| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Epidemiology (E.B.-M., L.H.K., M.P.), Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania 15261; and Division of Immunogenetics (S.P., Y.-J.Z., T.H., M.P.), Diabetes Institute, Department of Pediatrics, Rangos Research Center, Children’s Hospital of Pittsburgh, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15213
Address all correspondence and requests for reprints to: Emma Barinas-Mitchell, Ph.D., Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, 130 DeSoto Street, Pittsburgh, Pennsylvania 15261. E-mail: barinas{at}edc.pitt.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this study was to evaluate the relationship between GAD65AA and glucose tolerance status, current diabetes treatment, and clinical measures in older adults.
Design: GAD65AA were measured at baseline in 3318 participants from the Cardiovascular Health Study, a cohort study of 5888 individuals 65 or older.
Setting: The population-based cohort was recruited from four U.S. sites.
Patients: Patients included all Cardiovascular Health Study participants with known diabetes, newly diagnosed diabetes, impaired fasting glucose, impaired glucose tolerance, and a sample of normal glucose-tolerant participants.
Main Outcome Measures: GAD65AA, body mass index, fasting glucose and insulin levels, blood pressure, lipid levels, and diabetes treatment at baseline were measured.
Results: The prevalence of GAD65AA increased with decreasing glucose tolerance in both Blacks (n = 560) and Whites (n = 2730), being more pronounced in known diabetic individuals. GAD65AA were found in 2.3, 5.8, 7.8, and 8.3% of diabetic participants, reporting use of no diabetes medication, oral hypoglycemic agents, insulin only, and both oral hypoglycemic agents and insulin, respectively (P = 0.02, linear trend). Among diabetic participants, GAD65AA positivity was associated with diabetes treatment, higher fasting glucose, and lower body mass index.
Conclusions: Even among older individuals with diabetes, GAD65AA may be a useful marker in identifying a subgroup of autoimmune diabetes, serve as a marker of insulin requirement, and remain stable over years.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
AA to islet cell antigens, such as the 65-kDa isoform of glutamic acid decarboxylase (GAD65), are found in 70–80% of newly diagnosed type 1 diabetic patients and are considered surrogate markers of progression to insulin-requiring diabetes in first-degree relatives of probands with type 1 diabetes (7, 8). Furthermore, GAD65 AA appear to be present in up to 10% of adults clinically diagnosed with type 2 diabetes (6, 9, 10, 11). For the most part, in type 2 diabetic patients, positivity for GAD AA, as well as AA to other islet cell antigens, correlates with some of the phenotypic features consistent with those of type 1 diabetes, such as younger age at diagnosis, lower body mass index (BMI), and a loss of ß-cell function (12). This form of disease with initial type 2-like diabetes presentation and with serological evidence of islet cell autoimmunity has been termed latent autoimmune diabetes in adults and, in some populations, has been associated with a slowly progressive decline in ß-cell function before progression to insulin-requiring diabetes (10, 13, 14, 15, 16).
We reported in a pilot study of the Cardiovascular Health Study (CHS), a population-based longitudinal study of individuals 65 yr old or older, that the prevalence of GAD65AA was 12% in Whites and 10% in Blacks with diabetes and that the presence of GAD65AA was associated with current diabetes treatment (6). Furthermore, we recently reported a positive and independent association between GAD65AA positivity and lower BMI and C-peptide levels in non-Hispanic Whites and non-Hispanic Blacks, but not Mexican Americans, 40 yr of age or older from the Third National Health and Nutrition Examination Survey (NHANES III) (17). These cross-sectional findings, which are consistent with other studies, support the role of an autoimmune process leading to ß-cell damage and subsequent reduced insulin secretion as it occurs in a similar manner in type 1 diabetes (9, 11).
In the present study we expanded our initial evaluation of the CHS Pittsburgh cohort to the full CHS cohort in all four sites and have also included longitudinal data for analyses of GAD65AA stability. Thus, the objective of this study was to establish whether the marker of islet cell autoimmunity, GAD65AA, is present in a population of older individuals with type 2 diabetes and to determine to what extent this marker is related to impaired glucose control and to specific treatment of diabetes by insulin or oral agents. This is the first large epidemiology study of GAD65AA in a U.S. population-based cohort of older individuals.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Cardiovascular health study
The CHS is a population-based longitudinal study of risk factors and subclinical disease related to the incidence and natural history of cardiovascular disease in noninstitutionalized adults 65 yr and older (19). A detailed description of recruitment for CHS has been published (18). Participants were recruited from four communities in the United States: Forsyth County, North Carolina; Sacramento County, California; Washington County, Maryland; and Pittsburgh, Pennsylvania (Allegheny County). A randomized systematic sampling of the Medicare files in each site yielded a total of 5201 eligible participants in the original cohort, including 2955 women (153 Black) and 2246 men (91 Black). Potential participants were excluded if they were institutionalized, confined to a wheelchair in the home, or had severe illness that was expected to lead to early death. Baseline visit took place from June 1989 through May 1990 (yr 2 of CHS). To increase African-American representation, in 1992–1993 (yr 5 of CHS), a "new cohort" of almost exclusively African-American individuals (98%) was enrolled in a similar manner as the "original cohort". A total of 687 (424 Black women and 249 Black men) participants were enrolled as part of this new cohort, for a total population of 5,888. Informed consent explaining all procedures and assessments was obtained from all participants before enrollment.
All CHS participants had a clinical examination and provided a medical history at baseline visit (yr 2 for the original cohort and yr 5 for the new cohort). Physical and laboratory evaluations were performed at baseline and repeated at yr 5 for the original cohort. Anthropometric assessments are available for each year from baseline through yr 11 of the study for both cohorts. Blood specimens, after a 12-h fast, were collected at baseline, yr 5, and yr 9 for complete blood analysis of lipids, glucose, and insulin, as well as other biochemical markers of cardiovascular disease risk. Serum glucose and insulin levels 2 h after a 75-g oral glucose load were also measured at baseline and yr 9. Diabetic patients treated with insulin or oral hypoglycemic agents (OHGA) were not administered the oral glucose tolerance test (OGTT). The OGTT was also not administered yr 5; therefore, baseline measures related to the OGTT were not available for the new cohort. A detailed description of CHS laboratory techniques and quality assurance methods has been reported (20).
Study population
For the current study, available stored baseline serum samples (n = 3318) from all CHS participants classified into the four following groups were tested for GAD65AA: 1) known diabetes (n = 685), 2) newly diagnosed diabetes (n = 651), 3) impaired glucose tolerance only (IGT; n = 1046), and 4) impaired fasting glucose (IFG; n = 553). In the main CHS, 742, 695, 1092, and 580 participants were classified with known diabetes, newly diagnosed diabetes, IGT only, and IFG, respectively. Therefore, 92.3, 93.7, 95.8, and 95.3% of the samples, respectively, were tested for GAD65AA.
Figure 1
depicts the selection criteria for the study population tested for GAD65AA. CHS participants reporting a positive medical history of diabetes and/or current treatment with insulin or OHGA were classified as having known diabetes. CHS participants not classified with known diabetes were classified as having diabetes at baseline if they had a fasting glucose equal to or greater than 126 mg/dl or a 2-h postchallenge glucose during an OGTT equal to or greater than 200 mg/dl. The diagnosis of diabetes was based on both the American Diabetes Association and 1985 World Health Organization diabetes classification criteria (21, 22). IFG was defined as a fasting glucose equal to or greater than 110 mg/dl and less than 126 mg/dl, and not including diabetes as defined above. IGT was defined as a fasting glucose less than 110 mg/dl and a 2-h glucose equal to or greater than 140 and less than 200 mg/dl. This latter group did not include participants with diabetes or IFG, and all groups were mutually exclusive. Classification of all study groups was based on baseline visit data.
|
Baseline demographic and clinical characteristics of the study population by glucose tolerance and diabetes status are presented in Table 1. Mean (SD) age and BMI of the overall study population were 73 (6) yr and 27.4 (4.8) kg/m2, respectively. The study population for the current report consisted of 1438 (43.3%) men and 1880 (56.7%) women and 560 (16.9%) Black and 2730 (82.3%) White CHS participants. Twenty-eight (0.8%) participants were of "other" race. Analyses by race excluded these participants and were limited to 3290 (99.2% of study cohort) participants of Black or White race. Data on duration of diabetes were not collected at baseline visit as part of the main CHS.
|
Coded serum samples assayed for GAD65AA were collected at baseline visit and stored at a central laboratory at the University of Vermont (Burlington, VT) in a –70 C freezer. All samples assayed for GAD65AA were detected in triplicate using in vitro transcribed/translated 35S-[Met]-labeled recombinant human GAD65, as originally described (23, 24). The GAD65 construct used for this study was donated by Dr. Åke Lernmark (University of Washington, Seattle, WA).
The results are expressed as an index (index = sample cpm – negative control cpm/positive control cpm – negative control cpm) as previously reported (25). The cut-off point for the GAD65AA assay was 0.069 and was established as the 99th percentile of autoantibody values calculated using 280 healthy control subjects, which have been described previously and are younger than the CHS cohort (25).
The coefficient of variation of the GAD65AA assay was previously reported (25). In-house laboratory thresholds for positivity gave excellent performance in multiple proficiency workshops, organized by the University of Florida in Gainesville (1995, 1996, and 1997) and the Diabetes Autoantibody Standardization Program (2000, 2002, and 2003; Ref. 26). Proficiency workshops results are summarized as follows: 76–100% sensitivity, 90–100% specificity, and 100% validity for GAD65AA.
Statistical analysis
Data were analyzed using the Statistical Analysis System Software on Windows operating system (Release 8.2; SAS Institute Inc., Cary, NC). Data are presented as mean ± SD for normally distributed variables, as percentages for categorical variables, or as median (25th to 75th percentile) for nonnormal continuous variables. Comparisons of group means were performed using the one-way ANOVA for normally distributed variables and the Kruskal-Wallis test for nonnormal continuous variables. The 
2 and Fisher’s exact test were used to test between-group differences for categorical variables. A logistic regression model was run to assess variables associated with being autoantibody positive in known diabetic individuals. Variables in this model included age, gender, race, BMI, systolic blood pressure, fasting glucose and insulin levels, and diabetes medication. For the regression analysis, glucose and insulin were log-transformed. All statistical tests were two-tailed, and P values of less than 0.05 were considered to be statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
GAD65AA exceeded the limit of normal in 88 of 3318 (2.7%) participants tested, corresponding to 5.3% with known diabetes, 2.2% with newly diagnosed diabetes, 1.9% with IGT only (original cohort only), 2.0% with IFG, and 1.8% of the normal glucose tolerant controls (Table 1
). The prevalence of GAD65AA was statistically significantly higher in participants with known diabetes compared with normoglycemic controls (P = 0.006) and increased with decreasing glucose tolerance (P < 0.001).
There were no statistically significant differences in the prevalence of GAD65AA by glucose tolerance status within the nondiabetic groups (IGT, IFG, and normal glycemic groups; P = 0.95). Furthermore, the presence of GAD65AA was more pronounced in the group with known diabetes. Therefore, subsequent comparisons by diabetes status are presented with the three nondiabetic groups collapsed into one (Fig. 2).
|
). However, similar to results in the overall study population, GAD65AA prevalence differed by diabetes status in all gender, race, and age groups. For some of these groups, differences by diabetes status did not reach statistical significance (notably, for the greater than 75 age group). Some of the subgroup analyses were limited by small numbers. In Blacks and the younger age group (65–69), GAD65AA positivity was higher among both diabetes groups compared with the nondiabetic control group (although not statistically significant; Fig. 2). This is in contrast to the overall population, Whites, and other subgroups in which GAD65AA positivity is more pronounced in the group with known diabetes.
Among participants with known diabetes, the prevalence of GAD65AA varied by current diabetes treatment (Fig. 3). There were 214, 295, 128, and 48 diabetic participants who reported the use of no medications, OHGA only, insulin only, and both insulin and OHGA, respectively. GAD65AA were found in 2.3, 5.8, 7.8, and 8.3% of known diabetic participants, respectively (P = 0.02, linear trend). A similar trend was found by race, gender, and age (data not shown).
|
Among diabetic individuals, there were no significant differences by GAD65AA status in gender, age, BMI, waist circumference, fasting insulin, high-density lipoprotein cholesterol, triglycerides, and diastolic blood pressure measured at baseline (Table 2). Diabetic patients who were positive for GAD65AA had statistically significantly higher fasting glucose levels and were more likely to be of Black race compared with GAD65AA-negative diabetic individuals (P < 0.05; Table 2). The former group also had lower systolic blood pressure although not statistically significant (P = 0.07). GAD65AA-positive diabetic individuals (n = 31, 86.1%) were more likely to report use of any diabetes medication at baseline compared with GAD65AA-negative diabetic patients (n = 440, 67.8%; P = 0.02), which was mainly due to the higher use of insulin among GAD65AA-positive diabetic individuals. The presence of GAD65AA was not associated with higher levels of markers of inflammation (i.e. C-reactive protein and fibrinogen).
|
Stability of GAD65AA in older individuals with type 2 diabetes
We were able to test follow-up samples (n = 311) for GAD65AA to serve as a secondary and separate analysis of the stability of autoantibody patterns. The samples tested consisted of blood samples from 264 CHS participants who developed diabetes and 47 who did not (as of their last follow-up) during the CHS study period (baseline to yr 9) and for whom stored blood specimens were available at both baseline and at least one follow-up visit (yr 5 and/or yr 9). These analyses exclude samples from participants classified at baseline with known diabetes. These analyses were separate from the main analyses of this paper and, unlike the main analyses, include longitudinal data. Table 3 shows GAD65AA status at baseline and follow-up visit by diabetes status for this sample. Among newly diagnosed diabetic participants, 98.2% were concordant and 1.9% were discordant for GAD65AA status. Corresponding rates for nondiabetic controls were 97.9 and 2.1%, respectively. Of the four cases that converted from GAD65AA negative to positive, two were diagnosed with diabetes at baseline and two at follow-up (yr 9). The corresponding nondiabetic control in this group was classified as IGT at baseline but did not develop diabetes at follow-up. Only one individual converted from GAD65AA positive to negative during the study period.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Few data are available on GAD65AA in the elderly population. Comparable to our data, but in a somewhat younger population (50–74 yr old) from The Netherlands (n = 2350), GAD65AA positivity was more pronounced in patients with known history of diabetes (3.5% in known diabetes, 0% in newly diagnosed diabetes, 2.4% in IGT, and 0.7% in normal glucose tolerant individuals; Ref. 27). In the 55- to 65-yr-old subgroup of the UK Prospective Diabetes Study (UKPDS) cohort with newly diagnosed type 2 diabetes, there was a GAD prevalence of 7% (10). This is higher than the prevalence of GAD65AA found in the CHS for the newly diagnosed diabetes group; however, the CHS was older than the UKPDS cohort. Similar to our findings of no difference in GAD65AA positivity between normal controls and nondiabetic abnormal glucose tolerant groups, a Finnish study reported a GAD AA prevalence in 383 controls, 558 IGT, and 1122 type 2 diabetic individuals of 4.4, 3.6, and 9.3%, respectively (9). They also found that individuals with onset at more than 45 yr had a lower prevalence of GAD65AA (8.2 vs. 19.3% for individuals diagnosed 45 yr of age or younger). In the oldest group, the prevalence was almost 6%, similar to our results. In contrast, a Swedish study of newly diagnosed diabetes patients reported a much higher prevalence rate of GAD AA (15%) in their smaller cohort of patients (n = 231) 65 yr of age or older (11). They also reported that the frequency of GAD AA positivity remained high after 12 yr. These findings are consistent with ours that GAD65AA remain stable in older individuals with diabetes.
Do the GAD65AA-positive individuals have type 1 diabetes? We have no data on the time of diabetes onset or diabetes treatment duration among CHS diabetic participants. However, it is extremely unlikely that the majority of these 70-yr-old or older diabetic individuals had severe type 1 diabetes since childhood, i.e. in the 1920s to the present, or that they have been treated only with OHGA.
In our previous analysis from NHANES III, diabetic individuals with GAD65AA had lower C-peptide levels suggesting decreased insulin secretion (17). The CHS did not measure C-peptide; however, GAD65AA-positive diabetic participants had higher fasting glucose levels and were more likely to be on insulin or OHGA in combination with insulin, suggesting a more severe insulin deficiency, especially in the absence of significant differences in waist circumference or BMI, markers of severe insulin resistance. The higher prevalence of GAD65AA positivity among Blacks is also consistent with our findings from the NHANES III. The reasons for their higher prevalence, although not statistically significant in this sample, are unknown. Population-based studies of GAD65AA in older individuals in studies with larger minority cohorts are needed.
We have previously reported in the CHS that, in diabetic individuals on insulin therapy, higher blood insulin levels are associated with higher mortality (28). Insulin therapy among older diabetics to control blood sugar may be necessary because of severe insulin resistance and inability to maintain high insulin secretion, especially among overweight and obese diabetics (29). These individuals have relatively high blood insulin levels before insulin therapy. It is possible that, in the approximate 8% of diabetic insulin or insulin and OHGA users who are GAD65AA positive, this diabetes may represent a primary severe insulin secretory deficiency that may be more likely to benefit from insulin therapy with regards to reduction of complications primarily related to diabetes, especially microvascular disease, similar to results in type 1 diabetes (30). Similarly, the 5.8% of OHGA users who are GAD65AA positive may actually require insulin therapy to reduce the risk of complications. Larger clinical trials evaluating different therapies for older diabetic patients should include measures of GAD65AA.
Although the prevalence of GAD65AA was only 5.3% among CHS participants with known diabetes, we have noted from analysis of NHANES III that there are probably more older diabetic individuals (
50 yr old) with GAD65AA-positive diabetes in the population than younger traditional type 1 insulin-dependent GAD65AA-positive diabetes, despite much higher prevalence of GAD65AA positivity among younger diabetic individuals (17). Therefore, understanding the etiology of GAD65AA positivity among older diabetics is very important.
Consistent with the literature concerning islet cell autoimmunity in younger age groups, the presence of GAD65AA appears to be related to insulin therapy in type 2 diabetic patients 65 yr of age or older (10, 13, 31, 32, 33). In the UKPDS, it was reported that, in type 2 diabetic patients 45 yr of age or older, the presence of GAD AA or islet cell cytoplasm AA alone was a weaker predictor of insulin requirement as compared with patients carrying both AA (10). The possibility that multiple rather than single antibodies to islet autoantigens may correlate with a high risk of progression to insulin requirement cannot be excluded in elderly patients.
In summary, GAD65AA may be useful markers in identifying a subgroup of approximately 6% of autoimmune diabetes in older patients presenting with type 2 diabetes. This estimate might be even higher when additional markers of islet cell autoimmunity are studied. These observations may provide new insight toward an adequate classification and treatment of type 2 diabetes and in turn lead to better knowledge and understanding of the autoimmune mechanisms involved in autoimmune diabetes of the elderly.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure summary: E.B.-M., L.H.K., S.P., Y.-J.Z., T.H., and M.P. have nothing to declare.
First Published Online May 23, 2006
Abbreviations: AA, Autoantibodies; BMI, body mass index; CHS, Cardiovascular Health Study; GAD, glutamic acid decarboxylase; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; NHANES III, Third National Health and Nutrition Examination Survey; OGTT, oral glucose tolerance test; OHGA, oral hypoglycemic agent; OR, odds ratio; UKPDS, UK Prospective Diabetes Study.
Received December 8, 2005.
Accepted May 11, 2006.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Pietropaolo, E. Barinas-Mitchell, and L. H. Kuller The Heterogeneity of Diabetes: Unraveling a Dispute: Is Systemic Inflammation Related to Islet Autoimmunity? Diabetes, May 1, 2007; 56(5): 1189 - 1197. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, P < 0.001 for known DM vs. nondiabetic groups adjusting for gender (A), race (B), or age (C). P = not significant, females vs. males (A), Whites vs. Blacks (B), or across age groups (C). The nondiabetic group consists of the normal-glycemic nondiabetic controls, IGT and IFG groups combined.
, P = 0.05 for treatment group vs. no diabetes medications group (No meds).