When Should "Pre" Carry as Much Weight in the Diabetes Comorbidity Debate?

Insights From a Population-Based Survey

Negin Iranfar, MPH; Tyler C. Smith, MS, PhD


Prev Chronic Dis. 2018;15(3):e36 

In This Article


Prediabetes is an early indicator of diabetes and contributes to the worldwide pandemic.[3,7] Between 5% and 10% of people with prediabetes are estimated to progress annually to diabetes, depending on race/ethnicity and the detailed pathogenesis of their prediabetes.[2,9,11] Of the estimated 86 million individuals with prediabetes in the United States, only 8% to 11.6%, or between 7 and 10 million individuals, have received a diagnosis and are aware of their prediabetes condition. Furthermore, a consistent set of chronic diseases associated with diabetes is seen in people with diagnosed prediabetes, even at lower BMI. This is alarming and may indicate a greater need for more rigorous diagnosis of prediabetes. It also raises the question of whether current treatments and interventions for prediabetes, although successful in delaying progression to diabetes, sufficiently address other chronic diseases concomitant with prediabetes. Many of the chronic health conditions included in this study are closely related to obesity, and are most prevalent among populations with obesity who have diabetes (Figure). However, the increasing frequency of these conditions among people with diagnosed prediabetes at lower BMI (normal and overweight) may signify an unwelcome trend of increased risk of comorbidities at lower BMI in prediabetes.

In an extensive meta-analysis of 16 prospective cohort studies that included more than 890,000 participants, Huang et al found that people with prediabetes at baseline had a significantly increased risk of cancer.[18] Additional literature has associated increased risk for kidney disease,[19] CVD,[20] and arthritis[21] with prediabetes. Risk factors for diabetes and prediabetes (age, obesity, and physical inactivity) have been documented[22,23] and are confirmed in our study, with age and BMI being most highly predictive for both conditions. Conversely, regular annual checkups and access to physicians had a protective effect on diabetes. Accordingly, the focus has been on changing lifestyle habits among people with prediabetes and diabetes and using medication.[24,25]

Several international trials have demonstrated the reversion from prediabetes to normoglycemia, based on lifestyle and drug-based interventions. The Finnish Diabetes Prevention Study reported average weight loss of 4.2 kg during a 3-year period using lifestyle intervention and medication.[26] However, there are concerns that treating prediabetes with medication is an overtreatment of a nondisease condition and should be approached only in cases with other comorbidities, such as heart disease.[27] One of the many debates about treatments of prediabetes is the question of whether the focus should be on reversing the condition or simply delaying development of diabetes. Studies suggest that prolonged duration of prediabetes can result in both microvascular and macrovascular complications of diabetes, even in the absence of overt development of diabetes.[11] Our results concur with such concerns and add to the body of knowledge addressing the possible public health implications of an extended long-term prediabetes condition.

Our study has limitations. First, we used self-reported data, which were not confirmed by medical records or other health history information. Self-reported data may not reflect the continuum of disease and may better be assessed with a simple functional health assessment, which was outside the limits of this study. Furthermore, the survey questions were designed as "Have you ever…," eliminating any distinction between prevalence and those who may have reverted to normoglycemia, resulting in possible overestimation of current prevalence. Second, self-reported diabetes does not distinguish between type 1 and type 2 diabetes; however, it is generally accepted that more than 90% of diabetes in the United States is type 2.[15] Although limitations are inherent in the depth and accuracy of any self-reported survey data, it nonetheless allows us to identify consistencies in variables common in both diabetes and prediabetes. Third, although BRFSS data encompass a large cross-section of the population, including both cellular telephone and telephone landline surveys since 2011, they still exclude or could underrepresent certain groups and races/ethnicities with language limitations, telephone access limitations, or those who are institutionalized. Fourth, reporting frequencies on prediabetes-specific questions has been inconsistent among the 50 states and the District of Columbia; some states did not report on that survey question during 1 or more of the 5 periods of this study, and in particular 39 states did not collect data on prediabetes for the 2015 survey year. As such we expect the prevalence of diagnosed prediabetes to be an underestimation and not valid for geographic region comparisons. This has also precluded us from estimating prediabetes prevalence for 2015. Another possible bias is that prediabetes overall is more prevalent than diagnosed prediabetes. As such, there may be a differential misclassification for diagnosed prediabetes with concurrent comorbidities. More prospective data may provide an excellent source to isolate the effect size of any such bias. Fifth, the nature of the cross-sectional survey prevents any extrapolation of causal relationships between the various health conditions used in this study and diabetes or prediabetes. Therefore, it cannot be determined if impaired glucose metabolism is responsible for other health conditions or perhaps caused by some combination of comorbidities included in this study. However, it is generally accepted that obesity is a common cause for most chronic health conditions. Furthermore, the self-reported diagnosis of prediabetes is likely an underestimation of actual prediabetes in the United States, because the American Diabetes Association only recommends screening for this condition starting at age 45, and then only if there are other health factors;[15] similarly, adults younger than 50 may not be aware that they have diabetes. Lastly, BRFSS does not include any questions about frequency of testing the blood glucose level, glycated hemoglobin A1c, or any other screening or treatments of those diagnosed with prediabetes. As such, it is unknown if people with diagnosed prediabetes are getting the same or similar care as people diagnosed with diabetes.

Although much attention has been given to diagnosis of at-risk populations at the stage of prediabetes to reduce incidence of diabetes, efforts are focused on preventing prediabetes from progressing to diabetes. Implied in this attitude is the view that prediabetes has lower rates of morbidity compared with diabetes. However, the validity of this assumption is not clear. This study highlights that many chronic disease conditions are present at high rates in prediabetes and that a prolonged period of prediabetes does not necessarily reduce the risk of certain comorbidities compared with diabetes. Our results suggest that there may even be an increased risk at lower BMI among people with prediabetes to present with other chronic comorbid health conditions. In light of potential comorbidities that may occur in this at-risk population, substantial effort should be considered to identify prediabetes at a lower BMI and younger age, where rigorous attempts to reverse prediabetes to normoglycemia could prove far more beneficial in promoting public health.