Does Physical Activity Increase Life Expectancy? A Review of the Literature

Physical activity reduces many major mortality risk factors including arterial hypertension, diabetes mellitus type 2, dyslipidemia, coronary heart disease, stroke, and cancer. All-cause mortality is decreased by about 30% to 35% in physically active as compared to inactive subjects. The purpose of this paper was to synthesize the literature on life expectancy in relation to physical activity. A systematic PubMed search on life expectancy in physically active and inactive individuals was performed. In addition, articles comparing life expectancy of athletes compared to that of nonathletes were reviewed. Results of 13 studies describing eight different cohorts suggest that regular physical activity is associated with an increase of life expectancy by 0.4 to 6.9 years. Eleven studies included confounding risk factors for mortality and revealed an increase in life expectancy by 0.4 to 4.2 years with regular physical activity. Eleven case control studies on life expectancy in former athletes revealed consistently greater life expectancy in aerobic endurance athletes but inconsistent results for other athletes. None of these studies considered confounding risk factors for mortality. In conclusion, while regular physical activity increases life expectancy, it remains unclear if high-intensity sports activities further increase life expectancy.


Introduction
The most important causes of death in Western industrialized countries are cardio-and cerebrovascular diseases and malignancies. For instance, in Germany in 2008, 68.6% of all women and 65.9% of all men died from these diseases. In contrast, the third most frequent cause of death are respiratory diseases which cause less than 10% of deaths each year (Table 1). Important risk factors for cardio-and cerebrovascular diseases include smoking, arterial hypertension, obesity, diabetes mellitus, and dyslipidemia along with atrial fibrillation for ischemic strokes [1,2].
The relative risk of death is approximately 20% to 35% lower in physically active and fit persons compared to that in inactive and unfit persons [6,7]. Physical inactivity represents a major independent risk factor for mortality accounting for up to 10% of all deaths in the European region [3]. Hence, because a 40% lower mortality rate corresponds to an approximately 5-year higher life expectancy [8], one would expect an approximately 3.5-to 4.0-year higher life expectancy in physically active persons compared to that in inactive persons.
The purpose of this review was to synthesize the literature on life expectancy in relation to physical activity. Specifically, cohort studies on physically active and inactive subjects were reviewed to detect a possible difference in life expectancy between these subject groups. In addition, cohort studies on athletes and non-athletes were reviewed to detect a possible difference in life expectancy between these subject groups.

Method
To identify all relevant articles about cohort studies investigating the life expectancy of physically active versus inactive persons, a systematic literature search was conducted in the electronic bibliographical database PubMed (http://www.ncbi.nlm.nih.gov/pubmed/). We only searched for English-language peer-reviewed journal articles using the search terms "(life expectancy OR longevity) AND (physical activity OR exercise OR sport)" (last update January 3, 2012). A total of 1,932 articles were found. Using the abstracts and/or titles, forty-one of these articles were identified as cohort studies comparing mortality and/or life expectancy of physically active and inactive persons. However, of these, only 13 articles presented detailed data on life expectancy for both groups. Subsequently, a search was performed using the terms "(life expectancy OR longevity) AND athlete" to find articles on life expectancy in (former) athletes. Sixty-six articles were found. Additional publications were identified using the search term "(life expectancy OR longevity) AND (physical activity OR exercise OR sport)." Of all these articles, 21 articles investigated mortality and/or life expectancy of (former) athletes. Eleven of these articles presented detailed data on life expectancy of athletes compared to that of a control group.
The remaining life expectancy for physically active and inactive individuals or the difference in remaining life expectancy between the two groups, respectively, were reported in the articles. However, the attained ages of subject groups differed between studies. Because the differences of remaining life expectancies cannot be assumed to be independent from the attained age, we report the results stratified for attained age. Within each stratum of attained age, the results were presented for both sexes.
Only five articles [16][17][18][19][20] reported confidence intervals for the difference in remaining life expectancies between groups. In four articles [16,[18][19][20], a parametric bootstrap procedure was used, and in one publication [17] a nonparametric maximum likelihood approach was used. Consequently, we did not combine the results in a metaanalysis model.
Because a meta-analysis model was not appropriate, all data found in the literature search were reported despite some overlaps between the cohort studies. For instance, Jonker et al. [18] and Nusselder et al. [19] used data from the Framingham Heart Study and Paffenbarger et al. [21][22][23][24] used data from the Harvard Alumni cohort. While most studies reported results based on multivariate life-table analysis, the studies by Menotti et al. [25] and Pekkanen et al. [26] only reported the results of gained life expectancies for the cohorts based on classical survival analysis. Additionally, the study of Pekkanen et al. [26] reported the survival rate of men aged 40 to 65 years in the following 20 years and not the total life expectancy.
The median increase of life expectancy of men and women in the eight studies presenting data on both sexes amounted to 3.7 years each. Physical activity during leisure time seems to increase life expectancy more effectively than total physical activity (all-day, professional, or leisure time activity altogether; professional physical activity alone has not been studied): 3.4 added years due to total activities and 4.7 added years (median values) due to leisure time activities in women, 1.9 and 3.9 added years, respectively, in men. The number of studies, however, is too low for a statistical analysis. Furthermore, the description of the amount of physical activity in the active and inactive groups are too heterogeneous for any statistical correlation between the amount of activity and the added years of life.
The eleven case control studies on life expectancy of athletes, mostly elite athletes (Table 4), reported a mean life expectancy that was between 5.0 years lower and 8.0 years higher than that of the nonathlete control groups. Aerobic endurance sports resulted on average in a 4.3 to 8.0 years higher life expectancy and team sports activities on average in a 5.0 years lower to about 5 years higher life expectancy compared to that for normal physical activity. Only one study presented data on strength sports and reported a slightly higher life expectancy compared to that for normal physical activity. None of these studies considered any confounding factor that could affect life expectancy.

Discussion
The purpose of this review was to synthesize the literature on life expectancy in relation to physical activity. Being physically active indeed appears to be associated with a higher life expectancy. Samitz et al. [7] as well as Warburton et al. [11] reported a mean reduction of mortality of 31% to 35% in persons who participate in regular leisure-time or daily life physical activity compared to that in inactive persons. Assuming a 40% lower mortality corresponding to a 5-year higher life expectancy [8], regular physical activity should increase mean life expectancy by approximately (31% to 35%)/40% × 5 years or by 3.9 to 4.4 years. Indeed, the few articles that presented data on life expectancy in physically active individuals reported a 0.43-to 6.9-year higher life expectancy. High-quality studies considering confounding factors that could affect mortality reported a 0.43-to 4.21years higher life expectancy in physically active compared to inactive persons. The wide range of years added cannot be explained based on the published data.
The studies that standardized extended life estimates for confounding factors [16,[18][19][20][21][22][23][24][25]27] virtually calculated a net gain in life expectancy by being physically active. However, the actual increase in life expectancy should be much higher because of favorable effects of physical activity on other risk factors for mortality such as arterial hypertension, glucose and lipid metabolism, coronary heart disease, stroke, or malignancies (Table 1). In fact, nonsmoking, normal weight, and physically fit men live on average 12 years (95% confidence interval, 8.6 to 14.6 years) longer than smoking, overweight, and physically unfit control subjects [41]. Subjects who never smoked, follow a healthy diet, are adequately physically active, and consume only moderate alcohol have a mean life expectancy that is 11.1 years longer than those who practice none of these healthy life behaviors [42].
The mechanisms underlying the net effect of physical activity are speculative and include reduction of triglyceride and apolipoprotein B concentrations, increase of highdensity lipoproteins and tissue plasminogen activator activity, and reduction of coronary artery calcium resulting in reduced risks of vascular diseases, which carry the strongest mortality risk [43]. In addition, regular physical activity increases the endurance of cells and tissues to oxidative stress, vascularization, and energy metabolism [44].
According to the results of the meta-analysis on allcause mortality in relation to physical activity performed by Samitz et al. [7], vigorous physical activity (>6 metabolic equivalents (MET)) reduces mortality slightly, but this reduction is significantly more pronounced than that for moderate activity (3-6 MET).
A greater life expectancy is not associated with more years of being frail and depending on assistance. In contrary, Nusselder et al. [20] reported a gain of disability-free years of life with a higher life expectancy.
The few data available on life expectancy in athletes who were much more physically active than the average individual are inconclusive. All studies proved an increased life expectancy in endurance athletes ranging between 2.8 to 8.0 added years. This gain is probably higher than that found for persons performing vigorous physical activity in the cohort studies. In team sports and other sports disciplines, life expectancy may fall below or be above that of the control groups. However, data on health behaviors of these athletes other than their physical activity during their active sports career such as smoking, food, and alcohol consumption are not available. Thus, the effect of elite sports activities on life expectancy warrants further investigation.
In summary, as expected from numerous prospective cohort studies on all-cause mortality in physically active and physically inactive persons, estimates on life expectancy in relation to physical activity indicate additional years of life in active subjects: the conservative estimate of the net increase in life expectancy with physical activity is about 2-4 years but presumably even greater because of the positive influence of physical activity on major risk factors for mortality.     Year(s).