Diabetes mellitus is associated with increased incidence of heart failure,1 even in the absence of hypertension and coronary artery disease. Development of clinical heart failure in these individuals involves progression through several intermediate cardiac phenotypes including abnormal left ventricular remodeling and subclinical systolic and diastolic dysfunction.2,3 Subclinical diastolic dysfunction is one of the earliest manifestations of the effects of diabetes on cardiac function seen in up to 75% patients.4,5
Numerous studies have shown that higher level of cardiorespiratory fitness is associated with favorable cardiac structural and functional changes and lower risk of incident heart failure.6–8 Exercise training is a well-established, widely available, and low-cost intervention that has been shown to improve cardiorespiratory fitness levels in patients at risk for heart failure as well as those with established heart failure.9,10 Several studies have evaluated the impact of exercise training in individuals with diabetes who are at risk for future heart failure development.10,11 While, these studies have demonstrated improvement in cardiorespiratory fitness with training, the impact of exercise on diastolic performance has not been well defined.
A crucial step in evaluation of interventions such as exercise training as a potential strategy for prevention is to understand their impact on the intermediate cardiac phenotypes such as asymptomatic systolic and diastolic dysfunction, particularly in high-risk patients. Accordingly, we performed a meta-analysis to evaluate the effects of exercise training on cardiac function in patients with type 2 diabetes (T2D).
We conducted a comprehensive search of the following databases: MEDLINE, PubMed, EMBASE, and Cochrane library using the MeSH terms and key words included exercise training, cardiac rehabilitation, pulmonary rehabilitation, cardiopulmonary rehabilitation, exercise, walking, and diabetes. In addition, we also searched the reference lists of review articles to identify relevant studies. We conducted search from 1965 to February 2014 and updated in December 2017. We did not apply language restrictions. The protocol of our systematic review and meta-analysis was pre-specified but not published or registered.
We searched for all types of articles (randomized and nonrandomized study designs) that enrolled adult patients (age ≥18 y) with T2D and diastolic dysfunction who underwent exercise training versus control/no intervention. We excluded patients with metabolic syndrome, cardiovascular disease, and overt heart failure. Our initial search identified 23 315 studies after removing duplicates. To retain sensitivity and not to miss any important studies, we used a broad search strategy. Two authors carried out the process of study selection independently. We excluded 23 293 studies by review of title and abstract. The full text was obtained for 22 studies, out of which 16 studies were excluded after assessing them against the selection criteria. Other authors were consulted in case of disagreements or discrepancies. A flowchart of the study selection process was developed as suggested by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (see Supplemental Digital Content 1, available at: http://links.lww.com/JCRP/A76). Six studies were included in the final analysis. The primary outcome variables of the study were the earliest markers of diastolic (mitral annulus early diastolic velocity [é in cm/sec]) and systolic (myocardial strain expressed as %) functions. Secondary variables included others markers of diastolic function: E/A (mitral peak early [E] to late diastolic [A] filling velocity); E/é (mitral inflow to annular ratio); DT (deceleration time in msec); systolic velocity reported in cm/sec; and peak
O2 (peak oxygen uptake expressed in mL/kg/min). Studies failing to report at least 1 of the primary outcomes were excluded from analysis.
Two authors independently extracted data. If more than 1 publication of the same study existed, we grouped the reports together and used the publication with the most complete data in the analyses. The data extraction was carried out using a pre-defined data extraction form, including general characteristics of the trials, risk of bias, and characteristics of intervention and control.
ASSESSMENT OF RISK OF BIAS
Two authors independently assessed the risk of bias in included studies using the Cochrane Collaboration's tool for assessing risk of bias.12 Disagreements were resolved by discussion with other authors. We assessed the included studies for selection bias (sequence generation and allocation sequence concealment), performance bias (blinding of participants), detection bias (blinding of outcome assessors), attrition bias (incomplete outcome data), reporting bias (selective outcome reporting), and other biases (ie, baseline imbalance). Those findings of the final assessment for all included studies are summarized in Supplemental Digital Content 2, available at: http://links.lww.com/JCRP/A77.
DATA SYNTHESIS AND STATISTICAL ANALYSIS
All the analyses were performed using Stata (StataCorp) version 12.1 software. The meta-analysis has been reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.13 A fixed-effects model was used to conduct the meta-analysis of the outcomes. Heterogeneity was assessed using the I2 test and the random-effects model was used for significant heterogeneity (I2 > 50%). Data were analyzed using the standardized mean difference (SMD) with 95% CI for all the outcomes. To assess the effect of demographic factors, age, and body mass index (BMI) on primary outcomes of é and global longitudinal strain, a random-effects meta-regression model was used.
Six studies14–19 (n = 441 individuals) were included in the analyses. Mean age of the participants was 56 ± 10 y and 74.5% were males. Four of the included studies were randomized controlled trials, and 2 studies18,19 were nonrandomized controlled studies that were initially planned as randomized but, because of enrollment difficulties, were later changed to a nonrandomized controlled design. The mean duration of exercise training in the included studies was 12 mo. Baseline demographic and clinical characteristics of the study participants from the included studies are summarized in Table 1. All the included studies included participants with T2D without overt cardiovascular disease. The inclusion and exclusion criteria, exercise protocol, control group, duration of follow-up, and the outcomes measured in the included studies are summarized in Table 2. All the study participants had an echocardiographic assessment at baseline and follow-up.
QUALITY ASSESSMENT OF INCLUDED STUDIES
Overall, the quality of included studies was moderate to low (see Supplemental Digital Content 2, available at: http://links.lww.com/JCRP/A77). The random sequence generation was defined in 4 included trials.14–17 Allocation sequence was concealed in only 2 of the 4 randomized studies15,17 and was unclear in 2 studies.14,16 The risk of bias for allocation concealment was high in the nonrandomized studies. The participants were unblinded for the intervention in all the included studies due to the nature of the intervention (exercise); however, this is unlikely to introduce a bias. The outcomes assessors were blinded in 4 studies, reflecting a low risk of detection bias. The blinding of outcome assessors is unclear in 2 studies.14,18 Two studies had a high dropout rate and, therefore, had high risk of attrition bias.14,19 All included studies reported all the clinically relevant outcomes, except for 1,17 which did not report peak oxygen uptake, thereby reflecting a high risk of reporting bias. We also found significant other biases in all included studies: small sample size14–16; lack of true control group14; self-reporting of physical activity15; differences in baseline characteristics between the 2 groups16; and change in study design from randomized to nonrandomized controlled.18,19
Effect of exercise on é
Alterations in é being the earliest and most sensitive marker of diastolic dysfunction was our primary outcome. Five included studies14,15,17–19 reported this outcome at baseline and after exercise training and our meta-analysis showed a significant improvement in é with exercise (SMD = 0.58; 95% CI, 0.09-1.07) compared with the control group (Figure 1A). The analysis was performed using random-effects model because of significant heterogeneity between the studies (I2 = 77.4%). Metaregression analysis showed a nonsignificant effect of BMI (kg/m2) and age on the pooled SMD (BMI metaregression coefficient = −0.36; P = .24, and age: metaregression coefficient = −0.14; P = .10).
Effect of exercise on global longitudinal strain
Six included studies14–19 reported this outcome by either tissue Doppler imaging14–18 or 2-D Speckle-tracking echo19 and meta-analysis demonstrated a significant improvement from baseline to follow-up with exercise training compared with the control group (SMD = 0.62; 95% CI, 0.04-1.21). This analysis was also performed using random-effects model due to significant heterogeneity between the studies (I2 = 86.2%) (Figure 1B). Metaregression analysis showed a significant effect of age but nonsignificant effect of BMI on the pooled SMD (age metaregression coefficient, −0.32; P = .0012, and BMI metaregression coefficient, −0.43; P = .31).
There were nonsignificant differences noted between exercise and control groups when other parameters of diastolic dysfunction E/A, E/é and DT were analyzed from baseline to follow-up. Four of the 6 studies reported both E/A and E/é outcomes.14,17–19 Meta-analysis using random-effects model did not show a significant difference between exercise and control groups (SMD = 0.59; 95% CI, −0.40 to 1.58, and SMD = −0.34; 95% CI, −0.99 to 0.31, respectively) (Figures 2A and 2B). Pooling across the available studies17–19 was done using fixed-effects analysis for DT and found no significant difference between the groups (SMD = 0.14; 95% CI, −0.16 to 0.44; I2 = 39%) (Figure 2C).
Our meta-analysis showed a significant improvement in peak
O2 between exercise and control groups (SMD = 1.43; 95% CI, 0.51-2.35) from baseline to follow-up in patients with T2D (Figure 3A). The analysis was performed using random-effects model, since there was significant heterogeneity between the studies, I2 = 93.8%.
All 6 studies14–19 reported the effect of exercise on systolic velocity and pooling data from the studies using random-effects meta-analysis showed nonsignificant changes in systolic function (SMD = −0.66; 95% CI, −1.51 to 0.19; I2 = 93.2%) (Figure 3B). The weighted mean differences with standard errors for all outcomes are presented in Table 3. No adverse effects of exercise training were reported in any of the included studies.
In this meta-analysis, we observed several clinically relevant findings. First, exercise training was associated with improvement in cardiorespiratory fitness among individuals with diabetes, which is consistent with a previous study.20 Second, exercise training was associated with improvement in the earliest, preload independent measure of diastolic dysfunction (é) in patients with T2D.21 The improvement in é was noted even in patients with longer duration of diabetes.19 Third, exercise training was also associated with improvement in measures of systolic function as measured by global longitudinal strain. Finally and interestingly, the effect of exercise training on various measures of diastolic function was different, with improvements in é but no significant change in measures of E/A, E/é, or deceleration time.
We observed robust improvement in fitness, as noted by improvement in peak
O2, with exercise training in obese patients and patients with diabetes. This highlights the importance of exercise training in these individuals for improving fitness and potentially altering long-term cardiovascular risk.
Previous studies evaluating the impact of exercise training on diastolic function have demonstrated mixed results. A randomized controlled trial looking at the effects of exercise training in patients with diastolic heart failure showed significantly improved peak
O2 and markers of cardiac diastolic function including E/é, é, and left atrial volume index.22 This benefit of atrial reverse remodeling and improved left ventricular diastolic function also translated into improved quality of life in diastolic heart failure patients.22 Another study involving patients with systolic and diastolic dysfunction showed improvement in peak
O2 with exercise training but no improvement in echocardiographic diastolic parameters.23 Also, a randomized controlled trial in older patients with heart failure and preserved ejection fraction showed increased peak
O2 with exercise training; however, it did not assess changes in the diastolic function using novel tissue Doppler techniques.24
Our results are similar to the studies conducted in obese patients and patients with metabolic syndrome with subclinical diastolic dysfunction which showed favorable effects of exercise intervention on é,25,26 but no improvement in other markers of diastolic dysfunction. It is important to note that improvement in some of the markers of diastolic dysfunction was also observed with good glycemic control in 1 study27; however, the results were not reproducible in other studies,28,29 suggesting that glycemic control without exercise training is insufficient in producing improvement in diastolic function.
Strain and strain-rate imaging is a new echocardiographic method for comprehensive assessment of early myocardial dysfunction30; therefore, we assessed it as our coprimary outcome and observed improvements in systolic function measured as longitudinal strain. This is consistent with studies that have shown significantly higher left ventricular longitudinal strain and strain rates in athletes than in controls.31,32
The mechanisms underlying the improvement in cardiac function are unclear. Some of the proposed mechanisms include decreased afterload; improved cardiac sympathovagal balance18; cardiac remodeling as seen in an athletic heart33,34; enhanced endothelial function; alterations in cardiac lipid deposition; receptor for advanced glycation end products; and improved myofilament calcium sensitivity.16,20 However, these could not be assessed in the current meta-analysis because of lack of reporting of these outcomes in the included trials.
The limitations of this study include the number of included studies and inclusion of nonrandomized trials that were initially planned as randomized, reflecting the paucity of and challenges to conduct randomized controlled trials addressing this question. The high allocation bias seen in some of the included studies reflects the difficulties in allocation of participants in the control arm. Also, there is a high dropout rate and subjectivity in assessment during the maintenance phase of the intervention. There was also suspected detection bias in 2 included studies that did not specify whether the cardiologists reading the echocardiograms were blinded. In addition, the small sample size of some included trials and high dropout rate in a few included studies resulted in wide CIs for effect estimates. This emphasizes the significance of this systematic review and meta-analysis, which pooled all the available data.
It is important to note that there was, overall, a short duration follow-up in the included studies (mean = 12 mo), thereby probably leading to lack of significant changes in other parameters of diastolic dysfunction which may be affected by longer-term remodeling with consistent exercise. Longer duration trials are also needed to assess sustainability of effects and clinical outcomes such as development of heart failure and mortality. Other potential bias in this meta-analysis is the inclusion of only published studies.
Current studies do not allow us to draw conclusions regarding the evolution and maintenance of improvements in a broader panel of indices of diastolic function. Overall, our findings suggest that exercise training in patients with T2D in the pre-clinical stage, without overt coronary artery disease or congestive heart failure, improves early diastolic dysfunction, myocardial strain, and peak
O2 and thus may be used as a therapeutic intervention in this population. Long-term studies are needed to assess the benefits on other markers of diastolic dysfunction, quality of life, and overall mortality. This review of the limited data available suggests that sustained exercise training in patients with risk factors for key, early underpinnings of heart failure with preserved ejection fraction may enhance cardiac function and, hopefully, this would translate into sustained improvements in quality of life and delay the progression to symptomatic heart failure with preserved ejection fraction. Future studies with adequate sample sizes, longer follow-up duration, and relevant clinic end points are needed to establish the effects of exercise in this important patient population.
1. Nichols GA, Gullion CM, Koro CE, Ephross SA, Brown JB. The incidence of congestive heart failure in type 2 diabetes
: an update. Diabetes Care. 2004;27:1879–1884.
2. Borlaug BA. Fatness, fitness, stiffness, and age: how does it lead to heart failure?JACC Heart Fail. 2014;2:247–249.
3. Lam CS, Borlaug BA, Kane GC, Enders FT, Rodeheffer RJ, Redfield MM. Age-associated increases in pulmonary artery systolic pressure in the general population. Circulation. 2009;119:2663–2670.
4. Boyer JK, Thanigaraj S, Schechtman KB, Perez JE. Prevalence of ventricular diastolic dysfunction
in asymptomatic, normotensive patients with diabetes mellitus. Am J Cardiol. 2004;93:870–875.
5. Patil VC, Patil HV, Shah KB, Vasani JD, Shetty P, et al Diastolic dysfunction
in asymptomatic type 2 diabetes
mellitus with normal systolic function. J Cardiovasc Dis Res. 2011;2(4):213–222.
6. Berry JD, Pandey A, Gao A, et al Physical fitness and risk for heart failure and coronary artery disease. Circ Heart Fail. 2013;6:627–634.
7. Brinker SK, Pandey A, Ayers CR, et al Association of cardiorespiratory fitness
with left ventricular remodeling and diastolic function: the Cooper Center Longitudinal Study. JACC Heart Fail. 2014;2:238–246.
8. Pandey A, Garg S, Khunger M, et al Dose-response relationship between physical activity and risk of heart failure: a meta-analysis. Circulation. 2015;132:1786–1794.
9. Kitzman DW, Brubaker PH, Herrington DM, et al Effect of endurance exercise training
on endothelial function and arterial stiffness in older patients with heart failure and preserved ejection fraction: a randomized, controlled, single-blind trial. J Am Coll Cardiol. 2013;62:584–592.
10. Church TS, Blair SN, Cocreham S, et al Effects of aerobic and resistance training on hemoglobin A1c levels in patients with type 2 diabetes
: a randomized controlled trial. JAMA. 2010;304:2253–2262.
11. Church TS, Earnest CP, Skinner JS, Blair SN. Effects of different doses of physical activity on cardiorespiratory fitness
among sedentary, overweight or obese postmenopausal women with elevated blood pressure: a randomized controlled trial. JAMA. 2007;297:2081–2091.
12. Higgins JPT. Cochrane Handbook for Systematic Reviews of Interventions. The Cochrane Collaboration; 2011. http://www.cochrane-handbook.org
. Accessed December 2017.
13. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535.
14. Hollekim-Strand SM, Bjorgaas MR, Albrektsen G, Tjønna AE, Wisløff U, Ingul CB. High-intensity interval exercise effectively improves cardiac function in patients with type 2 diabetes
mellitus and diastolic dysfunction
: a randomized controlled trial. J Am Coll Cardiol. 2014;64:1758–1760.
15. Hordern MD, Coombes JS, Cooney LM, Jeffriess L, Prins JB, Marwick TH. Effects of exercise intervention on myocardial function in type 2 diabetes
. Heart. 2009;95:1343–1349.
16. Loimaala A, Groundstroem K, Rinne M, Nenonen A, Huhtala H, Vuori I. Exercise training
does not improve myocardial diastolic tissue velocities in type 2 diabetes
. Cardiovasc Ultrasound. 2007;5:32.
17. Ofstad AP, Johansen OE, Gullestad L, et al Neutral impact on systolic and diastolic cardiac function of 2 years of intensified multi-intervention in type 2 diabetes
: the randomized controlled Asker and Baerum Cardiovascular Diabetes (ABCD) study. Am Heart J. 2014;168:280–288.
18. Sacre JW, Jellis CL, Jenkins C, et al A six-month exercise intervention in subclinical diabetic heart disease: effects on exercise capacity, autonomic and myocardial function. Metabolism. 2014;63:1104–1114.
19. Schmidt JF, Andersen TR, Horton J, et al Soccer training improves cardiac function in men with type 2 diabetes
. Med Sci Sports Exerc. 2013;45:2223–2233.
20. Maiorana A, O'Driscoll G, Cheetham C, et al The effect of combined aerobic and resistance exercise training
on vascular function in type 2 diabetes
. J Am Coll Cardiol. 2001;38:860–866.
21. Sohn DW, Chai IH, Lee DJ, et al Assessment of mitral annulus velocity by Doppler tissue imaging in the evaluation of left ventricular diastolic function. J Am Coll Cardiol. 1997;30:474–480.
22. Edelmann F, Gelbrich G, Dungen HD, et al Exercise training
improves exercise capacity and diastolic function in patients with heart failure with preserved ejection fraction: results of the Ex-DHF (Exercise training
in Diastolic Heart Failure) pilot study. J Am Coll Cardiol. 2011;58:1780–1791.
23. Smart N, Haluska B, Jeffriess L, Marwick TH. Exercise training
in systolic and diastolic dysfunction
: effects on cardiac function, functional capacity, and quality of life. Am Heart J. 2007;153:530–536.
24. Kitzman DW, Brubaker PH, Morgan TM, Stewart KP, Little WC. Exercise training
in older patients with heart failure and preserved ejection fraction: a randomized, controlled, single-blind trial. Circ Heart Fail. 2010;3:659–567.
25. Fournier SB, Donley DA, Bonner DE, Devallance E, Olfert IM, Chantler PD. Improved arterial-ventricular coupling in metabolic syndrome after exercise training
: a pilot study. Med Sci Sports Exerc. 2015;47:2–11.
26. Wong CY, Byrne NM, O'Moore-Sullivan T, Hills AP, Prins JB, Marwick TH. Effect of weight loss due to lifestyle intervention on subclinical cardiovascular dysfunction in obesity (body mass index >30 kg/m2
). Am J Cardiol. 2006;98:1593–1598.
27. von Bibra H, Hansen A, Dounis V, Bystedt T, Malmberg K, Ryden L. Augmented metabolic control improves myocardial diastolic function and perfusion in patients with non-insulin dependent diabetes. Heart. 2004;90:1483–1484.
28. Jarnert C, Landstedt-Hallin L, Malmberg K, et al A randomized trial of the impact of strict glycaemic control on myocardial diastolic function and perfusion reserve: a report from the DADD (Diabetes mellitus And Diastolic Dysfunction
) study. Eur J Heart Fail. 2009;11:39–47.
29. Vintila VD, Roberts A, Vinereanu D, Fraser AG. Progression of subclinical myocardial dysfunction in type 2 diabetes
after 5 years despite improved glycemic control. Echocardiography. 2012;29:1045–1053.
30. Dandel M, Lehmkuhl H, Knosalla C, Suramelashvili N, Hetzer R. Strain and strain rate imaging by echocardiography—basic concepts and clinical applicability. Curr Cardiol Rev. 2009;5:133–148.
31. Simsek Z, Gundogdu F, Alpaydin S, et al Analysis of athletes' heart by tissue Doppler and strain/strain rate imaging. Int J Cardiovasc Imaging. 2011;27:105–111.
32. Vitarelli A, Capotosto L, Placanica G, et al Comprehensive assessment of biventricular function and aortic stiffness in athletes with different forms of training by three-dimensional echocardiography and strain imaging. Eur Heart J Cardiovasc Imaging. 2013;14:1010–1020.
33. Baggish AL, Wood MJ. Athlete's heart and cardiovascular care of the athlete: scientific and clinical update. Circulation. 2011;123:2723–2735.
34. Tumuklu MM, Ildizli M, Ceyhan K, Cinar CS. Alterations in left ventricular structure and diastolic function in professional football players: assessment by tissue Doppler imaging and left ventricular flow propagation velocity. Echocardiography. 2007;24:140–148.