A Literature Review on Physical Activity Interventions for the Prevention of Type 2 Diabetes

Introduction

Type 2 diabetes mellitus (T2DM) has become one of the most pressing challenges for public health systems worldwide, with prevalence rising in parallel with population ageing, urbanisation and the growing burden of overweight and obesity (International Diabetes Federation, 2021). Because the condition develops gradually through a prolonged phase of impaired glucose regulation, it offers an unusually clear window for preventive action. During this pre-diabetic period, defined variously by impaired glucose tolerance or elevated fasting glucose, the trajectory towards overt disease is neither fixed nor inevitable, and a substantial body of evidence now suggests that it can be slowed or interrupted. Physical activity has attracted particular interest in this respect, since it improves insulin sensitivity, supports weight management and influences several intermediate cardiometabolic pathways simultaneously (Colberg et al., 2016). Unlike single-target pharmacological agents, activity acts on a network of physiological systems at once, which is part of what makes it attractive as a population-level lever. Understanding what the accumulated evidence says about physical activity interventions, and where that evidence remains incomplete, is therefore central to designing prevention strategies that are both effective and equitable.

This review aims to synthesise the existing literature on physical activity interventions for the primary prevention of T2DM among adults at elevated risk. Its scope is conceptual rather than clinical: it does not offer guidance for individual patients, nor does it present primary data. Instead, it draws together landmark prevention trials, categorises the main intervention types, appraises evidence on effectiveness and long-term sustainability, and considers questions of equity and real-world implementation. The discussion is organised thematically. It begins with the foundational trials that established proof of concept, moves through the varied forms interventions have taken, examines how durable their benefits appear to be, and then turns to a critical appraisal of the methodological and translational gaps that continue to shape the field. Throughout, the emphasis is on population-level prevention rather than the management of established disease, and on the conditions under which preventive benefit is realised rather than merely on whether it exists.

Review Approach

The literature considered here was assembled through a structured but non-systematic search of the major bibliographic databases, supplemented by hand-searching the reference lists of key trials and reviews. Priority was given to large randomised controlled trials, their long-term follow-ups, and to systematic reviews and meta-analyses that pool across such studies, since these together carry most of the field’s evidential weight. Observational cohort evidence and implementation reports were drawn upon selectively to illuminate dose-response relationships and real-world delivery. The selection process is summarised illustratively in Figure 1; the counts shown are indicative of the funnel typically encountered in this literature rather than the product of a formal systematic protocol, and are presented to make the logic of inclusion transparent rather than to claim exhaustive coverage.

Thematic Synthesis

Landmark prevention trials

The intellectual foundations of the field rest on a small number of large randomised trials conducted from the late 1990s onwards. The Da Qing study in China was among the earliest, assigning communities of participants with impaired glucose tolerance to diet, exercise or combined lifestyle groups and reporting substantial reductions in the incidence of diabetes over six years (Pan et al., 1997). Its long follow-up subsequently became one of the most influential features of the literature, as later analyses suggested that benefits persisted for decades and extended to cardiovascular events, microvascular complications and all-cause mortality (Li et al., 2008; Gong et al., 2019). The Finnish Diabetes Prevention Study (DPS) reinforced these findings in a European setting, demonstrating that a structured lifestyle programme targeting modest weight loss, dietary change and increased physical activity reduced progression to diabetes by roughly half among high-risk adults (Tuomilehto et al., 2001). Importantly, secondary analyses of the DPS indicated that increases in physical activity, and moderate-to-vigorous activity in particular, contributed to risk reduction partly independently of weight change, strengthening the case for activity as a distinct preventive lever rather than a mere adjunct to dietary restriction (Laaksonen et al., 2005). A post-intervention follow-up of the DPS cohort further showed that the reduction in diabetes incidence endured after active counselling had ceased (Lindström et al., 2006).

The United States Diabetes Prevention Program (DPP) provided the largest and perhaps most widely cited evidence base. It compared an intensive lifestyle intervention with metformin and placebo, and found that the lifestyle arm — built around at least 150 minutes of moderate activity per week and a target of 7 per cent body-weight loss — outperformed pharmacological treatment in reducing diabetes incidence (Knowler et al., 2002). The long-term Diabetes Prevention Program Outcomes Study later showed that the relative benefit, although attenuated, was maintained over ten years and beyond of follow-up (Knowler et al., 2009; Diabetes Prevention Program Research Group, 2015). Crucially, these findings were not confined to affluent Western or East Asian settings. The Indian Diabetes Prevention Programme demonstrated that lifestyle modification was effective in a South Asian population at markedly elevated risk, though the additional benefit of combining lifestyle change with metformin was limited (Ramachandran et al., 2006). A smaller Japanese trial among men with impaired glucose tolerance similarly reported a large relative reduction in progression under intensive lifestyle intervention (Kosaka et al., 2005). The recurrence of comparable relative risk reductions across trials that differed in their diagnostic thresholds, target weight-loss goals, activity prescriptions and cultural settings is notable, because it suggests that the preventive signal is not an artefact of any single protocol. It also implies that the mechanism is unlikely to be narrowly specific to one modality of activity or one dietary pattern, but rather reflects a broadly shared metabolic response to sustained increases in energy expenditure and modest reductions in adiposity.

Taken together, these trials established a consistent and reproducible message across diverse populations: structured lifestyle interventions incorporating physical activity can meaningfully delay or prevent the onset of T2DM. This convergence of findings across geographic and ethnic contexts is one of the more robust features of the preventive literature and lends the core proposition an unusual degree of external validity (Gong et al., 2019). Table 1 summarises the principal trials, their designs and their headline findings.

Table 1. Landmark randomised trials of lifestyle intervention for type 2 diabetes prevention.

Intervention types and mechanisms

Beyond these foundational trials, the literature encompasses a wide range of intervention designs, and it is useful to distinguish between them. The classic programmes were multicomponent lifestyle interventions in which physical activity was combined with dietary counselling and behavioural support, making it difficult to isolate the contribution of exercise alone (Tuomilehto et al., 2001; Knowler et al., 2002). A distinct strand of research has examined structured exercise interventions in more controlled forms, comparing modalities such as aerobic training, resistance training and combined regimens for their effects on glycaemic markers and insulin sensitivity (Umpierre et al., 2011). Aerobic training improves cardiorespiratory fitness and enhances insulin-mediated glucose uptake, while resistance training increases skeletal muscle mass, the principal site of insulin-stimulated glucose disposal; the two therefore act through partially complementary pathways, which helps explain why combined regimens frequently yield the largest metabolic effects. More recently, attention has turned to time-efficient formats, including high-intensity interval training, on the grounds that limited time is a commonly reported barrier to participation, although evidence for its preventive superiority remains less mature and rests largely on short-term physiological endpoints rather than diabetes incidence (Jelleyman et al., 2015).

A further category concerns the setting and delivery mechanism rather than the exercise prescription itself. Community-based and group-delivered programmes, often adapted from the DPP curriculum, have been widely disseminated in an attempt to reach populations at scale (Ali et al., 2012). Digital and mobile-health interventions represent a newer and rapidly expanding strand, using applications, wearable devices and remote coaching to prompt behaviour change and monitor activity (Van Rhoon et al., 2020). These varied approaches reflect a broader shift in the field, from tightly controlled efficacy trials towards pragmatic models designed for routine service delivery. Alongside structured exercise, a related literature has stressed the importance of reducing sedentary behaviour and increasing incidental daily movement, suggesting that prevention need not depend solely on formal exercise sessions but can also be pursued by re-engineering the sedentary patterns of everyday life (Wilmot et al., 2012).

The mechanistic pathways through which these interventions are thought to lower diabetes risk are summarised in Figure 2. Their diversity is analytically important, because it implies that different intervention formats may exert benefit through partly different routes — some primarily by reducing adiposity, others by improving insulin sensitivity or dampening low-grade systemic inflammation independently of weight loss.

This proliferation of formats carries an important methodological implication for how the evidence should be read. Interventions differ not only in the type of activity prescribed but in the behaviour-change techniques that underpin them, such as goal-setting, self-monitoring, feedback and social support, and the literature increasingly treats these techniques as active ingredients in their own right (Van Rhoon et al., 2020). Two programmes may prescribe identical volumes of activity yet achieve very different levels of engagement depending on how participants are supported to sustain the behaviour. The distinction between exercise as a physiological stimulus and physical activity promotion as a behavioural intervention is therefore analytically significant, and reviews that conflate the two risk obscuring the mechanisms through which preventive benefit is actually produced. It also complicates any attempt to specify a single optimal “dose” of intervention, since the effective ingredient may be the behavioural scaffolding as much as the activity itself.

Effectiveness and sustainability

Evidence on effectiveness is generally favourable but nuanced. Systematic reviews and meta-analyses have consistently concluded that lifestyle interventions reduce the incidence of T2DM among high-risk adults, with pooled effects broadly comparable to those seen in the original trials (Gong et al., 2019). Structured exercise, examined in isolation, produces measurable improvements in glycated haemoglobin and insulin sensitivity, with combined aerobic and resistance training often yielding the largest metabolic benefits (Umpierre et al., 2011). The dose of activity appears to matter, and several analyses point to a graded relationship in which greater volumes of moderate-to-vigorous activity confer greater protection, though the precise shape of this relationship, and whether it plateaus or continues to decline at high volumes, remains debated (Aune et al., 2015). Observational dose-response work is valuable here precisely because trials cannot easily randomise participants to widely differing long-term activity levels, but it is also more vulnerable to residual confounding, since more active individuals differ systematically from sedentary ones in ways that are difficult to measure fully. A related interpretive difficulty is that the trials typically report effects for a package of activity, diet and weight loss, whereas the observational literature isolates activity; the two bodies of evidence are therefore complementary rather than directly comparable, and each compensates for a limitation of the other.

The more difficult question concerns sustainability. Intensive interventions frequently produce strong short-term results, but adherence tends to decline once structured support is withdrawn, and behavioural gains may erode over time (Dunkley et al., 2014). This pattern is consistent with a wider understanding of health behaviour, in which the initiation of activity and its long-term maintenance are governed by partly distinct psychological and environmental processes; the motivational and self-regulatory resources that carry a person through the first weeks are not necessarily those that sustain a habit across years. Programmes that succeed in the enthusiastic early weeks may falter over the months and years during which sustained protection is most needed, and relatively few studies have been designed with follow-up periods long enough to capture this maintenance phase adequately. Translational studies of DPP-derived programmes delivered in real-world settings have generally reported smaller effects than the original efficacy trials, reflecting lower intensity, reduced contact time, greater participant heterogeneity and the practical constraints of routine delivery (Ali et al., 2012).

Nonetheless, the long-term follow-ups of Da Qing, the DPS and the DPP suggest that even time-limited interventions can leave a durable imprint on risk, a phenomenon sometimes described as a legacy effect (Li et al., 2008; Knowler et al., 2009; Lindström et al., 2006). The mechanisms of such legacy effects are not fully understood, but they may reflect a lasting reduction in cumulative glycaemic exposure during the intervention period, partial retention of behaviour change, or favourable shifts in body composition that persist beyond the active phase. Reconciling the apparent durability of benefit in trial cohorts with the fragility of behaviour change in everyday practice remains a central tension in the literature. It suggests that the goal of prevention need not be permanent behavioural transformation; even a temporary delay in progression may carry meaningful long-term value if it defers the onset of disease and its complications. At the same time, the gap between efficacy and real-world effectiveness cautions against extrapolating trial results directly to routine services without attention to the intensity and fidelity of delivery.

Equity and implementation

A growing body of work has questioned whether the benefits demonstrated in trials are distributed equitably across the population. T2DM disproportionately affects socioeconomically disadvantaged groups and several minority ethnic communities, yet these populations are often under-represented in trials and may face structural barriers to participation, including limited access to safe and affordable spaces for activity, competing time and caring demands, insecure work and the direct and indirect costs of attendance (Whittle et al., 2020). There is therefore a risk that interventions which are effective in general may nonetheless widen inequalities if they are more readily taken up and sustained by advantaged groups, a concern described in the wider public health literature as intervention-generated inequality (Adams et al., 2016). This is not a hypothetical worry: agentic interventions that rely heavily on individual motivation and resources tend to benefit those who already possess them, whereas more structural or “downstream-light” approaches may be more equitable.

Implementation research has increasingly foregrounded these considerations. National-scale programmes, such as the Healthier You: NHS Diabetes Prevention Programme in England, have provided valuable evidence on reach, uptake and retention, while also exposing the difficulty of maintaining fidelity and intensity at scale and of converting referral into sustained participation (Barron et al., 2018). Frameworks that assess reach, effectiveness, adoption, implementation and maintenance have been advocated to ensure that prevention is judged not only by efficacy but by its real-world footprint, since a modestly effective programme that reaches many people may achieve greater population impact than a highly effective one that reaches few (Glasgow et al., 1999). Digital interventions have been promoted partly as a means of extending access and reducing per-participant cost, yet they raise their own equity concerns, since a digital divide in device access, connectivity and digital literacy may exclude precisely those groups at greatest risk (Van Rhoon et al., 2020). The literature thus increasingly frames prevention as a systems challenge, in which the design of the intervention is inseparable from the context in which it is delivered, and in which environmental and policy measures may need to complement individual behaviour change. A further implication is that evaluation itself must broaden: outcomes such as uptake among the most deprived quintile, retention across the full course, and equity of effect become as important as the average reduction in incidence. Without such metrics, a programme may appear successful in aggregate while quietly under-serving the very groups in whom the diabetes burden is concentrated, and the promise of population-level prevention will be only partly fulfilled.

Critical Appraisal and Gaps

Although the evidence base is substantial, several methodological limitations temper its interpretation. The multicomponent design of the foundational trials, while ecologically sensible, makes it difficult to attribute effects specifically to physical activity as opposed to dietary change or the weight loss that both produce (Laaksonen et al., 2005). Measurement of physical activity is a further weakness, since much of the literature relies on self-reported data that are prone to recall and social-desirability bias and that tend to overestimate activity; the growing use of accelerometers and wearable devices offers more objective assessment but has not been applied consistently across studies, and differences in devices, wear protocols and cut-points hamper comparison (Wilmot et al., 2012). Heterogeneity in intervention content, intensity, duration, comparator condition and outcome definitions also complicates synthesis, and meta-analyses frequently report considerable statistical heterogeneity that limits confidence in pooled estimates and cautions against a single summary figure (Dunkley et al., 2014). Publication and reporting biases, and the tendency for early efficacy trials to be conducted by highly motivated teams, may further inflate apparent effects relative to what routine services can achieve.

There are also notable gaps in coverage. The bulk of the strongest evidence derives from a relatively small number of high-income and East and South Asian settings, and the transferability of findings to low- and middle-income countries — where the diabetes burden is rising fastest and health-system resources are most constrained — is not well established (International Diabetes Federation, 2021). Longer-term cost-effectiveness evidence, particularly for newer digital modalities, remains comparatively thin, and few studies adequately capture the maintenance phase over which behaviour change so often falters, meaning that decision-makers must often extrapolate from short-term surrogate outcomes (Van Rhoon et al., 2020). The equity dimension, though increasingly acknowledged, is still under-researched, with limited data on how interventions perform among the most disadvantaged groups and how they might be adapted, resourced or supplemented to serve them better (Whittle et al., 2020). Finally, comparatively little work has examined the optimal integration of physical activity promotion with broader environmental and policy measures, such as the design of the built environment, active travel infrastructure and neighbourhood walkability, that may shape activity at a population level and reach those who never enter a formal programme (Sallis et al., 2016). A methodological priority, in this respect, is the wider adoption of pragmatic trial designs and hybrid effectiveness-implementation studies that evaluate clinical outcomes and delivery processes simultaneously, since these are better suited than classical efficacy trials to informing decisions about scale-up. Equally, greater use of core outcome sets and standardised activity metrics would reduce the heterogeneity that currently frustrates quantitative synthesis. Addressing these gaps would strengthen both the internal validity and the external relevance of the preventive evidence base, and would help move the field from demonstrating efficacy towards guiding equitable, sustainable delivery.

Conclusion

The literature offers a broadly consistent and encouraging picture: physical activity interventions, particularly when embedded within structured lifestyle programmes, can substantially reduce the risk of progression to type 2 diabetes among high-risk adults. The landmark trials in China, Finland, the United States, India and Japan provided robust and reproducible proof of concept across markedly different populations, and their long-term follow-ups suggest that the benefits can be durable, extending in some cases beyond glycaemic outcomes to cardiovascular events and mortality. At the same time, the field has matured from a focus on efficacy under controlled conditions towards a more demanding preoccupation with real-world effectiveness, sustainability and equity. Here the evidence is more equivocal, as translational programmes tend to yield smaller effects, behaviour change is difficult to maintain once structured support is withdrawn, and benefits may be unevenly distributed across the population in ways that risk widening rather than narrowing health inequalities.

The most productive directions for future scholarship therefore lie less in re-establishing that physical activity prevents diabetes — a proposition now well supported — than in understanding how, for whom and under what conditions preventive benefit can be realised and sustained at scale. Greater methodological consistency, more objective and comparable measurement of activity, deliberate inclusion of disadvantaged and under-served populations, longer follow-up that captures the maintenance phase, and closer attention to the policy and environmental contexts of behaviour would all help to close the remaining gaps. Viewed as a whole, the literature affirms physical activity as a cornerstone of diabetes prevention while underlining that its public health promise depends on the quality of implementation as much as on the strength of the underlying physiology.

References

Adams, J., Mytton, O., White, M. and Monsivais, P. (2016) ‘Why are some population interventions for diet and obesity more equitable and effective than others? The role of individual agency’, PLoS Medicine, 13(4), e1001990.

Ali, M.K., Echouffo-Tcheugui, J.B. and Williamson, D.F. (2012) ‘How effective were lifestyle interventions in real-world settings that were modeled on the Diabetes Prevention Program?’, Health Affairs, 31(1), pp. 67–75.

Aune, D., Norat, T., Leitzmann, M., Tonstad, S. and Vatten, L.J. (2015) ‘Physical activity and the risk of type 2 diabetes: a systematic review and dose–response meta-analysis’, European Journal of Epidemiology, 30(7), pp. 529–542.

Barron, E., Clark, R., Hewings, R., Smith, J. and Valabhji, J. (2018) ‘Progress of the Healthier You: NHS Diabetes Prevention Programme’, Diabetic Medicine, 35(4), pp. 513–518.

Colberg, S.R., Sigal, R.J., Yardley, J.E., Riddell, M.C., Dunstan, D.W., Dempsey, P.C., Horton, E.S., Castorino, K. and Tate, D.F. (2016) ‘Physical activity/exercise and diabetes: a position statement of the American Diabetes Association’, Diabetes Care, 39(11), pp. 2065–2079.

Diabetes Prevention Program Research Group (2015) ‘Long-term effects of lifestyle intervention or metformin on diabetes development and microvascular complications over 15-year follow-up: the Diabetes Prevention Program Outcomes Study’, The Lancet Diabetes & Endocrinology, 3(11), pp. 866–875.

Dunkley, A.J., Bodicoat, D.H., Greaves, C.J., Russell, C., Yates, T., Davies, M.J. and Khunti, K. (2014) ‘Diabetes prevention in the real world: effectiveness of pragmatic lifestyle interventions’, Diabetes Care, 37(4), pp. 922–933.

Glasgow, R.E., Vogt, T.M. and Boles, S.M. (1999) ‘Evaluating the public health impact of health promotion interventions: the RE-AIM framework’, American Journal of Public Health, 89(9), pp. 1322–1327.

Gong, Q., Zhang, P., Wang, J., Ma, J., An, Y., Chen, Y. et al. (2019) ‘Morbidity and mortality after lifestyle intervention for people with impaired glucose tolerance: 30-year results of the Da Qing Diabetes Prevention Outcome Study’, The Lancet Diabetes & Endocrinology, 7(6), pp. 452–461.

International Diabetes Federation (2021) IDF Diabetes Atlas. 10th edn. Brussels: International Diabetes Federation.

Jelleyman, C., Yates, T., O’Donovan, G., Gray, L.J., King, J.A., Khunti, K. and Davies, M.J. (2015) ‘The effects of high-intensity interval training on glucose regulation and insulin resistance: a meta-analysis’, Obesity Reviews, 16(11), pp. 942–961.

Knowler, W.C., Barrett-Connor, E., Fowler, S.E., Hamman, R.F., Lachin, J.M., Walker, E.A. and Nathan, D.M. (2002) ‘Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin’, New England Journal of Medicine, 346(6), pp. 393–403.

Knowler, W.C., Fowler, S.E., Hamman, R.F., Christophi, C.A., Hoffman, H.J., Brenneman, A.T. et al. (2009) ’10-year follow-up of diabetes incidence and weight loss in the Diabetes Prevention Program Outcomes Study’, The Lancet, 374(9702), pp. 1677–1686.

Kosaka, K., Noda, M. and Kuzuya, T. (2005) ‘Prevention of type 2 diabetes by lifestyle intervention: a Japanese trial in IGT males’, Diabetes Research and Clinical Practice, 67(2), pp. 152–162.

Laaksonen, D.E., Lindström, J., Lakka, T.A., Eriksson, J.G., Niskanen, L., Wikström, K. et al. (2005) ‘Physical activity in the prevention of type 2 diabetes: the Finnish Diabetes Prevention Study’, Diabetes, 54(1), pp. 158–165.

Li, G., Zhang, P., Wang, J., Gregg, E.W., Yang, W., Gong, Q. et al. (2008) ‘The long-term effect of lifestyle interventions to prevent diabetes in the China Da Qing Diabetes Prevention Study: a 20-year follow-up study’, The Lancet, 371(9626), pp. 1783–1789.

Lindström, J., Ilanne-Parikka, P., Peltonen, M., Aunola, S., Eriksson, J.G., Hemiö, K. et al. (2006) ‘Sustained reduction in the incidence of type 2 diabetes by lifestyle intervention: follow-up of the Finnish Diabetes Prevention Study’, The Lancet, 368(9548), pp. 1673–1679.

Pan, X.R., Li, G.W., Hu, Y.H., Wang, J.X., Yang, W.Y., An, Z.X. et al. (1997) ‘Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance: the Da Qing IGT and Diabetes Study’, Diabetes Care, 20(4), pp. 537–544.

Ramachandran, A., Snehalatha, C., Mary, S., Mukesh, B., Bhaskar, A.D. and Vijay, V. (2006) ‘The Indian Diabetes Prevention Programme shows that lifestyle modification and metformin prevent type 2 diabetes in Asian Indian subjects with impaired glucose tolerance (IDPP-1)’, Diabetologia, 49(2), pp. 289–297.

Sallis, J.F., Cerin, E., Conway, T.L., Adams, M.A., Frank, L.D., Pratt, M. et al. (2016) ‘Physical activity in relation to urban environments in 14 cities worldwide: a cross-sectional study’, The Lancet, 387(10034), pp. 2207–2217.

Tuomilehto, J., Lindström, J., Eriksson, J.G., Valle, T.T., Hämäläinen, H., Ilanne-Parikka, P. et al. (2001) ‘Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance’, New England Journal of Medicine, 344(18), pp. 1343–1350.

Umpierre, D., Ribeiro, P.A.B., Kramer, C.K., Leitão, C.B., Zucatti, A.T.N., Azevedo, M.J. et al. (2011) ‘Physical activity advice only or structured exercise training and association with HbA1c levels in type 2 diabetes: a systematic review and meta-analysis’, JAMA, 305(17), pp. 1790–1799.

Van Rhoon, L., Byrne, M., Morrissey, E., Murphy, J. and McSharry, J. (2020) ‘A systematic review of the behaviour change techniques and digital features in technology-driven type 2 diabetes prevention interventions’, Digital Health, 6, pp. 1–21.

Whittle, F., Blythe, A., Gibbs, R. and Griffin, S.J. (2020) ‘Socioeconomic inequalities in access to and effectiveness of diabetes prevention interventions’, Diabetic Medicine, 37(2), pp. 210–219.

Wilmot, E.G., Edwardson, C.L., Achana, F.A., Davies, M.J., Gorely, T., Gray, L.J. et al. (2012) ‘Sedentary time in adults and the association with diabetes, cardiovascular disease and death: systematic review and meta-analysis’, Diabetologia, 55(11), pp. 2895–2905.