Impact of cardiorespiratory fitness on survival in men with low socioeconomic status.

AIMS
Although both low socioeconomic status (SES) and poor cardiorespiratory fitness (CRF) are associated with increased chronic disease and heightened mortality, it remains unclear whether moderate-to-high levels of CRF are associated with survival benefits in low SES populations. This study evaluated the hypothesis that SES and CRF predict all-cause mortality and cardiovascular disease mortality and that moderate-to-high levels of CRF may attenuate the association between low SES and increased mortality.


METHODS
This study included 2368 men, who were followed in the Kuopio Ischaemic Heart Disease Study cohort. CRF was directly measured by peak oxygen uptake during progressive exercise testing. SES was characterized using self-reported questionnaires.


RESULTS
During a 25-year median follow-up, 1116 all-cause mortality and 512 cardiovascular disease mortality events occurred. After adjusting for potential confounders, men with low SES were at increased risks for all-cause mortality (hazard ratio 1.49, 95% confidence interval: 1.30-1.71) and cardiovascular disease mortality (hazard ratio1.38, 1.13-1.69). Higher levels of CRF were associated with lower risks of all-cause mortality (hazard ratio 0.54, 0.45-0.64) and cardiovascular disease mortality (hazard ratio 0.53, 0.40-0.69). In joint associations of SES and CRF with mortality, low SES-unfit had significantly higher risks of all-cause mortality (hazard ratio 2.15, 1.78-2.59) and cardiovascular disease mortality (hazard ratio 1.95, 1.48-2.57), but low SES-fit was not associated with a heightened risk of cardiovascular disease mortality (hazard ratio 1.09, 0.80-1.48) as compared with their high SES-fit counterparts.


CONCLUSION
Both SES and CRF were independently associated with subsequent mortality; however, moderate-to-high levels of CRF were not associated with an excess risk of cardiovascular disease mortality in men with low SES.


INTRODUCTION
1 Socioeconomic status (SES), typically evaluated using education, annual income, place 2 of residence, occupation, or combinations thereof, has been shown to be an important 3 determinant of several health outcomes. 1 Low SES is strongly associated with increased all-4 cause mortality (ACM) and cardiovascular disease (CVD) events and mortality, which have 5 been linked to established CVD risk factors and unhealthy lifestyle habits. 1, 2 6 Individuals with low SES are reported to have lower levels of leisure-time physical 7 activity when compared to their higher SES counterparts. 2 The lower levels of physical activity 8 in individuals with low SES may be due, at least in part, to reduced cardiorespiratory fitness 9 (CRF), which is an important determinant of cardiovascular outcomes and mortality in the 10 general population and predictive of prognosis and survival in diseased populations. 3-6 11 Accumulating studies suggest that CRF has a protective effect in attenuating ACM and 12 CVD mortality in 'at risk' populations. 7,8 A meta-analysis demonstrated that individuals with 13 low SES appear to have relatively low CRF levels compared to individuals with high SES. 9 14 Therefore, CRF may, at least in part, contribute to the above-referenced socioeconomic 15 differences in survival. However, it remains unclear whether the potential impact of CRF 16 confers survival benefits among underserved populations within the general population. The 17 present study evaluated the hypothesis that SES and CRF would predict ACM and CVD 18 mortality and that moderate-to-high levels of CRF would attenuate the association between low 19 SES and heightened mortality.  Socioeconomic status 8 SES was characterized using self-reported questionnaires via a summary index that 9 combined measures of income, education, occupational prestige, material standard of living, 10 and housing conditions. Income was divided into quintiles over the past 12 months. Education 11 was classified into four categories: less than an elementary education; completion of 12 elementary education; completion of middle school or a part of middle school; and completion 13 of high school or above. The occupational status (occupational prestige) of participants on the 14 basis of self-reported primary lifetime occupation was classified into three categories as white 15 collar (professional and managerial staff, low-paid clerical workers); blue collar (manual 16 laborers in construction, mining, manufacturing, or forestry); and farmer, including those who 17 spent most of their employed activities in the agricultural sector. Standard of living was 18 evaluated using a material possession index based on self-reports of ownership of 12 items 19 (color TV, video tape recorder, freezer, dish washer, car, motor cycle, telephone, summer 20 cottage, house trailer, motor boat, sailing boat, and ski mobile). The combined SES index 21 ranged from 0 to 25, with higher values indicating lower SES. 10, 11 SES index was classified by 22 tertiles: low SES (>11), middle SES (10-8), and high SES (<7), and classified into low SES 23 (>10) and high SES (<9) based on median values.

Cardiorespiratory fitness
1 Peak oxygen uptake (VO2peak), an objective marker of CRF, was directly assessed using 2 a computerized metabolic measurement system (Medical Graphics, St. Paul, MN, USA) during 3 progressive exercise testing to volitional fatigue on an electrically braked cycle ergometer. The 4 methodology for the direct measurement of VO2peak has been previously described. 12 Briefly, 5 the standardized testing protocol included a 3-min warm-up at 50 watts (W; 1 W = 6.12 6 kgm/min), followed by 20 W/min increases in workload with direct analyses of expired 7 respiratory gases. VO2peak was defined as the highest attained value for oxygen consumption 8 and/or a plateau in oxygen uptake at maximal exercise.

9
Categories of CRF were standardized based on methods as previously suggested. 13,14 10 Briefly, participants were stratified into four age groups: 42 to 47 years, 48-53 years, 54 to 59 11 years, and >60 years, and then we defined CRF categories by tertiles of VO2peak value within 12 each age group. We combined the individual CRF categories from each age group to form the 13 following age-specified CRF categories: lower (mean 22.2±4.7 ml/kg/min), moderate 14 (29.9±3.2 ml/kg/min), and higher (38.1±5.5 ml/kg/min). CRF was also classified into unfit 15 (24.3±5.1 ml/kg/min) and fit (36.0±5.8 ml/kg/min) based on median values of age-specific 16 VO2peak percentiles.

17
Other measurements 18 Resting blood pressure was measured twice using a random-zero sphygmomanometer in 19 the seated position following 5 and 10 minutes of quiet rest. The mean of these 2 values was 20 used as resting blood pressure. Body mass index was computed as the ratio of weight in with non-normal distribution and the chi-square (χ 2 ) test for categorical ones. We calculated 18 the hazard ratio (HR) and 95% confidence intervals (CI) via a multivariable Cox proportional 19 hazard model, adjusting for potential confounding variables (age, smoking, alcohol, body mass 20 index, systolic blood pressure, high-density lipoprotein cholesterol, low-density lipoprotein 21 cholesterol, glucose, diabetes, hypertensive medication, family history of coronary heart 22 disease, history of CVD, and physical activity) to determine the associations of SES and CRF, 23 using categorical (tertiles) and continuous variables, to the risk of death. The joint effects of SES and CRF on the risk of death were examined using 4 combined groups based on the median 1 values of SES and CRF (high SES-Fit, low SES-Fit, high SES-Unfit, and low SES-Unfit). Our 2 reference group was the high SES-Fit cohort. The survival probability for ACM and CVD 3 mortality in each group was presented using Kaplan-Meier survival curves. Statistical 4 significance was set at p<0.05. All tests for statistical significance were two-sided. Analyses 5 were conducted using the SPSS version 22.0 (SPSS, Armonk, NY).

7
Baseline characteristics of the study participants grouped by tertiles of SES score are 8 shown in Table 1. Men with low SES had higher age, history of CVD, incident hypertension 9 and resting systolic blood pressure, and low-density lipoprotein cholesterol, but lower alcohol 10 consumption, leisure time physical activity and peak oxygen uptake (all p<0.05) at baseline as 11 compared to men with high SES (Table 1).  with their high-level SES counterparts (reference). Each 1 score increase in SES was associated 19 with a 4-5% higher the risk of mortality after adjusting for covariates (table 2).
Each 1 ml/kg/min increment in CRF was associated with a 4-5% lower the risk of mortality 23 after adjusting for covariates.
Compared with high SES-fit as a reference, low SES-unfit had significantly higher risks 1 of ACM (HR 2.15, 1.78-2.59) and CVD mortality (HR 1.95, 1.48-2.57), but low SES-fit was 2 not associated with a heightened risk of CVD mortality (HR 1.09, 0.80-1.48). However, ACM In the present study, both low SES and high CRF were significantly associated with 8 higher and lower risks of ACM and CVD mortality, independent of potential confounding 9 variables. However, the novel findings from our study were that the risk of ACM and CVD 10 mortality associated with low SES were the highest in unfit men, whereas the risk of CVD 11 mortality was significantly attenuated in fit men with low SES. These findings suggest that 12 CRF may favorably modify the relationship between SES and the risk of mortality, highlighting 13 the prophylactic role of CRF to enhance survival in underserved populations. Thus, improving 14 CRF should be encouraged to reduce the risk of CVD mortality in individuals with low SES.

15
To our knowledge, only two previous studies have reported that CRF contributed to the 16 socioeconomic differential in ACM and survival after acute myocardial infarction. 15,16 The 17 former study suggested that differences in impaired estimated functional capacity explained as 18 much as 47% of the SES-mortality associations among patients with suspected coronary artery 19 disease. 16 The latter study reported that functional status using self-reported questionnaires 20 explained ~30% of the association between SES and long-term mortality after acute myocardial 21 infarction. 15 However, these studies did not directly measure CRF using gas analysis. Our results suggest that directly measured moderate-to-high levels of CRF attenuate the risk of 23 mortality associated with low SES.
Our results are consistent with previous studies demonstrating that patients with lower 1 SES have reduced functional capacity as determined by treadmill exercise testing. [17][18][19] 2 Furthermore, a meta-analysis found that individuals with low SES appear to have relatively 3 low CRF levels as compared with individuals of high SES, independent of physical activity. 9

4
In the present study, men with low SES had lower levels of directly measured peak oxygen 5 uptake (by 3.8 ml/kg/min at baseline) as compared to men with high SES. Considering that 6 each 1 metabolic equivalent (MET = 3.5 ml/kg/min) increase in functional capacity is 7 associated with a 13% and 15% lower risk of ACM and CVD events, 20 our findings suggest 8 that reduced CRF may contribute to social inequalities in health, and should be a target for 9 improving survival in socioeconomically disadvantaged populations.

10
Several potential mechanisms may serve to explain the role of CRF-mediated 11 alterations on the association between SES and mortality. Although CRF is partially explained 12 by genetic variation, CRF is highly reflective of the magnitude and intensity of leisure-time 13 physical activity, which is dose-dependent and is associated with a substantially reduced risk which may contribute to their reduced CRF.

18
In addition, the association between SES and increased mortality is largely mediated 19 by cardiovascular risk factors. 2 High CRF also attenuates the risk of mortality in 20 overweight/obese individuals, those with diabetes mellitus or metabolic syndrome and at any 21 given atherosclerotic CVD risk factor profile. 7, 8, 23, 24 Moreover, higher levels of CRF are 22 associated with a decreased prevalence of coronary artery calcium, carotid artery intima media 23 thickness, and pulse wave velocity, even in individuals with cardiometabolic risk factors. 25,26 Because the association between low SES and a heightened risk of subclinical and future 1 atherosclerosis 27 has been recently reported, 28 these data suggest that CRF-related decreased 2 surrogate markers of subclinical atherosclerosis may serve as a potential underlying mechanism 3 explaining the reduced mortality in fit, low SES individuals. Clearly, additional studies are 4 needed to further clarify the mechanisms underlying the protective effect of CRF on CVD 5 mortality in individuals with low SES. 6 We acknowledge several methodological limitations to our study. Our study population 7 included only Caucasian men, limiting the generalizability of our findings to women and other 8 races/ethnicities. SES was assessed using self-reported questionnaires, rather than review of 9 individual records. Moreover, we used a single measurement of SES and CRF at baseline to 10 predict the risk of death and did not correct for serial changes in SES (upward or downward 11 shift) and CRF over time or, for that matter, potential regression dilution bias. 29 Despite these 12 limitations, the strengths of this study included that we directly measured peak oxygen 13 consumption using metabolic gas analysis, which provides an objective and quantitative 14 measure of aerobic capacity that is widely accepted as the gold standard measure of CRF. 30

16
In this population-based study of middle-aged men, our findings indicate that both SES 17 and CRF were independently associated with ACM and CVD mortality. However, moderate-18 to-high levels of CRF appeared to attenuate the risk of CVD mortality in this cohort. These