Leisure time and occupational physical activity, resting heart rate and mortality in the Arctic region of Norway: The Finnmark Study

Abstract

Aims

This study examined the association of leisure time physical activity, occupational physical activity, and resting heart rate with all-cause and cardiovascular disease mortality in Sami and non-Sami populations.

Study design

This was a longitudinal, observational population-based study.

Methods

The Finnmark 3 study cohort was examined in 1987–1988 and followed for all-cause and cardiovascular disease mortality for 26 years. The cohort included 17,697 men and women with a mean age of 47.2 years at baseline. Leisure time physical activity and occupational physical activity were assessed with a validated questionnaire at baseline, whereas cause of death was obtained from the Norwegian Cause of Death Registry.

Results

A total of 1983 women and 3147 men died during follow-up. Leisure time physical activity was linearly and inversely associated with all-cause mortality, but not coronary heart disease mortality. Compared to inactive subjects, all-cause mortality was significantly reduced by 16% in the active leisure time physical activity group (hazard ratio 0.84; 95% confidence interval 0.76–0.92). Both for all-cause and cardiovascular disease mortality, we observed a U-shaped relationship with occupational physical activity, as participants in the walking and lifting group had significantly lower mortality than both the mostly sedentary and the heavy manual labour group (p < 0.05). An increase in resting heart rate by one beat per minute was associated with a 1.1% increase in all-cause mortality (hazard ratio 1.011; 95% confidence interval 1.009–1.013). The associations were similar in Sami and non-Sami subjects.

Conclusion

In this population-based study, leisure time physical activity was inversely associated with all-cause mortality, whereas resting heart rate was positively associated with all-cause and cardiovascular disease mortality. There was a U-shaped association between occupational physical activity and cardiovascular disease and all-cause mortality.

Introduction

The importance of regular physical activity to reduce the risk of cardiovascular disease (CVD) and death is well established.^1–3 Both physical activity^1,4 and exercise capacity^5,6 are inversely related to mortality. However, a majority of previous studies have included only leisure time physical activity (LTPA),^1,4 and the association between occupational physical activity (OPA) and mortality is less studied. Some studies suggest that high levels of OPA are associated with increased mortality, particularly among men,^7–9 whereas other studies show benefits of high levels of OPA.^10,11 Moreover, high levels of OPA in combination with low cardiorespiratory fitness seem to be associated with increased CVD mortality.¹²

When direct measures of exercise capacity are lacking, resting heart rate (RHR), which is an easily accessible clinical variable, may act as a proxy for physical fitness and physical activity in large cohorts.^13,14 Findings indicate that high RHR is independently associated with increased all-cause and cardiovascular mortality,¹⁵ and development of atherosclerosis.¹⁶ However, the independent role of RHR for mortality is disputed.¹⁵

We address these issues in a large cohort of men and women in Finnmark County in northern Norway, which is characterised by a large minority of indigenous Sami people. Historically, Finnmark county has had the highest mortality rates and incidence of coronary heart disease in Norway. The life expectancy for the population is still somewhat lower than the average in Norway, but the gap is slowly diminishing.¹⁷ We have previously shown that the Sami population has higher work physical activity levels and lower RHR than their Norwegian peers,¹⁸ which could suggest possible ethnic interactions in the association between physical activity and mortality.

The main aim of this study was to examine the associations of LTPA, OPA and RHR with all-cause and CVD mortality. Additionally, we wanted to examine whether these associations differed between Sami and non-Sami populations.

Methods

The Finnmark Study

The Finnmark Study is a longitudinal, population-based study in northern Norway conducted in 1974–1975 (Finnmark 1), 1977–1978. (Finnmark 2), and 1987–1988 (Finnmark 3).^19,20 The present analyses are based on data from the third Finnmark Study in 1987–1988, to which all residents in Finnmark County aged 40–62 years (born 1925–1947) were invited, together with a subsample among inhabitants aged 20–39 (born 1948–1967) (those invited to Finnmark 2 still living in Finnmark and 10% of those who were not invited to Finnmark 2). In total, 22,941 inhabitants were invited, and 17,821 men and women (77.7%) attended Finnmark 3. Invitations were sent by letter and the participants were asked to answer three questionnaires, which were presented in Norwegian and Sami languages. Participation rate increased with age and women had higher participation rates than men. Moreover, participation varied by community.²⁰

Selected sample for the present analyses

The present sample includes 17,697 men and women, aged 20–62 years at examination, with valid data on physical activity and covariates. Of these, 13,590 participants reported ethnic affiliation, among which 2813 (20.7%) were categorised as Sami and 10,777 (79.3%) as non-Sami, constituting a subsample.

Exposure assessment

LTPA levels at baseline were assessed by the Saltin-Grimby Physical Activity Level Scale²¹ with four mutually exclusive options. Due to a low number of participants (n = 224) answering the highest physical activity option, groups 3 and 4 were merged, leaving three groups for the analyses: ‘inactive', ‘moderate' (walking, bicycling, etc. ≥ 4 h a week) and ‘active' (recreational sports etc. ≥ 4 h a week or hard training or competitions several times a week).

OPA levels at baseline were also assessed by the Saltin-Grimby Physical Activity Level Scale²¹ with four mutually exclusive options: ‘mostly sedentary', ‘walking' (e.g. shop assistant, light industrial work, education), ‘walking and lifting' (e.g. mailman, heavy industrial work, construction work) and ‘heavy manual labour' (e.g. forestry work, heavy agriculture work, heavy construction work).

LTPA measured by the Saltin-Grimby Physical Activity Level Scale²¹ shows satisfactory rank validity when validated against maximal oxygen uptake (VO_2max) and accelerometry.¹³ In contrast to LTPA, no association between self-reported OPA and VO_2max was observed.¹³

RHR was measured during blood pressure monitoring (Dinamap, Criticon), sitting down after four minutes rest. Three measurements were taken with an interval of one minute, and the lowest of the three heart rate measurements was used in the analyses.

Covariates

Age was obtained from the National Population Registry. Data on daily smoking, previous CVDs (myocardial infarction, angina pectoris, stroke), diabetes mellitus, and treatment for hypertension was self-reported (yes/no). Non-fasting blood samples were collected and analysed for serum total cholesterol and triglycerides using an enzymatic method (Hitachi Auto Analyser, Roche Diagnostic, Switzerland). Height and weight of all subjects were measured by the screening nurse and recorded. Systolic and diastolic blood pressure (SBP and DBP) levels were measured automatically with the Dinamap (Criticon) blood pressure monitor.²² Three measurements were taken with an interval of one minute, and the mean value of the second and third measurements of blood pressure was used in the present analyses.

Ethnicity was dichotomised from four original categories into Sami (original categories Sami or Finnish/Sami) and non-Sami (original categories Norwegian or Finnish).²⁰ Classification was based on the two questions: ‘Are two or more of your grandparents of Sami origin?' and ‘Are two or more of your grandparents of Finnish origin?' as shown in Supplemental Material Table S1. Those who responded ‘unknown' were classified as missing.

Cause of death

Date and cause of death in the analytical cohort from date of attendance in 1987–1988 through 31 December 2013 were registered. CVD death was defined as International Classification Of Diseases (ICD)-9: 390–459 codes and ICD-10: I00–I99, using the underlying cause of death.

Statistical analysis

The associations of LTPA, OPA and RHR with CVD and all-cause mortality were estimated by Cox proportional hazard models with days-to-event as the time axis, with hazard ratio (HR) and 95% confidence interval (CI) as effect size. Proportional hazard assumptions were assessed by inspecting the log (-log) survival curves for the various physical activity categories. Model 1 tested the association between either LTPA, OPA or RHR as exposure and all-cause or CVD mortality as outcome, adjusted for age, sex, smoking status and body mass index (BMI) categories. Associations between LTPA and mortality were additionally adjusted for OPA, and vice versa. The model of RHR and mortality did not include OPA or LTPA. In Model 2, we additionally adjusted for self-reported angina pectoris, myocardial infarction, cerebral insult, diabetes and anti-hypertensive medication, which represent possible mediators in the association between physical activity and mortality. The analyses were repeated in a subsample (n = 13,590) with valid data on ethnicity, stratified by Sami or non-Sami.

By adding multiplicative interaction terms to the main multivariable models in Cox proportional hazard model, we assessed possible interactions between sex and LTPA, sex and OPA, ethnicity and LTPA, ethnicity and OPA, sex and RHR, ethnicity and RHR, and finally between LTPA and OPA, with OPA treated as a quadratic term. Values of p were two-sided with a significance level of 0.05. Data analyses were performed using IBM SPSS Statistics, version 24 (IBM Corporation, Armonk, New York, USA).

Ethical permission

The Norwegian Data Inspectorate approved the Finnmark Study. The present study was approved by the Regional Committee for Medical Research Ethics in Northern Norway (REK no. 2013/2249).

Results

In total, 17,697 participants with a mean age of 47.2 years at baseline were included in the analyses. During a mean of 23.3 years of follow-up, 5130 participants (1983 women and 3147 men) died, of which 1764 were due to CVD. The crude mortality rate was 12.2 per 1000 person-years.

Table 1 shows the characteristics of the study population at baseline (Finnmark 3, 1987–1988) by LTPA and OPA. Compared with the non-Sami, a higher percentage of the Sami were categorised as inactive in leisure time (p = 0.01), whereas the Sami were more active at work (p < 0.001). Values of BMI, DBP, treatment for hypertension and RHR decreased significantly with increasing levels of LTPA.

LTPA and mortality

We found an inverse linear association between LTPA and all-cause mortality after adjustments (p < 0.001) (Table 2, Model 1). Moderate LTPA was associated with a non-significant reduction in all-cause mortality, compared with the inactive group (HR 0.95; 95% CI 0.89–1.01; Model 1), and participants in the active LTPA group had 16% reduced risk (HR 0.84; 95% CI 0.76–0.92; Model 1). These inverse associations were slightly attenuated with further adjustment in Model 2, and found both in Sami and non-Sami subjects (Table 2). A non-significant inverse, linear association between LTPA and CVD mortality was observed in both Model 1 and 2 (Table 2). There were no interactions between sex or ethnicity and LTPA in any of the models. However, an interaction between LTPA and OPA was observed (p < 0.001), as mortality decreased with increasing LTPA in the two lowest OPA categories; however, in the two highest levels of OPA, LTPA did not seem to influence mortality (Supplemental Material Table S2).

OPA and mortality

The association of OPA with all-cause and CVD mortality was U-shaped (Table 3). Compared with the subjects with walking and lifting OPA, we found a 16% higher all-cause mortality in the mostly sedentary OPA group (HR 1.16; 95% CI 1.07–1.26; Model 1) and a 13% higher mortality in the heavy manual labour group (HR 1.13; 95% CI 1.01–1.26; Model 1) (Table 3). These associations were almost unchanged after further adjustments for CVD, diabetes and anti-hypertensive medication (Model 2, Table 3). The strength of the relationships was similar in Sami and non-Sami subjects.

There were no interactions with sex and ethnicity in any of the models. However, an interaction between OPA and LTPA was observed (p < 0.001). When stratifying by LTPA levels, the overall U-shaped association between OPA and mortality was found in the inactive and moderate LTPA groups. However, in the highest LTPA group (Active), we observed a different pattern, with a linearly increasing mortality with increasing OPA level (Supplemental Material Table S2).

RHR and mortality

All-cause mortality increased by 1.1% for each beat per minute increase in RHR (HR 1.011; 95% CI 1.009–1.013; Model 1). Similar results were observed for CVD mortality (HR 1.007; 95% CI 1.004–1.011; Model 1), and after further adjustments for CVD and diabetes. These trends were seen in both ethnic groups. There were no interactions between sex or ethnicity and RHR in any of the models.

Discussion

In this prospective study over 26 years, LTPA reduced all-cause mortality in a linear dose-response relationship, whereas the association between LTPA and CVD mortality was not statistically significant. We found a U-shaped association between OPA and all-cause and CVD mortality, with moderate OPA levels associated with the lowest mortality. Furthermore, we found a linear relationship between RHR and all-cause and CVD mortality. The association of LTPA, OPA and RHR with mortality did not differ between the Sami and non-Sami populations.

LTPA and mortality

Numerous studies have demonstrated an inverse association between LTPA level and all-cause mortality,^1,4,23 showing a 10–40% risk reduction with moderate physical activity. In the present study, participants with the highest LTPA level benefited most, as hard training of at least four hours a week showed a 16% reduction in all-cause mortality compared to the non-significant 5% risk reduction in the moderate LTPA group. Our finding that LTPA is associated with reduced mortality in a dose-response manner is also in accordance with results from pooled data from six studies in the National Cancer Institute Cohort Consortium¹ and a meta-analysis of 22 cohort studies.²⁴ One possible explanation for this finding could be better general health and socio-economic status (SES) in the groups with the highest LTPA levels. Unfortunately, data concerning SES were not available. On the other hand, numerous studies show strong evidence for an inverse and independent association between LTPA and mortality, supporting our findings.

Interestingly, decreasing mortality with increasing LTPA was only found in the two lowest OPA categories; in higher levels of OPA, LTPA did not seem to influence mortality (Supplemental Material Table S2).

The weaker association between LTPA and CVD mortality compared to all-cause mortality is in contrast to other studies.^1,3 One possible explanation could be low statistical power. Moreover, high prevalences of CVD in the Arctic region²⁵ may have influenced the effect of physical activity on CVD mortality in this population.²⁶

OPA and mortality

We found a 13–16% lower all-cause mortality in the walking and lifting group compared with the sedentary OPA group and the heavy manual labour group, suggesting a U-shaped association between OPA and mortality. However, this U-shape was found only in the inactive and moderate LTPA groups, whereas participants in the active LTPA groups showed a linearly increasing mortality with increasing OPA level.

The Danish National Work Environment Cohort study⁷ observed an increasing all-cause mortality with increasing OPA in men, whereas female workers showed a U-shaped association. The association between OPA and mortality is inconsistent in cohort studies, although many indicate an increased mortality rate among those with high OPA. This has led researchers to propose a ‘physical activity paradox',²⁷ suggesting that OPA and LTPA have opposite effects on CVD health and mortality. Possible reasons for this include the characteristics of OPA, such as low intensity, long duration, static postures and heavy lifting. Furthermore, OPA may elevate 24-hour heart rate and blood pressure.²⁷

RHR and mortality

Our finding that high RHR is related to a significant increase in all-cause and CVD mortality is in accordance with findings from several meta-analyses and large cohort studies showing that elevated RHR is independently associated with increased of all-cause and cardiovascular mortality,^15,28,29 even when controlling for familial factors.³⁰ A study showing that high RHR trajectories were associated with the highest risk of death, although only in men, further highlights these findings.³¹ A plausible explanation for the increased mortality risk with higher RHR is dysfunctional autonomic nervous activity¹⁵ and detrimental effects on the progression of atherosclerosis caused by higher RHR.¹⁶

Physical activity and mortality in Sami and non-Sami populations

The associations between physical activity and mortality did not differ with ethnicity in our study. In the 1980's, the percentage of employees in primary industries (reindeer industry, agriculture and fishing) was approximately 20% among the Samis compared to about 8% amongst the population at large.³² However, the numbers of employees in the reindeer industry decreased by 16% from 1990 to 2008,³³ indicating that the industries of the Sami and non-Sami populations became more similar during follow-up. This may have levelled out the differences in physical activity and RHR found between the Sami and non-Sami populations in 1986–1987.¹⁸ In addition, the Sami and non-Sami populations seem to be rather homogeneous regarding lifestyle and CVD risk factors.^34,35 To our knowledge, no previous studies have examined the association between physical activity and mortality among other indigenous peoples.

Limitations and strengths

Self-reported physical activity may be subject to misclassification errors. We did not have updated information on the exposure variables, which could lead to misclassification due to changes in exposure during follow-up. Moreover, we did not have information about SES. Physical activity levels increase with education level,³⁶ and higher social class is often associated with longevity,³⁷ thus introducing SES as potential confounder that could explain some of the association between physical activity and mortality. Moreover, the observed associations could be influenced by unmeasured confounders such as diet and genotype.

The strengths of this cohort study are the prospective design with 26 years follow-up time in a population with a large Sami minority, a large number of mortality cases and minimal loss to follow-up. The study involves rigorous outcome ascertainment, with a high degree of completeness of the Norwegian Cause of Death Registry, covering about 98% of all deaths in Norway.³⁸ The validity of CVD mortality in Norway shows substantial agreement with autopsy findings.³⁹

Conclusion

In this population-based study, LTPA was inversely associated with all-cause mortality, whereas a U-shaped association was observed between OPA and CVD and all-cause mortality. Elevated RHR was associated with higher all-cause and CVD mortality. The findings applied to both the Sami and non-Sami populations.

Acknowledgements

The Norwegian Institute of Public Health in collaboration with the University of Tromsø and local health authorities conducted the Finnmark surveys.

Author contribution

RH, BM and BKJ contributed to the design, statistical analysis and drafting of the manuscript. RH, BM, BKJ and MLL revised the manuscript critically. All of the authors approved the manuscript for publication and are accountable for all aspects of the work.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: Funding was provided by a scholarship from Finnmark Hospital Trust for 2014-2018 (140001/130016).

References