EAL

DM: Prevention of Type 2 Diabetes (2007)

Citation:

Yamaoka K, Tango T. Efficacy of lifestyle education to prevent type 2 diabetes: a meta-analysis of randomized controlled trials. Diabetes Care 2005;28(11):2780-2786.

PubMed ID: 16249558

Study Design:

Meta-analysis or Systematic Review

Class:

M - Click here for explanation of classification scheme.

Quality Rating:

POSITIVE: See Quality Criteria Checklist below.

Research Purpose:

To evaluate, by use of a meta-analysis of randomized controlled trials, the efficacy of lifestyle education for preventing type 2 diabetes in individuals at high risk.

Inclusion Criteria:

Randomized controlled trials of at least six months duration
Subjects were adults who were diagnosed to be at high risk for type 2 diabetes: impaired glucose tolerance (IGT), impaired fasting glucose (IFG), and borderline (per Japan Diabetes Society)
Studies with lifestyle (combined diet and exercise) or solely dietary education interventions were selected

Exclusion Criteria:

None noted by author.

Description of Study Protocol:

Recruitment

Searches in Medline and ERIC (Educational Resources Information Center) databases from January 1966 to November 2004 were searched (restricted to the English language). Search terms were free text terms, MeSH (Medical subject heading), and Medline and medical index terms. 123 studies were identified.

Design

Meta-analysis of nine randomized controlled trials.

Blinding used (if applicable)

Blinding of subjects was not possible due to the nature of the interventions.

Intervention (if applicable)

Lifestyle (combined diet and exercise) or solely dietary education interventions were selected. Control interventions were described as usual exercise with or without general information about diet or general dietary advice about healthy food choices.

Statistical Analysis

Primary data from the nine studies were used for the meta-anlaysis. Overall estimates were examined using a fixed-effects model (general variance-based method), a random-effects model (DerSimonian-Laird method), and a Bayesian model with noninformative priors (Monte Carlo Markov chain). A χ²test was used to assess heterogeneity among trials.

The measure of effect size for two-hour plasma glucose is given by the difference between the lifestyle education intervention group and control group (Δ) for each individual study, which is equal to Δ_i-Δ_c, where Δ_iandΔ_c are mean differences from baseline to end point in two-hour plasma glucose between, respectively, the lifestyle education intervention and control groups. No study reported r, and because the true value was unknown, r was set at 0.05 based on past study data. A sensitivity analysis was performed, using r=0.3 and r=0.7. If the 95% CI was shown instead of the SD, SD was calculated using the formula SD=(√n) (95%CI_upper-95% CI_lower) ÷ 4, where n denotes sample size of a group. A cumulative meta-analysis was also performed to deteremine at which point to demonstrate a beneficial lifestyle education intervention effect. Subgroup analysis by intervention type and follow-up duration (<1 vs. ≥two year, was conducted as a sensitivity analysis. The selection bias was visually examined using the funnel plot.

Data Collection Summary:

Timing of Measurements

Two-hour plasma glucose at baseline and difference from baseline at one year.

Dependent Variables

Baseline two-hour plasma glucose (PG)
Two-hour PG difference from baseline at one year

Independent Variables

Lifestyle education: combined diet and exercise or diet alone.

Control Variables

None noted by author

Description of Actual Data Sample:

Initial N: Nine studies met criteria - all studies had a lifestyle intervention

Attrition (final N): Nine studies

Age: Mean age ranged from 39 to 58 years in the studies.

Ethnicity: Not noted by author

Other relevant demographics: Mean BMI ranged from 24 to 34.

Anthropometrics Follow-up ranged from six months to six years.

Location: Study locations were not noted by the author.

Summary of Results:

Characteristics of the nine randomized controlled trials.

Studies	Follow-up duration (years)	Diabetes incidence (r/n)		2 hour PG (mmol/L)
		Control	Intervention	Baseline (means ± SD)	Difference from baseline at 1 year (means ± SD) (n)
Pan (n=577)	6	90/133	58/130	C: 9.03±0.89 L: 9.11±0.93	C: 3.96±3.82 (133) L: 1.65±3.16 (130)
Wein (n=200)	4.25	7/100	6/100	C: 9.8±0.74 L: 9.9±0.74	C: 0.1±1.94 (96) L: -0.1±2.19 (97)
Lindahl (n=186)	1	NA	NA	C: 8.0±11.09 L: 7.5±6.99	C: -0.30±2.75 (93) L: -0.68±1.95 (93)
Oldroyd (n=78)	0.5	NA	NA	C: 9.2±0.9 L: 9.1±0.9	C: -0.5±1.8 (32) L: -0.7±1.9 (35)
Tuomilehto (n=522)	6	51/257	22/265	C: 8.9±1.5 L: 8.9±1.5	C: -0.3±2.2 (250) L: -0.9±1.9 (256)
Swinburn (n=176)	1	NA	NA	C: 7.5±2.4 D: 7.9±2.5	C: 0.74±2.76 (70) D: 0.01±2.68 (66)
Mensink (n=114)	3	NA	NA	C: 8.6±1.48 D: 8.8±2.06	C: 0.2±2.23 (55) L: -0.8±2.06 (47)
Watanabe (n=173)	1	6/87	3/86	C: 7.3±1.7 D: 8.2±1.5	C: 0.67±1.74 (77) D: -0.76±1.36 (79)
DPPRG (n=3,234)	2.8	313/1,082	155/1,079	C: 9.1±0.9 L: 9.1±0.9	NA

C, control

L, lifestyle education

D, solely diet education intervention

Other Findings

Lifestyle education intervention reduced two hour PG by 0.84 mmol/L (95% CI 0.39-1.29) compared with the control group. The one year incidence of diabetes was reduced by approximately 50% (RR 0.55, 95% CI 0.44-0.69) compared with the control group.

Funnel plots revealed no bias.

Author Conclusion:

Lifestyle education was effective for individuals at high-risk of type 2 diabetes in reducing two-hour plasma glucose and relative risk (RR) and may be a useful tool in preventing diabetes.

Funding Source:

Reviewer Comments:

Many of the studies included had a small number of participants.
The meta-analysis was confined to English-language articles.
Only randomized controlled trials were included.
Publication bias may be a concern as well as selection bias with a meta-analysis.
There was variation in lifestyle education.

Quality Criteria Checklist: Review Articles
Relevance Questions
	1.	Will the answer if true, have a direct bearing on the health of patients?	Yes
	1.	Will the answer if true, have a direct bearing on the health of patients?	Yes
	2.	Is the outcome or topic something that patients/clients/population groups would care about?	Yes
	2.	Is the outcome or topic something that patients/clients/population groups would care about?	Yes
	3.	Is the problem addressed in the review one that is relevant to dietetics practice?	Yes
	3.	Is the problem addressed in the review one that is relevant to dietetics practice?	Yes
	4.	Will the information, if true, require a change in practice?	Yes
	4.	Will the information, if true, require a change in practice?	Yes

Validity Questions
	1.	Was the question for the review clearly focused and appropriate?	Yes
	1.	Was the question for the review clearly focused and appropriate?	Yes
	2.	Was the search strategy used to locate relevant studies comprehensive? Were the databases searched and the search termsused described?	Yes
	2.	Was the search strategy used to locate relevant studies comprehensive? Were the databases searched and the search termsused described?	Yes
	3.	Were explicit methods used to select studies to include in the review? Were inclusion/exclusion criteria specified andappropriate? Wereselectionmethods unbiased?	Yes
	3.	Were explicit methods used to select studies to include in the review? Were inclusion/exclusion criteria specified andappropriate? Wereselectionmethods unbiased?	Yes
	4.	Was there an appraisal of the quality and validity of studies included in the review? Were appraisal methodsspecified,appropriate, andreproducible?	No
	4.	Was there an appraisal of the quality and validity of studies included in the review? Were appraisal methodsspecified,appropriate, andreproducible?	No
	5.	Were specific treatments/interventions/exposures described? Were treatments similar enough to be combined?	No
	5.	Were specific treatments/interventions/exposures described? Were treatments similar enough to be combined?	No
	6.	Was the outcome of interest clearly indicated? Were other potential harms and benefits considered?	Yes
	6.	Was the outcome of interest clearly indicated? Were other potential harms and benefits considered?	Yes
	7.	Were processes for data abstraction, synthesis, and analysis described? Were they applied consistently acrossstudies and groups? Was thereappropriate use of qualitative and/or quantitative synthesis? Was variation in findings among studies analyzed? Were heterogeneity issued considered? If data from studies were aggregated for meta-analysis, was the procedure described?	Yes
	7.	Were processes for data abstraction, synthesis, and analysis described? Were they applied consistently acrossstudies and groups? Was thereappropriate use of qualitative and/or quantitative synthesis? Was variation in findings among studies analyzed? Were heterogeneity issued considered? If data from studies were aggregated for meta-analysis, was the procedure described?	Yes
	8.	Are the results clearly presented in narrative and/or quantitative terms? If summary statistics are used, are levels ofsignificance and/or confidence intervals included?	Yes
	8.	Are the results clearly presented in narrative and/or quantitative terms? If summary statistics are used, are levels ofsignificance and/or confidence intervals included?	Yes
	9.	Are conclusions supported by results with biases and limitations taken into consideration? Are limitations ofthe review identified anddiscussed?	Yes
	9.	Are conclusions supported by results with biases and limitations taken into consideration? Are limitations ofthe review identified anddiscussed?	Yes
	10.	Was bias due to the review's funding or sponsorship unlikely?	Yes
	10.	Was bias due to the review's funding or sponsorship unlikely?	Yes