Resurgence in Diabetes-Related Complications

The improved long-term outlook for adults after receiving a diagnosis of diabetes is one of the most important clinical and public health successes in recent decades. During the early 1990s, patients with diabetes had reductions in lifespan of 7 to 10 years and an increased risk oflower-extremityamputation(LEA)vsthosewithoutdiabetes (58 vs 3 cases/10 000 persons/year, respectively) and kidney failure (28 vs 2 cases/10 000 persons/year).1 Risk of cardiovascular events, which caused the most deaths, also was higher among persons with diabetes vs thosewithoutdiabetes(141vs38hospitalizationsforacute myocardial infarction [AMI] per 10 000 persons/year).1 But through multifaceted improvements in diabetes care, riskfactormanagement,self-managementeducationand support, and better integration of care, these risk differences were reduced by 28% to 68% across a range of complications between 1990 and 2010, with gains most notable for reductions in AMI, stroke, and death due to hyperglycemia.1,2 Although the excess morbidity risk remained too high and the reduction in cardiovascular disease mortality led to new types of complications and causesofdeath,3 acontinuedreductionintheoverallpublic health burden caused by diabetes seemed promising. However, in an unanticipated new challenge to these improvements, a resurgence of diabetes complications has appeared in national statistics and in the epidemiology literature. Between 2010 and 2015, an increase in diabetes-related LEAs occurred nationally, reversing more than one-third of the 20-year decline in only 5 years.4,5 For hyperglycemic crisis, annual emergency department visits almost doubled between 2009 and 2015 (from 16.2 to 29.4 per 1000), hospitalizations increased by 73% (from 15.3 to 26.6 per 1000), and deaths increased by 55% (from 15.7 to 24.2 per 1000).4 For endstage kidney disease, AMI, and stroke, the long-term improvements stalled after 2010. Updated national statistics indicate that the recent increase in complication rates is occurring in young and middle-aged adults (Figure), among whom the risk of hyperglycemic crisis, AMI, stroke, and LEAs each increased by more than 25% during only 5 years. Although the rebound in rates has been most apparent in young adults (aged 18-44 years), middle-aged adults (aged 45-64 years) have higher absolute rates and account for most of the increase. Among older adults, the long-term reductions in rates have reached a plateau (Figure). These changes may not be limited just to the population diagnosed with diabetes because improvements in AMI and stroke in the general population also have plateaued, and most counties in the United States have seen an increase in cardiovascular disease mortality among younger adults.6 Explaining these trends is difficult because of myriad potential factors inpreventivecare,patientcharacteristics, andsociety.Trackingincidenceofmajorchronicconditions relies on diagnoses and procedures that appear in administrative files, making systematic changes in coding a key consideration.Forexample,reimbursementincentivesthat affect upcoding of diabetes or changing diagnostic schemes could conceivably affect rates. However, there are no data to suggest that this has occurred, and the diversity of outcomes, consistency across data sources, and the use of 1 diagnostic classification system (International Classification of Diseases, Ninth Revision, Clinical Modification)pointstobroadercauses.Fourkeyplausiblesources of these changes warrant examination because understanding the modifiable etiology of this resurgence could be helpful in developing an effective response. First, there may be increasing heterogeneity or a changing profile for the population of new diabetes cases. The age divide in trends points toward the millennial generation (born roughly between 1980 and 2000) and their high obesity prevalence (approximately 35% with body mass index [calculated as weight in kilograms divided by heightinmeterssquared]>30),compoundedbyhigherlevels of smoking and poorer blood pressure and lipid level management seen among younger than older adults.2 These factors may be combining with chronic kidney disease to influence an earlier onset of macrovascular complications.Whethertrendsincomplicationratesvarybetween type1andtype2diabetesisnotclearduetodatalimitations in national surveys, but it is possible that a greater increase in type 2 diabetes than type 1 diabetes among nonwhites is gradually increasing the ratio of type 2 to type 1 diabetes and affecting the character of diabetes-related morbidity. For the middle-aged and older population in particular, the combinationofdecreasingmortalityamongthosewithdiabetesandrecentdecreasesindiabetesincidenceisincreasingtheaveragedurationofdiabetesinthepopulation.4 This shiftinthedistributionforthedurationofdiabetesmaynow be affecting the risk of complications. The effect of Medicaid expansion on detection of diabetes and the characteristics of the underlying population with diagnosed diabetes is also an open question. Second, there may be stagnation in preventive care, again most prominent among young adults. After encouraging reductions in hemoglobin A1c (HbA1c) levels among patients with diabetes through most of the 2000s, the proportion meeting individualized HbA1c targets declined from the 1990s to 2011-2014 by 6 percentage points overall and by 10 percentage points among young adults.7 The increase in suboptimal HbA1c levels occurred at a time when clinical organizations have encouraged better age and comorbidity personalization of treatment targets and prevention of hypoglycemia for older adults. Such efforts have been accompanied by a reduction in major hypoglycemic events,4 but also could be accompanied by an unintended relaxation of glycemic control targets for younger adults. The increases in LEAs could be driven more by clinical decision making about the management of VIEWPOINT