European Journal of Internal Medicine 27 (2016) 37–47

Contents lists available at ScienceDirect

European Journal of Internal Medicine journal homepage: www.elsevier.com/locate/ejim

Original Article

Efficacy of lifestyle interventions in patients with type 2 diabetes: A systematic review and meta-analysis Xiao-Li Huang ⁎, Jian-Hua Pan, Dan Chen, Jing Chen, Fang Chen, Tao-Tao Hu Department of Nephrology, The First Hospital of Wuhan, China

a r t i c l e

i n f o

Article history: Received 9 June 2015 Received in revised form 3 November 2015 Accepted 16 November 2015 Available online 3 December 2015 Keywords: Diabetes Cardiovascular Lifestyle Intervention Meta-analysis

a b s t r a c t Background: The current meta-analysis evaluated the outcomes of various lifestyle interventions, including diet modifications (DIET), physical activity (PA), and patient education (EDU) in reducing the risk of cardiovascular disease in patients with type 2 diabetes. Methods: Randomized clinical trials comparing lifestyle intervention with “usual care” (control) in type 2 diabetes patients were hand-searched from medical databases by two independent reviewers using the terms “diabetes, cardiovascular risk, lifestyle, health education, dietary, exercise/physical activities, and behavior intervention”. Results: Of the 235 studies identified, 17 were chosen for the meta-analysis. The average age of patients ranged from 50–67.3 years. Results reveal no significant difference between the groups, with respect to BMI, while PA and DIET yielded a greater reduction in HbA1c. Significant reduction in both systolic and diastolic pressures in the DIET group, and diastolic pressure in the PA group, was observed. HDL-c in the DIET group was significantly higher than the control group, while no change in LDL-c levels, was seen in all three intervention subtypes. There was no difference between the EDU vs. the control group in terms of HbA1c, blood pressure or HDL-c and LDL-c. Conclusion: DIET intervention showed an improvement in HbA1c, systolic/diastolic blood pressure and HDL-c, with an exception of LDL-c and BMI, suggesting that nutritional intervention had a significant impact on the quality of life by reducing the cardiovascular risk in type 2 diabetes patients. © 2015 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.

1. Introduction The global increase in the prevalence of diabetes seen in recent years has been attributed mostly to obesity, poor diet, and lack of physical activity. The World Health Organization projects diabetes as the 7th leading cause of death and estimates that there will be 366 million adults with diabetes in 2030 [1,2]. Of which, 90% of people will have type 2 diabetes, resulting from the body's ineffective use of insulin [3] and poor glycemic control. Reduced metabolic control of glucose can in turn increase the risk of cardiovascular diseases (CVD). The common cardiovascular risk factors associated with diabetes include, increased body weight, glycemia, serum lipids, and blood pressure [4]. Improvement of glycemic control and reduction of cardiovascular risk factors through diet and exercise has long been advocated [5,6]. However, with the introduction of many new oral hypoglycemic agents, such as the glucagon-like peptide-1 receptor (GLP-1R) agonist and dipeptidyl peptidase-4 (DPP-4) inhibitors, in addition to the conventional metformin mono-therapy and increased use of insulin, the nutritional control of diabetes seemed to have lost its relevance [7–9]. ⁎ Corresponding author at: Department of Nephrology, The First Hospital of Wuhan, Wuhan Qiaokou District Liji North Road No. 215, 430030 Wuhan, China. Tel.: + 86 13871101268. E-mail address: [email protected] (X.-L. Huang).

Furthermore, the expectation of strict adherence to dietary control, exercise regimen and other lifestyle changes is not as realistic as the drug treatment in many patients. Besides, the evidence for dietary modifications in glycemic control, though compelling, is often based on short term studies, mostly one year or less [10,11]. Clinical trials with a longer follow-up period are often necessary to confirm the positive impact of lifestyle changes in reducing the CVD in type 2 diabetes patients. Recently, the Look AHEAD study examined the effects of an intensive lifestyle intervention (with diet modification and physical activity) over a period of four years in a large cohort of overweight and obese individuals with type 2 diabetes [12]. The results indicate that patients who underwent intensive lifestyle intervention as opposed to diabetes support and education (control), had better glycemic control, blood pressure, high density lipoprotein-cholesterol (HDL-C) and triglyceride levels, thus lowering their cardiovascular disease risk. Though, the maximum benefits were seen at one year, the intervention group had greater improvements over the control group even at 4 years. In contrast, a long term behavioral intervention program, providing a regular, personalized exercise prescription did not improve glycemic control in sedentary, insulin treated type 2 patients during a 2 year intervention period [13], indicating that a more strictly supervised exercise training with personal coaching may be required to maximize the adherence and to increase the physical activity status. Similarly, a randomized clinical trial to determine the effect of case management in

http://dx.doi.org/10.1016/j.ejim.2015.11.016 0953-6205/© 2015 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.

38

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

the control of diabetes in type 2 diabetes patients did not yield any significant difference between the case management and the control groups [14]. There was no difference in the mean HbA1c level, or low density lipoprotein-cholesterol (LDL-C) or blood pressure (BP) among the groups, suggesting cautious use of resources on personalized interventions and case management in the treatment of type 2 diabetes. Though exercise has been shown to have a positive effect on cardiovascular health, the effect of exercise on BP reduction in type 2 diabetic patients is inconsistent. Recent findings suggest that though there was no reduction in BP, a modest reduction in the HbA1c levels (0.2%) was observed in the exercise group [15], further substantiating the protective effect of exercise on glycemic control. Reports elsewhere further corroborate the effectiveness of exercise intervention strategies in promoting physical activity and improving HbA1c and cardiovascular risk profile [16]. The above study also revealed that counseling alone, though effective in achieving the recommended level of physical activity (PA), was unsuccessful in minimizing the cardiovascular risk, suggesting the need for a larger volume of PA in high-risk patients. Conversely, Kirk et al. [17] reported that counseling improved glycemic control as well as the status of cardiovascular risk factors in type 2 diabetic patients. In reports elsewhere, regular drug-counseling through pharmacist care program significantly reduced the various CVD risk factors, including stroke [18]. The aforementioned studies along with various other reports indicate that structured reinforcement with diabetes health education, counseling, physical activity programs, and nutritional control can control the risk of CVD in patients with type 2 diabetes [19–26]. The aim of the present meta-analysis is to compare the outcomes of intensive exercise, dietary regimens, and comprehensive lifestyle interventions and its significance on clinical markers of cardiovascular disease in patients with type 2 diabetes. 2. Methods 2.1. Selection criteria We performed a literature search of PubMed Central and MEDLINE, the Cochrane Central Register of Controlled Trials (CENTRAL), EMBASE, and Google Scholar databases (until July 15, 2014) using the terms “diabetes, cardiovascular risk, lifestyle, health education, dietary, exercise/physical activities, and behavior intervention”. Prospective, randomized controlled trials comparing lifestyle interventions with the “usual care” control group were included in the current metaanalysis. Only articles in English were considered for analysis. We excluded studies that did not recruit patients with type 2 diabetes or those with no mention of lifestyle program/education relating to the patient dietary and exercise behavior/physical activities. 2.2. Study selection and data extraction Studies identified by the search strategy were hand-selected and data were extracted by two independent reviewers. Where there was uncertainty regarding eligibility, a third reviewer was consulted. The following information was extracted from studies that met the inclusion criteria: the name of the first author, year of publication, study design, number of participants in each treatment group, participants' age and gender, diagnostic criteria, intervention regiment for the study/control group, and results. The minimum number of participants was 30 in each study, while the minimum follow-up period was 6 months. 2.3. Outcomes The primary outcome was the reduction in the risk factors for cardiovascular disease, such as, body mass index, HbA1c, blood pressure, and the level of cholesterol.

2.4. Quality assessment The included studies were assessed for the risk of bias using the ‘Risk of Bias’ assessment tool, Review Manager 5.1. Recommendations for judging the risk of bias were provided in Chapter 8 of the Cochrane Handbook for Systematic Reviews Interventions [27]. 2.5. Statistical analysis The reduction in the risk factors for cardiovascular disease, such as, body mass index (BMI), HbA1c, systolic blood pressure (SBP), diastolic blood pressure (DBP), high-density lipoprotein-c (HDL-c), and lowdensity lipoprotein-c (LDL-c). BMI, HbA1c, SBP, DBP, HDL-c, and LDL-c were compared between participants having an intensive lifestyle intervention (intervention group) and the conventional intervention (control group). The intervention group was divided into 3 subgroups according to the type of intervention and data were analyzed separately according to 3 different programs, intensive physical activity (PA), intensive dietary (DIET), and education program (EDU). For each outcome measure, the standardized difference (Std. diff) in means with corresponding 95% confidence intervals was calculated for each individual and study. Std. diff in means of b0 indicates that the intervention was favored, showing a greater decrease in change of outcome than the control group. Conversely, Std. diff in means of N 0 indicates that the control group was favored, which means that the intervention group had less decrease in change of outcome than the control. Std. diff in means = 0 indicates that the intervention and control groups had a similar change in outcomes. Additionally, for changes in the levels of HDL-c, the effect was identified as the Std. diff in means of N 0, indicating that the intervention group was favored with greater increase in the change of HDL-c than the control; whereas Std. diff in means of b 0 indicates that the control group was favored, with greater increase in the change of HDL-c than the intervention group. A χ2 based test of homogeneity was performed using Cochran's Q statistic and I2. I2 illustrates the percentage of the total variability in effect estimates among trials resulting from heterogeneity rather than chance. Random effect models of analysis were used if heterogeneity was detected (I2 N 50%). Otherwise, fixed effect models were used. For each outcome measure, the standardized difference in means with corresponding 95% confidence intervals was calculated for each individual and study. A two-sided P value of b 0.05 indicated statistical significance for one comparison group over the other. Sensitivity analysis was carried out for the outcome HbA1c using the leave one-out approach. A funnel plot, the fail-safe N (which indicates whether the observed significance is spurious or not), and Egger's test (which detects whether the observed studies is asymmetry or not) were used to assess possible publication bias in EDU subgroup only. Since five or fewer studies are insufficient to detect funnel plot asymmetry [28], publication bias for the other two subgroups was not assessed. All analyses were performed using Comprehensive Meta-analysis statistical software, version 2.0 (Biostat, Englewood, NJ, USA). 3. Results 3.1. Literature search After the removal of duplicates, a total of 235 studies were identified through the database search, of which 167 studies were excluded due to lack of relevancy. After full text reviewing of 68 articles, we excluded 51 studies. The reasons for elimination being, eight articles were from the same trial that was included in the meta-analysis; while eight articles did not contain the outcome of interest; fourteen articles adopted intervention programs that were shorter than 6 months; and twelve articles had other interventions involved, for example, pharmacological intervention; while two articles only studied one gender in the trial; and

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

six articles were 3 or 4 arm RCTs. One study [29] was initially included, however, it later had to be excluded because of the variability in BMI, LDL-c and HDL-c analyses, as compared with other included studies. The inclusion of this study would have skewed the meta-analysis by increasing the discrepancy in both sensitivity and publication bias analyses. 17 studies were included in the final meta-analysis. The flowchart for the selection of trials is shown in Fig. 1. 3.2. Study characteristics The number of patients in each group per study ranged from 23 to 2575 (Table 1). The mean (SD) age was similar across the studies (range, 50 [12.4] to 67.3 [19]) and between the intervention and control groups (Table 1). The proportion of male patients ranged from 29% to 98%, and the length of the study ranged from 6 months to 8 years (Table 1). There were 10 studies in the EDU group, 5 in the PA group, and 3 in the DIET group. The Look AHEAD Research Group [12] study used intervention containing both intensive physical activity and dietary programs; therefore, the data in this study were included and analyzed separately in PA and DIET groups. The changes in outcome measures from the baseline for the studies included in this metaanalysis are summarized in the Supplementary Table S1. Fig. 2 represents the assessed outcomes of the included studies. For each trial, the risk of bias was detailed in the ‘risk of bias’ summary

39

(Fig. 2A). In addition, an overall assessment of risk of bias was presented in Fig. 2B. All studies utilized random sequence generation, thus avoiding any selection bias. However, only 65% (11/17) studies had allocation concealment, leaving the rest of the 35% of the studies with unclear risk of bias. Blinding of participants and personnel was not followed through in 15/17 (88%) studies, thus resulting in a higher risk of performance bias. In addition, some of these study results are further biased by not blinding the outcome assessment (detection bias; 6/17 studies; 38%). Only 47% of the studies included had intentto-treat analysis (Fig. 2A, B). 3.3. Clinical outcome measures 3.3.1. Body mass index (BMI) Twelve of the included studies reported change from baseline in BMI. There was heterogeneity for this outcome across studies in the PA group, but it was homogeneous in both DIET and EDU groups (PA: Q statistic = 35.77, I2 = 94.41%, P b 0.001; DIET: Q statistic = 5.54, I2 = 0%, P = 0.477; EDU: Q statistic = 5.20, I2 = 0%, P = 0.635). Therefore, the random effect analysis was applied to the PA group, and the fixed effect analyses in the DIET and EDU groups, respectively. The standardized differences in the means of change from baseline in BMI showed similar changes between the intervention and control groups for all three subgroups (PA: standardized difference in means, −0.77;

Fig. 1. A flowchart showing the selection of studies.

40

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

Table 1 Baseline characteristics of studies included in the meta-analysis. Study

Type of study

Totala

Intensive physical activity and dietary program The Look AHEAD RCT 5145 Research Group [12] Intensive physical activity program Dobrosielski DA [15] RCT

140

Group

Description of groups

Patientb

Age, years (mean ± SD)

Sex/males (n, %)

Length of program

Intervention Control

Intensive lifestyle intervention (ILI) Diabetes support and education (DSE)

2570 2575

58.6 ± 6.8 58.9 ± 6.9

1044, 40.7% 1038, 40.4%

4 years

70 70 32

57 ± 6 56 ± 6 54.3 ± 1.4

41, 59% 40, 57% 20, 62%

26 weeks 2 years

29 303

51.3 ± 1.8 NA

18, 62% NA

12 months

303 35

NA 57.6 ± 7.9d

NA 35, 50%d

12 months

Control

Supervised exercise General advice about physical activity Regular, structured and personalized exercise prescription Usual care Supervised training plus structured exercise counseling Standard care (counseling only) Physical activity counseling with a trained research assistant Standard exercise leaflet

Intervention Control Intervention Control

Intensive individualized dietary intervention Usual care Intensive individualized dietary intervention Usual care

45 48 40 46

56.6 ± 8.8 58.4 ± 8.8 NA NA

17,38% 21, 44% NA NA

Intervention

Pharmacist care package (about diabetes, its treatment and associated cardiovascular risk factors) Usual care (seen by a pharmacist only at the beginning of the study and then after 12

23

66.4 ± 12.7

10, 43.5%

23

66.8 ± 10.2

13, 56.5%

109

53.1 ± 12.4

40, 36.7%

12 months

181 51

50 ± 12.4 63.2 ± 9.5

50, 27.6% 30, 58.8%

9 months

54 132

61.7 ± 11.2 NA

28, 51.9% 38, 29%

6 months

133 94

NA 62.6 ± 10.3

46, 34.8% 71, 75.5%

18 months

95 300 300

60.3 ± 10.7 66.06 ± 8 67.28 ± 19

72, 75.8% 139, 46.2% 152, 50.7%

2 years

1017

58.5 ± 6.9

549, 54.0%

8 years

1016 90 88 123 123 56

58.6 ± 7.0 55.0 ± 9.0 56.0 ± 10.2 61 ± 10 61 ± 11 62.0 (35 to 80)c

538, 53.0% 44, 48.9% 34, 38.6% 121, 98% 117, 95% 27, 48.2%

56

61.0 (43 to 78)c

34, 60.7%

Wisse W [13]

RCT

74

Intervention Control Intervention

Balducci S [16]

RCT

606

Control Intervention

Kirk A [17]

RCT

70

Control Intervention

Intensive dietary program Coppell KJ [22] RCT

93

Uusitupa M [30]

RCT

86

Education program Ali M [25]

RCT

48

Control

Mohamed H [23]

RCT

290

Intervention

Chan CW [18]

RCT

120

Control Intervention

Sevick MA [26]

RCT

265

Control Intervention

Crasto W [20]

RCT

189

Control Intervention

Salinero-Fort MA [21]

RCT

608

Control Intervention Control

Sone H [35]

RCT

2033

Intervention

Ko GT [19]

RCT

180

Krein SL [14]

RCT

246

Trento M [24]

RCT

112

Control Intervention Control Intervention Control Intervention Control

months) Structured group counseling sessions in addition to educational toolkit Diabetes educational toolkit only Pharmacist care package (about drug-counseling and cardiovascular risks) Without pharmacist interventions Group counseling sessions over the diabetes self-management regimen Monthly contact with the study team Education medication optimization group (a structured self-management education program) Usual care by their own clinician Precede Health Promotion Education (PHPE) Conventional Health Promotion Education (CHPE) Lifestyle intervention (education on lifestyle modification) Conventional treatment Structured health program Usual care Care by a signed case manager Usual care Systemic education program in group and individual sessions Usual care

35

6 months 15 months

12 months

1 year 18 months 4 years

Abbreviations: RCT, randomized control trial; NA, not available. a Total number of patients enrolled. b Number of patients in each group. c Median (range: min. to max.) d Represented data for the entire group.

95% CI, − 1.94 to 0.39, P = 0.194; DIET: standardized difference in means, −0.20; 95% CI, −0.63 to 0.22, P = 0.354; EDU: standardized difference in means, −0.06; 95% CI, −0.14 to 0.01, P = 0.082; Figs. 3, 4, 5). 3.3.2. HbA1c All 17 studies provided data for the change in HbA1c from the baseline. Owing to the evidence of heterogeneity among the studies, a random effect analysis was applied to PA and EDU groups and a fixed

effect analysis to the DIET group (PA: Q statistic = 90.60, I2 = 95.59%, P b 0.001; DIET: Q statistic = 1.96, I2 = 0%, P = 0.375; EDU: Q statistic = 76.48 I2 = 88.23%, P b 0.001). The standardized difference in means of change from baseline in HbA1c significantly favored the intervention group, as compared with the control groups, in the PA and DIET intervention (PA: standardized difference in means, − 1.02; 95% CI, − 1.80 to − 0.23, P = 0.011; DIET: standardized difference in means, − 0.30; 95% CI, − 0.35 to − 0.24, P b 0.001) (Figs. 3 and 4).

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

41

Fig. 2. The quality assessment for each included study was summarized in (A) ‘risk of bias summary’ or (B) as percentages across all included studies in the ‘risk of bias graph’.

However, no difference between the intervention and control groups in the EDU patient group was seen (standardized difference in means, −0.08; 95% CI, −0.3 to 0.15, P = 0.509; Fig. 5).

3.3.3. Blood pressure (SBP and DBP) Sixteen of the 17 studies reported values for systolic (SBP) and diastolic (DBP) blood pressure at the baseline and following intervention.

42

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

Fig. 3. Forest plot comparing the intervention and control groups. Changes from the baseline (top to bottom) in BMI, HbA1c, SBP, DBP, LDL-c, and HDL-c for patients in the intensive physical activity program are given. Abbreviations: CI, confidence interval; std diff, standardized difference.

Additionally, in the EDU group, Krein study [14] DBP was not considered for analysis due to the observed 95% CI (−2.0 to 4.0) being not symmetry relative to the mean of 0.85. Hence, only 9 studies were retained for analysis. There was heterogeneity across the studies for both SBP and DBP in PA and EDU groups, but homogeneity in the DIET group (PA: Q statistic = 18.24, I2 = 83.56%, P b 0.001 for SBP; Q statistic = 49.55, I2 = 93.95%, P b 0.001 for DBP; DIET: Q statistic = 0.41, I2 = 0%, P = 0.817 for SBP; Q statistic = 2.94, I2 = 31.09%, P = 0.230 for DBP; EDU: Q statistic = 42.62, I2 = 78.88%, P b 0.001 for SBP; Q statistic = 24.06, I2 = 66.75%, P = 0.002 for DBP). Consequently, a random effect analysis was used for PA and EDU groups and a fixed analysis for the DIET group. The standardized difference in means of change from baseline in both SBP and DBP significantly favored the intervention in the DIET group (SBP: standardized difference in means, − 0.19: 95% CI, − 0.25 to − 0.14, P b 0.001; DBP: standardized difference in means: − 0.08, 95% CI = − 0.13 to −0.02, P = 0.005; Fig. 4). However, there was no significance in means of change from baseline in both SBP and DBP in the EDU group (SBP: standardized difference in means, −0.17:

95% CI, −0.34 to 0, P = 0.056; DBP: standardized difference in means, −0.08: 95% CI, −0.23 to 0.06, P = 0.267; Fig. 5), suggesting the limited impact of EDU alone strategy. In PA group, the intervention showed a significant decrease in DBP, but no significant change in SBP (SBP: standardized difference in means, −0.05: 95% CI, −0.46 to 0.35, P = 0.792; DBP: standardized difference in means, −0.76: 95% CI, −1.45 to −0.07, P = 0.030; Fig. 3). 3.3.4. LDL-c and HDL-c Of the 17 studies, 13 reported change from the baseline in LDL-c and 15 reported change from the baseline in HDL-c. In LDL-c, four studies belonged to the PA program, two in the DIET program, and eight in the EDU program. A random effect analysis was applied to LDL-c, as there was evidence of heterogeneity among the studies in all 3 groups. In HDL-c, four studies belonged to the PA program, three to the DIET program, and nine to the EDU program. Additionally, in EDU group, the Crasto study [20] was not considered, as the observed 95% CI (− 0.07, 0.04) was not symmetry relative to the mean of 0.01. Hence,

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

43

Fig. 4. Forest plot comparing intervention and control groups, for changes from the baseline in (top to bottom) BMI, HbA1c, SBP, DBP, LDL-c and HDL-c for those in the intensive dietary program are shown. Abbreviations: CI, confidence interval; std diff, standardized difference.

only 8 studies were retained for analysis in the EDU group. A fixed effect analysis was applied to HDL-c in the DIET group, but a random effect analysis was used in the PA and EDU groups (PA: Q statistic = 17.10, I2 = 82.46%, P = 0.001 for LDL-c; Q statistic = 13.54, I2 = 77.84%, P = 0.004 for HDL-c; DIET: Q statistic = 3.17, I2 = 68.42%, P = 0.075 for LDL-c; Q statistic = 0.97, I2 = 0%, P = 0.616 for HDL-c; EDU: Q statistic = 36.53, I2 = 80.84%, P b 0.001 for LDL-c; Q statistic = 25.42, I2 = 72.46%, P = 0.001 for HDL-c). According to the analysis, standardized difference in the mean change from the baseline favored the intervention group for HDL-c in the DIET program, suggesting that the intervention group had a higher increase in the HDL-c level, as compared with the control group (standardized difference in means, 0.026; 95% CI, 0.21 to 0.32, P b 0.001). But, there was no significant difference in the PA and EDU groups. The difference in LDL-c between the intervention and control groups was not significant in all 3 groups, either. 3.4. Sensitivity analysis Sensitivity analysis was performed in which the results were analyzed using leave-one-out approach, with each study removed in turn for HbA1c results for those in the PA, DIET, and the EDU programs, separately. The direction and magnitude of the combined estimates did not markedly change with the exclusion of individual studies, indicating that no one study dominated the findings (Fig. 6).

3.5. Publication bias Funnel plot analysis for publication bias was performed for the HbA1c education program only. It was found that the combined effect size yielded Z values of −0.53 for HbA1c (P = 0.597). The Eggar's test demonstrated marked evidence of symmetry, indicating that there was no significant evidence of publication bias (one tailed P = 0.428 for HbA1c; Fig. 7). The publication bias was not done in the physical activity and dietary programs, as at least five studies were required for funnel plot analysis.

4. Discussion The significance of intensive lifestyle intervention for weight loss and reduction in cardiovascular morbidity and mortality among type 2 diabetic or pre-diabetic patients have been demonstrated [30–32]. Though, comprehensive lifestyle interventions effectively decrease the incidence of type 2 diabetes in high-risk patients, its effect in patients who already have type 2 diabetes are variable among trials [33,34]. Furthermore, lowering blood glucose through lifestyle modification may improve cardio-metabolic risk factors, but may not affect CVD event rates [32]. Sone et al. have reported that lifestyle interventions had limited effects on typical cardiovascular risk factors, but had a significant effect on stroke incidence in patients with established type 2 diabetes

44

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

45

Fig. 6. Sensitivity analysis using the leave-one-out approach, showing the influence of each individual study on the pooled estimate of HbA1c in patients in the intensive physical activity, dietary, and education programs are given. Abbreviations: CI, confidence interval; std diff, standardized difference.

[35]. The present meta-analysis evaluated the effects of intensive physical activity (PA), diet control (DIET) and education and counseling (EDU) interventions, in reducing the risk of cardiovascular disease in patients with type 2 diabetes. Our analysis revealed that weight loss, as indicated by the reduction in BMI, was similar between the intervention and the control groups, irrespective of the nature of intervention (PA, DIET or EDU). Interestingly, the intensive PA and DIET intervention yielded a greater reduction in glycated hemoglobin (HbA1c) than the control group, while no significant difference was observed in the EDU group (Figs. 3, 4, 5). In terms of blood pressure, significant improvement in both systolic and diastolic blood pressure (SBP and DBP) was observed in the DIET intervention group, while intensive PA was able to reduce the DBP significantly, with little or no effect on the SBP levels. No significant difference in blood pressure was seen in the EDU intervention vs. the control group. Regarding serum lipids, the HDL-c level in the DIET group was significantly higher than the non-intervention control group, thereby lowering the risk of CVD in the DIET intervention group. However, there was no significant difference in LDL-c levels between the intervention and control groups in all three intervention subtypes. Our current analysis is in agreement with other studies where even small to moderate weight loss and a change in fatty acid composition of diet achieved a good metabolic control in the majority of obese, middle-aged patients with recent type 2 diabetes [30]. Likewise, the LOADD (Lifestyle Over and Above Drugs in Diabetes) study also have

demonstrated a significant difference in HbA1c, body weight, BMI, and waist circumference, after adjusting for age, sex and baseline measurements in type 2 diabetic patients [22]. Participants in the intensive lifestyle intervention in the Look AHEAD study (diet modification, calorie goal of 1200–1800 based on initial weight; and physical activity, at least 175 min/week) achieved and maintained significant weight loss and had better glycemic control, blood pressure, HDL-c, and triglycerides, thus reducing their CVD risk [11]. Our results also reveal that PA intervention had a significant impact on HbA1c and diastolic blood pressure (Fig. 3). However, unlike an Italian Diabetes and Exercise study (IDES) [16], it did not have any effect on other risk factors, including BMI, SBP, LDL-c, and HDL-c. Though the outcome of HDL-c favored intervention group with intensive physical activity, it was not statistically different from the control (Fig. 3). While two of the studies favored the intervention group [12,16], two other studies had no significant outcome [13,17]. The presence of heterogeneity along with small sample size, might have led to no significance in the pooled overall data. Besides, physical activity could be influenced by many factors, including motivation, reinforcement, intensity, and adherence to the PA program. In addition, reports elsewhere indicate that a personalized exercise prescription program provided by physical therapist every 6 weeks for 2 years did not alter the physical activity levels or improve glycemic control in sedentary, insulin treated type 2 diabetes patients [13]. Counseling alone, though successful in achieving the recommended amount of physical activity, was of limited efficacy on cardiovascular

Fig. 5. Forest plot comparing intervention and control groups, for changes from the baseline in (top to bottom) BMI, HbA1c, SBP, DBP, LDL-c and HDL-c for patients in the education program are shown. Abbreviations: CI, confidence interval; std diff, standardized difference.

46

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47

Fig. 7. Funnel plots for the publication bias in HbA1c in patients undergoing intensive education program. Abbreviations: std diff, standardized difference.

risk factors [16]. A strictly supervised exercise training program may not be always feasible on a long-term basis, however, it may improve the outcome and reduce the CVD risk in these patients. Contrary to the popular belief in patient education and counseling, no difference in glycated hemoglobin, blood pressure or triglyceride levels was seen between the intervention and control groups in the EDU alone group in our analysis (Fig. 5). This is in agreement with studies by Baker et al. where the absence of intensive individualized advice or ‘information only’ was very ineffective [36]. Further, results from case management studies also indicate that no significant improvement in outcomes was observed [14,16]. Though, some studies have reported positive effects of case management in the reduction of CVD risk factors [18,25]. It should be noted that multiple factors, like the ethnicity of patients, culture, and organizational structure of the program are critical factors in determining the effectiveness of an intervention. The existing evidence for multiple lifestyle interventions in reducing the CVD risk may be variable. Our meta-analysis included only RCTs in adults with type 2 diabetes. Clinical trials in children and adolescents were excluded to maintain the homogeneity of the studies. Care was also taken to choose only intervention programs longer than 6 months, thus avoiding the influence of short-term effect on the measured outcomes. Short-term interventions may not represent real life situations, where patients may fail to follow strict adherence and where compliance cannot be enforced beyond the intervention period. However, there are several limitations to the present analysis. The authors fully acknowledge that the difference in the formats of lifestyle intervention is diverse, and there was a high risk of performance bias in the studies included. Although data were analyzed according to the intervention subtypes, the delivery of the education programs was various, including group sessions, self-management, and pharmacist care, which may have a major impact on the cardiovascular risk. Moreover, the intensity of the exercise and dietary regimens might also affect the clinical outcomes. In addition, the wide range of variability in age, sex, and duration of follow-up included in this meta-analysis could have confounded our findings. Ideally, we would have used subgroup analysis to address this problem. However, due to the limitation of extracting sufficient data for each confounder, it was not possible to perform this type of analysis. Another factor which could affect the outcomes of this study is that the Look AHEAD study constituted the majority of the patients (5145 subjects) and had a longer follow-up period (4 years), thus reducing the overall impact of other studies included in the analysis. However, the leave-one-out sensitivity analysis indicates

that none of the studies dominated the current findings. Finally, though interventions involving pharmacological methods were excluded from this meta-analysis, the use of medications in routine medical care in most studies may have made the relative benefit of the intensive lifestyle interventions more difficult to demonstrate. In summary, the current analysis indicates that in patients with type 2 diabetes, nutritional intervention (DIET group) had a better impact on the quality of life by reducing their risk of CVD, as opposed to EDU or PA alone. DIET group showed a significant improvement in all major cardiovascular risk factors associated with diabetes, such as the HbA1c, SBP, DBP, and HDL-c, except for LDL-c and BMI. The PA intervention showed a significant impact on HbA1c and DBP, but not on the other risk factors (BMI, SBP, LDL-c, and HDL-c). The EDU program, on the other hand, did not show any difference in all the risk factors assessed in the study. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ejim.2015.11.016. Conflict of interests The authors state that they have no conflicts of interest. Acknowledgments None. References [1] Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med 2006;3, e442. [2] Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 2004;27:1047–53. [3] Puavilai G, Chanprasertyotin S, Sriphrapradaeng A. Diagnostic criteria for diabetes mellitus and other categories of glucose intolerance: 1997 criteria by the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus (ADA), 1998 WHO consultation criteria, and 1985 WHO criteria. World Health Organization, 44. Diabetes research and clinical practice; 1999. p. 21–6. [4] Hippisley-Cox J, Coupland C. Development and validation of risk prediction equations to estimate future risk of heart failure in patients with diabetes: a prospective cohort study. BMJ Open 2015;5, e008503. [5] Chen L, Pei JH, Kuang J, Chen HM, Chen Z, Li ZW, et al. Effect of lifestyle intervention in patients with type 2 diabetes: a meta-analysis. Metab Clin Exp 2015;64:338–47. [6] Rock CL, Flatt SW, Pakiz B, Taylor KS, Leone AF, Brelje K, et al. Weight loss, glycemic control, and cardiovascular disease risk factors in response to differential diet composition in a weight loss program in type 2 diabetes: a randomized controlled trial. Diabetes Care 2014;37:1573–80.

X.-L. Huang et al. / European Journal of Internal Medicine 27 (2016) 37–47 [7] Florez H, Temprosa MG, Orchard TJ, Mather KJ, Marcovina SM, Barrett-Connor E, et al. Metabolic syndrome components and their response to lifestyle and metformin interventions are associated with differences in diabetes risk in persons with impaired glucose tolerance. Diabetes Obes Metab 2014;16:326–33. [8] Yoo S, Chin SO, Lee SA, Koh G. Factors associated with glycemic variability in patients with type 2 diabetes: focus on oral hypoglycemic agents and cardiovascular risk factors. Endocrinol Metab (Seoul, Korea) 2015;30(3):352–60. [9] Stranges P, Khanderia U. Diabetes and cardiovascular disease: focus on glucagon-like peptide-1 based therapies. Ther Adv Drug Saf 2012;3:185–201. [10] Mann JI, De Leeuw I, Hermansen K, Karamanos B, Karlstrom B, Katsilambros N, et al. Evidence-based nutritional approaches to the treatment and prevention of diabetes mellitus. Nutr Metab Cardiovasc Dis 2004;14:373–94. [11] Look ARG, Pi-Sunyer X, Blackburn G, Brancati FL, Bray GA, Bright R, et al. Reduction in weight and cardiovascular disease risk factors in individuals with type 2 diabetes: one-year results of the look AHEAD trial. Diabetes Care 2007;30:1374–83. [12] Look ARG, Wing RR. Long-term effects of a lifestyle intervention on weight and cardiovascular risk factors in individuals with type 2 diabetes mellitus: four-year results of the Look AHEAD trial. Arch Intern Med 2010;170:1566–75. [13] Wisse W, Boer Rookhuizen M, de Kruif MD, van Rossum J, Jordans I, ten Cate H, et al. Prescription of physical activity is not sufficient to change sedentary behavior and improve glycemic control in type 2 diabetes patients. Diabetes Res Clin Pract 2010;88:e10–3. [14] Krein SL, Klamerus ML, Vijan S, Lee JL, Fitzgerald JT, Pawlow A, et al. Case management for patients with poorly controlled diabetes: a randomized trial. Am J Med 2004;116:732–9. [15] Dobrosielski DA, Gibbs BB, Ouyang P, Bonekamp S, Clark JM, Wang NY, et al. Effect of exercise on blood pressure in type 2 diabetes: a randomized controlled trial. J Gen Intern Med 2012;27:1453–9. [16] Balducci S, Zanuso S, Nicolucci A, De Feo P, Cavallo S, Cardelli P, et al. Effect of an intensive exercise intervention strategy on modifiable cardiovascular risk factors in subjects with type 2 diabetes mellitus: a randomized controlled trial: the Italian Diabetes and Exercise Study (IDES). Arch Intern Med 2010;170:1794–803. [17] Kirk A, Mutrie N, MacIntyre P, Fisher M. Effects of a 12-month physical activity counselling intervention on glycaemic control and on the status of cardiovascular risk factors in people with type 2 diabetes. Diabetologia 2004;47:821–32. [18] Chan CW, Siu SC, Wong CK, Lee VW. A pharmacist care program: positive impact on cardiac risk in patients with type 2 diabetes. J Cardiovasc Pharmacol Ther 2012;17: 57–64. [19] Ko GT, Li JK, Kan EC, Lo MK. Effects of a structured health education programme by a diabetic education nurse on cardiovascular risk factors in Chinese type 2 diabetic patients: a 1-year prospective randomized study. Diabet Med 2004;21:1274–9. [20] Crasto W, Jarvis J, Khunti K, Skinner TC, Gray LJ, Brela J, et al. Multifactorial intervention in individuals with type 2 diabetes and microalbuminuria: the Microalbuminuria Education and Medication Optimisation (MEMO) study. Diabetes Res Clin Pract 2011;93: 328–36. [21] Salinero-Fort MA, Carrillo-de Santa Pau E, Arrieta-Blanco FJ, Abanades-Herranz JC, Martin-Madrazo C, Rodes-Soldevila B, et al. Effectiveness of PRECEDE model for health education on changes and level of control of HbA1c, blood pressure, lipids, and body mass index in patients with type 2 diabetes mellitus. BMC Public Health 2011;11:267.

47

[22] Coppell KJ, Kataoka M, Williams SM, Chisholm AW, Vorgers SM, Mann JI. Nutritional intervention in patients with type 2 diabetes who are hyperglycaemic despite optimised drug treatment—Lifestyle Over and Above Drugs in Diabetes (LOADD) study: randomised controlled trial. BMJ 2010;341:c3337. [23] Mohamed H, Al-Lenjawi B, Amuna P, Zotor F, Elmahdi H. Culturally sensitive patientcentred educational programme for self-management of type 2 diabetes: a randomized controlled trial. Prim Care Diabetes 2013;7:199–206. [24] Trento M, Passera P, Bajardi M, Tomalino M, Grassi G, Borgo E, et al. Lifestyle intervention by group care prevents deterioration of type II diabetes: a 4-year randomized controlled clinical trial. Diabetologia 2002;45:1231–9. [25] Ali M, Schifano F, Robinson P, Phillips G, Doherty L, Melnick P, et al. Impact of community pharmacy diabetes monitoring and education programme on diabetes management: a randomized controlled study. Diabet Med 2012;29(9):e326–33. [26] Sevick MA, Korytkowski M, Stone RA, Piraino B, Ren D, Sereika S, et al. Biophysiologic outcomes of the Enhancing Adherence in Type 2 Diabetes (ENHANCE) trial. J Acad Nutr Diet 2012;112:1147–57. [27] Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343: d5928. [28] Sutton AJ, Duval SJ, Tweedie RL, Abrams KR, Jones DR. Empirical assessment of effect of publication bias on meta-analyses. BMJ 2000;320:1574–7. [29] Kadoglou NP, Iliadis F, Angelopoulou N, Perrea D, Ampatzidis G, Liapis CD, et al. The anti-inflammatory effects of exercise training in patients with type 2 diabetes mellitus. Eur J Cardiovasc Prev Rehabil 2007;14:837–43. [30] Uusitupa M, Laitinen J, Siitonen O, Vanninen E, Pyorala K. The maintenance of improved metabolic control after intensified diet therapy in recent type 2 diabetes. Diabetes Res Clin Pract 1993;19:227–38. [31] Norris SL, Zhang X, Avenell A, Gregg E, Bowman B, Schmid CH, et al. Long-term effectiveness of weight-loss interventions in adults with pre-diabetes: a review. Am J Prev Med 2005;28:126–39. [32] Koivula RW, Tornberg AB, Franks PW. Exercise and diabetes-related cardiovascular disease: systematic review of published evidence from observational studies and clinical trials. Curr Diab Rep 2013;13:372–80. [33] Schellenberg ES, Dryden DM, Vandermeer B, Ha C, Korownyk C. Lifestyle interventions for patients with and at risk for type 2 diabetes: a systematic review and meta-analysis. Ann Intern Med 2013;159:543–51. [34] Angermayr L, Melchart D, Linde K. Multifactorial lifestyle interventions in the primary and secondary prevention of cardiovascular disease and type 2 diabetes mellitus—a systematic review of randomized controlled trials. Ann Behav Med 2010;40:49–64. [35] Sone H, Tanaka S, Iimuro S, Tanaka S, Oida K, Yamasaki Y, et al. Long-term lifestyle intervention lowers the incidence of stroke in Japanese patients with type 2 diabetes: a nationwide multicentre randomised controlled trial (the Japan Diabetes Complications Study). Diabetologia 2010;53:419–28. [36] Baker MK, Simpson K, Lloyd B, Bauman AE, Singh MA. Behavioral strategies in diabetes prevention programs: a systematic review of randomized controlled trials. Diabetes Res Clin Pract 2011;91:1–12.