Prediabetes—A Risk Factor Twice Removed

The concept of prediabetes (and similar concepts of impaired fasting glucose and impaired glucose tolerance) formally emerged in the 1990s to better characterize the incremental pathophysiology leading to the eventual development of diabetes.¹ It is defined by ranges of disordered glucose values between normal and diabetes (eg, hemoglobin A_1C levels, 5.7%-6.4% [to convert to the proportion of total hemoglobin, multiply by 0.01]). We treat it as a risk factor for diabetes; in midlife, those with a hemoglobin A_1C level of 6.0% to 6.5% have 20 times the risk of developing diabetes compared with those with a hemoglobin A_1C level of 5.0%.²

In this issue of JAMA Internal Medicine, Rooney and colleagues³ explore the association between prediabetes and outcomes in older adults. Using data from the Atherosclerosis Risk in Community Study, the authors examined the 6-year natural progression of glycemic measures in 3412 older adults aged 70 to 91 years (mean age, 75 years) without diabetes. They found that prediabetes was present in up to 59% of individuals depending on the definition used (hemoglobin A_1C levels of 5.7%-6.4% or fasting blood glucose levels of 100-125 mg/dL [to convert to millimoles per liter, multiply by 0.0555]). Prediabetes was a risk factor for diabetes compared with normoglycemia (hazard ratio, 3.0-3.2; range depending on the definition of prediabetes), but participants were more likely to regress to normoglycemia (13%-44%) or die (16%-19%) than progress to diabetes (8%-9%). Furthermore, all-cause mortality for those with prediabetes was similar to those with normoglycemia (hazard ratio, 0.83-1.07). Thus, prediabetes appears to be a weaker risk factor for diabetes in older adults.

The results of Rooney et al³ suggest that the concept of prediabetes may be of limited importance for older adults. While diabetes was an acute, symptomatic, and invariably fatal disease when originally described, it has become an asymptomatic chronic condition because of early detection and the wide-spread availability of effective treatment.⁴ For most patients, it is the end-organ vascular complications that results from years of poorly controlled diabetes that cause symptoms. Therefore, the modern definition of diabetes is conceptually closer to being a risk factor itself (eg, something that portends future disease) than an illness (eg, something that patients experience). Prediabetes, then, is a risk factor twice removed; it is a risk factor for diabetes, which itself may be most accurately described as a risk factor for end-organ vascular disease.

The 2020 guidelines from the American Diabetes Association advocate that all patients with prediabetes be monitored annually for the development of type 2 diabetes, and that patients be referred to Diabetes Prevention Programs for weight loss and physical activity.⁵ Therefore, the results reported by Rooney and colleagues³ have important implications for these guidelines while raising broader issues on how we define and address risk factors in older patients.

First, in older adults with frailty and limited life expectancy, prediabetes is irrelevant and can safely be ignored. Because the benefits of prediabetes management are most likely accrued 10 or more years in the future, older adults with frailty and limited life expectancy are unlikely to benefit from prediabetes management. Guidelines should clarify that prediabetes is a concept that should be reserved for healthier, middle-aged adults rather than older adults with frailty.

Second, in healthy adults older than 75 years (the mean age of the study participants), we should recognize that prediabetes, as a risk factor twice removed, should be lower priority than symptomatic conditions (which are affecting patients immediately) or traditional risk factors. Older persons may spend up to 2 hours a day engaging in health care-related activities.⁶ Diagnosing prediabetes and then expending time and effort discussing management strategies should not come at the expense of attending to other issues of immediate importance to the patient. For all but the healthiest of older adults older than 75 years, the current recommendations for annual monitoring and weight loss are likely low yield. Future long-term cohort studies should explicitly examine whether prediabetes increases the risk of mortality and morbidity in healthy older adults. In the interim, guideline management of prediabetes in persons older than 75 years should be individualized, much in the way that cancer screening is individualized based on life expectancy, expected benefit, and patient values and preferences.⁷

Third, if the diagnostic thresholds for prediabetes developed in middle-aged adults are less applicable to older adults, we should reexamine whether diabetes diagnostic thresholds, which are also developed in middle-aged adults, are valid for older adults. If most older adults with prediabetes revert to normoglycemia, how many older adults with mild diabetes (eg, hemoglobin A_1C levels of 6.5%-7.0%) revert to prediabetes or even normoglycemia? Additional studies are needed to determine whether newly diagnosed mild diabetes in older adults leads to adverse outcomes if left untreated. If it does not, shifting the cutoffs for diagnosing diabetes in older adults would help us focus treatment on those older adults for whom diabetes is likely to result in symptomatic end-organ damage, while avoiding identifying many older adults for whom diabetes is unlikely to cause harm.

For middle-aged adults, a new diagnosis of diabetes can lead to substantial morbidity and mortality; thus, focusing on risk factors, such as prediabetes, is high-value and appropriate in a middle-aged population. However, for many older adults, new-onset diabetes will often be mild and asymptomatic and only one of many potentially life-threatening conditions. This study shows that identifying prediabetes in older adults should be regarded as a low priority, as it rarely leads to incident diabetes or adverse outcomes. To ensure high-value care for older adults, we should focus our care and research on what matters most to older adults and deprioritize twice-removed risk factors, such as prediabetes.