Original Article 857 Fasting Plasma Glucose and Incidence of Diabetes --- Implication for the Threshold for Impaired Fasting Glucose: Results from the Population-Based Omiya MA Cohort Study Masayuki Kato, Mitsuhiko Noda 2, Hiroshi Suga 3, Masahiko Matsumoto 3, and Yasunori Kanazawa 4 for the Omiya MA Cohort Study Group Japan Foundation for the Promotion of International Medical Research Cooperation, Tokyo, Japan 2 Department of Diabetes and Metabolic Medicine, International Medical Center of Japan, Tokyo, Japan 3 Omiya Medical Association, Saitama, Japan 4 Japan Diabetes Foundation, Tokyo, Japan Aim: In 23, the American Diabetes Association recommended that the threshold for diagnosing impaired fasting glucose (IFG) should be lowered from 6. mmol/l ( mg/dl) to 5.6 mmol/l ( mg/dl). To discuss the diagnostic threshold for IFG, the association between fasting plasma glucose (FPG) and the risk of future diabetes must be known; however, data regarding this relation in the Japanese population are scarce. The aim of this study was to determine the relation between FPG and the risk of future diabetes in the Japanese general population. Methods: A retrospective cohort study was conducted using data from annual health check-ups performed in Omiya city. A total of,369 subjects between the ages of 4 79 years who were not diabetic at baseline were followed for seven years. Diabetes was defined as FPG 26 mg/dl or self-report. Results: The incidence of diabetes increased as the baseline FPG level increased and a similar pattern was observed irrespective of sex or age. The hazard ratios compared with subjects with FPG 85 mg/ dl adjusted for possible confounding factors were 3.83 (95% confidence interval (95% CI); 2.4 6.8) for subjects with to 4 mg/dl FPG and 7.97 (95% CI; 4.98 2.4) for subjects with 5 to 9 mg/dl FPG. Conclusions: Subjects with 9 mg/dl FPG have an appreciable risk of diabetes that cannot be considered as normal and should be notified of their potential risk of developing diabetes. J Atheroscler Thromb, 29; 6:857-86. Key words; Diabetes, Impaired fasting glucose Introduction The impaired fasting glucose (IFG) category was first introduced in 997 in the revised diagnostic criteria for diabetes of the American Diabetes Association (ADA). This category was introduced as an intermediate group of subjects whose glucose levels, although not meeting criteria for diabetes, are never- Address for correspondence: Mitsuhiko Noda, Department of Diabetes and Metabolic Medicine, International Medical Center of Japan, -2- Toyama, Shinjuku-ku, Tokyo 62-8655, Japan E-mail: mnoda@imcj.hosp.go.jp Received: February 7, 29 theless too high to be considered altogether normal and was defined as a fasting plasma glucose (FPG) level 6. mmol/l but 7. mmol/l ( mg/dl and 26 mg/dl) ). The same cut-off values were used in the definition of IFG published by the World Health Organization (WHO) in 999 2). In 23, the ADA recommended that the threshold for diagnosing IFG should be lowered to 5.6 mmol/l ( mg/dl) 3) ; however, this new criterion has not been accepted by other groups. At its 25 Annual Meeting, the European Diabetes Epidemiology Group, a subgroup of the European Association for the Study of Diabetes, resolved to review the evidence for lowering the FPG threshold from 6. mmol/l to 5.6 mmol/l and recommended that the diagnostic cut-off point of 6.
858 mmol/l should be retained for IFG 4). WHO and the International Diabetes Federation also recommended that the FPG cut-off point for IFG should remain at 6. mmol/l 5). The Japan Diabetes Society (JDS) has defined normal type as those who have remained non-diabetic after follow-up for several years 6) ; therefore, the lower limit of IFG could be reasonably defined as the point at which the risk of diabetes increases and is appreciable, since the lower limit of the IFG category is also the upper limit of the normal category. To discuss the diagnostic threshold for IFG, the association between FPG and the risk of future diabetes must be known. Several papers have reported this association 7-) ; however, few data are available about the Japanese population. The aim of this study was to determine the association between FPG and the risk of future diabetes in the Japanese general population. Subjects and Methods This retrospective cohort study was conducted using data from annual health checkup performed in Omiya city by the Omiya Medical Association between 2 (baseline) and 27. The annual health checkup included a short questionnaire about medical conditions and lifestyle. The questionnaire asked about the status (not present, under treatment, cured, or left untreated) of several medical conditions, such as hypertension, cardiovascular disease, cancer and diabetes. Health check-up participants in 2, 67% of whom were women (median age for men and women was 65 and 6 years, respectively), were included in the present analysis if they were between the ages of 4 79 years and if their FPG data were available at baseline (in 2). Health check-up participants were not included in the analysis if they had any of the following conditions at baseline: cardiovascular disease, chronic liver disease, kidney disease or any type of cancer (n 3,44). Because the present study examined the incidence of diabetes, any health check-up participants with diabetes at baseline were also not included (n,772). Subjects who did not undergo an annual health checkup in 2 were subsequently excluded from the study because they had no follow-up data. The remaining cohort consisted of,369 subjects (3,35 men and 8,54 women; 7% of the whole subjects). Subjects were regarded as incident cases of diabetes if they became diabetic (FPG level 26 mg/dl or answered under treatment, cured, or left untreated to the question regarding diabetes status) for the first time during the course of the study. Subjects were Table. Baseline characteristics of subjects Age BMI FPG (mg/dl) History of Hypertension (yes) Alcohol Never Ex-drinker Occasional drinker Habitual drinker Smoking Never Ex-smoker Current Smoker (n,369) 62 (55 68) 22.8 (2.9) 92 (87 99) 28.7 52.6. 26.8 9.5 76.5 9. 4.5 regarded as censored cases if they missed annual health check-up, or underwent check-up without fasting, or did not answer the question regarding diabetes status for the first time during the course of the study. To examine the association between FPG and the risk of future diabetes, we calculated the incidence of diabetes categorizing baseline FPG with 5 mg/l intervals ( 85, 85 9, 9 94, 95 99, 4, 5 9, 4, 5 9 and 2 25 mg/dl). As the data regarding diabetes were obtained at one-year intervals, the data were analyzed using grouped survival time models. We used a complementary log-log regression model, which can be viewed as a grouped survival time version of the proportional hazard model 2), and calculated the hazard ratio adjusted for sex, age (categorized as 4 49, 5 59, 6 69, and 7 79 years), body mass index (categorized as 9, 9 2.9, 2 22.9, 23 24.9, 25 26.9, 27 28.9 and 29 kg/m 2 ), history of hypertension, family history of diabetes (yes, no, and missing), alcohol intake (never, ex-drinker, occasional drinker, and habitual drinker) and smoking status (never, ex-smoker, and current smoker). Results (n 3,35) 65 (6 7) 23.3 (2.7) 95 (89 2) 33.9 2. 2.2 28. 49.6 42.3 25. 32.6 (n 8,54) 6 (54 67) 22.6 (3.) 9 (86 98) 26.6 66..6 26.3 7. 9.5 2.4 7. Age and FPG are represented as the median (interquartile range), BMI is represented as the mean (standard deviation), and the other characteristics are represented as proportions. FPG, fasting plasma glucose BMI, body mass index 95% CI, 95% confidence interval The baseline characteristics of the subjects are shown in Table. During the 7-year follow-up period,
859 Density Density Density.3.2..2..2. <85 85-89 9-94 95-99 - 4 5-9 - 4 5-9 Baseline Fasting Plasma Glucose (mg/dl) developed diabetes NOT developed diabetes 2-25 Fig.. Distribution of baseline fasting plasma glucose levels of subjects. there were 72 incident cases (37 men and 395 women) of diabetes. The baseline FPG values of the subjects who developed diabetes and those who did not develop diabetes during the follow up period are shown in Fig.. The best cut-off point (maximizing sensitivity plus specificity) separating subjects who developed diabetes from subjects who did not develop diabetes was mg/dl (3 mg/dl for men and 99 mg/dl for women). The sensitivities and specificities of these values were.7 and.82 at mg/dl, respectively, and.47 and.97 at mg/dl, respectively (Fig. 2). The incidence rate according to the baseline FPG categories is shown in Fig. 3. The incidence rate increased as the baseline FPG value increased. A similar pattern was observed irrespective of sex or age. The hazard ratios according to the baseline FPG categories adjusted for sex, age, body mass index, history of hypertension, family history of diabetes, alcohol drinking status and smoking status are shown in Table 2. Compared with subjects with a baseline FPG 85 mg/dl, both men and women with an FPG of 95 to 99 mg/dl had a significantly greater risk of diabetes; the hazard ratios were 3.83 (2.52 for men and 4.57 for mewon) for subjects with an FPG Sensitivity.8.6.4.2 FPG mg/dl FPG mg/dl FPG mg/dl.2.4.6.8 -Specificity Fig.2. Receiver operating characteristic curve for developing diabetes against baseline fasting plasma glucose. level of to 4 mg/dl and 7.87 (5.55 for men and 9.8 for women) for subjects with an FPG level of 5 to 9 mg/dl. Discussion Originally, the IFG category was introduced as an intermediate group between diabetes and normal; however, defining normal is rather difficult. In the context of diabetes, it seems reasonable to define normal according to the risk of developing diabetes 6). Another way to define normal is to focus on the risk of cardiovascular disease (CVD) and CVD mortality; however, an increased CVD risk is one of the characteristics of both diabetes and IFG and is neither a necessary nor sufficient condition for diagnosing diabetes or IFG. Moreover, a recent study showed that CVD mortality in individuals with IFG can be explained by their conversion to a diabetic status 3) ; therefore, defining IFG by focusing on the risk of diabetes seems reasonable. In our analysis, the risk of diabetes increased continuously as the FPG level increased and at no point did the risk of diabetes increase rapidly; therefore, it was impossible to define the lower limit of IFG as such a point. However, the risk of diabetes in subjects with an FPG level of 4 mg/dl was about four-fold and that of subjects with an FPG level 5 9 mg/dl was about eight fold greater than that of
86 Incidence Rate (per person-years) 2 8 6 4 2 Age 4-59 Age 6-79 Age 6-79 Age 4-59 <85 85-89 9-94 95-99 -4 5-9 -4 5-9 2-25 Baseline Fasting Plasma Glucose (mg/dl) Fig.3. Incidence rate of diabetes mellitus according to baseline fasting plasma glucose levels. Table 2. ratios for the incidence of diabetes mellitus according to the baseline fasting plasma glucose levels Baseline FPG (mg/dl) n cases ratio 95% CI n cases ratio 2 95% CI n cases ratio 2 95% CI 85 (reference) 85 9 9 94 95 99 4 5 9 4 5 9 2 25,765 2,366 2,622,979,24 675 348 224 86 24 42 64 79 8 9 22.8.5 2.33 3.83 7.84 22. 32.9 54.9.72.96.94 2.4.47 3.7 2.4 6.8 4.98 2.4 4. 34.4 2.9 5.8 34.9 86.3 322 547 736 656 434 29 5 88 92 9 4 28 38 33 45 54 43 53.87.25.93 2.53 5.55 4.5 22.3 34..38 2.4.59 2.66.93 4..2 5.32 2.7.5 7. 29.7.8 46.2 6.6 69.7,443,89,886,323 77 385 98 36 94 5 28 36 4 47 45 68 58 57.34.54 2.35 4.57 9.8 26.9 4. 75.8.72 2.52.84 2.8.29 4.26 2.54 8.24 5.8 6.6 5.2 47.6 22.4 7.6 42. 36.4 adjusted for sex, age (4 49, 5 59, 6 69, and 7 79 years), body mass index ( 9, 9 2.9, 2 22.9, 23 24.9, 25 26.9, 27 28.9 and 29 kg/m 2 ), history of hypertension, family history of diabetes (yes, no, and missing), alcohol intake (never, ex-drinker, occasional drinker, and habitual drinker) and smoking status (never, ex-smoker, and current smoker) 2 adjusted for same variables (except sex) as subjects with an FPG level of 85 mg/dl; thus it does not seem reasonable to consider subjects with FPG levels within the former ranges as normal. Moreover, 7% of the subjects who developed diabetes during the course of the study had a baseline FPG level of to 25 mg/dl, while 47% had a baseline FPG level of to 25 mg/dl, that is to say, about one-fourth of the newly developed diabetes originated from a baseline FPG level of to 9 mg/dl. One of the main problems with lowering the FPG cut-off point for the diagnosis of IFG to mg/dl is the marked increase in the number of subjects with IFG. In this study, the prevalence of IFG increased to three-fold when the FPG cut-off point was lowered from to mg/dl. Despite this problem, subjects with FPG levels of 9 mg/dl had an appreciable risk of diabetes that cannot be considered normal, and such individuals should be notified of their potential risk of developing diabetes. One way of dealing with this problem would be to establish a risk stratification using other risk factors for diabetes. FPG is not the only risk factor for diabetes and risk stratification to identify subjects with a higher risk of diabetes using FPG levels and other risk factors seems feasible.
86 Although risk stratification using an oral glucose tolerance test may be best, several other risk scores for identifying individuals with a high risk of future diabetes that do not require an oral glucose tolerance test are also available 4-6). The present study had several limitations. First, this study was performed retrospectively using data from annual health check-ups, and information on subjects who did not undergo health check-ups on an annual basis was not available. We have treated these subjects as censored cases; however, this procedure may produce a bias if an association exists between the incidence of diabetes and participation in annual health check-ups (for example, subjects who become diabetic between annual health check-ups may tend either to undergo or to miss their next annual health check-up). Also, we defined diabetes as an FPG level of 26 mg/dl or more on one occasion. This definition might have included subjects with a transiently high FPG level, although diabetes should be defined as persistent hyperglycemia. We also analyzed the data by defining diabetes as two consecutive FPG levels 26 mg/dl, but the pattern of the resulting risk curve was similar to that shown in Fig. 3. Despite these limitations, this study is one of only a few studies to reveal an association between FPG and the future risk of diabetes in the Japanese general population. Acknowledgements This study was supported by a Community Medicine Research Grant from Saitama city and a grant from the Japan Diabetes Foundation. The authors are grateful to Haruhito Masuda (Omiya Medical Association) for his assistance. Study Group Members Members of the Omiya MA Cohort Study Group, Omiya Medical Association: H. Suga (Chair), M. Matsumoto, C. Satomura, MD, T. Nakamura, MD, K. Nakayama, MD, R. Kawaguchi, MD, M. Mitani, MD, and outside members: M. Noda, M. Kato, and Y. Kanazawa. References ) Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care, 997; 2: 83-97 2) World Health Organization: Definition, Diagnosis and Classification of Diabetes Mellitus and its Complications. Geneva: World Health Organization, 999 3) Genuth S, Alberti KG, Bennett P, Buse J, Defronzo R, Kahn R, Kitzmiller J, Knowler WC, Lebovitz H, Lernmark A, Nathan D, Palmer J, Rizza R, Saudek C, Shaw J, Steffes M, Stern M, Tuomilehto J, Zimmet P: Follow-up report on the diagnosis of diabetes mellitus. Diabetes Care, 23; 26: 36-367 4) Forouhi NG, Balkau B, Borch-Johnsen K, Dekker J, Glumer C, Qiao Q, Spijkerman A, Stolk R, Tabac A, Wareham NJ: The threshold for diagnosing impaired fasting glucose: a position statement by the European Diabetes Epidemiology Group. Diabetologia, 26; 49: 822-827 5) World_Health_Organization: Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia. Geneva: World Health Organization, 26 6) Kuzuya T, Nakagawa S, Satoh J, Kanazawa Y, Iwamoto Y, Kobayashi M, Nanjo K, Sasaki A, Seino Y, Ito C, Shima K, Nonaka K, Kadowaki T: Report of the Committee on the classification and diagnostic criteria of diabetes mellitus. Diabetes Res Clin Pract, 22; 55: 65-85 7) Dinneen SF, Maldonado D, Leibson CL, Klee GG, Li H, Melton LJ, Rizza RA: Effects of changing diagnostic criteria on the risk of developing diabetes. Diabetes Care, 998; 2: 48-43 8) Wareham NJ, Byrne CD, Williams R, Day NE, Hales CN: Fasting proinsulin concentrations predict the development of type 2 diabetes. Diabetes Care, 999; 22: 262-27 9) Shaw JE, Zimmet PZ, Hodge AM, de Courten M, Dowse GK, Chitson P, Tuomilehto J, Alberti KG: Impaired fasting glucose: how low should it go? Diabetes Care, 2; 23: 34-39 ) Gabir MM, Hanson RL, Dabelea D, Imperatore G, Roumain J, Bennett PH, Knowler WC: The 997 American Diabetes Association and 999 World Health Organization criteria for hyperglycemia in the diagnosis and prediction of diabetes. Diabetes Care, 2; 23: 8-2 ) Nichols GA, Hillier TA, Brown JB: Normal fasting plasma glucose and risk of type 2 diabetes diagnosis. Am J Med, 28; 2: 59-524 2) Prentice RL, Gloeckler LA: Regression analysis of grouped survival data with application to breast cancer data. Biometrics, 978; 34: 57-67 3) Rijkelijkhuizen JM, Nijpels G, Heine RJ, Bouter LM, Stehouwer CD, Dekker JM: High risk of cardiovascular mortality in individuals with impaired fasting glucose is explained by conversion to diabetes: the Hoorn study. Diabetes Care, 27; 3: 332-336 4) Stern MP, Williams K, Haffner SM: Identification of persons at high risk for type 2 diabetes mellitus: do we need the oral glucose tolerance test? Ann Intern Med, 22; 36: 575-58 5) Lindström J, Tuomilehto J: The diabetes risk score: a practical tool to predict type 2 diabetes risk. Diabetes Care, 23; 26: 725-73 6) Schmidt MI, Duncan BB, Bang H, Pankow JS, Ballantyne CM, Golden SH, Folsom AR, Chambless LE: Identifying individuals at high risk for diabetes: The Atherosclerosis Risk in Communities study. Diabetes Care, 25; 28: 23-28