This study showed that complications add significantly to the cost of medical care when hyperglycemia levels were below the threshold for diabetes, and impaired glucose tolerance independently contributed to costs.
: To estimate the prevalence of complications associated with diabetes in patients with hyperglycemia below the threshold for diabetes, and to evaluate the associated medical costs.
: Retrospective observational cohort study.
: We used fasting and random glucose test results, and a previously validated predictive equation to assign 26,111 nondiabetic patients to the following categories: normoglycemia, isolated impaired fasting glucose (I-IFG), isolated impaired glucose tolerance (I-IGT), or IFG with IGT (IFG/ IGT). We identified microvascular complications (retinopathy, neuropathy, nephropathy) and macrovascular complications (cardiovascular disease, stroke, peripheral vascular disease, heart failure) commonly associated with diabetes from electronic medical records. We then calculated and compared the impacts of hyperglycemia and its complications in terms of age/sexâ€“adjusted mean annual medical care costs.
: Complications were most prevalent among the I-IGT and IFG/IGT patientsâ€”more than half (51.1% in each group) had at least 1 complication compared with 33.9% of normoglycemic patients (P <.001 for both comparisons). Macrovascular complications added $3863 (P <.0001) to annual age/sexâ€“adjusted per-person medical costs; microvascular complications added $1874 (P <.0001). I-IGT ($716; P <.0001) and IFG/IGT ($438; P = .009) independently added to costs after controlling for presence of any complication.
: For many patients, complications associated with hyperglycemia appear to develop before diabetes diagnosis. Complications add significantly to the cost of medical care at hyperglycemic levels below the threshold for diabetes. However, the increased prevalence of complications did not completely explain the observed differences in age/sexâ€“adjusted medical care costs.
(Am J Manag Care. 2008;14(12):791--798)
Nondiabetic patients with impaired fasting glucose and/or impaired glucose tolerance often have complications normally associated with diabetes, but these complications do not fully explain why their costs are higher than patients with normal glycemic levels.
Microvascular and macrovascular complications including retinopathy, neuropathy, nephropathy, and cardiovascular disease afflict the majority of patients with diabetes. The American Association of Clinical Endocrinologists recently reported that 58% of patients with diabetes have at least 1 complication, and about 25% have 2 or more.1 That report also estimated that complications added about $20 billion to the cost of diabetes treatment in the United States in 2006. The American Diabetes Association estimated that direct medical expenditures for diabetes in 2007 totaled $116 billion, of which $58 billion was attributed to the chronic complications of diabetes.2
Many of these costs may be avoidable through the prevention of diabetes and more intensive diabetes management. However, microvascular and macrovascular complications of hyperglycemia often predate the diabetes diagnosis. Greater prevalence of retinopathy and nephropathy has been associated with increasing hyperglycemia below the threshold for diabetes diagnosis.3 In addition, data from the Framingham Heart Study offspring cohort showed an increased risk of chronic kidney disease among patients with impaired fasting glucose (IFG) or impaired glucose tolerance (IGT).4 Two meta-analyses demonstrated that nondiabetic hyperglycemia is a risk factor for cardiovascular disease.5,6
These studies all indicate that nondiabetic hyperglycemia is a risk factor for microvascular and macrovascular complications. To our knowledge, however, no study to date has simultaneously estimated the prevalence of these complications in a general population of patients with hyperglycemia below the diagnostic threshold for diabetes. Furthermore, the medical costs of hyperglycemic complications before diabetes onset have not been described previously. We recently reported that medical costs were higher among patients with dysglycemia than among those with normoglycemia.7 The objectives of the current study were to assess the prevalence of hyperglycemic complications among patients with IFG, IGT, or both relative to patients with normal glucose tolerance, and to determine the extent to which cost differences between patients with and without dysglycemia are explained by differences in microvascular and macrovascular complications.
Of the 26,111 patients, 59.8% (n = 15,629) were defined as normoglycemic; they did not screen as dysglycemic using the Tabaei et al equation8 and had a fasting glucose test result of <100 mg/dL (). Another 21.9% (n = 5713) did not screen as dysglycemic, but had a fasting glucose test result in the 100 to 125 mg/dL range and were therefore classified as I-IFG. Of those who did screen as dysglycemic, about half (n = 2552; 9.8% of total) had a fasting glucose result of <100 mg/dL and were classified as having I-IGT. The remainder (n = 2217; 8.5% of total) had a fasting glucose result of 100 to 125 mg/dL and were categorized as having IFG/IGT.
Age and sex were different across the 4 groups (P <.001 for all comparisons). With the exception of retinopathy and macular edema, statistically significant differences in all microvascular and macrovascular hyperglycemic complications were found among the groups. In general, complications were most prevalent among the I-IGT and IFG/IGT patients—more than half (51.1% in each group) had at least 1 complication. In contrast, 33.9% of patients with normoglycemia had at least 1 complication (P <.001 for both comparisons). For I-IGT and IFG/IGT patients versus patients with normoglycemia, prevalence of macrovascular complications (37.1% and 37.2% vs 23.9%; P <.001) and microvascular complications (27.7% and 25.9% vs 14.8%; P <.001) followed a similar pattern. Although adjustment for age and sex reduced the differences in prevalence of all complications between glycemic categories, statistical significance did not change (with the exception of peripheral vascular disease, which became nonsignificant).
Descriptive statistics for unadjusted total annual medical costs are provided in . The high standard deviation and large difference between mean and median costs indicates substantial variability and the nonnormal distribution commonly seen in medical cost data. However, very few patients had no costs in any of the glycemia groups.
Age/sex–adjusted annual medical costs for each of the individual complications are displayed in (model 1). Peripheral vascular disease ($3824; P <.001 for logged costs), congestive heart failure ($4401; P <.001), and retinopathy ($4433; P = .021) were the most expensive complications, but all complications added significantly to the medical cost burden. After controlling for the individual complications, patients with I-IGT ($584; P <.001) or IFG/IGT ($297; P = .018) had additional medical costs relative to patients with normoglycemia. Model 2 shows that any macrovascular complication added $3863 (P <.001) to medical costs, whereas any microvascular complication added $1874 (P <.001). IIGT ($639; P <.001) and IFG/IGT ($380; P = .015) independently added to costs after controlling for any microvascular or macrovascular complication. Model 3 indicates that the presence of any complication increased costs by $3496 P <.001). Again, I-IGT ($716; P <.001) and IFG/IGT ($438; P = .009) significantly increased costs after controlling for any complication.
displays the annual age/sex–adjusted medical costs across the 4 glycemic status groups, stratified by presence or absence of hyperglycemic complications. Among patients with no complications, there were no statistically significant differences in medical costs according to glycemic status. However, among patients with any complication, those with I-IGT or IFG/IGT incurred significantly greater costs than those with normoglycemia or I-IFG ($9617 and $9314 vs $8525 and $8155; P <.001 for both comparisons).
Among patients with macrovascular complications, there were no statistically significant differences in total medical costs according to glycemic status (). However, inpatient costs for those with macrovascular complications differed significantly between I-IGT patients and those with I-IFG or normoglycemia. Among patients with microvascular complications (), those with IFG/IGT ($9710) incurred significantly greater total costs than those with normoglycemia ($8657; P = .001) or I-IFG ($8046; P <.001). The I-IGT patients also incurred higher total costs than the normoglycemic or I-IFG patients (P = .008 and .003, respectively). Inpatient costs were largely responsible for the significant differences.
DISCUSSIONIn this observational cohort study of more than 26,000 individuals, we found that patients with IGT (either isolated or in combination with IFG) had significantly more microvascular and macrovascular complications than patients with normoglycemia. The prevalence of complications also was increased among patients with I-IFG relative to patients with normoglycemia, though not to the same degree. The observed differences between glycemic status groups were most distinct for microvascular complications, particularly for renal disease. However, macrovascular disease was more prevalent than microvascular disease in all glycemic status groups. We found that 43% to 51% of patients with nondiabetic dysglycemia had at least 1 complication, a prevalence rate only slightly lower than the 58% recently estimated among patients with diabetes.1 Thus, for many patients, microvascular and macrovascular complications commonly associated with diabetes appear to develop before diabetes diagnosis.
The increased prevalence of complications among patients with dysglycemia did not completely explain the observed differences in age/sex–adjusted medical care costs. Although costs for patients with I-IFG were similar to those for normoglycemic patients, I-IGT and IFG/ IGT independently accounted for additional costs after controlling for the presence of complications. This effect appeared to be concentrated in patients with microvascular complications; a subanalysis of patients with no complications and another of patients with macrovascular complications yielded no significant differences in comparisons of the 4 glycemic status groups. It is also important to note that other comorbidities did not explain costs differences. With the exception of cataracts (8.9%) and arthritis (5.0%), no specific condition accounted for more than 3% of hospitalizations, and the incidence of those 2 conditions was equal across groups.
Despite excluding patients with chronic renal failure, renal dysfunction accounted for the majority of microvascular complications in our data. Thus, the elevated costs we observed may be largely attributable to chronic kidney disease, a finding consistent with a previous report in this setting.12 There is strong evidence that intensive glycemic control can delay the onset of microalbuminuria and prevent progression from microalbuminuria to macroalbuminuria in patients with diabetes.13-15 Given the enormous cost of renal failure, early attention to hyperglycemia appears warranted.
In developing the predictive equation we used, Tabaei et al reported specificity of 90%, sensitivity of 55%, and positive predictive value of 65%.8 Even if the equation performed as well as or better in our population, we undoubtedly misclassified some individuals. Because we did not have oral glucose tolerance test results available, we could not assess the accuracy of the equation in our own data. Nevertheless, the clinical and demographic characteristics of the groups we created were consistent with the known characteristics of patients with IFG and IGT.16 Furthermore, misclassification is most likely to occur among normoglycemic patients who should in fact be classified as having IGT. If so, the complication prevalences and costs we reported for the patients with normal glycemia would be artificially elevated. Thus, our results may well be conservative.
An important limitation of the current study is that we required patients to have both a random and a fasting glucose test no more than 1 year apart. Because the tests typically did not occur on the same day (the mean number of days between tests was 210), it is possible that glucose homeostasis changed from one test to the other, which could result in misclassification. Furthermore, by requiring 2 tests that rarely occurred simultaneously, we may have identified patients who were seeking an explanation for ill health, and were therefore at greater risk of adverse health events. That may have inflated our prevalence estimates of hyperglycemic complications. Indeed, 34% of our dysglycemic study sample had at least 1 macrovascular complication, which is slightly higher than the 31% of patients with diabetes found to have cardiovascular disease in the same setting.17 On the other hand, previous reports indicate higher prevalences of peripheral vascular disease18,19 and retinopathy20,21 than those we found, whereas our estimate of the prevalence of peripheral neuropathy was similar to that in 1 prior study.19 Because so few population-based hyperglycemia studies are available, and none assessed all hyperglycemic complications simultaneously, it is difficult to determine the precision of our estimates.
Consistent with prior studies, we found that microvascular and macrovascular complications were more prevalent in patients with glycemic levels below the threshold for diabetes relative to normoglycemic patients. Complication rates were particularly high when IGT was present, a hyperglycemic state not typically assessed in the United States. It is not yet known whether early detection and treatment of patients at risk of developing diabetes can prevent additional complications from developing. However, if emerging research suggests complications can be avoided with diabetes prevention, earlier detection and treatment may provide substantial economic benefit to society and improved health for individual patients.
Author Affiliations: From the Center for Health Research (GAN), Kaiser Permanente, Portland, OR; GlaxoSmithKline, Inc (BA), Philadelphia, PA; and the Department of Endocrinology (WHH), University of Michigan, Ann Arbor.
Funding Source: This study was funded by a research grant from GlaxoSmithKline, Inc. By contract, GlaxoSmithKline, Inc, was allowed to review and comment on the manuscript. Also by contract, however, all final decisions on study design, data presentation and interpretation, and manuscript text were the responsibility of the primary author, Dr Nichols.
Author Disclosure: Dr Nichols reports having received grant support from GlaxoSmithKline, Inc. Dr Arondekar is an employee of GlaxoSmithKline, Inc, a company that manufactures treatments for this condition, and reports owning stock in that company. Dr Herman reports having received consulting fees and having served on the advisory boards for GlaxoSmithKline, Inc.
Authorship Information: Concept and design (GAN, BA, WHH); acquisition of data (GAN); analysis and interpretation of data (GAN); drafting of the manuscript (GAN); critical revision of the manuscript for important intellectual content (BA, WHH); statistical analysis (GAN); obtaining funding (GAN, BA); and administrative, technical, or logistic support (WHH).
Address correspondence to: Gregory A. Nichols, PhD, Center for Health Research, Kaiser Permanente, 3800 N Interstate Ave, Portland, OR 97227-1098. E-mail: email@example.com.
1. American Association of Clinical Endocrinologists. State of Diabetes Complications in America. Jacksonville, FL: American Association of Clinical Endocrinologists; 2007:1-13. www.gsk.com/media/press-kits/avandia-state-of-diabetes-Report-30july2007.pdf. Accessed November 21, 2008.
3. Gabir MM, Hanson RL, Dabelea D, et al. Plasma glucose and prediction of microvascular disease and mortality: evaluation of 1997 American Diabetes Association and 1999 World Health Organization criteria for diagnosis of diabetes. Diabetes Care. 2000;23(8):1113-1118.
5. Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident cardiovascular events: a meta-regression analysis of published data from 20 studies of 95,783 individuals followed for 12.4 years. Diabetes Care. 1999;22(2):233-240.
7. Nichols GA, Arondekar B, Herman WH. Medical care costs one year after identification of hyperglycemia below the threshold for diabetes. Med Care. 2008;46(3):287-292.
9. Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med. 1999;130(6):461-470.
11. Brown JB, Pedula KL, Bakst AW. The progressive cost of complications in type 2 diabetes. Arch Intern Med. 1999;159(16):1873-1880.
13. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group [published correction appears in Lancet. 1999;354(9178):602]. Lancet. 1998;352(9131):837-853.
15. Retinopathy and nephropathy in patients with type 1 diabetes four years after a trial of intensive therapy. The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Research Group [published correction appears in N Engl J Med. 2000;342(18):1376]. N Engl J Med. 2000;342(6):381-389.
17. Nichols GA, Brown JB. The impact of cardiovascular disease on medical care costs in subjects with and without type 2 diabetes. Diabetes Care. 2002;25(3):482-486.
19. Tapp RJ, Shaw JE, de Courten MP, et al. Foot complications in type 2 diabetes: an Australian population-based study. Diabet Med. 2003;20(2):105-113.
21. Tapp RJ, Shaw JE, Harper CA, et al. The prevalence of and factors ssociated with diabetic retinopathy in the Australian population. Diabetes Care. 2003;26(6):1731-1737.