© 2004 by the American Diabetes Association, Inc.
Diabetes and Renal Disease in Veterans
1 Seattle Epidemiologic Research and Information Center, Department of Veterans Affairs, Puget Sound Health Care System, Seattle, Washington Address correspondence and reprint requests to Bessie A. Young, MD, MPH, VA Puget Sound Health Care System (152-E), 1660 S. Columbian Way, Seattle, WA 98108. E-mail: youngb{at}u.washington.edu
OBJECTIVETo describe the frequency and sequelae of diabetic renal disease in veterans who receive health care from the Veterans Administration (VA). RESEARCH DESIGN AND METHODSVeterans with a diagnosis of diabetes, diabetic nephropathy, other kidney diseases, and end-stage renal disease (ESRD) were identified by diagnosis codes from national VA databases for FY1998. Data were obtained and analyzed for prevalence of renal disease, comorbid conditions, and 1-year mortality. RESULTSA total of 44,671 (10.7%) of the 415,910 veterans with diabetes had a concomitant diagnosis of any renal disease. The average age was 67 years; 98% were male and 60% were white. The prevalence of diabetic nephropathy was 6.0% (n = 25,263). ESRD secondary to diabetes was present in 4.2% (17,636) of subjects. The age-standardized prevalence of diabetes and any renal disease was 72.6/1,000 persons and differed by race (white 76.1/1,000, black 103.4/1,000 persons). Diabetes-associated ESRD prevalence was higher among black versus white veterans and male versus female veterans. One-year age-standardized mortality was 10.7%. CONCLUSIONSNephropathy is prevalent in veterans with diabetes. Greater mortality is observed among those with renal disease compared with those without renal disease. Additional surveillance is needed to identify persons likely to progress to diabetic nephropathy and to plan for appropriate and timely health care for these individuals.
Abbreviations: ESRD, end-stage renal disease VA, Veterans Administration
Diabetes is a prevalent disease in the U.S. Approximately 17 million Americans are estimated to have diabetes, but only 11.1 million are aware that they have the disease (1). Approximately 510% of persons with diabetes will have type 1 diabetes and 9095% will have type 2 diabetes (1,2). Among those with diabetes, the prevalence of kidney disease ranges from 20 to 40% and varies depending on the type and duration of diabetes (311). Diabetic nephropathy is now the leading cause of end-stage renal disease (ESRD) in the U.S. and accounts for 40.3% of all new cases of dialysis (12). Because the prevalence of type 2 diabetes is substantially higher than type 1 diabetes, most persons diagnosed with diabetic ESRD will have type 2 diabetes (1214). Limited population-based data and no national registry data describe the incidence or prevalence of diabetic renal disease before the onset of ESRD (5,911,1517). Most studies have evaluated the incidence and/or prevalence of microalbuminuria or "incipient diabetic nephropathy" in confined geographic areas or for a specific homogeneous population; however, few studies describe the prevalence of diagnosed diabetic nephropathy or associated renal disease before initiation of dialysis. Determining the number of individuals diagnosed with diabetic nephropathy in a population-based setting before ESRD is difficultyet necessary to project for future resource utilizationbecause not every person with diabetes progresses to diabetic nephropathy or ESRD. The Veterans Administration (VA) routinely collects national data from which these questions can be answered. The purpose of this article is to describe the characteristics, morbidity, and mortality for veterans with diabetic renal disease receiving VA care.
Veterans with diabetes were identified through national VA databases at the Austin Automation Center, Austin, Texas. Inpatient and outpatient data were extracted from the Patient Treatment Files (PTFs) and Outpatient Care Files (OPCs) using ICD-9 and CPT codes as outlined in Table 1 (18,19).
Veterans with diabetes were identified by having at least one outpatient visit with an ICD-9 diagnosis code of 250.xx for FY1998 and a total of three or more visits for that year. Persons with diabetes and renal disease were defined by ICD-9 code 250.4 (diabetic nephropathy) or a code for diabetes (250.xx) coupled with additional secondary renal conditions, including nephropathy, glomerulonephritis, or glomerulonephropathy. Veterans with a diagnosis of diabetes and other renal diseases (interstitial nephritis, acute renal failure, cystic disease, secondary vasculitis, hypertensive disease, neoplasms, glomerulonephritis from another cause, or other renal diseases) were identified through additional codes. Veteran mortality was determined using the Beneficiary Identification and Records Locator System (BIRLS) death file.
Data were initially downloaded as SAS data files (20). Data analyses were performed using SPSS and STATA statistical software packages (21,22). Renal disease prevalence rates were calculated for veterans with diabetes, as diagnosed by ICD-9 code. Standardized prevalence rates were compiled using the direct method; rates were age adjusted to the 1990 U.S. population of individuals aged 18 years and older (23). One-year, age-standardized mortality was computed using the direct method (23). Among veterans with diabetes and renal disease, analyses of differences in comorbid characteristics were conducted using the independent Students t test for continuous data and the
Prevalence of diabetes and renal disease The median age for veterans with diabetes and renal disease was 67 years; 98.4% were male (Table 2). White veterans with renal disease comprised 59.9% of the population, followed by black (22.3%) and Hispanic (6.2%) veterans. Race or ethnicity was unknown for 10.6%. Compared with diabetic individuals without renal disease, veterans with diabetes and renal disease were significantly older, more likely to be of black race, and more likely to have comorbid conditions. Cardiovascular disease, stroke, high blood pressure, and chronic obstructive pulmonary disease were all significantly more common in veterans with diabetes and renal disease than those without renal disease.
Of the 415,910 veterans identified with diabetes, 44,671 carried a diagnosis of any type of renal diseases; of those, 56.6% (25,263) had a diagnosis of diabetic nephropathy and 39.5% (17,636) had a diagnosis of diabetic ESRD (Table 3). ESRD among those with diabetic nephropathy was present in 70% of veterans. In veterans with diabetes and other renal diseases, interstitial disease (10.5%) was the most frequent diagnosis, and acute renal failure accounted for 4.3%. Less than 10% of the renal diagnoses were secondary to tumor, hypertension, or cystic disease. Renal diseases secondary to congenital abnormality, hereditary abnormality, sickle cell disease, or "not otherwise specified" were included in the "other" category (20.9%).
Prevalence of renal disease The age-standardized prevalence of diabetic renal disease was 72.6/1,000 persons. The prevalence of diabetic nephropathy excluding ESRD was 13.6/1,000 persons. The prevalence of diabetic nephropathy including ESRD was 27.2/1,000 persons. When evaluated by race/ethnicity, black veterans with diabetes had a higher age-adjusted prevalence of renal disease (103.4/1,000 persons) compared with white veterans (76.1/1,000 persons). In addition, the prevalence of diabetic nephropathy with ESRD was higher in black (44.9/1,000) compared with white (26.5/1,000) veterans. The frequency of acute renal failure was also higher in black (6.3/1,000) compared with white (2.7/1,000) veterans.
Sex differences
Mortality
Diabetes is a pervasive disease and affects 12.5% of veterans served in VA outpatient clinics (25). We found prevalent renal disease in 10.7% of veterans with diabetes, which occurred at a frequency of 72.6/1,000 patients. The prevalence of comorbid conditions was significantly greater in veterans with diabetes and renal disease than in veterans with diabetes without renal disease. Among those with diabetes and renal disease, the most common recorded renal diagnosis was diabetic nephropathy and ESRD. Black veterans were disproportionately affected with ESRD compared with white veterans, as were men compared with women. Using ICD-9 diagnoses of renal disease to identify patients, we observed a greater prevalence of renal disease than reported previously. The prevalence of diabetic renal disease in population-based studies has varied widely. In a population of patients with diabetes, Klein and colleagues (9,10) reported a similar prevalence of microalbuminuria in younger patients (21%) and macroalbuminuria in older patients (18.4%) in Wisconsin. Prevalence varied by insulin usage and duration of diabetes. Since microalbuminuria and macroalbuminuria may represent an earlier end of the spectrum of diabetic nephropathy than that used in the current study, their results may not be comparable to those of the current study. In addition, Humphrey and Ballard (6), using a definition of chronic renal failure as a serum creatinine >4.0 mg/dl and/or proteinuria, reported a 3% prevalence of chronic renal failure among their diabetic patients. Age- and sex-adjusted incidence rates of diabetic nephropathy were 235/100,000 person-years for those with type 1 diabetes and 50/100,000 person-years for those with type 2 diabetes, values which are substantially lower than that observed in the current study. This difference may be due to differences in our case definition and patient population, or it may represent an increase in secular trends in the development of diabetic renal disease (12). We also observed that the prevalence of diabetic nephropathy in a heterogeneous population differed from that reported by others in more homogeneous populations. Several authors (1517) have evaluated the epidemiology of early diabetic nephropathy in population-based settings; however, data on more advanced disease are limited. Pima Indians, for whom the risk of diabetic ESRD is greatest (16), have elevated prevalence of incipient diabetic nephropathy, which varies widely by age (4761%) (26,27). Population-based prevalence of diabetic nephropathy before the initiation of dialysis among Pima Indians has not been reported. In addition, compared with the current study, a similar prevalence of diabetic nephropathy (12.5%) was found in a population-based Norwegian cohort study; however, diabetic nephropathy or advanced renal disease prevalence was not reported. The current study found that renal disease prevalence differed significantly by race. Our study confirms that of previous reports, most of which have shown that minority groups tended to have higher rates of ESRD than whites (13,16,28,31). Ethnic differences in the risk of renal disease were not evaluated in the current study. In addition to the above findings, the current study found that women veterans treated in the VA were less affected by renal disease than men. In regard to whether this discrepancy represents differences in the biology of diabetic renal disease by sex among veterans, the small proportion of women veterans, or selection factors leading to noncomparability between male and female veterans with regard to risk of diabetic renal disease cannot be determined from this analysis. Finally, we observed that the overall age-standardized mortality over 1 year was 10.7% for all veterans with diabetes and a diagnosis of renal disease. National race- and age-adjusted mortality rates are not available for persons with diabetes and renal disease before the onset of ESRD. The age-, race-, and sex-adjusted standardized mortality rate reported by the U.S. Renal Data System (USRDS) (23.2/100 patient-years) is greater than that reported by our study. It is likely that the USRDS population includes more type 1 diabetic patients, who may have greater disease severity and higher mortality than a population eligible for military service (12). This study has a number of strengths, one of which includes evaluation of a large cohort of veterans who received primary health care in a national setting. However, a number of limitations exist. Because the current study used administrative data to identify persons with diabetes, selection bias may have occurred. In addition, the potential for misclassification of diabetic renal disease and ESRD was also possible since the diagnosis of diabetic nephropathy relied exclusively on ICD-9 codes and access to clinical, pharmacy, or histologic data was not available for disease confirmation. Further, the lack of clinical data, such as microalbuminuria, proteinuria, or creatinine levels, may have led to an underestimation of the true number of veterans with diabetic renal disease. An additional limitation was the inability to distinguish between persons with type 1 and type 2 diabetes. Veterans with type 1 diabetes might have different renal disease outcomes compared with those with type 2 diabetes, although no data that directly compare the course of renal disease between type 1 and type 2 diabetic patients currently exist. However, the prevalence of type 1 diabetes is thought to be low in veterans, as type 1 diabetes is an exclusion criterion for military service. Finally, this study was limited in its ability to determine significant differences based on sex. Because women only represented 2% of the diabetic veteran population, the study lacked power to evaluate sex differences in diabetic renal disease rates stratified by race or ethnicity. Sex differences in type 1 diabetes renal disease prevalence have been described in a population of type 1 diabetic patients, results of which varied as function of duration of diabetes (7). Duration of diabetes was not available in our database. In summary, we found that renal disease was present in a large number of veterans with diabetes who received VA medical care during FY1998. Prevalence of renal disease varied by race and sex. The most frequently identified type of renal disease in this population was diabetic nephropathy. Veterans with diabetes and renal disease were older, more likely to have associated comorbid conditions, and less likely to survive for 1 year. This information may provide assistance to persons involved in health care planning who care for people with similar demographic features, as well as for identifying potential areas for further investigation.
The Department of Veterans Affairs, Veterans Health Administration, Epidemiologic Research and Information Center, and Health Services Research and Development, Seattle, Washington, supported the research reported and outlined in this article. At the time of this research, B.A.Y. was a fellow in the Health Services Research and Development Program at VA Puget Sound. B.A.Y. is currently supported by an American Diabetes Association Career Development Award and a Robert Wood Johnson Minority Faculty Development Fellowship. G.R. is supported by a VA Health Services Research and Development Career Scientist Award.
Funding for this supplement was provided by The Seattle Epidemiologic Research and Information Center and the VA Cooperative Studies Program. A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances. Received for publication July 1, 2003. Accepted for publication July 25, 2003.
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||