fluctuation in the proportions, which have also varied also by areas and years. Also, among women, there were large regional differences in BMI.

Discussion

The long-term population risk factor trends have been very favourable and substantial, but the most recent 5-year changes raise some concern about increased total cholesterol and diastolic blood pressure. On the other hand, positive trends seem to continue in smoking and systolic blood pressure.

An upward trend in total cholesterol after a long decreasing trend has also been reported in Sweden, starting between 2008 and 2010. ¹³ Whether a more global phenomenon of diverting trends is taking place is too early to speculate, as most of the previous time trend studies do not cover the 2010s. ^14,15 In Finland, the consumption of butter and mixed butter-margarine products has increased during the past years, ¹⁶ suggesting an increased intake of saturated fatty acids in diet. Supporting this individual change, the Food Balance sheets report a 20% increase in consumption of butter in 2010–2011 and a decrease in consumption of skimmed milk. ¹⁷ When looking at long-term changes in the use of these products, however, Finland has still experienced a substantial decline in the consumption of butter and a substantial increase of skimmed milk. ^18,19 Another plausible explanation for the increase in total cholesterol could be a change in the use of cholesterol-lowering medication, but the use, during the past 5 years, has been stable among FINRISK participants ²⁰ and even increased in whole population according to Finnish Statistics on Medicines. ²¹

These recent dietary changes thus raise concern in the public discussion. It is too early to say whether the observations in Finland and Sweden are due to some specific issues in the public discussion (low carb discussion) or whether it can be expected that after long decline some backlash will occur. In such situations, a good monitoring system is important. It can detect these phenomena and bring them to public discussion and decision-making.

Although systolic blood pressure has continued to decline, diastolic blood pressure decline has levelled off. This adverse and puzzling result could be explained by the changing intake of sodium in the population. Increased sodium intake and related resistance in peripheral circulation might affect first the diastolic blood pressure while the remarkably decreased atherosclerosis in population keeps systolic blood pressure declining. ^22,23 Current Finnish guidelines regarding the intake of sodium recommend a maximum of 6 g of salt per day in women and 7 g in men, while the World Health Organization recommends a maximum of 5 g of salt per day. ²⁴ Intake of sodium has decreased in Finland over time. ²⁵ but the past 5 year’s changes are not reported yet to evaluate the potential effect on diastolic blood pressure. However, it needs to be taken into account that the assessment of the fifth Korotkoff sound is more difficult to detect than the first sound and there is also a possibility of measurement bias. The distribution of blood pressure has systematically changed to better direction with less and less people having hypertension over time. ²⁶ However, in the last 5-year period, the proportion of hypertensive people no longer decreased. The use of anti-hypertensive medication has remained stable during the last 5 years among the FINRISK participants ²⁷ or has increased in whole population according to the Finnish Statistics on Medicines. ²¹

Smoking in Finnish men has steadily decreased over the study years, while in women the number of smokers increased until 2002 and decreased thereafter. This difference in time trends by sex has commonly been recognized. ²⁸ The improvements in smoking rates have been supported by legislative actions that have restricted the use of tobacco products in 1976, 1995, 2000 and 2010. ²⁹ At present, Finland prohibits all forms of tobacco promotion; smoking in public indoor places, restaurants and workplaces; sale of tobacco products to persons under 18 years and sale of smokeless tobacco. Warnings on tobacco packages are obligatory and 0.5% of the tobacco tax is designated for tobacco control. The current tobacco law states that its aim is a smoke-free Finland.

Consistently with our findings on no longer increasing BMI, there is some indication of similar trends elsewhere, ³⁰ yet these are among the first results showing levelling off of the obesity epidemic. However, we see area differences in the prevalence of obesity; lower levels of overweight and obesity in the metropolitan area of Helsinki/Vantaa and higher levels in North Karelia. Although the flattening trends are promising news, it should be born in mind that more than half of the population is overweight or obese. Therefore, the National Obesity Programme ³¹ was launched at the National Institute for Health and Welfare in 2012, aiming to reduce overweight and obesity at population level by promotion of healthy diet and physical activity.

Throughout the years, intensive public health action has taken place, including many related policy decisions. There has been successful interaction with the private and public sector, as well as with non-governmental organizations. ³² Results on the CVD risk factors presented in this article show some adverse changes since the launch of the North Karelia Project. This emphasizes the need for continued support for national risk factor reduction activities and respective monitoring. Most of the health education regarding heart-healthy diets, from the North Karelia Project and the national program, is 20–40 years old. New generations have come and the modern society offers a much more challenging commercial and media environment. Globalization influences increasingly the markets, both in healthy and unhealthy ways, and consumers are puzzled by increasing choices and powerful commercial messages. In spite of these concerns, the Finnish diet has overall greatly improved. Today’s Finns eat three times more vegetables and four times more fruit than in the 1970s ¹⁸ and much less saturated fat and salt. The Heart Symbol, signifying heart-healthiness in food products, is well recognized and used by majority of the population. ³³

A couple of important methodological issues should be mentioned. Firstly, participation rates have declined considerably over the years. We know that those who do not participate in the health examination are younger and male have lower educational level, more unhealthy lifestyles and more severe illnesses. ³⁴ Non-response may affect our results by underestimating some risk factors. ⁷ The change in participation rate from 2007 to 2012 has not, however, been large, suggesting good comparability between the last 5 years. Secondly, the sampling frame has changed during the survey years, but not for 2007 and 2012 where the changes were found. The risk factor trends have followed the same pattern over the survey years, suggesting no major bias in the interpretation of the time trends.

Conclusions

After 35 years of favourable changes in the risk factors levels in the Finnish population, some adverse changes are seen in serum total cholesterol and diastolic blood pressure between 2007 and 2012. The trends are still decreasing for smoking and systolic blood pressure and levelling off for the increase in BMI. Although the big picture is still very positive, the results emphasize the need for continuous support and monitoring.

Acknowledgements

This work was supported by the National Institute for Health and Welfare and the Academy of Finland [136895 and 263836 to S.M. and 139635 to V.S.].

Conflicts of interest : None declared.

Key points

• Population trends in cardiovascular risk factors have shown development towards better from 1972 until 2007 in Finland, particularly in the levels of serum cholesterol, blood pressure and smoking.

• After three decades of favourable development, the population risk factor levels showed some increase in total cholesterol and diastolic blood pressure.

• Our findings emphasize the need for continued efforts towards national disease prevention and community level action.