- Email: [email protected]
Diabetes in older adults: comparison of 1997 American Diabetes Association classification of diabetes mellitus with 1985 WHO classification
ARTICLES
Diabetes in older adults: comparison of 1997 American Diabetes Association classification of diabetes mellitus with 1985 WHO classification Patricia W Wahl, Peter J Savage, Bruce M Psaty, Trevor J Orchard, John A Robbins, Russell P Tracy
Summary Background We aimed to compare the prevalence of abnormal glucose tolerance identified by the 1985 WHO and the 1997 American Diabetes Association (ADA) diagnostic categories based on information collected in the Cardiovascular Health Study, an epidemiological study of elderly people. Methods We recruited 672 African-Americans in 1992–93. We measured glucose concentrations during fasting and 2 h after a 75 g oral glucose-tolerance test in participants aged 65–100 years in the Cardiovascular Health Study. From a 1989 cohort, we analysed the glucose measurements of 4515 individuals without a previous diagnosis of diabetes and of 262 additional measurements. Findings In the 1989 cohort, the prevalence of untreated diabetes with ADA diagnostic fasting criteria was 7·7% versus a prevalence of 14·8% by the WHO criteria. In the AfricanAmerican cohort, the prevalence of untreated diabetes was 2·7% with ADA criteria and 11·8% with WHO criteria. 3509 (77%) of the 4515 participants in the 1989 cohort had normal glucose concentrations according to ADA fasting criteria, compared with 2401 (53·2%) according to WHO criteria. In the African-American cohort, the corresponding numbers were 239 (91·2%) versus 153 (58·4%). All differences in prevalence of abnormal glucose tolerance between ADA and WHO classifications were significant (p<0·0001). Interpretation Among elderly individuals, there was a significant difference in the prevalence of diabetes identified by the WHO diagnostic criteria and the ADA fasting criteria. Consequently, many individuals currently classified as nondiabetic according to ADA criteria would previously have had a diagnosis of diabetes according to WHO criteria. Longitudinal studies are needed to assess the value of ADA criteria in the identification of individuals at increased risk of diabetesassociated chronic complications.
Lancet 1998; 352: 1012–15
Departments of Biostatistics (Prof P W Wahl PhD) and Medicine and Epidemiology (Prof B M Psaty MD), School of Public Health and Community Medicine, University of Washington, Seattle, WA, USA; Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Seattle (P J Savage MD); Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh (Prof T J Orchard MMed); Department of Medicine, School of Medicine, University of California, Sacramento (Prof J A Robbins MD); and Departments of Pathology and Biochemistry, School of Medicine, University of Vermont, Vermont (Prof R P Tracy PhD) Correspondence to: Dr Patricia W Wahl, Cardiovascular Health Study Coordinating Center, Century Square, Suite 2025, Seattle, WA 98101, USA (e-mail: [email protected])
1012
Introduction In 1997, an Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, sponsored by the American Diabetes Association (ADA), published new guidelines recommending aetiology-based criteria for the diagnosis of diabetes mellitus.1 A major advantage of the new criteria is that fasting glucose concentrations can be used alone, and these are easier to obtain and more stable than the 2 h values used in the WHO2 and National Diabetes Data Group3 criteria. The approach used to establish the new ADA diagnostic criteria was to identify the fasting glucose value that would achieve about the same estimates of diabetes prevalence as the 1985 WHO diagnostic criteria, which included results from the 2 h glucose tolerance test. The ADA report presented results from NHANES III based on glucose measurements in individuals aged 40–74 years and gave the prevalence of undiagnosed diabetes as 6·34% with the 1985 WHO criteria and 4·35% with the new ADA criteria.1,4 The decrease in identified diabetes was judged acceptable because the new classification scheme based on fasting glucose alone would be more widely used. The difference between the two classification criteria, however, was greater among older adults, in whom the prevalence of mild hyperglycaemia is very high. Furthermore, the ADA report did not give age-specific data for the elderly (75 years). Thus, if the effect of fasting for 2 h changes with age, the results of the new ADA fasting criteria may be different in older and middleaged people. With the ageing of the population in the USA and many other developed countries, diabetes and its cardiovascular consequences will have an increasing impact on the public’s health.5–7 Thus, it is important to assess the effect of use of ADA and WHO diagnostic criteria in an elderly cohort since diabetes is most common in this age-group. The Cardiovascular Health Study, which includes a group of African-Americans, measured fasting and 2 h glucose concentrations in more than 4000 men and women aged 65–100 years, and provides an opportunity to assess the effects of the different diabetic criteria in a large elderly population. Our aim was to compare the prevalence of abnormal glucose tolerance in this cohort with WHO diagnostic criteria versus ADA fasting criteria.
Methods Participants 5201 participants in the Cardiovascular Health Study were identified through a sample of Medicare eligibility lists, stratified by age and sex, from four communities: Forsyth County, North Carolina; Sacramento County, California; Washington County, Maryland; and Pittsburgh, Pennsylvania. We excluded individuals who were living in institutions, wheelchair-bound, or undergoing active treatment for cancer. Details of the Cardiovascular Health Study design and recruitment have been reported elsewhere.8,9 A supplemental cohort of 672 exclusively African-American participants of similar age was recruited 3 years after the main cohort, and their fasting and 2 h glucose concentrations were measured 5 years after recruitment.
THE LANCET • Vol 352 • September 26, 1998
ARTICLES
Methods After an overnight fast, venepuncture was done within the first hour in a clinic visit and 2 h after a 75 g oral glucose load was given to participants who were not taking medication for diabetes. Serum concentrations of glucose were measured at the University of Vermont Cardiovascular Health Study Central Blood Laboratory with a Kodak Ektachem 700 Analyser (Eastman Kodak, Rochester, NY, USA). A control pool with a mean glucose serum concentration of 5·05 mmol/L was used to calculate an average monthly coefficient of variation of 0·93% and an overall coefficient variation of 1·86%.10 The 1985 WHO criteria define diabetes as a fasting glucose concentration of greater than or equal to 7·8 mmol/L or a 2 h glucose concentration of 11·1 mmol/L or higher. According to the WHO criteria, individuals with fasting glucose concentrations of less than 7·8 mmol/L have impaired glucose tolerance (IGT) if their 2 h glucose is between 7·8 mmol/L and 11·05 mmol/L. The 1997 ADA fasting criteria classify diabetes as a fasting glucose concentration of greater than or equal to 7·0 mmol/L, and impaired fasting glucose (IFG) is defined as fasting glucose between 5·8 mmol/L and 6·9 mmol/L. A participant was judged to have prevalent cardiovascular disease at the time of the baseline clinical examination if they had had myocardial infarction, bypass surgery, angioplasty, angina, stroke, or transient ischaemic attack. Participants were classified as diabetic if they were identified during a medication inventory as users of insulin, or oral hypoglycaemic agents. We included in our analyses only those participants who had fasted for at least 9 h, had both fasting and 2 h glucose measurements, and were not receiving antidiabetic treatment. In the 1989 cohort, we excluded 386 (7·4%) participants who were receiving treatment for diabetes (296 on oral hypoglycaemic agents, 100 on exogenous insulin, and ten on both). 91 (13·5%) participants with treated diabetes were excluded from the AfricanAmerican cohort (55 on oral hypoglycaemic agents, 37 on exogenous insulin, and one on both). Thus, 4515 participants in the 1989 cohort and 262 participants in the African-American cohort were included in our analysis.
Statistical analysis Our ability to detect differences in prevalence estimates of at least 5% was greater than 80% in the 1989 cohort but less than 50% in the African-American cohort. To assess the degree of agreement between the WHO and ADA classification schemes, we used the statistic, although the glucose-impaired category (IGT or IFG) is not identical in the two schemes. Values of the statistic greater than 0·75 indicate good agreement, whereas values less than 0·25 represent poor agreement.11 We estimated the correlations between age and fasting and 2 h glucose values by means of Spearman’s correlation coefficient; tests for correlation differences were based on Fisher’s Z transformation.12 We used McNemar’s test to test the difference in the prevalence of diabetes between the two diagnostic classifications. The 2 test was used to analyse the association between WHO and ADA diagnostic categories, sex, and age (65–69, 70–74, 75–79, 80 years).12
Results 1989 cohort The mean age of the participants was 73 (SD 5·6) years. There were 2574 women and 1941 men; 4304 of the participants were white. The mean fasting glucose was 5·77 (1·22) mmol/L and the median value was 5·55 mmol/L. The mean 2 h glucose was 8·16 (3·18) mmol/L and the median value was 7·54 mmol/L. The correlations of age with fasting glucose and with 2 h glucose were 0·020 and 0·103, respectively; the two correlations differed significantly (p<0·0001). There were large discrepancies in the prevalence of diabetes between the WHO and ADA diagnostic criteria.
THE LANCET • Vol 352 • September 26, 1998
1985 WHO criteria
1997 ADA criteria
WHO total/ prevalence
Normal
IFG
New diabetes
Normal Row % Column % IGT Row % Column New diabetes Row % Column %
2198 91·5% 62·6% 1102 76·1% 31·4% 209 31·4% 6·0%
185 7·7% 28·2% 301 20·8% 45·8% 171 25·7% 26·0%
18 0·7% 5·2% 45 3·1% 12·9% 286 42·9% 81·9%
ADA total/prevalence
3509/77·7%
657/14·6%
349/7·7%
2401/53·2%
1448/32·1%
666/14·8%
4515
Table 1: Comparison of WHO and ADA diagnostic categories in 1989 cohort
Table 1 shows the glycaemic category of the 4515 individuals in the 1989 cohort cross-tabulated by the WHO (rows) and ADA (columns) classification schemes. The prevalence of normoglycaemia was 77·7% with ADA fasting criteria, versus 53·2% with WHO criteria. By contrast, the prevalence of IFG with ADA fasting criteria was 14·6%, which was less than half the 32·1% prevalence of IGT with WHO criteria. Similarly, for untreated diabetes the prevalence was 7·7% with ADA criteria versus 14·8% with WHO criteria.
WHO classification Of 1448 participants who had IGT according to WHO criteria, 1102 (76·1%) were normoglycaemic according to ADA fasting criteria. Of the 666 participants classified as having untreated diabetes by WHO criteria, only 286 (42·9%) were classified as diabetic, 171 (25·7%) had IFG, and 209 (31·4%) were normoglycaemic according to ADA fasting criteria. There was a significant association between WHO diagnostic categories and age (p<0·0001), whereas there was no association with sex (p=0·662; table 2). ADA classification Of 3509 participants classified as normoglycaemic according to ADA fasting criteria, 1102 (31·4%) had IFG and 209 (6·01%) had untreated diabetes according to WHO criteria. Of 657 participants with IFG according to ADA fasting criteria, 185 (28·2%) were normoglycaemic, 301 (45·9) had IGT, and 171 (26%) had diabetes according to WHO criteria (table 1). Of the 349 participants with diabetes according to ADA criteria, 286 (82%) were similarly classified by WHO criteria. There was no association between the ADA glycaemic categories and age (p=0·207), whereas there was a significant association with sex (p<0·0001; table 2). Overall, among participants who had had no previous treatment for diabetes, substantially fewer men and women were identified with untreated diabetes by ADA fasting criteria than by WHO criteria. 1102 (76%) participants who had IGT according to WHO criteria were classified as Age (years)
Women (n=2589)
Men (n=1926)
Total (n=4151)
65–69 Row % Column % 70–74 Row % Column % 75–79 Row % Column % 80 Row % Column %
978 62·7% 37·8% 799 55·9% 30·9% 515 56·3% 19·9% 297 48·7% 11·5%
583 37·3% 30·3% 631 44·1% 32·8% 399 43·7% 20·7% 313 51·3% 16·3%
1561 (34·6%)
1430 (31·7%)
914 (20·2%)
610 (13·5%)
Table 2: Age categories by sex for 1989 cohort
1013
ARTICLES
1985 WHO criteria
WHO total/ prevalence
1985 WHO criteria
0
153/58·4%
Normal IGT New diabetes
533 (24%) 337 (31%) 64 (31%)
46 (25%) 91 (30%) 53 (31%)
4 (22%) 16 (36%) 96 (34%)
0
78/29·8%
ADA total
934 (27%)
190 (29%)
116 (33%)
1997 ADA criteria Normal
IFG
New diabetes
Normal Row % Column % IGT Row % Column New diabetes Row % Column %
150 98·0% 62·8% 74 94·9% 31·0% 15 48·4% 6·3%
3 2·0% 18·8% 4 5·1% 25·0% 9 29·0% 56·3%
ADA total/prevalence
239/91·2%
16/6·1%
7 22·6% 100·0%
31/11·8%
7/2·7%
262
normoglycaemic by ADA fasting criteria. The statistic to measure agreement between the two classification schemes was low: 0·275 (SE 0·012).
African-American cohort The mean age of the African-American participants was 74 (5·1) years; there were 165 women and 97 men. The mean fasting glucose was 5·27 (0·75) mmol/L and the median value was 5·16 mmol/L. The mean 2 h glucose was 7·83 (2·85) mmol/L and the median was 7·33 mmol/L. The correlations of age with fasting glucose and 2 h glucose were 0·044 and 0·060, respectively; the difference between the correlations was not significant (p>0·05), but the sample size was small. Table 3 compares the glycaemic status of AfricanAmericans with each classification scheme. The WHO classification results were similar to those in the 1989 cohort, whereas the results with ADA fasting criteria had a lower prevalence of IFG and diabetes. 74 (94·9%) of African-American participants had IGT according to WHO criteria (table 3), but were classified as normal by ADA fasting criteria. There was no association between WHO or ADA categories and age or sex, but the power to detect such differences was low (table 4). The statistic to measure agreement between the two classification schemes was low: 0·136 (SE 0·039). Thus, differences of similar size magnitude and direction were seen in both cohorts. Cardiovascular disease Table 5 shows the proportion of participants in the 1989 cohort who had cardiovascular disease at the time of the baseline clinic visit. For each classification scheme, the rate of cardiovascular disease increased from 24% in the normal glycaemic category to 32% in the new diabetes category, according to WHO criteria and from 27% to 33% with ADA fasting criteria. Of the 209 participants in the 1989 who had diabetes according to WHO criteria and were classified as normal with ADA criteria, 64 (31%) had cardiovascular disease. By contrast, of the 18 participants who had diabetes Age (years)
Women (n=165) 57 37·6% 34·5% 53 57·0% 32·1% 35 77·8% 21·2% 20 80·0% 12·1%
Men (n=97) 42 42·4% 43·3% 40 43·0% 41·2% 10 22·2% 10·3% 5 20·0% 5·2%
Total (n=262) 99 (37·8%)
93 (35·5%)
45 (17·2%)
25 (9·5%)
Table 4: Age categories by sex for African-American cohort
1014
Normal
WHO total
IFG
New diabetes 583 (24%) 444 (31%) 213 (32%) 4515
386 with known diabetes.
Table 3: Comparison of WHO and ADA diagnostic categories in African-American cohort
65–69 Row % Column % 70–74 Row % Column % 75–79 Row % Column % 80 Row % Column %
1997 ADA criteria
Table 5: Prevalence of cardiovascular disease by WHO and ADA diagnostic categories in 1989 cohort
according to ADA criteria who were classified as normal with WHO criteria, four (22%) had cardiovascular disease.
Discussion Among elderly men and women, use of the 1985 WHO diagnostic criteria, which have been widely used in population studies, and the 1997 ADA fasting criteria for the diagnosis of diabetes and abnormal glucose tolerance gave widely different diagnoses. The 1989 AfricanAmerican cohorts in the Cardiovascular Health Study provided two independent estimates of the prevalence of new diabetes. In both cohorts, our estimates of diabetes prevalence based on WHO classification criteria were about two to four times higher than the corresponding estimates with ADA fasting criteria. One reason for a different prevalence of diabetes among the elderly may be the increased sensitivity of the 2 h glucose values to ageing. Although neither fasting nor 2 h glucose measurements were strongly associated with age in either of our age-limited cohorts, the correlation of age with 2 h glucose was stronger than that of age with fasting glucose. Also, in the 1989 cohort, WHO diagnostic categories were significantly associated with age, whereas ADA categories, based on fasting glucose values, were not. The sensitivity of the oral glucose tolerance test to age was reported by Andres13 in 1971 when he suggested use of a higher cut-off point for the diagnosis of diabetes in the elderly. This sensitivity has also been noted by Harris and colleagues14 who found that, among individuals participating in the National Health and Nutrition Examination Survey, fasting glucose values were not related to age, but 2 h glucose values increased significantly with age. These observations suggest that WHO criteria, which use 2 h glucose values, might be expected to give higher estimated prevalence of untreated diabetes in our older participants Stolk and colleagues15 suggested that the use of fasting blood glucose alone would be most appropriate for epidemiological studies. Such studies focus on accurate estimates of the prevalence of diabetes, rather than in making a diagnosis in a patient for possible treatment. Our aim was to compare prevalence estimates of untreated diabetes according to WHO criteria and ADA fasting criteria. We conclude that, compared with WHO criteria, ADA fasting criteria significantly underestimate the prevalence of untreated diabetes in older adults. To obtain equivalent WHO and ADA diabetes prevalence estimates in our population, we would have to increase the cut-off point of 2 h glucose values to 13·32 mmol/L or reduce the fasting glucose cut-off point to 6·38 mmol/L. If the observed differences between WHO and ADA fasting criteria found in the 1989 Cardiovascular Health Study cohort were to be applied to the US population, which in 1995 included 33 532 000 people aged 65 years and older, there would be a reduction in the estimated prevalence of diabetes of about 2 million. This example shows how important it is to determine just what degree of
THE LANCET • Vol 352 • September 26, 1998
ARTICLES
glucose intolerance needs to be identified and treated. If the Diabetes Prevention Trial that is currently under way indicates benefit in treating patients with IGT, this finding will underscore the need for accurate identification of the large number of early individuals at risk of diabetes. We compared WHO criteria with only the ADA fasting criteria. The new ADA recommendations allow use of fasting glucose, 2 h glucose, or any blood glucose over 11·1 mmol/L. If we had used the additional ADA criteria in the analysis, the ADA estimates would have been higher. The effect of the potential misclassification on the diagnosis of diabetes is unknown. Although some misclassification may arise from the higher variation of 2 h glucose values than of fasting glucose within an individual, the stronger association of 2 h glucose with age seems to be the major cause of this discrepancy in the elderly.16,17 Of possible concern are those older individuals who have diabetes according to WHO classification but who are not classified as diabetic by ADA fasting criteria. However, if the adverse effects of hyperglycaemia on excess microvascular and macrovascular disease are concentrated among individuals with diabetes according to ADA fasting criteria, the new recommendations are an improvement in specificity that will enable treatment to be targeted at those individuals at greatest risk of diabetes. If, however, substantial excess complications occur in individuals who are classified as diabetic by WHO criteria but not by ADA criteria, WHO criteria will still be of clinical value. In general, the consequences of diabetes among the elderly are less well understood than among younger individuals. The selection of precise cut-off points is, of necessity, somewhat arbitrary because of the scarcity of data. On the basis of our results, we make no recommendation as to the use of one set of criteria over the other. Our purpose is to point out that the substantial difference between WHO and ADA diagnostic criteria in estimating diabetes prevalence and in the identification of people with diabetes in an elderly population. The overall prevalence of cardiovascular disease in individuals classified as newly diabetic by either classification scheme is similar, although we found 31% of participants classified as non-diabetic by ADA criteria would be classified as diabetic by WHO criteria and would have cardiovascular disease. However, it is not possible to decide which criteria are better on the basis of prevalence estimates. From the point of view of public health, the classification system that best predicts incident microvascular and macrovascular disease will serve to identify the patients in most need of treatment. Additional analyses to compare the incidence of macrovascular and microvascular disease with the two sets of criteria are under way. Contributors Each investigator participated actively in the production of this paper. All are investigators on the Cardiovascular Health Study, a longitudinal study of cardiovascular disease in 5688 adults 65 years of age and older, funded by the National Heart, Lung and Blood Institute. Peter Savage contributed a national NIH perspective on the implications of our findings, and helped with the estimates of how many potential cases of diabetes could be missed under the ADA criteria. Bruce Psaty contributed to the initial structure and content of the paper. Trevor Orchard helped identify some of the most recent studies on current thinking on the project. John Robbins provided some of the clinical interpretation and made valuable suggestions to the tables and text. Russell Tracy did the glucose assays and provided an invaluable interpretation of the values.
John Chen, Beverly Tucker); and Bowman Gray School of Medicine-EKG Reading Center (Farida Rautaharju, Pentti Rautaharju). Sacramento County, CA: University of California, Davis (William Bommer, Charles Bernick, Andrew Duxbury, Mary Haan, Calvin Hirsch, Lawrence Laslett, John Robbins, Richard White). Washington County, MD: The Johns Hopkins University (M Jan Busby-Whitehead, Joyce Chabot, George W Comstock, Adrian Dobs, Linda P Fried, Joel G Hill, Steven J Kittner, Shiriki Kumanyika, David Levine, Joao A Lima, Neil R Powe, Thomas R Price, Jeff Williamson, Moyses Szklo, Melvin Tockman); and MRI Reading Center-Washington County, MD, The Johns Hopkins University (R Nick Bryan, Norman Beauchamp, Carolyn C Meltzer, Naiyer Iman, Douglas Fellows, Melanie Hawkins, Patrice Holtz, Michael Kraut, Grace Lee, Larry Schertz, Cynthia Quinn, Earl P Steinberg, Scott Wells, Linda Wilkins, Nancy C Yue). Allegheny County, PA: University of Pittsburgh (Diane G Ives, Charles A Jungreis, Laurie Knepper, Lewis H Kuller, Elaine Meilahn, Peg Meyer, Roberta Moyer, Anne Newman, Richard Schulz, Vivienne E Smith, Sidney K Wolfson); Echocardiography Reading Center (Follow-Up) Georgetown Medical Center (John Gottdiener, Eva Hausner, Stephen Kraus, Judy Gay, Sue Livengood, Mary Ann Yohe, Retha Webb); Ultrasound Reading Center—Geisinger Medical Center (Daniel H O’Leary, Joseph F Polak, Laurie Funk); Central Blood Analysis Laboratory, University of Vermont (Edwin Bovill, Elaine Cornell, Mary Cushman, Russell P Tracy); Respiratory Sciences, University of Arizona-Tucson (Paul Enright); Coordinating Center, University of Washington, Seattle (Alice Arnold, Annette L Fitzpatrick, Bonnie K Lind, Richard A Kronmal, Bruce M Psaty, David S Siscovick, Lynn Shemanski, Will Longstreth, Patricia W Wahl, David Yanez, Paula Diehr, Maryann McBurnie, Chuck Spiekerman, Scott Emerson, Cathy Tangen, Priscilla Velentgas); and NHLBI Project Office (Diane E Bild, Robin Boineau, Teri A Manolio, Peter J Savage, Patricia Smith).
Acknowledgments This work was supported by the National Heart, Lung, and Blood Institute Contract N01-HC-87079.
References 1
2 3
4
5
6 7 8 9
10
11 12 13 14
15 16
Participating institutions and principal staff Forsyth County, NC: Bowman Gray School of Medicine of Wake Forest University (Gregory L Burke, Sharon Jackson, Alan Elster, Walter H Ettinger, Curt D Furberg, Gerardo Heiss, Dalane Kitzman, Margie Lamb, David S Lefkowitz, Mary F Lyles, Cathy Nunn, Ward Riley,
THE LANCET • Vol 352 • September 26, 1998
17
American Diabetes Association. Report of the Expert Committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 1997; 7: 1183–97. WHO. WHO Expert Committee on diabetes mellitus. Second report. (Tech Rep Serv no 646). Geneva: WHO, 1980. Natonal Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979; 28: 1039–57. Harris MI, Eastman RC, Cowie CC, Flegal KM, Eberhardt MS. Comparison of diabetes diagnostic categories in the US population according to 1997 American Diabetes Association and 1980–1985 World Association and 1980–1985 World Health Organization diagnostic criteria. Diabetes Care 1997; 29: 1859–62. Savage PJ. Cardiovascular complications of diabetes mellitus: what we know and what we need to know about their prevention. Ann Intern Med 1996; 124: 123–36. Harris MI. Undiagnosed NIDDM: clinical and public health issues. Diabetes Care 1993; 16: 642–52. Harris MI, Eastman RC. Early detection of undiagnosed non-insulindependent diabetes mellitus. JAMA 1996; 276: 1261–62. Fried LP, Borhani NO, Enright P, et al. The Cardiovascular Health Study: design and rationale. Ann Epidemiol 1991; 1: 263–76. Tell G, Fried LP, Hermanson B, Manolio TA, Newman AB, Borhani NO. Recruitment of adults 65 years and older as participants in the Cardiovascular Health Study. Ann Epidemiol 1993; 3: 358–66. Cushman M, Cornell ES, Howard PR, Boville EG, Tracy RP. Laboratory methods and quality assurance in the Cardiovascular Health Study. Clin Chem 1995; 41: 264–70. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas 1960; 20: 37–46. Rosner B. Fundamentals of biostatistics, 4th edn. London: Duxbury Press, 1997: 514–15. Andres R. Aging and diabetes. Symposium on diabetes mellitus. Med Clin North Am 1971; 55: 835–46. Harris MI, Hadden WC, Knowler WC, Bennett PH. Prevalence of diabetes and impaired glucose tolerance and plasma glucose levels in US population aged 20–74 yr. Diabetes 1987; 36: 523–34. Stolk RP, Orchard TJ, Grobbee DE. Why use the oral glucose tolerance test? Diabetes Care 1995; 18: 1045–49. McCance DR, Hanson RL, Pettett DJ, Bennett PH, Hadden DR, Knowler WC. Diagnosing diabetes mellitus—do we need new criteria? Diabetologia 1997; 40: 247–55. Mooy JM, Gootenbuis PA, de Vries H, et al. Intra-individual variation of glucose, specific insulin and proinsulin concentrations measured by two oral glucose tolerance tests in a general Caucasian population: the Hoorn study. Diabetologia 1996; 39: 298–305.
1015
Diabetes in older adults: comparison of 1997 American Diabetes Association classification of diabetes mellitus with 1985 WHO classification Patricia W Wahl, Peter J Savage, Bruce M Psaty, Trevor J Orchard, John A Robbins, Russell P Tracy
Summary Background We aimed to compare the prevalence of abnormal glucose tolerance identified by the 1985 WHO and the 1997 American Diabetes Association (ADA) diagnostic categories based on information collected in the Cardiovascular Health Study, an epidemiological study of elderly people. Methods We recruited 672 African-Americans in 1992–93. We measured glucose concentrations during fasting and 2 h after a 75 g oral glucose-tolerance test in participants aged 65–100 years in the Cardiovascular Health Study. From a 1989 cohort, we analysed the glucose measurements of 4515 individuals without a previous diagnosis of diabetes and of 262 additional measurements. Findings In the 1989 cohort, the prevalence of untreated diabetes with ADA diagnostic fasting criteria was 7·7% versus a prevalence of 14·8% by the WHO criteria. In the AfricanAmerican cohort, the prevalence of untreated diabetes was 2·7% with ADA criteria and 11·8% with WHO criteria. 3509 (77%) of the 4515 participants in the 1989 cohort had normal glucose concentrations according to ADA fasting criteria, compared with 2401 (53·2%) according to WHO criteria. In the African-American cohort, the corresponding numbers were 239 (91·2%) versus 153 (58·4%). All differences in prevalence of abnormal glucose tolerance between ADA and WHO classifications were significant (p<0·0001). Interpretation Among elderly individuals, there was a significant difference in the prevalence of diabetes identified by the WHO diagnostic criteria and the ADA fasting criteria. Consequently, many individuals currently classified as nondiabetic according to ADA criteria would previously have had a diagnosis of diabetes according to WHO criteria. Longitudinal studies are needed to assess the value of ADA criteria in the identification of individuals at increased risk of diabetesassociated chronic complications.
Lancet 1998; 352: 1012–15
Departments of Biostatistics (Prof P W Wahl PhD) and Medicine and Epidemiology (Prof B M Psaty MD), School of Public Health and Community Medicine, University of Washington, Seattle, WA, USA; Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Seattle (P J Savage MD); Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh (Prof T J Orchard MMed); Department of Medicine, School of Medicine, University of California, Sacramento (Prof J A Robbins MD); and Departments of Pathology and Biochemistry, School of Medicine, University of Vermont, Vermont (Prof R P Tracy PhD) Correspondence to: Dr Patricia W Wahl, Cardiovascular Health Study Coordinating Center, Century Square, Suite 2025, Seattle, WA 98101, USA (e-mail: [email protected])
1012
Introduction In 1997, an Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, sponsored by the American Diabetes Association (ADA), published new guidelines recommending aetiology-based criteria for the diagnosis of diabetes mellitus.1 A major advantage of the new criteria is that fasting glucose concentrations can be used alone, and these are easier to obtain and more stable than the 2 h values used in the WHO2 and National Diabetes Data Group3 criteria. The approach used to establish the new ADA diagnostic criteria was to identify the fasting glucose value that would achieve about the same estimates of diabetes prevalence as the 1985 WHO diagnostic criteria, which included results from the 2 h glucose tolerance test. The ADA report presented results from NHANES III based on glucose measurements in individuals aged 40–74 years and gave the prevalence of undiagnosed diabetes as 6·34% with the 1985 WHO criteria and 4·35% with the new ADA criteria.1,4 The decrease in identified diabetes was judged acceptable because the new classification scheme based on fasting glucose alone would be more widely used. The difference between the two classification criteria, however, was greater among older adults, in whom the prevalence of mild hyperglycaemia is very high. Furthermore, the ADA report did not give age-specific data for the elderly (75 years). Thus, if the effect of fasting for 2 h changes with age, the results of the new ADA fasting criteria may be different in older and middleaged people. With the ageing of the population in the USA and many other developed countries, diabetes and its cardiovascular consequences will have an increasing impact on the public’s health.5–7 Thus, it is important to assess the effect of use of ADA and WHO diagnostic criteria in an elderly cohort since diabetes is most common in this age-group. The Cardiovascular Health Study, which includes a group of African-Americans, measured fasting and 2 h glucose concentrations in more than 4000 men and women aged 65–100 years, and provides an opportunity to assess the effects of the different diabetic criteria in a large elderly population. Our aim was to compare the prevalence of abnormal glucose tolerance in this cohort with WHO diagnostic criteria versus ADA fasting criteria.
Methods Participants 5201 participants in the Cardiovascular Health Study were identified through a sample of Medicare eligibility lists, stratified by age and sex, from four communities: Forsyth County, North Carolina; Sacramento County, California; Washington County, Maryland; and Pittsburgh, Pennsylvania. We excluded individuals who were living in institutions, wheelchair-bound, or undergoing active treatment for cancer. Details of the Cardiovascular Health Study design and recruitment have been reported elsewhere.8,9 A supplemental cohort of 672 exclusively African-American participants of similar age was recruited 3 years after the main cohort, and their fasting and 2 h glucose concentrations were measured 5 years after recruitment.
THE LANCET • Vol 352 • September 26, 1998
ARTICLES
Methods After an overnight fast, venepuncture was done within the first hour in a clinic visit and 2 h after a 75 g oral glucose load was given to participants who were not taking medication for diabetes. Serum concentrations of glucose were measured at the University of Vermont Cardiovascular Health Study Central Blood Laboratory with a Kodak Ektachem 700 Analyser (Eastman Kodak, Rochester, NY, USA). A control pool with a mean glucose serum concentration of 5·05 mmol/L was used to calculate an average monthly coefficient of variation of 0·93% and an overall coefficient variation of 1·86%.10 The 1985 WHO criteria define diabetes as a fasting glucose concentration of greater than or equal to 7·8 mmol/L or a 2 h glucose concentration of 11·1 mmol/L or higher. According to the WHO criteria, individuals with fasting glucose concentrations of less than 7·8 mmol/L have impaired glucose tolerance (IGT) if their 2 h glucose is between 7·8 mmol/L and 11·05 mmol/L. The 1997 ADA fasting criteria classify diabetes as a fasting glucose concentration of greater than or equal to 7·0 mmol/L, and impaired fasting glucose (IFG) is defined as fasting glucose between 5·8 mmol/L and 6·9 mmol/L. A participant was judged to have prevalent cardiovascular disease at the time of the baseline clinical examination if they had had myocardial infarction, bypass surgery, angioplasty, angina, stroke, or transient ischaemic attack. Participants were classified as diabetic if they were identified during a medication inventory as users of insulin, or oral hypoglycaemic agents. We included in our analyses only those participants who had fasted for at least 9 h, had both fasting and 2 h glucose measurements, and were not receiving antidiabetic treatment. In the 1989 cohort, we excluded 386 (7·4%) participants who were receiving treatment for diabetes (296 on oral hypoglycaemic agents, 100 on exogenous insulin, and ten on both). 91 (13·5%) participants with treated diabetes were excluded from the AfricanAmerican cohort (55 on oral hypoglycaemic agents, 37 on exogenous insulin, and one on both). Thus, 4515 participants in the 1989 cohort and 262 participants in the African-American cohort were included in our analysis.
Statistical analysis Our ability to detect differences in prevalence estimates of at least 5% was greater than 80% in the 1989 cohort but less than 50% in the African-American cohort. To assess the degree of agreement between the WHO and ADA classification schemes, we used the statistic, although the glucose-impaired category (IGT or IFG) is not identical in the two schemes. Values of the statistic greater than 0·75 indicate good agreement, whereas values less than 0·25 represent poor agreement.11 We estimated the correlations between age and fasting and 2 h glucose values by means of Spearman’s correlation coefficient; tests for correlation differences were based on Fisher’s Z transformation.12 We used McNemar’s test to test the difference in the prevalence of diabetes between the two diagnostic classifications. The 2 test was used to analyse the association between WHO and ADA diagnostic categories, sex, and age (65–69, 70–74, 75–79, 80 years).12
Results 1989 cohort The mean age of the participants was 73 (SD 5·6) years. There were 2574 women and 1941 men; 4304 of the participants were white. The mean fasting glucose was 5·77 (1·22) mmol/L and the median value was 5·55 mmol/L. The mean 2 h glucose was 8·16 (3·18) mmol/L and the median value was 7·54 mmol/L. The correlations of age with fasting glucose and with 2 h glucose were 0·020 and 0·103, respectively; the two correlations differed significantly (p<0·0001). There were large discrepancies in the prevalence of diabetes between the WHO and ADA diagnostic criteria.
THE LANCET • Vol 352 • September 26, 1998
1985 WHO criteria
1997 ADA criteria
WHO total/ prevalence
Normal
IFG
New diabetes
Normal Row % Column % IGT Row % Column New diabetes Row % Column %
2198 91·5% 62·6% 1102 76·1% 31·4% 209 31·4% 6·0%
185 7·7% 28·2% 301 20·8% 45·8% 171 25·7% 26·0%
18 0·7% 5·2% 45 3·1% 12·9% 286 42·9% 81·9%
ADA total/prevalence
3509/77·7%
657/14·6%
349/7·7%
2401/53·2%
1448/32·1%
666/14·8%
4515
Table 1: Comparison of WHO and ADA diagnostic categories in 1989 cohort
Table 1 shows the glycaemic category of the 4515 individuals in the 1989 cohort cross-tabulated by the WHO (rows) and ADA (columns) classification schemes. The prevalence of normoglycaemia was 77·7% with ADA fasting criteria, versus 53·2% with WHO criteria. By contrast, the prevalence of IFG with ADA fasting criteria was 14·6%, which was less than half the 32·1% prevalence of IGT with WHO criteria. Similarly, for untreated diabetes the prevalence was 7·7% with ADA criteria versus 14·8% with WHO criteria.
WHO classification Of 1448 participants who had IGT according to WHO criteria, 1102 (76·1%) were normoglycaemic according to ADA fasting criteria. Of the 666 participants classified as having untreated diabetes by WHO criteria, only 286 (42·9%) were classified as diabetic, 171 (25·7%) had IFG, and 209 (31·4%) were normoglycaemic according to ADA fasting criteria. There was a significant association between WHO diagnostic categories and age (p<0·0001), whereas there was no association with sex (p=0·662; table 2). ADA classification Of 3509 participants classified as normoglycaemic according to ADA fasting criteria, 1102 (31·4%) had IFG and 209 (6·01%) had untreated diabetes according to WHO criteria. Of 657 participants with IFG according to ADA fasting criteria, 185 (28·2%) were normoglycaemic, 301 (45·9) had IGT, and 171 (26%) had diabetes according to WHO criteria (table 1). Of the 349 participants with diabetes according to ADA criteria, 286 (82%) were similarly classified by WHO criteria. There was no association between the ADA glycaemic categories and age (p=0·207), whereas there was a significant association with sex (p<0·0001; table 2). Overall, among participants who had had no previous treatment for diabetes, substantially fewer men and women were identified with untreated diabetes by ADA fasting criteria than by WHO criteria. 1102 (76%) participants who had IGT according to WHO criteria were classified as Age (years)
Women (n=2589)
Men (n=1926)
Total (n=4151)
65–69 Row % Column % 70–74 Row % Column % 75–79 Row % Column % 80 Row % Column %
978 62·7% 37·8% 799 55·9% 30·9% 515 56·3% 19·9% 297 48·7% 11·5%
583 37·3% 30·3% 631 44·1% 32·8% 399 43·7% 20·7% 313 51·3% 16·3%
1561 (34·6%)
1430 (31·7%)
914 (20·2%)
610 (13·5%)
Table 2: Age categories by sex for 1989 cohort
1013
ARTICLES
1985 WHO criteria
WHO total/ prevalence
1985 WHO criteria
0
153/58·4%
Normal IGT New diabetes
533 (24%) 337 (31%) 64 (31%)
46 (25%) 91 (30%) 53 (31%)
4 (22%) 16 (36%) 96 (34%)
0
78/29·8%
ADA total
934 (27%)
190 (29%)
116 (33%)
1997 ADA criteria Normal
IFG
New diabetes
Normal Row % Column % IGT Row % Column New diabetes Row % Column %
150 98·0% 62·8% 74 94·9% 31·0% 15 48·4% 6·3%
3 2·0% 18·8% 4 5·1% 25·0% 9 29·0% 56·3%
ADA total/prevalence
239/91·2%
16/6·1%
7 22·6% 100·0%
31/11·8%
7/2·7%
262
normoglycaemic by ADA fasting criteria. The statistic to measure agreement between the two classification schemes was low: 0·275 (SE 0·012).
African-American cohort The mean age of the African-American participants was 74 (5·1) years; there were 165 women and 97 men. The mean fasting glucose was 5·27 (0·75) mmol/L and the median value was 5·16 mmol/L. The mean 2 h glucose was 7·83 (2·85) mmol/L and the median was 7·33 mmol/L. The correlations of age with fasting glucose and 2 h glucose were 0·044 and 0·060, respectively; the difference between the correlations was not significant (p>0·05), but the sample size was small. Table 3 compares the glycaemic status of AfricanAmericans with each classification scheme. The WHO classification results were similar to those in the 1989 cohort, whereas the results with ADA fasting criteria had a lower prevalence of IFG and diabetes. 74 (94·9%) of African-American participants had IGT according to WHO criteria (table 3), but were classified as normal by ADA fasting criteria. There was no association between WHO or ADA categories and age or sex, but the power to detect such differences was low (table 4). The statistic to measure agreement between the two classification schemes was low: 0·136 (SE 0·039). Thus, differences of similar size magnitude and direction were seen in both cohorts. Cardiovascular disease Table 5 shows the proportion of participants in the 1989 cohort who had cardiovascular disease at the time of the baseline clinic visit. For each classification scheme, the rate of cardiovascular disease increased from 24% in the normal glycaemic category to 32% in the new diabetes category, according to WHO criteria and from 27% to 33% with ADA fasting criteria. Of the 209 participants in the 1989 who had diabetes according to WHO criteria and were classified as normal with ADA criteria, 64 (31%) had cardiovascular disease. By contrast, of the 18 participants who had diabetes Age (years)
Women (n=165) 57 37·6% 34·5% 53 57·0% 32·1% 35 77·8% 21·2% 20 80·0% 12·1%
Men (n=97) 42 42·4% 43·3% 40 43·0% 41·2% 10 22·2% 10·3% 5 20·0% 5·2%
Total (n=262) 99 (37·8%)
93 (35·5%)
45 (17·2%)
25 (9·5%)
Table 4: Age categories by sex for African-American cohort
1014
Normal
WHO total
IFG
New diabetes 583 (24%) 444 (31%) 213 (32%) 4515
386 with known diabetes.
Table 3: Comparison of WHO and ADA diagnostic categories in African-American cohort
65–69 Row % Column % 70–74 Row % Column % 75–79 Row % Column % 80 Row % Column %
1997 ADA criteria
Table 5: Prevalence of cardiovascular disease by WHO and ADA diagnostic categories in 1989 cohort
according to ADA criteria who were classified as normal with WHO criteria, four (22%) had cardiovascular disease.
Discussion Among elderly men and women, use of the 1985 WHO diagnostic criteria, which have been widely used in population studies, and the 1997 ADA fasting criteria for the diagnosis of diabetes and abnormal glucose tolerance gave widely different diagnoses. The 1989 AfricanAmerican cohorts in the Cardiovascular Health Study provided two independent estimates of the prevalence of new diabetes. In both cohorts, our estimates of diabetes prevalence based on WHO classification criteria were about two to four times higher than the corresponding estimates with ADA fasting criteria. One reason for a different prevalence of diabetes among the elderly may be the increased sensitivity of the 2 h glucose values to ageing. Although neither fasting nor 2 h glucose measurements were strongly associated with age in either of our age-limited cohorts, the correlation of age with 2 h glucose was stronger than that of age with fasting glucose. Also, in the 1989 cohort, WHO diagnostic categories were significantly associated with age, whereas ADA categories, based on fasting glucose values, were not. The sensitivity of the oral glucose tolerance test to age was reported by Andres13 in 1971 when he suggested use of a higher cut-off point for the diagnosis of diabetes in the elderly. This sensitivity has also been noted by Harris and colleagues14 who found that, among individuals participating in the National Health and Nutrition Examination Survey, fasting glucose values were not related to age, but 2 h glucose values increased significantly with age. These observations suggest that WHO criteria, which use 2 h glucose values, might be expected to give higher estimated prevalence of untreated diabetes in our older participants Stolk and colleagues15 suggested that the use of fasting blood glucose alone would be most appropriate for epidemiological studies. Such studies focus on accurate estimates of the prevalence of diabetes, rather than in making a diagnosis in a patient for possible treatment. Our aim was to compare prevalence estimates of untreated diabetes according to WHO criteria and ADA fasting criteria. We conclude that, compared with WHO criteria, ADA fasting criteria significantly underestimate the prevalence of untreated diabetes in older adults. To obtain equivalent WHO and ADA diabetes prevalence estimates in our population, we would have to increase the cut-off point of 2 h glucose values to 13·32 mmol/L or reduce the fasting glucose cut-off point to 6·38 mmol/L. If the observed differences between WHO and ADA fasting criteria found in the 1989 Cardiovascular Health Study cohort were to be applied to the US population, which in 1995 included 33 532 000 people aged 65 years and older, there would be a reduction in the estimated prevalence of diabetes of about 2 million. This example shows how important it is to determine just what degree of
THE LANCET • Vol 352 • September 26, 1998
ARTICLES
glucose intolerance needs to be identified and treated. If the Diabetes Prevention Trial that is currently under way indicates benefit in treating patients with IGT, this finding will underscore the need for accurate identification of the large number of early individuals at risk of diabetes. We compared WHO criteria with only the ADA fasting criteria. The new ADA recommendations allow use of fasting glucose, 2 h glucose, or any blood glucose over 11·1 mmol/L. If we had used the additional ADA criteria in the analysis, the ADA estimates would have been higher. The effect of the potential misclassification on the diagnosis of diabetes is unknown. Although some misclassification may arise from the higher variation of 2 h glucose values than of fasting glucose within an individual, the stronger association of 2 h glucose with age seems to be the major cause of this discrepancy in the elderly.16,17 Of possible concern are those older individuals who have diabetes according to WHO classification but who are not classified as diabetic by ADA fasting criteria. However, if the adverse effects of hyperglycaemia on excess microvascular and macrovascular disease are concentrated among individuals with diabetes according to ADA fasting criteria, the new recommendations are an improvement in specificity that will enable treatment to be targeted at those individuals at greatest risk of diabetes. If, however, substantial excess complications occur in individuals who are classified as diabetic by WHO criteria but not by ADA criteria, WHO criteria will still be of clinical value. In general, the consequences of diabetes among the elderly are less well understood than among younger individuals. The selection of precise cut-off points is, of necessity, somewhat arbitrary because of the scarcity of data. On the basis of our results, we make no recommendation as to the use of one set of criteria over the other. Our purpose is to point out that the substantial difference between WHO and ADA diagnostic criteria in estimating diabetes prevalence and in the identification of people with diabetes in an elderly population. The overall prevalence of cardiovascular disease in individuals classified as newly diabetic by either classification scheme is similar, although we found 31% of participants classified as non-diabetic by ADA criteria would be classified as diabetic by WHO criteria and would have cardiovascular disease. However, it is not possible to decide which criteria are better on the basis of prevalence estimates. From the point of view of public health, the classification system that best predicts incident microvascular and macrovascular disease will serve to identify the patients in most need of treatment. Additional analyses to compare the incidence of macrovascular and microvascular disease with the two sets of criteria are under way. Contributors Each investigator participated actively in the production of this paper. All are investigators on the Cardiovascular Health Study, a longitudinal study of cardiovascular disease in 5688 adults 65 years of age and older, funded by the National Heart, Lung and Blood Institute. Peter Savage contributed a national NIH perspective on the implications of our findings, and helped with the estimates of how many potential cases of diabetes could be missed under the ADA criteria. Bruce Psaty contributed to the initial structure and content of the paper. Trevor Orchard helped identify some of the most recent studies on current thinking on the project. John Robbins provided some of the clinical interpretation and made valuable suggestions to the tables and text. Russell Tracy did the glucose assays and provided an invaluable interpretation of the values.
John Chen, Beverly Tucker); and Bowman Gray School of Medicine-EKG Reading Center (Farida Rautaharju, Pentti Rautaharju). Sacramento County, CA: University of California, Davis (William Bommer, Charles Bernick, Andrew Duxbury, Mary Haan, Calvin Hirsch, Lawrence Laslett, John Robbins, Richard White). Washington County, MD: The Johns Hopkins University (M Jan Busby-Whitehead, Joyce Chabot, George W Comstock, Adrian Dobs, Linda P Fried, Joel G Hill, Steven J Kittner, Shiriki Kumanyika, David Levine, Joao A Lima, Neil R Powe, Thomas R Price, Jeff Williamson, Moyses Szklo, Melvin Tockman); and MRI Reading Center-Washington County, MD, The Johns Hopkins University (R Nick Bryan, Norman Beauchamp, Carolyn C Meltzer, Naiyer Iman, Douglas Fellows, Melanie Hawkins, Patrice Holtz, Michael Kraut, Grace Lee, Larry Schertz, Cynthia Quinn, Earl P Steinberg, Scott Wells, Linda Wilkins, Nancy C Yue). Allegheny County, PA: University of Pittsburgh (Diane G Ives, Charles A Jungreis, Laurie Knepper, Lewis H Kuller, Elaine Meilahn, Peg Meyer, Roberta Moyer, Anne Newman, Richard Schulz, Vivienne E Smith, Sidney K Wolfson); Echocardiography Reading Center (Follow-Up) Georgetown Medical Center (John Gottdiener, Eva Hausner, Stephen Kraus, Judy Gay, Sue Livengood, Mary Ann Yohe, Retha Webb); Ultrasound Reading Center—Geisinger Medical Center (Daniel H O’Leary, Joseph F Polak, Laurie Funk); Central Blood Analysis Laboratory, University of Vermont (Edwin Bovill, Elaine Cornell, Mary Cushman, Russell P Tracy); Respiratory Sciences, University of Arizona-Tucson (Paul Enright); Coordinating Center, University of Washington, Seattle (Alice Arnold, Annette L Fitzpatrick, Bonnie K Lind, Richard A Kronmal, Bruce M Psaty, David S Siscovick, Lynn Shemanski, Will Longstreth, Patricia W Wahl, David Yanez, Paula Diehr, Maryann McBurnie, Chuck Spiekerman, Scott Emerson, Cathy Tangen, Priscilla Velentgas); and NHLBI Project Office (Diane E Bild, Robin Boineau, Teri A Manolio, Peter J Savage, Patricia Smith).
Acknowledgments This work was supported by the National Heart, Lung, and Blood Institute Contract N01-HC-87079.
References 1
2 3
4
5
6 7 8 9
10
11 12 13 14
15 16
Participating institutions and principal staff Forsyth County, NC: Bowman Gray School of Medicine of Wake Forest University (Gregory L Burke, Sharon Jackson, Alan Elster, Walter H Ettinger, Curt D Furberg, Gerardo Heiss, Dalane Kitzman, Margie Lamb, David S Lefkowitz, Mary F Lyles, Cathy Nunn, Ward Riley,
THE LANCET • Vol 352 • September 26, 1998
17
American Diabetes Association. Report of the Expert Committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 1997; 7: 1183–97. WHO. WHO Expert Committee on diabetes mellitus. Second report. (Tech Rep Serv no 646). Geneva: WHO, 1980. Natonal Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979; 28: 1039–57. Harris MI, Eastman RC, Cowie CC, Flegal KM, Eberhardt MS. Comparison of diabetes diagnostic categories in the US population according to 1997 American Diabetes Association and 1980–1985 World Association and 1980–1985 World Health Organization diagnostic criteria. Diabetes Care 1997; 29: 1859–62. Savage PJ. Cardiovascular complications of diabetes mellitus: what we know and what we need to know about their prevention. Ann Intern Med 1996; 124: 123–36. Harris MI. Undiagnosed NIDDM: clinical and public health issues. Diabetes Care 1993; 16: 642–52. Harris MI, Eastman RC. Early detection of undiagnosed non-insulindependent diabetes mellitus. JAMA 1996; 276: 1261–62. Fried LP, Borhani NO, Enright P, et al. The Cardiovascular Health Study: design and rationale. Ann Epidemiol 1991; 1: 263–76. Tell G, Fried LP, Hermanson B, Manolio TA, Newman AB, Borhani NO. Recruitment of adults 65 years and older as participants in the Cardiovascular Health Study. Ann Epidemiol 1993; 3: 358–66. Cushman M, Cornell ES, Howard PR, Boville EG, Tracy RP. Laboratory methods and quality assurance in the Cardiovascular Health Study. Clin Chem 1995; 41: 264–70. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas 1960; 20: 37–46. Rosner B. Fundamentals of biostatistics, 4th edn. London: Duxbury Press, 1997: 514–15. Andres R. Aging and diabetes. Symposium on diabetes mellitus. Med Clin North Am 1971; 55: 835–46. Harris MI, Hadden WC, Knowler WC, Bennett PH. Prevalence of diabetes and impaired glucose tolerance and plasma glucose levels in US population aged 20–74 yr. Diabetes 1987; 36: 523–34. Stolk RP, Orchard TJ, Grobbee DE. Why use the oral glucose tolerance test? Diabetes Care 1995; 18: 1045–49. McCance DR, Hanson RL, Pettett DJ, Bennett PH, Hadden DR, Knowler WC. Diagnosing diabetes mellitus—do we need new criteria? Diabetologia 1997; 40: 247–55. Mooy JM, Gootenbuis PA, de Vries H, et al. Intra-individual variation of glucose, specific insulin and proinsulin concentrations measured by two oral glucose tolerance tests in a general Caucasian population: the Hoorn study. Diabetologia 1996; 39: 298–305.
1015