impaired glucose tolerance
With the global escalation of diabetes mellitus it is predicted that 200 million individuals worldwide will have diabetes by 2010(Reference Bastaki1) and 300 million by 2025(Reference King, Aubert and Herman2). The cumulative lifetime risk of developing diabetes for individuals born in the USA in 2000 is 38·5% for women and 32·8% for males(Reference Narayan, Hoskin and Kozak3). However, there is impelling evidence from many studies that subjects with impaired fasting glucose or impaired glucose tolerance (IGT) can be delayed by lifestyle modification(Reference Makrilakis and Katsilambros4).
The Nurses' Study, in which 84 941 female nurses were followed from 1980 to 1996 was one of the first cohorts to provide evidence on the importance of lifestyle and development of type 2 diabetes(Reference Hu, Manson and Stamper5). A low-risk group for development of diabetes was defined as one with five variables: BMI <25 kg/m2; diet high in cereal fibre and PUFA and low in trans-fatty acids with a low glycaemic load; moderate-to-vigorous physical activity (PA) for ≥30 min/d; not smoking; consumption of on average ≥15 ml spirits or ≥125 ml wine daily. Furthermore, overweight or obesity was found to be the single most important predictor of diabetes. Lack of PA, unhealthy diet, smoking, and abstinence from alcohol were all shown to be associated with an increased risk of diabetes even after adjusting for BMI.
It is generally accepted that β-cell dysfunction and insulin resistance are major factors involved in development of diabetes. It is proposed that subjects at risk of diabetes have β-cell dysfunction before they develop IGT, while insulin resistance can be explained by the presence of obesity(Reference Chiasson and Rabasa-Lhoret6). Furthermore, insulin resistance in turn accelerates the progression to diabetes in those subjects with a propensity to β-cell failure(Reference Chiasson and Rabasa-Lhoret6). Data on lifestyle modification interventions suggest that reducing insulin resistance protects and preserves β-cell function.
The primary objective of the present review is to assess lifestyle interventions with a dietary component that have aimed at the prevention of type 2 diabetes in order to make practical and sustainable recommendations for primary prevention.
Relationship between lifestyle risk factors and the development of diabetes
Overweight–obesity has been recognised as one of the strongest risk factors for the development of diabetes. Prediabetes (IGT and impaired fasting glucose) dramatically increases the risk of diabetes. Consequently, weight loss for individuals with prediabetes may delay or prevent the progression to type 2 diabetes(Reference Norris, Zhang and Avenell7). Even modest weight loss has been associated with a reduced risk of diabetes.
Numerous studies have shown a decrease in insulin sensitivity with diets high in fat while some have demonstrated positive associations between intake of saturated fat and hyperinsulinaemia(Reference Hu, van and Liu8). The Nurses' Health Study has additionally reported a positive association between trans-fats (trans-fatty acids) and development of diabetes(Reference Hu, Manson and Stamper5). Furthermore, the Nurses' Health Study and the Health Professional's Follow-up Study has found an inverse relationship between cereal fibre and development of diabetes(Reference Bazzano, Serdula and Liu9). Hence, whole-grain intake has been recommended as part of lifestyle modifications.
Numerous studies and specifically the Health Professional's Follow-up Study(Reference Manson, Nathan and Krolewski10) have shown that individuals with a physically-active lifestyle are less likely to develop IGT or type 2 diabetes mellitus than those who are sedentary. Results suggest that exercise training increases tissue sensitivity to insulin(Reference Bazzano, Serdula and Liu9).
Numerous studies have shown that cigarette smoking may increase the risk of diabetes(Reference Bazzano, Serdula and Liu9). Smokers tend to have a higher HbA1c concentration. In the Nurses' Study it was shown that women who smoke twenty-five or more cigarettes per d have a 42% greater risk of developing diabetes(Reference Rimm, Manson and Stampfer11).
Lifestyle modification of diet, physical activity and weight status
The European Prospective Investigation into Cancer was a prospective study that has evaluated the association between the incidence of diabetes in 25 155 adults aged 40–79 years and the achievement of five behavioural goals: BMI <25 kg/m2; PA >4 h/week; total fat <30% energy intake; SFA <10% energy intake; fibre intake ≥15 g/4184 kJ. Follow up was at a mean of 4·6 years(Reference Simmons, Harding and Jakes12). Their results show that only 20% of participants met three or more goals and diabetes incidence was inversely related to the number of goals achieved (P<0·0001). None of the participants who met all five goals developed diabetes. The researchers estimate that if the entire population were able to meet one or more goals the total incidence of diabetes would be predicted to fall by 20%.
The present objective was to evaluate intervention research programmes that were aimed at preventing type 2 diabetes mellitus. Inclusion criteria were: peer-reviewed studies from 1975 to 2008; a sample of at least fifty subjects; a healthy eating and/or PA component; prevention of diabetes as a primary goal. Generally, the participants were in a high-risk category for the development of diabetes or had prediabetes. Studies of subjects with diabetes were not included nor were interventions aimed only at weight loss or CVD. Studies with pharmacological agents were only included if there was a lifestyle modification arm. Outcomes were evaluated at two points in time (pre- and post-intervention) in terms of knowledge, behaviour change (dietary intake and behaviour) and clinical improvement, which included non-invasive measures (weight, blood pressure, BMI, body fat, waist circumference or waist:hip ratio and physiological and biochemical measures).
A search was undertaken on Medline with the terms ‘diabetes’, ‘intervention’ and ‘diet*’ and/or ‘physical activity’. The primary diabetic journals were also individually searched on line: Diabetes; Diabetes Care; Diabetic Medicine; Clinical Diabetes; Diabetologia; The Diabetes Educator; Diabetes Research and Clinical Practice; Diabetes and Metabolism; Diabetes, Obesity and Metabolism. The resulting 634 articles were then individually scanned to ensure that they complied with the search criteria, resulting in 113 publications on interventions. The final group comprised thirty-four interventions complying with the search and inclusion criteria (Table 1). Of these interventions four are large cohort studies, four are in ethnic minorities, eight are in schools and the rest (seventeen) involve adults (in an overweight, high-risk or IGT category). The search also yielded six reviews on diabetes prevention.
RCT, randomised controlled trial; PA, physical activity; IGT, impaired glucose tolerance; DM, diabetes mellitus; IR, insulin resistance; IS, insulin sensitivity; BP, blood pressure; GOAL, Good Ageing in Lahti Region; SLIM, Study on lifestyle-intervention and impaired glucose tolerance Maastricht; OGTT, oral glucose tolerance test.
* ‘Wellness’ in Swahili.
† Area under the curve of insulin level:area under curve of the glucose level during the OGTT.
Existing reviews on interventions for diabetes prevention
A few reviews have evaluated the effectiveness of lifestyle modification and prevention of diabetes. It has been found that Orlistat, sibutramine and metformin appear to be beneficial in the treatment of obesity in adults(Reference Avenell, Broom and Brown13). PA appears to improve weight loss when added to a dietary regimen. Low-fat diets together with increased PA are associated with prevention of type 2 diabetes and hypertension.
In a review of school-based diabetes prevention programmes(Reference Gittelsohn and Kumar14) most of the six programmes were aimed at minority groups in the USA and Canada(Reference Saksvig, Gittelsohn and Harris15–Reference Rosenbaum, Nonans and Weil20). All these studies were considered to be pilot studies and were done in one or two schools and few had control groups or randomly-assigned subjects to treatment. Two of the programmes were successful in improving insulin sensitivity, while others improved dietary or social behaviours. Only two programmes involved the community and/or environment.
A meta-analysis has evaluated nine randomised controlled trials of lifestyle education to patients at high risk of developing diabetes(Reference Yamaoka and Tango21). Lifestyle education intervention was found to reduce 2 h plasma glucose by 0·84 mmol/l compared with the control group. The 1-year incidence of diabetes was found to be reduced by about 50% (relative risk 0·55).
In a review of recent trials designed to test whether lifestyle, pharmacological interventions or both reduce the incidence of type 2 diabetes(Reference Kanaya and Narayan22) it was reported that four major studies (Diabetes Prevention Program(23–Reference Hamman, Wing and Edelstein26) in the USA; Da Qing IGT and Diabetes Study(Reference Pan, Li and Hu27, Reference Li, Hu and Yang28) in China; Finnish Diabetes Prevention Study(Reference Eriksson, Lindstrom and Valle29–Reference Lindstrom, Peltonen and Eriksson34) in Finland; Malmo Feasibility Study(Reference Eriksson and Lingarde35) in Sweden) of lifestyle intervention have found a clear benefit for diet and exercise intervention compared with usual care.
A recent review of the literature on the influence of lifestyle and diet in the development of diabetes has provided evidence that type 2 diabetes mellitus, which comprises 95% of diabetes cases, is a preventable disease(Reference Bazzano, Serdula and Liu9). It concludes that in order to reduce the burden of this devastating disease, prevention programmes must target not only the affected individuals but also families, workplaces, schools and communities.
Sixteen evidence-based research intervention studies over the last 10 years that have successfully targeted modifiable risk factors for obesity in children and adolescents has been summarised(Reference Nwobu and Johnson36). These interventions have shown the effectiveness of health education, dietary modification, a decrease in sedentary behaviour and incorporation of moderate-to-vigorous physical activity in schools in reducing unhealthy behaviours among youth.
Findings from trials of pharmacological agents such as metformin, acarbose and troglitazone have been encouraging(Reference Irons, Mazzolini and Greene37); however, the American Diabetes Association recommend that drug therapy should not be used routinely to prevent diabetes until there is more information on the cost-effectiveness of such interventions(38).
Diabetes prevention interventions
Large cohort studies
Four studies have clearly illustrated the positive health benefits of lifestyle modification in the prevention of type 2 diabetes. These studies were randomised controlled trials that included high-risk individuals, had large samples and a long follow up over time.
Diabetes Prevention Program
The Diabetes Prevention Program Research Group have reported on a 3-year lifestyle intervention programme(23, 24). A total of 3234 participants without diabetes (mean age 51 years) were randomly assigned to a placebo, metformin (850 mg twice daily) or lifestyle programme group. Initially, there was also a troglitazone therapy arm. However, because of concern with its liver toxicity it was discontinued. The lifestyle group had goals of ≥7% weight loss and ≥150 min PA/week. Goal achievement was assessed at the end of a sixteen-session core curriculum over the first 6 months. Thereafter, the participants attended at least once every 2 months for individual or group sessions.
After an average follow up of 2·8 years the incidence of diabetes was found to be 11·0, 7·8 and 4·8 cases per 100 person-years in the placebo, metformin and lifestyle groups respectively. The incidence was reduced by 58 (95% CI 48, 66) % in the lifestyle group and by 31 (95% CI 17, 43) % in the metformin group. Furthermore, it was projected that incorporation of the Diabetes Prevention Program intervention into clinical practice in five developed countries would lead to an increase in diabetes-free years of life, improvements in life expectancy and cost savings(Reference Palmer, Roze and Valentine25).
Analysis of another arm of the intervention (n 1079) has shown that weight loss is the dominant predictor of reduced diabetes incidence; for every 1 kg weight loss there was a 16% reduction in risk, adjusted for changes in activity and diet(Reference Hamman, Wing and Edelstein26).
Da Qing IGT and Diabetes Study
In 1986 a total of 110 660 adults from thirty-three health clinics in the city of Da Qing in China were screened for IGT and type 2 diabetes(Reference Pan, Li and Hu27, Reference Li, Hu and Yang28). Of these individuals 577 subjects who were classified as having IGT were randomly assigned to a clinical trial or a control group over a period of 6 years. There were three treatment groups: diet; PA; diet+PA. At 6 years the cumulative incidence of diabetes (%) was found to be 67·7; 43·8, 41·1 and 46 in the control group, diet group, PA group and diet+PA groups respectively. Furthermore, it was reported that the diet, PA and diet+PA groups were associated with 31, 46, and 42% reductions in risk of developing diabetes respectively. Both insulin resistance and β-cell function were found to be predictors of diabetes in Chinese with IGT.
Multiple Risk Factor Intervention Trial
A 2-year randomised controlled trial was undertaken in 11 827 men aged 35–57 years (without diabetes or IGT) and they were followed for 6–7 years(Reference Smith, Bracha and Svendsen39). The intervention group received nutritional counselling aimed at reducing saturated fat and cholesterol intake. Smokers participated in a cessation programme. Men who were ≥115% expected weight were required to follow a moderate-PA programme and to reduce dietary energy intake. Results indicate a significantly higher hazard ratio for diabetes among smokers (1·26) v. non-smokers (0·82). Their results imply that in non-smokers a lifestyle intervention programme comprising nutrition counselling to produce reductions in weight and serum cholesterol reduces the risk for diabetes.
The Finnish Diabetes Prevention Study
An intensive diet and PA programme measured the effectiveness of preventing or delaying type 2 diabetes in 523 overweight subjects with IGT aged 40–64 years(Reference Eriksson, Lindstrom and Valle29–Reference Lindstrom, Peltonen and Eriksson34). An intensive individualised programme was presented by nutritionists during individual sessions. The objectives were: to decrease fat to <30% energy intake; SFA <10% energy intake; minimum of 15 g fibre/4184 kJ; to have ≥4 h PA/week; >5% weight reduction. At 1 year most of the clinical variables were found to be different when compared with the control group. At 1 and 2 years 43·4 and 41·8% of participants were found to have maintained a weight loss of ≥5 kg and this loss was maintained at 2 years. Most of the beneficial changes in risk factors were also maintained at 2 years.
After 3 years individuals with a low-fat intake and high-fibre intake were found to have lost more weight compared with those consuming a high-fat low-fibre diet (3·1 kg v. 0·7 kg). Hazard ratios for diabetes incidence at 4·1 years were found to be highest for fat intake and saturated fat intake(Reference Lindstrom, Peltonen and Eriksson34).
Kahnawake Schools Diabetes Prevention Project
This project is a 3-year primary diabetes prevention programme targeted at children from grade 1 through to grade 6 (6–12 years of age) in a Mohawk community near Montreal in Canada(Reference Paradis, Levesque and Macaulay16,40–Reference Cargo, Levesque and Macaulay42). A health education programme was introduced into the curriculum comprising ten 45 min lessons each year in each grade. The lessons covered healthy diet and PA and teachers were trained to deliver the programme themselves by year 3. The intervention focused not only on the school but included activities for the parents and the entire community and was planned taking the local culture into perspective. Results of an 8-year evaluation were reported to be disappointing when compared with control schools; however, the findings can provide many lessons about such interventions(Reference Paradis, Levesque and Macaulay16). Despite showing some improvements in the short term, these benefits (i.e. PA, fitness and television watching) were not found to be maintained in the long term. The authors ascribe this outcome to some extent to the fact that children are exposed to 20 000–40 000 food commercials annually, many of which promote foods high in sugar and fat by aggressive marketing strategies. They believe that it may be possible to have more beneficial outcomes with a longer intervention of greater intensity coupled with community participation(Reference Paradis, Levesque and Macaulay16).
NEEMA (‘wellness’ in Swahili) is a 14-week diabetes prevention programme tailored for African-American children with a before and after component and no control schools(Reference Shaw-Perry, Horner and Trevino18). The intervention comprises four focus areas: a curriculum (diet and physical activity); home component; food service at school; after school activities (fitness clubs). The programme is based on the Bienestar programme(Reference Trevino, Pugh and Hernandez43, Reference Trevino, Yin and Hernandez44), which had been implemented with some successful outcomes in Mexican-American children. Follow up of the 14-week programme has shown decreases in fasting capillary glucose and percentage body fat and increased fitness laps. However, the shortness of the evaluation programme and lack of control groups make it difficult to predict long-term sustainability(Reference Shaw-Perry, Horner and Trevino18).
Zuni Pueblo of New Mexico
A 4-year high school-based intervention programme was tested in the Zuni Pueblo of New Mexico and compared with an Anglo (English-speaking Caucasians) group having no intervention(Reference Ritenbaugh, Teufel-Shone and Aickin17, Reference Teufel and Ritenbaugh45). The programme included an educational component targeting decreased consumption of sugared beverages, knowledge of diabetes risk factors and a youth-orientated fitness centre. A healthier school meal service was also achieved with increased fruit and vegetables and decreased sugar and fat in the school meals. Over the intervention period sugared beverage consumption was found to have decreased and use of the fitness centre created at the school increased. No significant changes were found in BMI or between the Anglo and the Zuni students. However, reductions were found in fasting and 30 min plasma insulin levels in the Zuni group, implying that a lifestyle intervention may suppress markers of type 2 diabetes risk. It should also be noted that sugar beverages were replaced by ‘diet’ beverages in the vending machines at the school and that water was commonly available from water cooler bottles. The study illustrates the positive outcomes from a simple measure such as replacing sugar beverages with water.
Junior High School (Hispanic Middle School)
Eighth grade students in a predominantly Hispanic public school in New York City participated in this randomised controlled trial(Reference Rosenbaum, Nonans and Weil20). A 3-month intervention comprising fourteen 45 min classroom sessions on diet and PA were provided by the investigators. The intervention also comprised a PA programme offered three times weekly. Participation in the intervention was associated with reductions in body fat, insulin resistance and circulating concentrations of IL-6 and C-reactive protein, suggesting that a short-term lifestyle programme can be beneficial in decreasing risk factors of type 2 diabetes(Reference Rosenbaum, Nonans and Weil20). However, the long-term effects of this intervention are not known and should be evaluated before recommendations can be made.
Sandy Lake Diabetes Prevention Programme
This programme was undertaken in 3rd–5th grade Objiway-Cree students (age range 7–14 years) from a Native North American reserve in Ontario, Canada over a 1-year period(Reference Saksvig, Gittelsohn and Harris15). The programme comprised: a curriculum component of sixteen weekly 45 min teacher-led lessons; a family component; a peer component; an environmental component; a healthy school lunch programme. The intention was to improve students' knowledge, self-efficacy and skills relating to diet and PA. The curriculum was based on the Child and Adolescent Trial for Cardiovascular Health Study(Reference Webber, Osganian and Feldman46) and the Kahnawake programmes(Reference Paradis, Levesque and Macaulay16). It was shown that mean BMI and percentage body fat increased between 1998 and 1999. However, increases were found in dietary intention, dietary preference, knowledge and dietary self-efficacy between baseline and follow up. Unfortunately, this study did not include a control group in the study design and the PA component was not strong.
Bienestar (Well-Being) Health Programme: Diabetes Risk-Factor Prevention Pilot Program
A pilot study was undertaken in 102 low-income 4th grade American-Mexican children in San Antonio, TX, USA over a period of 7 months(Reference Trevino, Pugh and Hernandez43). All children were overweight and had a family history of diabetes. The intervention was aimed at decreasing dietary fat intake and reducing body fat and involved a curriculum of twenty-eight lessons. Additionally, there was a parental component and the school cafeteria complied by selling healthy food options. Some promising results were reported but with insufficient evidence to make recommendations. Subsequently, the researchers repeated the study using a control group and including a larger sample (intervention group 619, control group 602)(Reference Trevino, Yin and Hernandez44). They included a PA component and a family component. In total fifty health sessions were provided. Modest improvements were found, including decreased fasting glucose levels, increased fitness scores and increased dietary fibre intake. Positive results are thought to be the outcome of culturally-appropriate materials used, multiple systems of delivery and frequency of contacts. It is unknown whether outcomes will be sustained over time.
An action research programme was undertaken in five inner-city secondary schools serving a predominantly South Asian population in Leicester, UK(Reference Khunti, Stone and Bankart47). Baseline dietary and PA data were collected from children aged 11–15 years, followed by focus groups with children and teachers to provide them with feedback on the baseline results and to identify ideas for interventions and to evaluate barriers to implementation. No preset intervention was planned and it was intended that the schools themselves would initiate and implement interventions based on their participation in the focus groups. At the end of the year the children and staff were re-evaluated. Suboptimal diet and poor activity habits were identified at baseline. Overall, these habits were found to have persisted at follow up, although a few limited positive lifestyle changes were identified.
Jump into Action
A 3-month study (Jump into Action) was undertaken in a school district with a predominantly Hispanic population on the Texas–Mexico border(Reference Holcomb, Lira and Kingery48). The study included two groups of teachers and their 5th grade students (n 1114). Teachers were trained to implement the intervention on thirteen campuses. The programme was designed by an interdisciplinary team to improve students' knowledge, self-efficacy and behaviours relating to diabetes prevention with a focus on diet and PA. Positive effects were observed for knowledge and self-efficacy gains and for healthy dietary and exercise-related behaviour changes from pre- to post-test. However, the long-term sustainability of the programme is unknown.
Summary on school studies
The school studies were all undertaken on indigenous or ethnic minorities. Unfortunately, none of the studies were undertaken in developing countries. Only two of eight studies were randomised controlled trials and sample sizes were generally <100 children. Overall, the intervention periods were short (<1 year) and not all included a longer period of follow up. Most school studies showed positive benefits in knowledge, attitudes and/or behaviour. However, few positive physical and/or clinical benefits were shown.
Studies with indigenous adult communities
Indigenous American Indian population in British Columbia
This community-based 2-year programme was targeted at the Indian population of rural Okanagan region in Canada(Reference Daniel, Green and Marion49). A single intervention community was matched to two comparison communities. The programme used a participatory approach and included strategies to improve the environment and to change behaviour. Cohorts (105 high-risk individuals) were tracked over 16 months. It was reported that the project yielded few changes in quantifiable outcomes, which was thought to be a result of the short planning and intervention phases and because the level of penetration was too limited(Reference Daniel, Green and Marion49).
Pilot study in Pima Indians
A lifestyle randomised clinical trial over 12 months was undertaken in Pima Indians(Reference Narayan, Hoskin and Kozak3). A group of ninety-five obese normo-glycaemic adults were randomly assigned to an action (intervention) group or a pride (control) group. In the intervention group participants were required to walk 10–12 h/month or the equivalent activity output. Participants were advised by a dietitian to reduce fat and alcohol intake and to increase their fibre intake in line with the American Diabetes Association recommendations at the time (macronutrient recommendations the same as the 2007 recommendations(38)). Subjects also participated in weekly meetings that covered a range of nutrition activities. In the control group participants met once monthly to discuss aspects of Pima culture and history. Participants also received printed materials on healthy eating and PA. After 12 months both groups were reported to have increased levels of PA. However, BMI, blood pressure, 2 h glucose and 2 h insulin were shown to have increased in the intervention group to a greater extent than in the control group.
A 2-year church-based programme lifestyle intervention to prevent diabetes was undertaken in Western Samoans in South Auckland, New Zealand(Reference Simmons, Fleming and Voyle50). The study involved two complete church congregations (intervention group n 78; control group n 144). The intervention was designed to be culturally appropriate and was provided by a nurse specialist and two Samoans, one of whom was trained to give diabetes education and the other to do PA. The intervention comprised individual and group activities, leaflets and a video. PA sessions were held once weekly in the first year and twice weekly in the second year. Weight was reported to have remained stable in the intervention group but to have increased in the control group. The intervention group was found to have a decrease in waist circumference and intake of key fatty foods compared with the control group. Both diabetes knowledge and proportion exercising regularly was reported to have increased in the intervention group.
Looma healthy lifestyle
This intervention was undertaken in the Looma Aborigine community (n 199) over 4 years in Western Australia(Reference Rowley, Su and Cincotta51). The intervention comprised a healthy lifestyle programme aimed at decreasing SFA and sugar intakes, increasing fruit and vegetable intakes and increasing PA. Intervention activities included store policy changes, health promotion activities, regular PA classes and nutrition education by a diabetes nurse educator. Results show an increase in the supply of fresh fruit, vegetables and carotenoids in the food supply at the local store over the period. A reduction was found in the prevalence of hypercholesterolaemia and mean plasma homocysteine concentrations of the participants.
Summary of indigenous minorities
Overall, there were few trials in this category. However, the studies undertaken all showed promise and some positive outcomes. Perhaps the most important lessons to be learned from these studies are the fact that they involved the community in planning and second they developed culturally-appropriate materials.
Studies in overweight and high-risk adults
Eat Well Live Well Programme
This peer-led programme was tailored to the participant's stage of change for individual dietary patterns(Reference Auslander, Haire-Joshu and Houston52). The sample comprised 294 overweight African-American women aged 25–55 years who were at risk of diabetes. These women were assigned to intervention and control groups. The peer educators were trained by a multi-disciplinary team. The manual-based programme comprised six group sessions and six individual sessions per peer educator integrated over a 3-month intervention phase. The primary focus of the intervention was to reduce dietary fat intake and increase low-fat dietary patterns by tailoring the intervention to stage of readiness. Reductions were found in fat intake among women in the intervention group and these reductions were maintained at 3-month follow up. Similarly, changed dietary patterns were reported, which were maintained(Reference Auslander, Haire-Joshu and Houston52).
Vasterbotten Intervention Programme
A total of 186 Swedish adults with IGT and obesity were randomly assigned to an intensive lifestyle group or a usual-care group(Reference Lindahl, Nilson and Jansson53). The intervention group (n 93) stayed at a wellness centre for 1 month, where they were subjected to an intensive lifestyle programme: behaviour modification counselling; a low-fat high-fibre diet; PA sessions. Results after 12 months were reported to show improvements in body weight, systolic blood pressure, waist:hip ratio, fasting plasma glucose, fibrinogen, O2 consumption and physical fitness. Despite the very promising outcomes this type of intervention will not be cost-effective or sustainable to implement for the general population.
Overweight individuals with a family history of diabetes
Overweight participants (n 154) with a family history of diabetes were randomly assigned to a 2-year intervention of diet, PA, diet+PA or no treatment(Reference Wing, Venditti and Jakicic54). The intervention comprised weekly group meetings in the first 12 months led by a multi-disciplinary team and refresher courses during the second 12 months. A dietitian and a behaviour therapist served on the team. It was found that only the diet+PA group maintained a weight loss from baseline to 2 years, with other variables showing no sustained loss over 2 years.
Good Ageing in Lahti Region Study
The aim of the study was to evaluate whether the results obtained in the Finnish Diabetes Prevention Study(Reference Eriksson, Lindstrom and Valle29) could be replicated in routine health care(Reference Absetz, Valve and Oldenburg55). The trial was undertaken in 352 middle-aged participants with a high risk for diabetes (i.e. family history and clinical risk). The intervention was based on social cognitive theory and comprised six group counselling sessions of 2 h each over a period of 1 year. The groups were led by public health nurses who had received training from a dietitian. Local sports officers introduced the groups to local sports facilities and guided one exercise session. All clinical and nutritional data were collected by the nurses. The intervention was delivered by the nurses as part of their existing schedule. Results were reported to be similar to those of the Finnish Diabetes Prevention Study, with 20% of participants meeting at least four of the five goals. The trial demonstrates that lifestyle counselling can be effective and feasible in real-world situations(Reference Absetz, Valve and Oldenburg55).
Lifestyle intervention in relatives of patients with diabetes in Sweden
Seventy-seven first-degree relatives of patients with diabetes who did not themselves have diabetes were randomised to diet or diet+PA interventions or a control group(Reference Brekke, Jansson and Lenner56). The diet group received group counselling from a dietitian (1–2 h) on reducing SFA, increasing n-3 fat intake, fruit and vegetable intake and foods with a high glycaemic index. Those patients in the diet+PA group were required to increase walking to ≥30 min four to five times weekly. Dietary changes were found to be significant at 1 year (P<0·05) and to a large extent were sustained at 2 years. At 1 year the diet only group were reported to show a reduction in LDL-cholesterol:HDL-cholesterol (P=0·028) while the diet+PA group were found to have decreased their body weight by 2·7% and increased HDL v. controls. At 2 years cholesterol levels were found to be reduced within the diet only group. Fasting insulin was found to be reduced in the diet+PA group when compared with the diet only group.
Adults with impaired glucose tolerance
Study on lifestyle intervention and impaired glucose tolerance Maastricht
The study was undertaken as a 3-year lifestyle intervention in subjects with IGT(Reference Mensink, Blaak and Corpeleijn57, Reference Mensink, Corpeleijn and Feskens58). Subjects were randomly assigned to an intervention group and control group and eighty-eight subjects completed 2 years of intervention. The intervention comprised dietary and PA components. Dietary recommendations based on the Dutch guidelines for a healthy diet were given at regular intervals by a dietitian. Participants were required to increase PA to ≥30 min five times weekly. The control group received oral and written information and no individual appointments. By 2 years fifty-five subjects in the intervention group were found to have reduced their BMI, waist circumference and SFA intake and improved their aerobic capacity. Plasma glucose levels at 2 h were found to have decreased from 8·7 mm to 8·0 mm in the intervention group and increased from 8·6 mm to 9·4 mm in the control group. Subjects who adhered to both the diet and the exercise components were found to show the largest reduction in 2 h plasma glucose levels.
Lifestyle intervention in Japanese males
A randomised controlled trial of lifestyle modification was undertaken in Japanese males with IGT identified at a health screening centre over a period of 4 years(Reference Kosaka, Noda and Kuzuya59). The intervention group (n 102) were required to lose weight or maintain weight at BMI <22 kg/m2 while the control group (n 356) were to reduce or remain at BMI <24 kg/m2. To achieve this objective the intervention group participants were screened and provided with dietary advice according to individual lifestyle every 2–3 months. Second, they were required to walk for 30–40 min/d. The 4-year incidence of diabetes was reported to be 9·3% in the control group and 3% in the intervention group and the reduction risk of diabetes was 67·4% (P<0·001).
Japanese male workers
Males (n 173) with a high risk (borderline diabetes fasting plasma glucose values) for type 2 diabetes were randomly assigned to an intervention group or control group over 1 year(Reference Watanabe, Yamaoka and Yokotsuka60). The intervention comprised a new dietary education programme based on information on the individual's dietary energy intake for breakfast, lunch and dinner, i.e. the actual dietary practices of the participants as obtained by a specific FFQ. The participant was further motivated to improve dietary practices and helped to realise his need for behaviour modification. A nutritionist encouraged the subject to recognise latent dietary problems and to set his own goals for improvement, i.e. the concept of self-management. The control group received the usual nutrition education component. The intervention group were found to have a decreased 2 h plasma glucose after 1 year, while the value increased in the control group; the percentage change was significant (P<0·001)(Reference Watanabe, Yamaoka and Yokotsuka60).
Malmo Feasibility Study
Forty-one subjects with early-stage type 2 diabetes and 181 subjects with IGT were selected from 6956 men aged 47–49 years in Malmo, Sweden to take part in a programme of diet and PA(Reference Eriksson and Lingarde35). At a mean follow up of 6 years the accumulated incidence of diabetes was reported to be 10·6%, with 50% of the patients with diabetes in remission. Blood pressure, lipids and hyperinsulinaemia were found to be reduced. Improvement in glucose tolerance was shown to be correlated with weight reduction and increased fitness (P<0·02). Body weight was reported to be reduced by 2·3–3·7% among participants whereas values were increased by 0·5–1·7% in subjects not undergoing the intervention and normal control subjects. The researchers conclude that long-term intervention is possible even on a large scale.
A reduced-fat diet v. usual diet was followed by adults (n 136) with glucose intolerance(Reference Swinburn, Metcalf and Ley61). For 1 year the intervention group were required to attend monthly small-group education sessions on reduced-fat eating. It was found that weight decreased at 1 year in the intervention group compared with the control group but this difference was no longer present at 5 years. The same scenario was found for glucose tolerance. However, the most-compliant 50% of patients in the intervention group were shown to have maintained lower fasting glucose levels at 5 years.
Diet and physical activity intervention in adults with impaired glucose tolerance
Adults with IGT (n 78) from Newcastle upon Tyne, UK were randomly assigned to either an intervention group or a control group(Reference Oldroyd, Unwin and White62). The intervention group received individual dietary advice from a dietitian and advice on PA from a physiotherapist. The dietitian used motivational interviewing and the stages of change theory as a model for behaviour change. Twelve individual appointments were scheduled over the 24-month period. A decrease in total fat intake and BMI in the intervention group v. the control groups was found at follow up, with whole-body insulin sensitivity improved after 12 months in the intervention group.
A group of sixty-four Japanese-American men and women with IGT followed the American Heart Association step 2 diet of fat <30% energy intake, SFA <7% energy intake and cholesterol <200 mg (intervention group) and step 1 (control) diets with three 1 h endurance exercise sessions per week(Reference Liao, Asberry and Shofer63). After 6 months the PA was home-based. At 24 months improvements were found in percentage body fat, BMI and subcutaneous fat. This type of intervention may be effective in retarding the development of type 2 diabetes.
Women in Melbourne
Women with IGT (n 200) were randomised to a trial of intensive v. routine dietary advice(Reference Wein, Beischer and Harris64). The annual incidence rate of diabetes for the two groups were reported to be 6·1% in the intervention group and 7·3% in the control group. Overall, a return to normal glucose tolerance was found in 44% of patients. It was shown by multivariate analysis that BMI, fasting and 2 h plasma glucose levels at trial entry are associated with an increased risk of diabetes.
Interventions using technology
This intervention was undertaken in Alberta, Canada at five workplaces(Reference Plotnikoff, McCargar and Wilson65). The intervention, which was conducted by email messages over 12 weeks, targeted the 1566 participants with one message per week. The messages were aimed at prevention of diabetes by promoting a healthy diet and increasing PA. Differences in self-reported healthy eating practices, balanced diet and enjoyment of meals as well as PA measures were reported. This programme may be a cost-effective sustainable programme for worksites where employees have access to emails. Tailoring messages to individuals and increasing message frequency and intervention duration may improve efficacy of intervention. Longer-term effectiveness needs to be evaluated.
Use of the internet
There is increasing recognition that because of the growing burden of diabetes globally it will become more important to develop self-management strategies in order to reach larger segments of the population. Some studies have found limited success in the use of specifically-designed internet programmes(Reference Tate, Wing and Winett66, Reference Tate, Jackvony and Wing67), while others found that long-term maintenance of a weight-loss programme is not as effective as ‘in-person’ therapist support(Reference Harvey-Berino, Pintuaro and Gold68). More recently, a 10-week interactive internet programme has been developed that is unique by virtue of being based on the goals and needs of the participants and as such can be considered to be a personalised programme(Reference McCoy, Couch and Duncan69). The exercise and diet plans are based on the individual's health risk appraisal. A total of 808 participants registered for the programme, with 683 completing at least one online health risk appraisal and 56% of the participants agreeing that the programme helped them achieve their goals. It was reported that the average participant used the homepage twenty-nine times, suggesting that the website was used with a high frequency. However, because there are no hard outcomes it is difficult to conclude whether the programme made a difference to clinical outcomes.
Summary of adult interventions
Overall, the majority of lifestyle interventions evaluated had very positive health benefits and most appeared to be sustainable. However, as with all interventions certain crucial aspects need to be clarified, specifically the duration of the intervention period and the period of follow up and re-evaluation. Mostly, the latter were done soon after the evaluation and the longer-term benefits are not known.
Studies on cost effectiveness of interventions
An assessment has been made in Australia, France, Germany, Switzerland and the UK of whether implementing active treatments (metformin or intensive lifestyle change) in the Diabetes Prevention Program would be cost effective(Reference Palmer, Roze and Valentine25). Assuming only within-trial effects, and costs of interventions, both metformin and intensive lifestyle change were found to improve life expectancy v. control.
Using a Markov Model(Reference Sonnenberg and Beck70) and, based on the probabilities, the incorporation of the Diabetes Prevention Program interventions into clinical practice in the five developed countries is projected to lead to an increase in diabetes-free years of life, improvements in life expectancy and either cost savings or minor cost increases compared with standard lifestyle advice in a population with IGT.
A systematic review of economic evaluations of preventive interventions in type 2 diabetes reported between January 1990 and May 2004 (twenty-three studies) has found that cost effectiveness seems possible for strict blood pressure control, with a high number of studies showing cost-effectiveness ratios ranging from cost saving to very low cost per life year gained(Reference Vijgen, Hoogendoorn and Baan71). Medication to reduce weight and hyperglycaemia simultaneously also seems to be cost effective compared with conventional therapies.
In another evaluation of the economic performance of ten nutrition interventions to prevent diabetes eight of the nutrition interventions were estimated to be highly cost-effective based on societal norms(Reference Dalziel and Sedal72). Population-based interventions were not found to necessarily provide better value for money than targeted interventions. In this exercise targeted interventions were found to be associated with more certain benefits.
Overall, there is a sizeable amount of evidence showing that lifestyle interventions under a number of conditions are beneficial in delaying or reducing type 2 diabetes. Health policymakers need to plan lifestyle modification interventions as part of routine primary health care and in settings such as the school and workplace in order to provide long-term solutions to the burgeoning diabetes epidemic. Furthermore, lifestyle modification in subjects without diabetes should be built on evidence-based principles such as those recommended by the American Diabetes Association(38):
1. programmes that emphasise lifestyle change including weight loss (7% body weight) and regular physical activity (150 min/week) with dietary strategies (e.g. reduced intake of fat) to reduce energy intake;
2. individuals at high risk should have a dietary fibre intake of ≥14 g/1000 kJ. Low-GI foods should thus be encouraged, based on their high fibre content.
The authors declare no conflict of interest.