Look AHEAD to Precision Prevention in Type 2 Diabetes

Precision medicine aspires to reduce error and improve accuracy in medical decisions and health recommendations. The pillars of precision medicine include disease prediction (of primary and secondary events), prevention, diagnosis, and treatment (1).

The idea of precision medicine in the context of disease prevention is predicated on the failure of standard approaches to adequately slow or prevent disease onset in many people. Although it is often claimed that conventional lifestyle interventions work well for the prevention of diabetes and its complications, studies such as the Diabetes Prevention Program (DPP) illustrate that most people at risk of developing diabetes eventually will, with the average delay in disease onset with lifestyle and metformin a mere 36 and 12 months, respectively (2); the DPP also showed that neither intensive lifestyle intervention nor metformin extend life expectancy beyond standard advice (3). The effects of intensive lifestyle intervention in diabetes complications have also been studied. The Look AHEAD trial, for instance, investigated the effect of the same intensive lifestyle intervention used in the DPP on a combined cardiovascular end point in about 5000 adults with type 2 diabetes (T2D). After about a decade of intervention, there was no discernable difference in cardiovascular event rates between the two intervention arms (hazard ratio in the intervention group, 0.95; 95% CI, 0.83-1.09; P = .51), and the trial was stopped because of futility (4).

In this edition of the Journal, Bancks et al report a reanalysis of the Look AHEAD trial (5), motivated by their earlier work on a closely related topic (6). The authors previously showed that cardiovascular event rates in Look AHEAD vary considerably depending on baseline characteristics, and that in some participants, the intensive lifestyle intervention may have caused harm; specifically, while the Look AHEAD lifestyle intervention conveyed cardiovascular risk protection in many participants, event rates were elevated in people with poor glycemic control assigned to the lifestyle intervention, a subgroup within which the incidence was roughly double that seen in participants with dysglycemia randomly assigned to the control arm (6).

In this new publication (5), Bancks et al report detailed analysis of glycemic control and specific cardiovascular end points. The authors conclude that “among individuals with [glycated hemoglobin A_1c] HbA_1c greater than 8.7%, who are at high absolute risk for cardiovascular events, the evidence-based Look AHEAD intervention was not associated with [cardiovascular disease] CVD prevention and was associated with higher CVD risk.” This finding contradicts a widely espoused view that one-size-fits-all lifestyle interventions are adequate for cardiovascular risk reduction in people with diabetes, highlighting a potential role for precision lifestyle interventions for cardiovascular risk reduction in diabetes.

The authors found that the least weight was lost in those with the worst baseline glycemic control, particularly when randomly assigned to the intensive lifestyle intervention. Moreover, providing the 2-year weight loss goal was met, the lifestyle intervention protected against cardiovascular disease (CVD) in all participants regardless of their baseline glycemic control. This suggests that CVD risk may be primarily driven by excess adiposity and secondarily by poor glycemic control in people with T2D, while also highlighting that even with intensive lifestyle intervention, sustained weight loss is a major challenge for many people with diabetes. Importantly, the intensive lifestyle intervention lowered all-cause mortality, cardiovascular mortality, and heart failure, even in those with the worst glycemic control, providing the 7% weight loss goal at 2 years was met. Nevertheless, when receiving the lifestyle intervention, this glycemic subgroup was more prone to myocardial infarction, angina, stroke, coronary artery bypass grafting, and carotid endarterectomy than those randomly assigned to the control intervention. It may be, therefore, that the specific type of lifestyle/weight loss intervention for people with diabetes should be conditional on baseline glycemic control, highlighting the need for precision lifestyle interventions.

An important and often overlooked challenge of behavioral intervention trials is that it is very difficult to mask which intervention a participant has been randomly assigned to, which may drive confounding and other types of bias. For instance, when a participant is aware of the intervention they are receiving, it might cause unintended changes in behavior, so some participants receiving the active lifestyle intervention may decrease activity levels outside the intervention sessions, and some participants allocated to the control intervention may feel inspired to adopt healthy behaviors. With objective behavioral assessment it may be possible to make statistical adjustments that help minimize this type of bias, although no such corrections were undertaken in the analysis by Bancks et al (5). Moreover, in their analysis, behavioral compensation might have differed by level of baseline glycemic control, which is suggested by differences in weight change across HbA_1c categories. It is also possible that participants with poor baseline glycemic control tended to have the most adverse genetic risk profiles, indicated by the higher prevalence of family history of diabetes in this group, or varied in other ways that influenced the efficacy of the intensive lifestyle intervention. This might also confound the interpretation of the results in this paper. A further consideration is that HbA_1c, albeit the clinical standard, is a relatively crude measure of glycemic control; more granular assessment methods (eg, continuous glucose monitoring), might help further elucidate how best to design precision lifestyle interventions in people with T2D. It may also be that, as the authors point out, glycemic control may be merely a marker of other pathobiological processes, useful for prediction and treatment stratification, rather than a treatment target per se.

Notwithstanding these limitations, the work reported by Bancks et al (5) highlights considerable heterogeneity in cardiovascular adaptations to standard lifestyle interventions intended to mitigate cardiovascular risk, with apparently detrimental effects in a subgroup of the Look AHEAD cohort with poor glycemic control at baseline. The reasons for this are speculative but may relate to the nature of the diet intervention in Look AHEAD, which focused on reduced caloric intake, largely by lowering dietary fat; an unintended consequence of which is that the dietary carbohydrate-to-fat ratio was increased.

More than 60 years ago, Reaven et al (7) observed that even in the absence of established diabetes people who cannot adequately metabolize dietary carbohydrates have elevated atherogenic risk profiles, and that this may be a key component of the etiology of atherosclerotic heart disease. More recently, studies focusing on precision nutrition interventions have shown that in people with T2D (8), postprandial glycemic excursions can be diminished; the key components of these precision nutrition diets include the meal's carbohydrate content and carbohydrate-to-fat ratio.

What remains unknown, and should perhaps be prioritized for future research, is whether people of similar characteristics to those Look AHEAD participants with poor baseline glycemic control would enjoy cardiovascular benefits from well-designed precision nutrition interventions.

In summary, Bancks et al (5) have highlighted that conventional lifestyle interventions might convey undesirable cardiovascular health consequences in some people with diabetes, specifically those with poor glycemic control. Prospective trials involving carefully designed precision nutrition interventions are needed to test this hypothesis.

Disclosures

P.W.F. is director of the Department of Translational Medicine at the Novo Nordisk Foundation, a purely philanthropic enterprise foundation based in Denmark. From 2017 to 2022, P.W.F. was on the scientific advisory board of Zoe Ltd. This commentary was written in P.W.F.'s academic capacity. The views expressed in this commentary are not necessarily those of the Novo Nordisk Foundation.

References

Abbreviations

 
• CVD

  cardiovascular disease

 
• DPP

  Diabetes Prevention Program

 
• HbA_1c

  glycated hemoglobin A_1c

 
• T2D

  type 2 diabetes