OBJECTIVE

The aims of the study were to assess how baseline leisure-time physical activity (LTPA) and its exercise components intensity, duration, and frequency are associated with all-cause and cardiovascular mortality in patients with type 1 diabetes 1) overall, 2) stratified by presence or absence of chronic kidney disease (CKD), and 3) stratified by sex.

RESEARCH DESIGN AND METHODS

The study design was prospective and observational and included 2,639 patients with type 1 diabetes from the ongoing nationwide multicenter Finnish Diabetic Nephropathy (FinnDiane) Study. Mean follow-up time was 11.4 ± 3.5 years. LTPA was assessed by using a validated self-report questionnaire. Three hundred ten patients (11.7%) had CKD defined as an estimated glomerular filtration rate of ≤60 mL/min/1.73 m2.

RESULTS

During follow-up, 270 deaths occurred. LTPA and all its components were associated with all-cause mortality, even after adjustment for the potential confounders sex, diabetic nephropathy, duration of diabetes, age at onset of diabetes, systolic blood pressure, triglycerides, BMI, and HbA1c. Only exercise intensity was associated with cardiovascular mortality after adjustment for the confounders. Of the patients with CKD, 127 died during follow-up. The total amount of LTPA and exercise frequency were independently associated with lower risk of all-cause mortality when adjusted for covariates.

CONCLUSIONS

Exercise is associated with a lower risk of premature all-cause and cardiovascular mortality in patients with type 1 diabetes. This study also demonstrates that physical activity is associated with a lower risk of mortality in patients with type 1 diabetes and CKD.

A high level of leisure-time physical activity (LTPA) or fitness is associated with a reduced risk of premature mortality in the general population and in individuals with type 2 diabetes (1,2), but whether this is true for patients with type 1 diabetes has not yet been determined because the available evidence is limited. Only a few prospective studies have explored the association of LTPA and mortality in type 1 diabetes, and none have been performed in patients with type 1 diabetes and chronic kidney disease (CKD).

We have previously demonstrated that the intensity of LTPA is associated with the development and progression of diabetic nephropathy in patients with type 1 diabetes (3). Similarly, a higher amount of LTPA was associated with a lower risk of incident cardiovascular events (4). Because we have previously shown that diabetic nephropathy accounts for the increased mortality observed in type 1 diabetes (5), we hypothesized that exercise also reduces the risk of premature mortality in these patients.

Of the few previous studies that explored mortality and type 1 diabetes, the Pittsburgh Insulin-Dependent Diabetes Mellitus (IDDM) Morbidity and Mortality Study demonstrated that sedentary men with type 1 diabetes have a higher mortality rate than active men. A similar pattern, although not statistically significant, was seen in women (6,7). Moreover, a report from the EURODIAB Prospective Complications Study demonstrated a borderline inverse association between physical activity and all-cause mortality in type 1 diabetes when both sexes were combined (8).

Exercise is already recommended for patients with CKD and those who undergo dialysis, although conclusive end point studies are scarce (912), and even more so for individuals with type 1 diabetes. Therefore, the aim of the current study was to fill in this gap and to assess how baseline LTPA and its various components are associated with all-cause and cardiovascular mortality in patients with type 1 diabetes with and without CKD as well as when stratified by sex.

This prospective and observational study comprised 2,369 patients participating in the nationwide Finnish Diabetic Nephropathy (FinnDiane) Study, which has been previously described in detail (13). The FinnDiane Study was initiated in 1997, and to date, >5,000 patients have participated. The LTPA questionnaire was introduced in the beginning of 2000, and in this substudy, we included all patients with available LTPA data. Type 1 diabetes was defined as a diagnosis of diabetes before the age of 40 years and permanent insulin treatment initiated within 1 year of diagnosis. Patients with not-yet-classifiable renal disease were excluded (n = 93). Before participation, all patients gave informed consent, and the study protocol was approved by the Ethics Committee of the Helsinki University Hospital and confirmed by the local ethics committees. The study was conducted according to the Declaration of Helsinki.

The primary end point of this study was death as a result of any cause through the end of 2014 and identified through a search of the Finnish National Death Registry and the local center databases. A major cardiovascular event was the cause of death in 28% of the patients. A cardiovascular event was defined as clinically verified myocardial infarction (ICD-8 and ICD-9 codes 410–412, ICD-10 codes I21–I23) or either an ischemic or a hemorrhagic stroke (ICD-8 and ICD-9 codes 430–434, ICD-10 codes I60–I64).

Renal function (estimated glomerular filtration rate [eGFR]) was estimated by using the Chronic Kidney Disease Epidemiology Collaboration equation (14). CKD was defined as an eGFR ≤60 mL/min/1.73 m2 or end-stage renal disease (ESRD) (defined as undergoing dialysis or having received a kidney transplant). However, patients with a kidney transplant and eGFR >60 mL/min/1.73 m2 (n = 53) were not defined as having CKD. Patients with CKD were grouped as having CKD without ESRD, undergoing dialysis, or having a kidney transplant. The baseline urinary albumin excretion rate (UAER) was defined as follows: normal UAER as <20 μg/min and <30 mg/24 h (n = 1,633), microalbuminuria as ≥20 and <200 μg/min or ≥30 mg/24 h and <300 mg/24 h (n = 312), and macroalbuminuria as UAER ≥200 μg/min or ≥300 mg/24 h (n = 271) in at least two of three consecutive 24-h or timed overnight urine collections.

Data on medication, cardiovascular status, and diabetic complications were registered by a standardized questionnaire and verified from the medical files by the patient’s attending physician. Data on smoking status were collected by a self-administered questionnaire. Anthropometric data were collected by a trained nurse. Blood pressure was measured twice after a 10-min rest with a 2-min interval, and the mean values were used in the analyses. Blood samples were drawn to determine and HbA1c, serum creatinine, and lipid levels by routine methods as previously described (13).

LTPA was assessed by a self-report questionnaire, which comprised various items regarding exercise type, intensity, duration, and frequency. The details and validation of the questionnaire have been previously described (1517) (Supplementary Table 1). The total amount of LTPA is reported in metabolic equivalents (MET-h/week), and patients were categorized as follows: sedentary (<10 MET-h/week), moderately active (10–40 MET-h/week), and active (>40 MET-h/week). We also analyzed other LTPA-related exposures: 1) exercise intensity (low: no self-reported subjective shortness of breath and no sweating; moderate: a moderate degree of self-reported subjective shortness of breath and sweating; high: a high degree of subjective shortness of breath and sweating), 2) single session duration (low: ≤30 min/session; moderate: 31–60 min/session; and high: >60 min/session), and 3) exercise frequency (low: fewer than one session per week; moderate: one to two sessions per week; high: more than two sessions per week).

Statistical Analyses

Data were analyzed by using SPSS version 22.2 statistical software (IBM Corporation, Armonk, NY). Normally distributed continuous variables are expressed as mean ± SD; otherwise, data are expressed as median with interquartile range. Categorical variables are given as percentages. Between-group differences were assessed by using ANOVA for normally distributed variables and the Kruskal-Wallis test for nonnormally distributed variables. Categorical variables were analyzed by using the χ2 test. The cumulative mortality was assessed by the Kaplan-Meier method, and the log-rank test was used to test between-group differences.

Follow-up started from the baseline visit, and the person-years at risk were calculated until death or the end of 2014. The association between the various LTPA components and mortality was analyzed by using univariable and multivariable Cox proportional hazards regression models. Because there was no interaction between the LTPA components and sex or diabetic nephropathy (defined as the presence of micro/macroalbuminuria or ESRD), all additional analyses were conducted by pooling all individuals together. The multivariable analyses first included the static confounders nephropathy status, sex, duration of diabetes, and age at onset of diabetes and then included in the final models the dynamic covariates HbA1c, systolic blood pressure (SBP), triglycerides, smoking status, and BMI. In the subanalyses of patients with CKD, all CKD groups were analyzed together because the interaction term between the LTPA components and CKD groups was not significant. However, additional analyses were conducted to include CKD groups as a covariate. P < 0.05 was considered statistically significant.

This longitudinal study (mean follow-up time 11.4 ± 3.5 years) included 2,369 patients. At baseline, 310 patients had CKD, 48.5% were men, mean age was 40.1 ± 12.6 years, BMI 25.2 ± 3.6 kg/m2, SBP 135 ± 19 mmHg, HbA1c 8.3 ± 1.4%, and duration of diabetes 23.3 ± 12.8 years. At baseline, the median LTPA was 16.4 MET-h/week (5.3–32.9 MET-h/week), 44.8% of the patients had a history of former smoking (20.7%) or current smoking (24.1%), 68.9% had a normal UAER, and 6.5% had a cardiovascular disease event at baseline.

During follow-up, 270 deaths as a result of any cause occurred. Of these, 75 (27.8%) were cardiovascular deaths. Table 1 shows the baseline clinical characteristics of the patients who died as a result of any cause or cardiovascular events compared with those still alive during follow-up. The patients who died as a result of any cause were more often men, former or current smokers, older, heavier, and sedentary. These patients also had a higher SBP, longer duration of diabetes, more frequent use of antihypertensive drugs, worse lipid profile, and worse glycemic control. The baseline risk factors for a cardiovascular death were nearly identical to those for any cause, apart from the history of smoking that was not statistically significant for cardiovascular mortality.

Table 1

Baseline characteristics according to vital status during follow-up and stratified by death as a result of any cause and cardiovascular death

Death
CharacteristicAliveAny causeP valueCardiovascularP value
Participants (n2,099 270  75  
Male sex (%) 47.0 60.7 <0.001 56.0 0.124 
Age (years) 38.8 ± 12.2 50.1 ± 10.5 <0.001 48.5 ± 9.5 <0.001 
Duration of diabetes (years) 21.9 ± 12.4 33.6 ± 10.8 <0.001 34.2 ± 9.1 <0.001 
BMI (kg/m225.2 ± 3.6 25.5 ± 4.2 0.151 25.7 ± 4.1 0.22 
WHR men 0.92 ± 0.07 0.95 ± 0.08 <0.001 0.96 ± 0.08 <0.001 
WHR women 0.82 ± 0.06 0.86 ± 0.08 <0.001 0.86 ± 0.05 <0.001 
SBP (mmHg) 134 ± 18 146 ± 23 <0.001 147 ± 22 <0.001 
DBP (mmHg) 79 ± 10 78 ± 11 0.235 79 ± 12 0.747 
HbA1c (%) 8.3 ± 1.4 8.6 ± 1.4 <0.001 9.0 ± 1.4 <0.001 
HbA1c (mmol/mol) 67 ± 15 70 ± 15 <0.001 75 ± 15 <0.001 
Total cholesterol (mmol/L) 4.79 ± 0.86 4.91 ± 0.99 0.027 4.83 ± 0.91 0.661 
HDL cholesterol (mmol/L) 1.44 ± 0.40 1.38 ± 0.45 0.047 1.23 ± 0.43 <0.001 
Triglycerides (mmol/L) 0.96 (0.74–1.33) 1.25 (0.92–1.69) <0.001 1.42 (1.07–2.02) <0.001 
AHT (%) 34.3 78.7 <0.001 85.3 <0.001 
β-Blockers (%) 10.0 45.6 <0.001 54.7 <0.001 
Ever smokers (%) 44.2 59.0 <0.001 52.1 0.185 
LTPA (MET-h/week) 17.2 (6.7–33.8) 8.6 (0–26.7) <0.001 8.0 (0–25.9) <0.001 
Low exercise intensity (%) 23.8 54.9 <0.001 54.8 <0.001 
Low exercise duration (%) 12.4 26.3 <0.001 22.7 0.013 
Low exercise frequency (%) 13.8 33.0 <0.001 26.0 0.003 
Death
CharacteristicAliveAny causeP valueCardiovascularP value
Participants (n2,099 270  75  
Male sex (%) 47.0 60.7 <0.001 56.0 0.124 
Age (years) 38.8 ± 12.2 50.1 ± 10.5 <0.001 48.5 ± 9.5 <0.001 
Duration of diabetes (years) 21.9 ± 12.4 33.6 ± 10.8 <0.001 34.2 ± 9.1 <0.001 
BMI (kg/m225.2 ± 3.6 25.5 ± 4.2 0.151 25.7 ± 4.1 0.22 
WHR men 0.92 ± 0.07 0.95 ± 0.08 <0.001 0.96 ± 0.08 <0.001 
WHR women 0.82 ± 0.06 0.86 ± 0.08 <0.001 0.86 ± 0.05 <0.001 
SBP (mmHg) 134 ± 18 146 ± 23 <0.001 147 ± 22 <0.001 
DBP (mmHg) 79 ± 10 78 ± 11 0.235 79 ± 12 0.747 
HbA1c (%) 8.3 ± 1.4 8.6 ± 1.4 <0.001 9.0 ± 1.4 <0.001 
HbA1c (mmol/mol) 67 ± 15 70 ± 15 <0.001 75 ± 15 <0.001 
Total cholesterol (mmol/L) 4.79 ± 0.86 4.91 ± 0.99 0.027 4.83 ± 0.91 0.661 
HDL cholesterol (mmol/L) 1.44 ± 0.40 1.38 ± 0.45 0.047 1.23 ± 0.43 <0.001 
Triglycerides (mmol/L) 0.96 (0.74–1.33) 1.25 (0.92–1.69) <0.001 1.42 (1.07–2.02) <0.001 
AHT (%) 34.3 78.7 <0.001 85.3 <0.001 
β-Blockers (%) 10.0 45.6 <0.001 54.7 <0.001 
Ever smokers (%) 44.2 59.0 <0.001 52.1 0.185 
LTPA (MET-h/week) 17.2 (6.7–33.8) 8.6 (0–26.7) <0.001 8.0 (0–25.9) <0.001 
Low exercise intensity (%) 23.8 54.9 <0.001 54.8 <0.001 
Low exercise duration (%) 12.4 26.3 <0.001 22.7 0.013 
Low exercise frequency (%) 13.8 33.0 <0.001 26.0 0.003 

Data are mean ± SD, median (interquartile range), or %. P values were calculated from univariable Cox proportional hazards regression models with a new event (any-cause death/cardiovascular death) as outcome. AHT, antihypertensive medication; DBP, diastolic blood pressure; WHR, waist-to-hip ratio.

Table 2 shows the 10-year cumulative all-cause mortality rates grouped by total amount of LTPA and its components' intensity, frequency, and exercise duration. In these univariable models, the total amount of LTPA and all the exercise components were associated with all-cause mortality during follow-up.

Table 2

Ten-year cumulative incidence rates for all-cause mortality by LTPA and by exercise intensity, duration, and frequency

Low*Moderate*High*P value
LTPA     
 Incidence 14.4 (12.2, 16.6) 6.6 (5.1, 8.1) 4.8 (2.7, 6.8) <0.001 
 Participants (n833 1,109 427  
 Events (n141 95 34  
Exercise intensity     
 Incidence  17.7 (15.1, 20.2) 6.4 (5.1, 7.7) 2.3 (1.0, 3.6) <0.001 
 Participants (n631 1,224 459  
 Events (n141 101 15  
Exercise frequency     
 Incidence  19.9 (16.5, 23.1) 6.6 (2.9, 10.2) 6.7 (5.6, 7.8) <0.001 
 Participants (n375 158 1,813  
 Events (n87 16 161  
Exercise duration     
 Incidence  16.6 (12.9, 20.1) 6.6 (5.1, 8.1) 6.4 (4.7, 8.0) <0.001 
 Participants (n300 1,149 726  
 Events (n59 103 62  
Low*Moderate*High*P value
LTPA     
 Incidence 14.4 (12.2, 16.6) 6.6 (5.1, 8.1) 4.8 (2.7, 6.8) <0.001 
 Participants (n833 1,109 427  
 Events (n141 95 34  
Exercise intensity     
 Incidence  17.7 (15.1, 20.2) 6.4 (5.1, 7.7) 2.3 (1.0, 3.6) <0.001 
 Participants (n631 1,224 459  
 Events (n141 101 15  
Exercise frequency     
 Incidence  19.9 (16.5, 23.1) 6.6 (2.9, 10.2) 6.7 (5.6, 7.8) <0.001 
 Participants (n375 158 1,813  
 Events (n87 16 161  
Exercise duration     
 Incidence  16.6 (12.9, 20.1) 6.6 (5.1, 8.1) 6.4 (4.7, 8.0) <0.001 
 Participants (n300 1,149 726  
 Events (n59 103 62  

Data are % (95% CI) unless otherwise indicated.

*LTPA: low <10 MET-h/week; moderate 10–40 MET-h/week, and high >40 MET-h/week. Intensity: low (no self-reported subjective shortness of breath and no sweating), moderate (a moderate degree of self-reported subjective shortness of breath and sweating), and high (a high degree of subjective shortness of breath and sweating). Frequency: low fewer than one session/week, moderate one to two sessions/week, and high more than two sessions/week. Duration: low ≤30 min/session, moderate 31–60 min/session, and high >60 min/session.

The multivariable Cox proportional hazards regression models are shown in Table 3. Comparisons were made for low and moderate versus high (the reference group) for the amount of LTPA as well as its components with regard to all-cause mortality. Model 1 is a univariable model that included the exercise components and all-cause mortality. After adjustment for history of smoking and the static confounders sex, diabetic nephropathy, duration of diabetes, and age at onset of diabetes, all the exercise components were associated with all-cause mortality (model 2). In model 3, we added the dynamic risk factors of SBP, triglycerides, BMI, and HbA1c that themselves could be affected by exercise. After this final adjustment, all exercise components were still associated with all-cause mortality.

Table 3

Cox proportional hazards regression models for low and moderate versus high total LTPA and exercise intensity, frequency, and duration for all-cause mortality

Model 1Model 2Model 3
LTPA    
 Low* 2.49 (1.71, 3.62) 2.07 (1.40, 3.06) 1.92 (1.29, 2.86) 
 Moderate* 1.11 (0.75, 1.64) 1.34 (0.89, 2.02) 1.37 (0.91, 2.07) 
 High* 1.00 1.00 1.00 
 Participants (n2,369 2,315 2,274 
 Events (n270 261 255 
Exercise intensity    
 Low 7.83 (4.60, 13.33) 2.78 (1.57, 4.90) 2.39 (1.34, 4.25) 
 Moderate 2.55 (1.48, 4.39) 1.42 (0.80, 2.50) 1.34 (0.76, 2.38) 
 High 1.00 1.00 1.00 
 Participants (n2,314 2,261 2,221 
 Events (n257 249 244 
Exercise frequency    
 Low 2.92 (2.25, 3.79) 2.35 (1.79, 3.09) 2.03 (1.53, 2.70) 
 Moderate 1.13 (0.68, 1.89) 1.45 (0.86, 2.43) 1.33 (0.79, 2.24) 
 High 1.00 1.00 1.00 
 Participants (n2,346 2,292 2,251 
 Events (n264 255 249 
Exercise duration    
 Low 2.50 (1.75, 3.57) 1.86 (1.29, 2.68) 1.79 (1.23, 2.58) 
 Moderate 1.08 (0.79, 1.48) 1.01 (0.73, 1.39) 1.09 (0.78, 1.51) 
 High 1.00 1.00 1.00 
 Participants (n2,175 2,126 2,092 
 Events (n224 217 214 
Model 1Model 2Model 3
LTPA    
 Low* 2.49 (1.71, 3.62) 2.07 (1.40, 3.06) 1.92 (1.29, 2.86) 
 Moderate* 1.11 (0.75, 1.64) 1.34 (0.89, 2.02) 1.37 (0.91, 2.07) 
 High* 1.00 1.00 1.00 
 Participants (n2,369 2,315 2,274 
 Events (n270 261 255 
Exercise intensity    
 Low 7.83 (4.60, 13.33) 2.78 (1.57, 4.90) 2.39 (1.34, 4.25) 
 Moderate 2.55 (1.48, 4.39) 1.42 (0.80, 2.50) 1.34 (0.76, 2.38) 
 High 1.00 1.00 1.00 
 Participants (n2,314 2,261 2,221 
 Events (n257 249 244 
Exercise frequency    
 Low 2.92 (2.25, 3.79) 2.35 (1.79, 3.09) 2.03 (1.53, 2.70) 
 Moderate 1.13 (0.68, 1.89) 1.45 (0.86, 2.43) 1.33 (0.79, 2.24) 
 High 1.00 1.00 1.00 
 Participants (n2,346 2,292 2,251 
 Events (n264 255 249 
Exercise duration    
 Low 2.50 (1.75, 3.57) 1.86 (1.29, 2.68) 1.79 (1.23, 2.58) 
 Moderate 1.08 (0.79, 1.48) 1.01 (0.73, 1.39) 1.09 (0.78, 1.51) 
 High 1.00 1.00 1.00 
 Participants (n2,175 2,126 2,092 
 Events (n224 217 214 

Data are HR (95% CI) unless otherwise indicated. Model 1: exercise components and all-cause mortality. Model 2: model 1 plus sex, duration of diabetes, smoking status, age at onset of diabetes, and diabetic nephropathy. Model 3: model 2 plus SBP, triglycerides, BMI, and HbA1c.

*LTPA: low <10 MET-h/week; moderate 10–40 MET-h/week, and high >40 MET-h/week. Intensity: low (no self-reported subjective shortness of breath and no sweating), moderate (a moderate degree of self-reported subjective shortness of breath and sweating), and high (a high degree of subjective shortness of breath and sweating). Frequency: low fewer than one session/week, moderate one to two sessions/week, and high more than two sessions/week. Duration: low ≤30 min/session, moderate 31–60 min/session, and high >60 min/session.

The 10-year cumulative cardiovascular mortality rates are shown in Supplementary Table 2. The 10-year cumulative cardiovascular mortality rates were 4.7% (95% CI 3.2%, 6.2%) in the low, 1.9% (1.1%, 2.7%) in the moderate, and 1.8% (0.4%, 3.1%) in the high LTPA groups (P = 0.001). Similarly, the 10-year cumulative cardiovascular mortality rates were 6.7% (4.7%, 8.7%) in the low, 1.9% (1.1%, 2.7%) in the moderate, and 0.2% (0.0%, 0.6%) in the high exercise intensity groups (P < 0.001) and 5.5% (3.1%, 7.9%) in the low, 2.8% (0.1%, 5.4%) in the moderate, and 2.2% (1.4%, 3.0%) in the high exercise frequency groups (P = 0.01). The corresponding rates for exercise duration were 5.4% (2.6%, 8.0%) in the low, 2.0% (1.2%, 2.8%) in the moderate, and 2.7% (1.5%, 3.8%) in the high groups (P = 0.02).

Supplementary Table 3 shows the multivariable Cox proportional hazards regression models for all exercise components and cardiovascular mortality. After adding the static confounders and the history of smoking (model 2), exercise frequency and the intensity were still associated with cardiovascular mortality. However, after the final adjustment for all potential confounders (model 3), only exercise intensity was associated with cardiovascular mortality.

We repeated the multivariable analyses for men and women separately with respect to total amount of LTPA, its components, and all-cause mortality (Supplementary Tables 4 and 5). When adjusted for all confounders in model 3, both the total amount of LTPA and all the exercise components (frequency, duration, and intensity) were associated with all-cause mortality in men. In women, on the other hand, only exercise frequency and duration were associated with all-cause mortality in model 3.

Finally, we analyzed the relationship between LTPA and all-cause mortality in patients with CKD (n = 310). Of these patients, 127 died during follow-up as a result of any cause; 48.1% were men; and mean age was 48.8 ± 10.3 years, duration of diabetes 34.0 ± 9.9 years, BMI 26.0 ± 4.2 kg/m2, SBP 146 ± 23 mmHg, and HbA1c 8.4 ± 1.3%. Of the patients with CKD, 64 had received a kidney transplant, and 36 were undergoing dialysis. Table 4 presents the Cox proportional hazards regression models for this population by LTPA, its components, and all-cause mortality. In the univariable model (model 1), all exercise components were associated with all-cause mortality, and this association remained significant in model 2 when adjusted for the static risk factors and history of smoking. In model 3, however, when adjusted for the dynamic confounders, only the total amount of LTPA and exercise frequency were independently associated with all-cause mortality. Additional analyses that included CKD groups as a covariate did not change the results. Only the effect of duration of exercise on mortality became borderline significant (P = 0.06) in model 2. When eGFR was included as a covariate in the final Cox proportional hazards regression model, the association between exercise frequency and mortality remained significant (P = 0.045), but the total amount of LTPA did not. We repeated the same regression analyses in the patients without CKD, and the results were no different from the entire group of patients with type 1 diabetes; that is, the total amount of LTPA and the all exercise components were associated with all-cause mortality when adjusted for all the previous confounders (data not shown). Separate additional Cox proportional hazards regression models in patients undergoing dialysis or patients who had received a kidney transplant are shown in Supplementary Tables 6 and 7.

Table 4

Cox proportional hazards regression models for low versus moderate/high* total LTPA and exercise intensity, frequency, and duration for all-cause mortality in patients with type 1 diabetes and CKD

Model 1Model 2Model 3
LTPA    
 Mortality 1.89 (1.33, 2.69) 1.72 (1.20, 2.46) 1.47 (1.02, 2.12) 
 Participants (n310 303 296 
 Events (n127 124 119 
Exercise intensity    
 Mortality 2.10 (1.45, 3.03) 1.80 (1.23, 2.64) 1.39 (0.92, 2.09) 
 Participants (n297 291 285 
 Events (n119 117 113 
Exercise frequency    
 Mortality 2.39 (1.65, 3.46) 2.20 (1.48, 3.26) 1.90 (1.26, 2.87) 
 Participants (n306 299 292 
 Events (n124 121 116 
Exercise duration    
 Mortality 1.93 (1.24, 3.01) 1.69 (1.07, 2.67) 1.49 (0.92, 2.42) 
 Participants (n264 259 256 
 Events (n99 97 95 
Model 1Model 2Model 3
LTPA    
 Mortality 1.89 (1.33, 2.69) 1.72 (1.20, 2.46) 1.47 (1.02, 2.12) 
 Participants (n310 303 296 
 Events (n127 124 119 
Exercise intensity    
 Mortality 2.10 (1.45, 3.03) 1.80 (1.23, 2.64) 1.39 (0.92, 2.09) 
 Participants (n297 291 285 
 Events (n119 117 113 
Exercise frequency    
 Mortality 2.39 (1.65, 3.46) 2.20 (1.48, 3.26) 1.90 (1.26, 2.87) 
 Participants (n306 299 292 
 Events (n124 121 116 
Exercise duration    
 Mortality 1.93 (1.24, 3.01) 1.69 (1.07, 2.67) 1.49 (0.92, 2.42) 
 Participants (n264 259 256 
 Events (n99 97 95 

Data are HR (95% CI) unless otherwise indicated. Model 1: exercise components and all-cause mortality. Model 2: model 1 plus sex, duration of diabetes, smoking status, age at onset of diabetes, and diabetic nephropathy. Model 3: model 2 plus SBP, triglycerides, BMI, and HbA1c.

*LTPA: low <10 MET-h/week; moderate 10–40 MET-h/week; and high >40 MET-h/week. Intensity: low (no self-reported subjective shortness of breath and no sweating), moderate (a moderate degree of self-reported subjective shortness of breath and sweating), and high (a high degree of subjective shortness of breath and sweating). Frequency: low fewer than one session/week, moderate one to two sessions/week, and high more than two sessions/week. Duration: low ≤30 min/session, moderate 31–60 min/session, and high >60 min/session.

In this prospective study, we show that baseline LTPA is associated with reduced all-cause mortality in type 1 diabetes. Of note, physical activity is associated with a lower risk of premature mortality in patients with type 1 diabetes and CKD. In addition, we show that all exercise components are associated with reduced mortality rates during follow-up. This finding is unaffected by any adjustment for potential confounders.

The observation that physical activity reduces mortality in patients with CKD is novel, and to our knowledge, no other studies have investigated the effect of exercise on mortality in this patient group. The finding is not unexpected given that we previously showed that exercise, and in particular intensive exercise, is associated with a reduced risk of progression of diabetic nephropathy in individuals with type 1 diabetes (3). Patients with CKD are expected to be more physically inactive and to have a lower level of fitness and muscle function (912), but exercise appears to be safe and to have significant health benefits in these patients, and 30 min of daily moderate-intensity exercise is therefore recommended (9). In the current study, total LTPA as well as all its components were associated with all-cause mortality when adjusted for smoking and static risk factors. However, when the dynamic confounders were added to the Cox proportional hazards regression model, only the total amount of LTPA and a higher exercise frequency were associated with a reduced mortality rate. These findings support the current recommendations for CKD management and suggest that exercise should always be a part of the treatment regimen.

With respect to exercise and mortality in individuals with type 1 diabetes, only a few previous longitudinal studies have been performed. The Pittsburgh IDDM Morbidity and Mortality Study showed that sedentary men were more likely to die during follow-up than active men, also after controlling for potential confounders such as age, BMI, insulin dose, presence of diabetic complications, cigarette smoking, and current alcohol drinking (7). Of note, participation in team sports during high school or college was associated with a reduced mortality risk in men, although this association was no more significant after controlling for other potential confounders such as age and year of type 1 diabetes onset, smoking status, alcohol consumption, hypertension, renal disease, and education level (6). The EURODIAB Prospective Complications Study showed an inverse association between moderate or vigorous physical activity once or more a week at baseline and all-cause mortality in both sexes. However, this association was reduced to a nonsignificant level after adjustment for potential confounders (8).

The current finding of an association between exercise and all-cause mortality in the entire study cohort not only is in line with but also extends the observations from the EURODIAB Prospective Complications Study and the Pittsburgh IDDM Morbidity and Mortality Study (6,8). We also observed that the association is true for both sexes with respect to duration and frequency, but the associations with total amount of LTPA and intensity were significant only in men after adjustment for confounders. However, on the basis of previous studies and the current study, sex-specific differences regarding the association between baseline physical activity and all-cause mortality are unlikely.

The data regarding the effect of physical activity on cardiovascular mortality are in line with the data for all-cause mortality, although after adjustment for all potential confounders, only exercise intensity remained associated with cardiovascular mortality. This finding probably is due to lost power because the number of cardiovascular deaths was only 75 compared with 270 deaths as a result of any cause. We have previously shown that exercise, and particularly intensive and frequent exercise, is associated with a reduced risk of incident cardiovascular events in type 1 diabetes (4). Moreover, higher exercise intensity was associated with lower risk of diabetic nephropathy, a main driver of the risk of cardiovascular disease (3). Given that cardiovascular disease is the most common cause of death among patients with type 1 diabetes (18), that exercise intensity also is associated with a lower risk of cardiovascular mortality is not surprising. The reason why only intensity remains significant after adjustment for dynamic confounders is not known, but it can be hypothesized that exercise intensity is the most important component of LTPA because of its impact on heart rate and autonomic nervous function. We have previously shown that patients with type 1 diabetes display reduced baroreflex sensitivity (19), which is clinically detectable as a higher pulse rate and, thus, potentially reduced exercise tolerance and fitness. Another possibility is that the dynamic confounders are themselves affected by exercise.

The major strengths of this study are the substantially larger nationwide study cohort compared with previous studies and the prospective design. Furthermore, the LTPA questionnaire has been previously validated in the Finnish population and is custom-made for Finnish conditions (17). An important feature of the study compared with previous longitudinal studies is that we were able to analyze not only the total amount of LTPA but also its main components. Furthermore, to our knowledge, no prospective data are available in the literature about exercise and mortality in patients with type 1 diabetes and CKD. Finally, the mortality data are complete, and no patients were lost to follow-up.

The study also has some limitations. Exercise was self-reported, and because work-related exercise was not assessed, some overestimation or underestimation of the true activity level of the patients exists. The ideal method to assess physical activity would be to use more accurate objective measurements, such as pedometers or accelerometers (20), but although such methods are feasible for smaller study populations, they are difficult to apply to nationwide cohorts, such as ours. Exercise was assessed only at baseline, and change of exercise habits over time, therefore, was not evaluated. On the other hand, the questionnaire has been shown to be representative and to have a small intraindividual variability (1517). Unfortunately, we were not able to evaluate the interaction between exercise and other lifestyle factors, such as nutrition or socioeconomic status, in the current study.

In conclusion, the findings indicate that physical activity is associated with reduced all-cause and cardiovascular mortality in patients with type 1 diabetes. In addition, we show an association between physical activity and reduced all-cause mortality in patients with type 1 diabetes and CKD. Thus, exercise can be recommended to all patients with type 1 diabetes, including those with CKD.

This article is featured in a podcast available at http://www.diabetesjournals.org/content/diabetes-core-update-podcasts.

Acknowledgments. The authors acknowledge the important role of all the FinnDiane physicians and nurses at each participating center for patient recruitment and collection of samples and data (Supplementary Table 8).

Funding. The FinnDiane Study Group was supported by grants from Folkhälsan Research Foundation, Wilhelm and Else Stockmann Foundation, Liv och Hälsa Foundation, Helsinki University Central Hospital Research Funds, Academy of Finland, Novo Nordisk Foundation, Finska Läkaresällskapet, and Tekes. None of these bodies played a role in data collection, analysis, or preparation of the manuscript.

Duality of Interest. P.-H.G. has received investigator-initiated research grants from Eli Lilly and Roche; is an advisory board member for AbbVie, AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Medscape, Merck Sharp & Dohme, Novartis, and Sanofi; and has received lecture fees from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Genzyme, Novartis, Novo Nordisk, Merck Sharp & Dohme, and Sanofi. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. H.T.-D. and V.H. analyzed the data. H.T.-D. and P.-H.G. drafted the manuscript. J.W., C.F., L.M.T., M.S., N.E., H.O.T., and P.-H.G. were involved in data collection. All authors contributed to and approved of the final submitted version of the manuscript. P.-H.G. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 53rd European Association for the Study of Diabetes Annual Meeting, Lisbon, Portugal, 11–15 September 2017.

1.
Nocon
M
,
Hiemann
T
,
Müller-Riemenschneider
F
,
Thalau
F
,
Roll
S
,
Willich
SN
.
Association of physical activity with all-cause and cardiovascular mortality: a systematic review and meta-analysis
.
Eur J Cardiovasc Prev Rehabil
2008
;
15
:
239
246
[PubMed]
2.
Sigal
RJ
,
Kenny
GP
,
Wasserman
DH
,
Castaneda-Sceppa
C
,
White
RD
.
Physical activity/exercise and type 2 diabetes: a consensus statement from the American Diabetes Association
.
Diabetes Care
2006
;
29
:
1433
1438
[PubMed]
3.
Wadén
J
,
Tikkanen
HK
,
Forsblom
C
, et al.;
FinnDiane Study Group
.
Leisure-time physical activity and development and progression of diabetic nephropathy in type 1 diabetes: the FinnDiane Study
.
Diabetologia
2015
;
58
:
929
936
[PubMed]
4.
Tikkanen-Dolenc
H
,
Wadén
J
,
Forsblom
C
, et al.;
FinnDiane Study Group
.
Frequent and intensive physical activity reduces risk of cardiovascular events in type 1 diabetes
.
Diabetologia
2017
;
60
:
574
580
[PubMed]
5.
Groop
PH
,
Thomas
MC
,
Moran
JL
, et al.;
FinnDiane Study Group
.
The presence and severity of chronic kidney disease predicts all-cause mortality in type 1 diabetes
.
Diabetes
2009
;
58
:
1651
1658
[PubMed]
6.
Dorman
JS
,
Laporte
RE
,
Kuller
LH
, et al
.
Pittsburgh Insulin-Dependent Diabetes Mellitus (IDDM) Morbidity and Mortality Study: mortality results
.
Diabetes
1984
;
33
:
271
276
[PubMed]
7.
Moy
CS
,
Songer
TJ
,
LaPorte
RE
, et al
.
Insulin-dependent diabetes mellitus, physical activity, and death
.
Am J Epidemiol
1993
;
137
:
74
81
[PubMed]
8.
Tielemans
SM
,
Soedamah-Muthu
SS
,
De Neve
M
, et al
.
Association of physical activity with all-cause mortality and incident and prevalent cardiovascular disease among patients with type 1 diabetes: the EURODIAB Prospective Complications Study
.
Diabetologia
2013
;
56
:
82
91
[PubMed]
9.
Howden
EJ
,
Coombes
JS
,
Isbel
NM
.
The role of exercise training in the management of chronic kidney disease
.
Curr Opin Nephrol Hypertens
2015
;
24
:
480
487
[PubMed]
10.
Gould
DW
,
Graham-Brown
MP
,
Watson
EL
,
Viana
JL
,
Smith
AC
.
Physiological benefits of exercise in pre-dialysis chronic kidney disease
.
Nephrology (Carlton)
2014
;
19
:
519
527
[PubMed]
11.
Stack
AG
,
Molony
DA
,
Rives
T
,
Tyson
J
,
Murthy
BV
.
Association of physical activity with mortality in the US dialysis population
.
Am J Kidney Dis
2005
;
45
:
690
701
[PubMed]
12.
Beddhu
S
,
Baird
BC
,
Zitterkoph
J
,
Neilson
J
,
Greene
T
.
Physical activity and mortality in chronic kidney disease (NHANES III)
.
Clin J Am Soc Nephrol
2009
;
4
:
1901
1906
[PubMed]
13.
Thorn
LM
,
Forsblom
C
,
Fagerudd
J
, et al.;
FinnDiane Study Group
.
Metabolic syndrome in type 1 diabetes: association with diabetic nephropathy and glycemic control (the FinnDiane study)
.
Diabetes Care
2005
;
28
:
2019
2024
[PubMed]
14.
Levey
AS
,
Stevens
LA
,
Schmid
CH
, et al.;
CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration)
.
A new equation to estimate glomerular filtration rate
.
Ann Intern Med
2009
;
150
:
604
612
[PubMed]
15.
Wadén
J
,
Tikkanen
H
,
Forsblom
C
, et al.;
FinnDiane Study Group
.
Leisure time physical activity is associated with poor glycemic control in type 1 diabetic women: the FinnDiane study
.
Diabetes Care
2005
;
28
:
777
782
[PubMed]
16.
Lakka
TA
,
Salonen
JT
.
Intra-person variability of various physical activity assessments in the Kuopio Ischaemic Heart Disease Risk Factor Study
.
Int J Epidemiol
1992
;
21
:
467
472
[PubMed]
17.
Salonen
J
,
Lakka
T
.
Assessment of physical activity in population studies: validity and consistency of the methods in the Kuopio Ischemic Heart Disease Risk Factor Study
.
Scand J Sports Sci
1987
;
9
:
89
95
18.
Morrish
NJ
,
Wang
SL
,
Stevens
LK
,
Fuller
JH
,
Keen
H
.
Mortality and causes of death in the WHO multinational study of vascular disease in diabetes
.
Diabetologia
2001
;
44
(
Suppl. 2
):
S14
S21
[PubMed]
19.
Rosengård-Bärlund
M
,
Bernardi
L
,
Fagerudd
J
, et al.;
FinnDiane Study Group
.
Early autonomic dysfunction in type 1 diabetes: a reversible disorder
?
Diabetologia
2009
;
52
:
1164
1172
[PubMed]
20.
Trost
SG
,
Pate
RR
,
Freedson
PS
,
Sallis
JF
,
Taylor
WC
.
Using objective physical activity measures with youth: how many days of monitoring are needed
?
Med Sci Sports Exerc
2000
;
32
:
426
431
[PubMed]
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.

Supplementary data