|
A correction has been published: N Engl J Med 2006;354(22):2401.
|
Background The normal fasting plasma glucose level was recently defined as less than 100 mg per deciliter (5.55 mmol per liter). Whether higher fasting plasma glucose levels within this range independently predict type 2 diabetes in young adults is unclear.
Methods We obtained blood measurements, data from physical examinations, and medical and lifestyle information from men in the Israel Defense Forces who were 26 to 45 years of age.
Results A total of 208 incident cases of type 2 diabetes occurred during 74,309 person-years of follow-up (from 1992 through 2004) among 13,163 subjects who had baseline fasting plasma glucose levels of less than 100 mg per deciliter. A multivariate model, adjusted for age, family history of diabetes, body-mass index, physical-activity level, smoking status, and serum triglyceride levels, revealed a progressively increased risk of type 2 diabetes in men with fasting plasma glucose levels of 87 mg per deciliter (4.83 mmol per liter) or more, as compared with those whose levels were in the bottom quintile (less than 81 mg per deciliter [4.5 mmol per liter], P for trend <0.001). In multivariate models, men with serum triglyceride levels of 150 mg per deciliter (1.69 mmol per liter) or more, combined with fasting plasma glucose levels of 91 to 99 mg per deciliter (5.05 to 5.50 mmol per liter), had a hazard ratio of 8.23 (95 percent confidence interval, 3.6 to 19.0) for diabetes, as compared with men with a combined triglyceride level of less than 150 mg per deciliter and fasting glucose levels of less than 86 mg per deciliter (4.77 mmol per liter). The joint effect of a body-mass index (the weight in kilograms divided by the square of the height in meters) of 30 or more and a fasting plasma glucose level of 91 to 99 mg per deciliter resulted in a hazard ratio of 8.29 (95 percent confidence interval, 3.8 to 17.8), as compared with a body-mass index of less than 25 and a fasting plasma glucose level of less than 86 mg per deciliter.
Conclusions Higher fasting plasma glucose levels within the normoglycemic range constitute an independent risk factor for type 2 diabetes among young men, and such levels may help, along with body-mass index and triglyceride levels, to identify apparently healthy men at increased risk for diabetes.
Our investigation, which involved the use of data from the Metabolic, Lifestyle, and Nutrition Assessment in Young Adults (MELANY) study, assessed whether fasting plasma glucose levels can help to identify young, healthy, normoglycemic persons at increased risk for type 2 diabetes.
Methods
The Melany Study
The MELANY study has been conducted at the Israel Defense Forces Staff Periodic Examination Center, to which all career service personnel older than 25 years of age are referred every three to five years. A computerized database established in 1992 is the source of data for MELANY, which was designed to investigate risk factors for common diseases in young adults. At each visit to the Staff Periodic Examination Center, participants completed a detailed questionnaire assessing demographic, nutritional, lifestyle, and medical factors. Thereafter, blood samples were drawn after a 14-hour fast and analyzed. A trained medical technician measured height and weight, and a physician at the center performed a complete physical examination. Primary care for all Israel Defense Forces personnel between scheduled visits to the center is obtained at designated military clinics, and all medical information was recorded in the same central database, thereby facilitating ongoing, tight, and uniform follow-up.
The institutional review board of the Israel Defense Forces Medical Corps approved this study on the basis of strict maintenance of participants' anonymity during database analyses. Data from subjects were recorded anonymously, and no individual consent was obtained. The authors are solely responsible for the design of the study, analysis and interpretation of the data, and writing of the manuscript, without any form of censorship or limitation by the Israel Defense Forces.
Inclusion and Exclusion Criteria
Included in the study have been 13,163 men with fasting plasma glucose levels of less than 100 mg per deciliter at their initial Staff Periodic Examination Center visit, for whom follow-up data have been available through either a subsequent scheduled visit or visits (average number of visits per person, 2.5; range, 2 to 6) or from the primary physician for men who have received a diagnosis of diabetes. The ongoing cohort of the MELANY study currently includes 9538 additional men for whom follow-up data are not yet available. Patients were excluded from the study if they had confirmed type 1 or type 2 diabetes at the time of enrollment. Women were not included, since only 11 new cases of diabetes were diagnosed among 1961 normoglycemic women, an insufficient number of incident cases to facilitate meaningful analysis. Glucose-tolerance tests, shown rarely to be impaired in people with fasting glucose levels of less than 100 mg per deciliter,16 were not performed, a decision consistent with current clinical guidelines for the diagnosis of diabetes in young, asymptomatic, normoglycemic persons.17
Outcome Definitions
The diagnosis of type 2 diabetes was defined as the primary end point of the study. All cases of diabetes were diagnosed according to the criteria published by the American Diabetes Association expert committee.18 The diagnoses of all 208 new cases of diabetes in the MELANY study were made on the basis of two fasting plasma glucose levels of 126 mg per deciliter (7.00 mmol per liter) or more. Since the diagnostic criteria for diabetes were changed during the follow-up period, all fasting glucose values were revised to identify subjects with fasting plasma glucose levels of 126 to 140 mg per deciliter (7.00 to 7.77 mmol per liter). Of the 208 subjects with diabetes, 26 received a diagnosis before July 1997, and three cases were detected when the new diagnostic criteria were applied to the population already enrolled in the study at that time. During the follow-up period, two patients received a diagnosis of type 1 diabetes and were excluded from the study. An end-point determination was made at each sequential Staff Periodic Examination Center visit according to measurements of fasting plasma glucose. Alternatively, between visits, the diagnosis of diabetes was made by the Israel Defense Forces primary care physician on the basis of the same diagnostic criteria described above.18 Three army physicians reviewed and confirmed each of the cases before recording them in the central medical corps database.
Laboratory Methods
Biochemical analyses of blood were performed on fresh samples in a core laboratory facility that handles 1.2 million samples per year. The laboratory is authorized to perform tests according to the international quality standard ISO-9002. Periodic assessment of quality control (by the British company National External Quality Assessment Service) has been performed on a regular basis. To ensure that venous fasting plasma glucose levels were reliably determined, blood samples were collected in tubes containing sodium fluoride and delivered to the laboratory within two hours. All measured biochemical markers were identified with the use of a BM/Hitachi917 automated analyzer (Boehringer Mannheim). Blood pressure measurements were performed by medical technicians with the use of mercury sphygmomanometers.
Statistical Analysis
We excluded 3784 of the 16,947 young men at enrollment, 403 because of preexisting diabetes and 3381 because of impaired fasting plasma glucose levels (100 mg per deciliter or more). For analysis, we included 13,163 normoglycemic men with baseline fasting glucose levels of less than 100 mg per deciliter. A general linear model was used to assess the age-adjusted means and proportions of the population's characteristics across quintiles of fasting glucose levels and to fit the median of the quintiles as a continuous variable to estimate the trend of variables across quintiles. We conducted a Cox proportional-hazards analysis during each interval of follow-up to estimate the hazard ratios and 95 percent confidence intervals for the development of type 2 diabetes. We added the values for body-mass index (the weight in kilograms divided by the square of the height in meters) and triglyceride levels separately to the age-adjusted model to evaluate the potential role of each as a confounder of the tested association between fasting plasma glucose level and diabetes. In the final multivariate model, we controlled for age, family history of diabetes, body-mass index, serum triglyceride levels, physical-activity level, and smoking status. We tested for effect modification with stratified analyses of body-mass index, triglyceride levels, and family history of diabetes, all of which remained independent risk factors for diabetes in the multivariate model. Interaction terms were computed by modeling the quintile medians as continuous variables. Next, we evaluated the joint risk attributed to fasting plasma glucose levels (categorized according to the bottom two quintiles, the median quintile, and the top two quintiles) with either body-mass index (<25, 25 to 29.9, and 
30) or with triglyceride levels (<150 or 150 mg per deciliter [1.69 mmol per liter]). We calculated the population attributable risk as previously described.19 All statistical analyses were performed with the use of SAS statistical software, version 8.0.
Results
Data from 13,163 apparently healthy men (mean age, 32 years; range, 26 to 45) with fasting plasma glucose levels of less than 100 mg per deciliter at baseline were analyzed. Age-adjusted values for body-mass index, triglyceride levels, and the proportion of men with a family history of diabetes were more likely to increase across quintiles of fasting glucose levels (Table 1).
|
|
In the multivariate model, serum triglyceride levels and family history of diabetes, in addition to fasting plasma glucose levels and body-mass index, remained independent risk factors for the development of diabetes (data not shown). Thus, we further assessed the association between fasting plasma glucose levels and the occurrence of type 2 diabetes among strata of these independent risk factors (Table 3). In the multivariate models, increased levels of normal fasting plasma glucose were more strongly associated with diabetes among overweight and obese men (those with a body-mass index of 25 or more) than among leaner men (P for interaction, 0.03). The trend of increased risk of type 2 diabetes across increasing quintiles of normal fasting plasma glucose levels appeared to be similar among subgroups classified according to triglyceride levels and family-history status (P for interaction >0.05).
|
|
In this follow-up study of 13,163 apparently healthy young adult men, we found an increased risk of type 2 diabetes across quintiles of fasting plasma glucose levels within the newly defined normal range; this increase was independent of other traditional risk factors for diabetes. Our findings suggest that among young adults, who generally have a relatively low incidence of diabetes, elevated normal fasting plasma glucose levels may predict type 2 diabetes.
Several limitations of this study warrant consideration. First, the MELANY cohort may be considered representative of a unique group of healthy young men. However, the characteristics of the population are strikingly similar to those of cohorts in published studies of young men from various industrialized countries,20,21,22,23,24 and the relatively homogeneous environment to which participants in our study were exposed might reduce the effect of unknown confounders. Second, although they did not compromise the outcome definition, measurements of circulating insulin, C-peptide, or both were not obtained in this study, limiting our ability to assess the role of insulin resistance in the association between normal fasting plasma glucose levels and diabetes. Finally, we did not measure glycosylated hemoglobin levels or perform glucose-tolerance tests. Although the current definition of normal fasting plasma glucose levels resulted in a substantial increase in the overlap with normal glucose tolerance, as defined by glucose-tolerance testing,8 we may have missed men with normal fasting plasma glucose levels who were already glucose intolerant at enrollment. To limit this possibility, we confirmed our results by performing a secondary analysis in which a two-year lag between enrollment and outcome was used. The strengths of the MELANY study include the detailed, uniform, and systematic follow-up and outcome definition; the use of measured (rather than reported) values for the body-mass index; the availability of reliable determinations of glucose levels in fresh venous blood; and the direct measurements of lipids.
The identification of a high-normal fasting plasma glucose level as a risk factor for type 2 diabetes may help to identify young, healthy men for whom preventive interventions might be considered. Indeed, a number of strategies, including lifestyle modification14 and medications such as metformin,14 thiazolidinediones,13 acarbose,11,12 and orlistat15 have been reported as efficient interventions that may delay the onset of diabetes in selected groups that have classic risk factors for the disease. If such strategies are also found to be efficacious in preventing diabetes in young men with high-normal fasting plasma glucose levels, the findings of our study may facilitate efforts to halt the diabetes pandemic that is increasingly affecting people in the third to fifth decades of life.10
An impaired fasting plasma glucose level is a known risk factor for diabetes, along with other traditional risk factors such as a family history, sedentary lifestyle, central adiposity, dyslipidemia, and hypertension.25 However, the definition of a normal fasting plasma glucose level was recently revised to be less than 100 mg per deciliter.2 It is interesting to note that a few studies have reported the absence of a threshold in the association between fasting plasma glucose levels and the risk of diabetes in cohorts with a wide age range. Furthermore, a fasting plasma glucose level of 94 mg per deciliter (5.22 mmol per liter) was suggested as an optimal point of specificity and sensitivity for predicting type 2 diabetes.8,9 Our results suggest that in young men, fasting plasma glucose levels within the normoglycemic range can predict type 2 diabetes. Consistent with our findings is the observation that elevated fasting plasma glucose levels within the normoglycemic range can predict cardiovascular, cerebrovascular, and overall mortality risks in persons 45 years of age or older.26,27 Thus, subcategories within the range defined as normal for fasting plasma glucose levels contain information relevant for the assessment of the risk of various diseases,28,29 as indicated here for type 2 diabetes.
More than half of the entire study population had fasting plasma glucose levels exceeding 90 mg per deciliter (5.00 mmol per liter), which were associated with a significantly increased risk of diabetes during the mean follow-up of nearly six years. The absolute incident risk of type 2 diabetes among men who had fasting plasma glucose levels of 91 to 99 mg per deciliter was 2.3 percent during this follow-up period. Therefore, designating a fasting plasma glucose level of more than 90 mg per deciliter as the sole marker of imminent diabetes is unlikely to be useful. Alternatively, the use of an individualized definition of a normal fasting plasma glucose level, which incorporates the compound effect observed with body-mass index and triglyceride levels (Figure 1), may prove to be of greater clinical value. Indeed, among normoglycemic obese subjects with fasting plasma glucose levels of more than 90 mg per deciliter, the incidence of diabetes was 5.7 percent, as compared with 0.4 percent in lean men with glucose levels of 86 mg per deciliter or less. On the basis of population attributable risk, 60.5 percent of the cases might be preventable by a joint reduction in the risks associated with obesity and high-normal fasting plasma glucose levels. Risk stratification for the definition of normoglycemia is reminiscent of the current guidelines for antidyslipidemic and antihypertensive interventions.30
Our study raises potential testable hypotheses with regard to mechanisms. The fasting plasma glucose level is largely determined by hepatic glucose production.31 Thus, the observation that a high-normal fasting plasma glucose level predicts type 2 diabetes suggests that a relative overproduction of hepatic glucose already exists early in the natural history of diabetes and is exaggerated by obesity (Figure 1B). Obese persons who do not have diabetes consistently exhibit an enhanced rate of glucose production.32 This enhanced rate may emanate from elevated levels of free fatty acids that directly accelerate the rate of hepatic gluconeogenesis,33 combined with desensitization of the hepatic regulatory loop involving hypothalamic sensing of fatty acids.34 Obesity-associated altered secretion of adipocytokines from adipocytes, macrophages in fat tissue, or both has been suggested as the mechanism involved in mediating such dysregulated "crosstalk" between fatty tissue and the liver.35,36,37,38 Understanding the operative mechanisms that regulate fasting plasma glucose levels may bring us closer to finding new and effective measures to prevent type 2 diabetes in young adults.
Supported by the Israel Defense Forces Medical Corps.
We are indebted to Drs. Ilana Harman-Boehm and Shimon Weitzman of the Soroka Medical Center and Ben-Gurion University, Beer-Sheva, Israel, for their valuable discussions and careful reading of the manuscript; to Dr. Itamar Raz, Hadassah Medical Center, Jerusalem, for his support and encouragement; and to Ms. Yehudit Mish for her valuable assistance with data collection.
Source Information
From the Medical Corps Headquarters (A.T., E.I., T.S., I.K.) and the Center for Medical Services (D.T.-M.), Israel Defense Forces Medical Corps; the Department of Internal Medicine A, Sheba Medical Center, Tel-Hashomer (A.T.); the S. Daniel Abraham International Center for Health and Nutrition (I.S., A.R.), the Department of Epidemiology (I.S.), and the Department of Clinical Biochemistry (A.R.), Faculty of Health Sciences, Ben-Gurion University of the Negev, Beer-Sheva; and the Department of Internal Medicine A, Meir Hospital, Sapir Medical Center, Kfar-Sava (D.P.) — all in Israel.
Drs. Tirosh and Shai contributed equally to the study.
Address reprint requests to Dr. Tirosh at the Department of Internal Medicine A, Sheba Medical Center, Tel-Hashomer, Israel, or at amirt{at}bgumail.bgu.ac.il.
References
The following are members of the Israeli Diabetes Research Group Investigators, all in Israel: Soroka Medical Center, Beer-Sheva — I. Harman-Bohem; Hillel Yaffe Medical Center, Hadera — A. Jaffe; Rambam Medical Center, Haifa — E. Karnieli and N. Shehadeh; Lin Medical Center, Haifa — O. Minuchin; Wolfson Medical Center, Holon — J. Wainstein; A. Clalit Health Services, Jaffa — E. Stern; Hadassah–Hebrew University Hospital, Jerusalem — B. Glaser and I. Raz; Clalit Health Services, Jerusalem — A. Tsur; Western Galilee Hospital, Nahariya — T.A. Herskovits; Sheba Medical Center, Tel-Hashomer — O. Kalter-Leibovici; Kaplan Medical Center, Rehovot — H. Knobler; Assaf Harofeh Medical Center, Zerifin — A. Buchs and M. Rapoport; Maccabi Healthcare Services, Rishon-Le-Ziyon — J. Cohen; Souraski Medical Center, Tel-Aviv — A. Robinshtein; Clalit Health Services, Tel-Aviv — Y. Yerushalmy.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Normal Fasting Plasma Glucose Levels and Type 2 Diabetes
Gambino R., Reichberg S., Schwartz J. G., Tirosh A., Shai I., Rudich A.
Extract |
Full Text |
PDF
N Engl J Med 2006;
354:87-88, Jan 5, 2006.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
60 or >60 minutes per week or missing information); family history of diabetes (positive, negative, or missing information); and smoking status (current smoker, non-current smoker, or missing information). CI denotes confidence interval.