Introduction

Any degree of dysglycaemia, as measured by fasting glycaemia or HbA1c, is an established risk factor for various cardiovascular disease (CVD) outcomes including ischaemic and haemorrhagic stroke, myocardial infarction and death [1, 2]. Although regular exercise is established as being beneficial for improving glycaemic control and various health outcomes in diabetes [3], the joint association of physical activity and glycaemic control on mortality risk is unclear. The purpose of this analysis is to examine the protective effect of physical activity on all-cause and CVD mortality risk in individuals across all levels of glycaemic control as measured by HbA1c in the US population.

Research design and methods

Study design and population

The study sample was obtained from the Third National Health and Nutrition Examination Survey (NHANES ΙΙΙ), which is a nationally representative study of 33,994 US citizens between 1988 and 1994, with mortality follow-up through 31 December 2006. Cause of death was assessed using ICD-9 and -10 codes. All participants gave their informed written consent before participation in the examination, and the study protocol was approved by the National Center for Health Statistics. Individuals were excluded if they were pregnant, had a BMI of less than 18.5 kg/m2, were less than 20 years of age or had missing data for metabolic variables or physical activity. This left a final sample of 10,352 men and women.

Measurements

At baseline, participants completed a questionnaire regarding the frequency of their participation over the preceding month in the following leisure-time activities: walking, jogging, bicycling, swimming, aerobics or aerobic dancing, other dancing, calisthenics or exercises, gardening or yard work, and weight lifting. Participants could also list up to four additional activities. Individuals reporting no leisure-time activity were classified as ‘inactive’ and individuals who reported engaging in a leisure-time activity one or more times per week were classified as being ‘active’. Work and transportation physical activity was not assessed.

Age, sex, income, ethnicity, smoking status, alcohol intake, dietary fat intake (>30% or ≤30% of kilocalories from fat), medication use (type 2 diabetes, CVD, dyslipidaemia or hypertension) and a self-reported physician diagnosis of hypertension, heart attack, stroke and congestive heart failure were also assessed by questionnaire. The physical examination included measurement of height, weight and blood pressure, and a fasting blood draw for assessment of glucose, insulin, total cholesterol, HDL cholesterol, HbA1c and insulin resistance (HOMA). HOMA was derived from the following formula: (fasting insulin [pmol/l]/6) × (fasting glucose [mmol/l]/22.5) [4]. HbA1c was classified as <5.7% (39 mmol/mol), 5.7–6.4% (39–46 mmol/mol), 6.5–6.9% (48–52 mmol/mol) and ≥7% (53 mmol/mol) [5].

Statistical analysis

Baseline characteristics were stratified by physical activity and weighted to be representative of the US population. Differences between physical activity groups were assessed using independent Student’s t tests and χ 2 tests. HR and 95% CI for all-cause and CVD mortality risk stratified by physical activity group and HbA1c level were estimated by Cox proportional hazard regressions. Inactive individuals with HbA1c <5.7% (39 mmol/mol) were the referent. Models were adjusted for age, sex, ethnicity, income, smoking status, fatty diet, alcohol intake, history of CVD, hypertension, total cholesterol, HDL cholesterol, BMI and insulin resistance (HOMA). Analyses were performed using SAS version 9.2 (SAS Institute, Inc., Cary, NC, USA).

Results

Baseline characteristics stratified by physical activity group are presented in Table 1. The majority of the population was categorised as active, with only 20% reporting no leisure-time physical activity. Active individuals had a lower BMI, lower total cholesterol, were more insulin-sensitive and were less likely to have CVD, type 2 diabetes, or hypertension or be on medications for these conditions (p < 0.05).

Table 1 Baseline characteristics according to physical activity level
Full size table

During follow-up of 13.4 ± 3.9 years, there were 2,463 deaths, with 1,113 due to cardiovascular causes. Physical activity was associated with a decreased risk of all-cause (HR 0.74 [95% CI 0.67, 0.81]) and CVD (HR 0.71 [95% CI 0.62, 0.82]) mortality. In comparison with individuals with normal glycaemic control, higher levels of HbA1c were associated with increased mortality risk. HbA1c ≥7% (53 mmol/mol) was associated with the highest risk for all-cause (HR 1.54 [95% CI 1.30, 1.82]) and CVD (HR 1.93 [95% CI 1.52, 2.45]) mortality.

Figure 1 presents the combined effects of physical activity and HbA1c level on all-cause and CVD mortality risk. Across all levels of HbA1c, active individuals had a lower risk of mortality compared with inactive individuals with the same HbA1c level. In addition, individuals with HbA1c ≥7% (53 mmol/mol) were not at an increased risk for all-cause mortality if they were physically active. For CVD mortality, active individuals with HbA1c ≥7% (53 mmol/mol) were still at an increased risk compared with inactive euglycaemic individuals; however, their risk for CVD death was substantially lower than their inactive counterparts with the same level of dysglycaemia (HR 1.38 [95% CI 1.03, 1.84] vs HR 1.98 [95% CI 1.34, 2.92]).

Fig. 1
figure 1

The joint association of physical activity and HbA1c level with all-cause (a) and CVD (b) mortality risk. Black bars indicate inactive; white bars indicate active. Models adjusted for age, sex, income, ethnicity, smoking status, alcohol intake, dietary fat intake, history of CVD, hypertension, total cholesterol, HDL-cholesterol, BMI and insulin resistance (HOMA). *Significantly different from referent category: HbA1c <5.7% inactive (p < 0.05). All-cause mortality: activity × HbA1c: p > 0.05, activity main effect: p < 0.05; HbA1c main effect: p < 0.05. CVD mortality: activity × HbA1c: p > 0.05, activity main effect: p < 0.05; HbA1c main effect: p < 0.05. To convert values for HbA1c in % to mmol/mol, subtract 2.15 and multiply by 10.929

Full size image

When physical activity was divided by intensity (electronic supplementary material [ESM] Table 1), light and moderate/vigorous activity was still protective against all-cause and CVD mortality for any category of HbA1c, although no consistent pattern between exercise intensity and mortality risk was observed.

Conclusions

This analysis demonstrates that physical activity is associated with both lower all-cause and CVD mortality risk across the range of glycaemic control. Our findings suggest that engaging in physical activity and achieving normal levels of glycaemic control may both be important factors for the prevention of early mortality. If optimal glycaemic control is not achieved, regular exercise may still have a protective effect on overall mortality, independent of traditional metabolic risk factors. Alternatively, if physical activity is not achieved, then good glycaemic control may be critical in lowering morality risk.

A physically active lifestyle is associated with decreased risk of mortality in individuals with normoglycaemia as well as those with impaired glucose tolerance or type 2 diabetes [6, 7]. Given that increased HbA1c levels are associated with an increased risk of cardiovascular events and mortality in individuals with [8] and without [2] type 2 diabetes, it is important to examine whether active individuals have reductions in mortality risk, even if HbA1c levels remain elevated. The current findings extend previous research and demonstrate that for all levels of HbA1c, active individuals may receive substantial mortality risk reductions compared with inactive individuals with the same HbA1c level. Interestingly, this protective effect of physical activity was observed after adjustment for traditional risk factors such as insulin resistance, BMI and the lipid profile. Our results also highlight that physical activity is associated with a reduction in mortality risk above and beyond normalisation of glycaemic control. The physiological mechanisms responsible for the activity-induced reduction in mortality risk remains unclear. However, animal models suggest that aerobic exercise can induce adaptations that restore mitochondrial alterations associated with ageing and metabolic diseases [9]. In addition, physical activity can increase skeletal muscle mass, insulin sensitivity, and cardiac and endothelial function; all factors that may increase longevity [10]. Conversely, it is plausible that the lack of metabolic diseases allowed the individuals to be physically active thus avoiding early mortality events.

In the clinical setting, there is a large emphasis for all patients with diabetes mellitus to achieve a HbA1c <7% (53 mmol/mol) as a way to minimise microvascular and macrovascular complications and prevent early death [5]. Our findings suggest that individuals with good glycaemic control are protected compared with those with poor glycaemic control, whether they are active or not. Active individuals with HbA1c ≥7% (53 mmol/mol) were not at an increased risk for all-cause mortality and their risk for death was comparable with inactive individuals with normal glycaemic control. This suggests that engaging in an active lifestyle may provide similar protection against all-cause mortality as achieving glycaemic targets. Although active individuals with HbA1c ≥7% (53 mmol/mol) still had significantly increased risk for CVD mortality compared with euglycaemic individuals, their risk was substantially lower than their inactive counterparts with the same level of dysglycaemia. It is unclear what intensity of physical activity is required to obtain these benefits, but we provide preliminary evidence that even light physical activity may be sufficient for individuals with HbA1c ≥7%. These results highlight the clinical importance of prescribing physical activity for not only improving glycaemic control, but perhaps for also preventing early death.

Limitations of the current analysis warrant mention. Physical activity levels were based on self-reported questionnaire data, which did not include non-leisure-time activity (i.e. labour, transportation) or the duration of time spent in each activity, hence participants’ physical activity levels are subject to misclassification. Thus, we were unable to determine the optimal volume and intensity of physical activity necessary to lower mortality risk in this population. Also, given the cross-sectional design of the study, causality cannot be inferred. Nevertheless, a long length of follow-up with a large number of deaths and the use of a large sample representative of the general US population are significant strengths to the study.

In conclusion, this study demonstrates that physical activity is associated with lower all-cause and CVD mortality risk in individuals across all levels of HbA1c, independent of traditional metabolic risk factors. Therefore, physical activity should be strongly recommended for the prevention of early death in all individuals, especially those with poor glycaemic control.