OBJECTIVE—To describe the characteristics and vital prognosis of men with diabetes diagnosed by one fasting plasma glucose (FPG) concentration ≥7.0 mmol/l, with diabetes diagnosed by one isolated postchallenge hyperglycemia (IPH) (FPG <7.0 mmol/l and a 2-h plasma glucose concentration ≥11.1 mmol/l), or with impaired glucose tolerance (IGT).

RESEARCH DESIGN AND METHODS—This study involved a cohort of 6,881 Caucasian nondiabetic men from the Paris Prospective Study, aged 44–55 years, who were followed for cause of death for 20 years.

RESULTS—Diabetes was diagnosed in 4.3% of the men (1.0% diabetes diagnosed by IPH), and IGT was diagnosed in 9% of the men. At baseline, the men with diabetes diagnosed by IPH had a lower cardiovascular risk profile than those with diabetes diagnosed by FPG, as did the men with IGT and a normal fasting glucose level (<6.1 mmol/l, IGT and normal fasting glucose), compared with men with impaired fasting glucose (6.1–6.9 mmol/l, IGT and impaired fasting glucose [IFG]). At 20 years of follow-up, all-cause and cancer death rates were higher in men with diabetes diagnosed by IPH than in men with diabetes diagnosed by FPG (55 vs. 44%, P < 0.1 and 31 vs. 17%, P < 0.01, respectively) but were not significantly different for coronary causes (6 vs. 11%). Men with IGT and normal fasting glucose also had significantly higher cancer death rates than men with IGT and IFG.

CONCLUSIONS—The most likely explanation for the high cancer and low coronary death rates is that men with diabetes diagnosed by IPH consumed alcohol; the men in this study drank 49 g of pure alcohol on average per day, equivalent to 0.6 l of wine. If these results are confirmed by other prospective studies, screening subjects for isolated postchallenge hyperglycemia may not be worthwhile.

The currently recommended method for screening and diagnosing diabetes is based on fasting plasma glucose (FPG), measured on at least two occasions (1,2). Although subjects with hyperglycemia 2 h after undergoing standard 75-g oral glucose tolerance test are still considered to have diabetes, this test is unlikely to be performed, considering the recommendations from the American Diabetes Association, even if the World Health Organization was less categorical about not using the oral glucose tolerance test (1,2). In fact, even though the previous recommendations for the diagnosis of diabetes were current (3), the oral glucose tolerance test was rarely performed in practice. In France, its use was limited; in 1998, for ambulatory laboratory tests, there were 17,000,000 prescriptions for fasting glucose determinations and only 77,000 0- to 2-h oral glucose tolerance tests (4). Therefore, in subjects with normal FPG but postchallenge hyperglycemia, diabetes was often not diagnosed; in the future, there will be even less chance that diabetes will be diagnosed in these subjects.

Since the 1985 report by the World Health Organization (3), most epidemiological studies have used the oral glucose tolerance test. The reported prevalences of diabetes have often been based only on the 2-h glucose evaluation, as recommended by the World Health Organization, because the fasting status of subjects in epidemiological studies is difficult to ensure. In some studies, FPG was not determined at all.

We have very limited knowledge about the characteristics and prognosis of subjects in whom diabetes is diagnosed by an isolated elevated 2-h plasma glucose, also known as isolated postchallenge hyperglycemia (IPH) (2-h plasma glucose ≥11.1 mmol/l and FPG <7.0 mmol/l) or about subjects in whom impaired glucose tolerance (IGT) is diagnosed (2-h plasma glucose between 7.8 and 11.1 mmol/l and FPG <7.0 mmol/l).

The aim of this study was to describe the characteristics of men with diabetes diagnosed by IPH and those with IGT, with and without impaired fasting glucose (IFG) (fasting plasma glucose between 6.1 and 6.9 mmol/l), and to study their risks and causes of early death, compared with men with normal glucose tolerance without IGT and those in whom diabetes was diagnosed by FPG concentration ≥7.0 mmol/l.

Subjects

This study analyzed the 6,881 Paris policemen, all Caucasians aged 44–55 years, who had no cardiovascular disease and in whom diabetes had not been diagnosed during the first annual follow-up examination of the Paris Prospective Study cohort in 1968–1974. These men were all followed for causes of death for 20 years.

Methods

All men participating in the study underwent a 2-h, 75-g oral glucose tolerance test. Glucose and insulin concentrations were determined using both the fasting and 2-h plasma samples, and cholesterol, triglyceride, and nonesterified fatty acid concentrations were determined using the fasting sample. The mean corpuscular volume was used as a marker of excessive alcohol consumption (5). The men were questioned about smoking habits, blood pressure was measured, and BMI was determined.

We have classed the men participating in the study into five groups:

1. Normal 2-h glucose and normal or impaired fasting glucose (normal/IFG): 2-h plasma glucose <7.8 mmol/l and FPG <7.0 mmol/l (1,2).

2. Impaired glucose tolerance and normal fasting glucose (IGT and non-IFG): 2-h plasma glucose ≥7.8 and <11.1 mmol/l and FPG <6.1 mmol/l.

3. Impaired glucose tolerance and impaired fasting glucose (IGT and IFG): 2-h plasma glucose ≥7.8 and <11.1 mmol/l and FPG ≥6.1 and <7.0 mmol/l.

4. Diabetes diagnosed by IPH: 2-h plasma glucose ≥11.1 mmol/l and FPG <7.0 mmol/l.

5. Diabetes diagnosed by FPG: FPG ≥7.0 mmol/l, regardless of the 2-h glucose concentration.

The men were studied after 10 and 20 years of follow-up. Inquiries were made through official sources to ascertain the date of death of deceased subjects. The International Classification of Diseases, 8th and 9th revisions (ICD-8 and ICD-9) (6), was used to code the causes of death, which, until the end of 1988, were based on information from the treating physician, hospital records, and the family of the deceased. For those in which the causes of death was missing before this date and for deaths after 1988, the officially certified causes of death were used. Coronary heart disease (CHD) is defined as customary in the Paris Prospective Study (7) by the ICD-8 codes: 410–414 (ischemic heart disease), 427.0 (congestive heart failure), 427.1 (left ventricular failure), 519.1 (acute edema of lung), 782 (symptoms referable to cardiovascular and lymphatic system), 795.0 (sudden death); cancers: 140–209; cancers mainly associated with alcohol: 141 (tongue), 143–146 (mouth and oropharynx), 148–150 (hypopharynx, pharynx, and esophagus), 155 (liver), and 157 (pancreas); all alcohol-related diseases: the alcohol-related cancers listed above and 291 (alcoholic psychosis), 303 (alcoholism), 571.0 (alcoholic cirrhosis), and 784.5 (hematemesis).

Statistical analysis

The characteristics of the men are presented as percentages or as means (SD); if the parameter was skewed, it was logarithmically transformed and the geometric mean and the 95% CI are given. Death rates have been calculated using the actuarial method. Comparisons of means, percentages, and death rates were performed using Student’s t test, χ2 test, and log-rank test. SAS software (version 8; SAS Institute, Cary, NC) was used for all analyses.

Distribution of subjects at baseline

Of the 6,881 men, diabetes was diagnosed in 297 (4.3%): 3.3% by FPG ≥7.0 mmol/l (1.3% had both FPG and 2-h hyperglycemia and 2.0% had an isolated fasting hyperglycemia testing) and 1.0% by IPH. Therefore, diabetes would not have been diagnosed in nearly one in four of these diabetic men if only the FPG concentration had been measured. Furthermore, 627 (9.1%) of the men had IGT: 263 (3.8%) had IFG, and 364 (5.3%) had non-IFG.

Characteristics of the subjects at baseline

Compared with the 5,957 men with normal 2-h glucose and normal or impaired fasting glucose, the 68 men with diabetes diagnosed by IPH had a significantly lower BMI but higher systolic and diastolic blood pressures, fasting and 2-h plasma glucose concentrations, 2-h insulin and fasting nonesterified fatty acids concentrations, and more frequently, a high mean corpuscular volume (Table 1). There was no difference in age, family history of diabetes, smoking status, and fasting insulin, cholesterol, or triglyceride levels.

The 68 men with diabetes diagnosed by IPH differed from the 229 men with diabetes diagnosed by FPG. They had a significantly lower frequency of a family history of diabetes, lower BMI, lower FPG (by definition), lower triglyceride, and lower insulin concentrations. However, the concentration of nonesterified fatty acids and the frequency of a high mean corpuscular volume were significantly higher in the men with diabetes diagnosed by IPH.

In the 627 men with IGT, all parameters were significantly higher compared with the 5,957 men with normal/IFG, except for age, family history of diabetes, and smoking status (data not shown). The differences were mainly due to the 263 men with IGT and IFG, who, compared with the 364 men with IGT and non-IFG, had significantly higher mean BMI, systolic blood pressure, fasting glucose concentration, fasting and 2-h insulin concentrations, and less frequently, high mean corpuscular volume (Table 1). There was no difference in age, family history of diabetes, smoking status, diastolic blood pressure, cholesterol, and triglyceride and nonesterified fatty acid levels.

All-cause deaths

After 10 years of follow-up, 7.5% of the men had died, and after 20 years, 22.1% of the men had died (Table 2). At both follow-up times, men with diabetes diagnosed by IPH had the highest all-cause death rates of all five groups studied, and although it was not significantly different from the death rate of men with diabetes diagnosed by FPG, it was significantly higher than for the men with normal/IFG (P < 0.0001 at both 10 and 20 years). The men with IGT also had significantly higher all-cause death rates than the men with normal/IFG (P < 0.0001, data not shown in table); this was mainly due to the high death rates of the men with IGT and non-IFG, which were significantly higher than in the men with IGT and IFG (P < 0.01 at 10 years and P < 0.02 at 20 years).

Causes of death

CHD death rates were highest in the men with diabetes diagnosed by FPG but were not significantly different from any of the other groups, because of the low number of coronary deaths, even after 20 years of follow-up. For the men with diabetes diagnosed by IPH, the coronary death rates were close to and between those of the men with normal glucose tolerance as well as IGT and IFG.

The death rates from cancer for the men with diabetes diagnosed by IPH and IGT and non-IFG were two to three times higher than for the men with normal/IFG (P < 0.05 at 10 years and P < 0.0001 at 20 years). These death rates were higher in the men with IGT and non-IFG than in the men with IGT and IFG (P < 0.03) and higher in the men with diabetes diagnosed by IPH than in men with diabetes diagnosed by FPG (P < 0.01 at both 10 and 20 years).

After 10 years of follow-up, the deaths rates for alcohol-related cancers were six times higher in the men with diabetes diagnosed by IPH than in the men with diabetes diagnosed by FPG or the men with normal/IFG (P < 0.01 for both). At 20 years of follow-up, there was still a sixfold excess compared with the men with normal glucose tolerance (P < 0.0001), but the excess was only twofold that in the men with diabetes diagnosed by FPG (P < 0.06).

The death rates for all alcohol-related causes of death were highest in the men with diabetes diagnosed by IPH, significantly higher than in the men with normal/IFG (P < 0.0003 at 10 years and P < 0.0001 at 20 years), and significantly higher than in the men with diabetes diagnosed by FPG only at 20 years (P < 0.03). In contrast, the men with IGT and IFG and men with IGT and non-IFG had similar death rates for all alcohol-related causes.

To further illustrate the causes of death in men with diabetes diagnosed by IPH and with diabetes diagnosed by FPG, we divided the latter group into the 139 men with diabetes diagnosed by an isolated fasting hyperglycemia (FPG ≥7.0 mmol/l and 2-h plasma glucose <11.1 mmol/l) and the 90 men with diabetes diagnosed by combined fasting and 2-h hyperglycemia (FPG ≥7.0 mmol/l and 2-h plasma glucose ≥11.1 mmol/l) (Fig. 1). For all causes of death, men with a 2-h plasma glucose ≥11.1 mmol/l had similar death rates that were independent of whether the fasting plasma glucose level was higher or lower than 7.0 mmol/l (54.6 and 51.1%, respectively); the rates were 20% higher than in men with diabetes diagnosed by an isolated fasting hyperglycemia (diabetes diagnosed by IPH versus diabetes diagnosed by isolated fasting hyperglycemia, P < 0.03) and more than twofold higher than in the nondiabetic men (P < 0.0001). In contrast, for CHD mortality, death rates in men with diabetes diagnosed by IPH were only a little higher than in the nondiabetic men (5.8 and 4.5%) and lower than in men with diabetes diagnosed by isolated fasting hyperglycemia (9.2%) and men with diabetes diagnosed by fasting and 2-h hyperglycemia (13.2%). However, for death from cancer, the death rate in men with diabetes diagnosed by IPH was very high: three times higher than in nondiabetic men (31.3 and 10.4%, P < 0.0001) and two times higher than in men with diabetes diagnosed by isolated fasting hyperglycemia or diabetes diagnosed by fasting and 2-h hyperglycemia (18.2%, P < 0.03 and 15.1%, P < 0.01, respectively).

Should we try to find subjects with diabetes diagnosed by IPH? First, these men were 1% of the total population, and although this is not a large percentage, it was 23% of the subjects diagnosed with diabetes at baseline. Second, the cardiovascular risk profile was higher in the men with diabetes diagnosed by FPG than those with diabetes diagnosed by IPH, which was, in turn, higher than that in men with normal/IFG. Therefore, it is not surprising that their coronary death rate at 20 years of follow-up showed a decreasing trend over these three groups, even though the differences were not statistically significant. Thirdly, death by cancer, by alcohol-associated cancer, and by all alcohol-related diseases was significantly higher in men with diabetes diagnosed by IPH than in men with normal/IFG or diabetes diagnosed by FPG. In this population, we would implicate a heavy alcohol intake as the main reason for the high death rates in men with diabetes diagnosed by IPH; treatment to lower the blood glucose level is unlikely to change the prognosis in these patients.

Both death rate and conversion to diabetes are known to be higher in subjects with IGT than in those with normal/IFG (3). In this study, 9% of the men had IGT, 4% had IGT and non-IFG, and 5% had IGT and IFG. The group with IGT and non-IFG was metabolically close to the men with normal/IFG, but the risk of death in those with IGT and non-IFG was almost two times higher, which was largely due to cancer mortality. The group with IGT and IFG had more factors associated with insulin resistance, but their vital prognosis was very similar to that in the men with normal/IFG.

To our knowledge, no study has analyzed subjects with an IGT, and few studies have analyzed subjects with diabetes diagnosed by IPH. The first was the Rancho-Bernardo Study of 1,858 subjects (average age 70 years); in this study, diabetes was diagnosed by IPH in 72% of the diabetic women and 48% of the diabetic men (8). However, in the DECODE Study of 29,108 subjects (average age 53 years), this percentage was 35% (9), and in the subgroup of 3,991 men and 2,149 women aged 60–90 years, diabetes was diagnosed by IPH in 34% of men and 37% of women (10). In the elderly subjects from the Cardiovascular Health Survey, diabetes was diagnosed by IPH in 52% of diabetic subjects (11). In younger populations (average age 36–43 years) from Mauritius, Fiji, and Nauru, diabetes was diagnosed by IPH in 22, 20, and 8%, respectively, of the patients with newly diagnosed diabetes (12). The most recent information comes from the Third National Health and Nutrition Examination Survey of men and women aged 40–74 years (13); diabetes was diagnosed by IPH in 41% of the diabetic subjects, and this percentage increased with age. Diabetes was diagnosed by IPH in 36% of patients aged 45–54 years (age-group of our study) and in 54% of patients aged 70–74 years. These percentages differ widely and depend on the age, sex, community, and ethnic group of the subjects, and perhaps more importantly, on the frequency of routine screening in the population and the number of subjects in the population in whom diabetes has already been diagnosed.

The first publication on the mortality of subjects with diabetes diagnosed by IPH was from the Rancho-Bernardo Study (8). Compared with nondiabetic men, the men had a relative risk of cardiovascular death of 0.7 (0.3–1.6), after adjustment for cardiovascular risk factors. In contrast, the women had a significantly increased risk compared with nondiabetic women; the relative risk was 2.6 (1.4–4.7). The risk was slightly higher for ischemic heart disease mortality. No information is given in this study about other causes of death or of the risk of death by all causes. The second study published was in the subgroup of elderly men and women aged 60–80 years from the DECODE Study (10); men and women with diabetes diagnosed by IPH had a significantly higher death rate, relative risk 1.6 (1.1–2.3), for a follow-up of ≤10 years. The most complete study comes from the islands of Mauritius, Fiji, and Nauru, which included more than 8,000 men and women in whom diabetes had not been previously diagnosed; most subjects were from Mauritius (12). The relative risk of death by cardiovascular disease was significantly higher in the men and women with diabetes diagnosed by IPH than in the nondiabetic subjects: 2.3 (1.2–4.2) and 2.6 (1.3–5.1), respectively. However, in concordance with our study, the relative risk of death from cancer was high, with a risk ratio of 8.0 (3.6–17.9) in men and 2.2 (0.8–6.5) in women.

One possible explanation for our observation that diabetes diagnosed by IPH (or even in men with IGT and non-IFG) is associated with death from cancer is that the accompanying hyperinsulinemia promotes growth of cancer (14). However, this does not seem to be the explanation in the Paris Prospective Study, in which the men with diabetes diagnosed by IPH who died of cancer had lower fasting and 2-h insulin concentrations at baseline than the men who did not die because of cancer.

The men with diabetes diagnosed by IPH or the men with IGT and non-IFG had the two highest levels of nonesterified fatty acids, which could be a cancer promoter, an early catabolic effect of cancer, or a marker of alcohol and tobacco consumption. A recent study of this cohort showed that nonesterified fatty acids were an independent risk factor for alcohol- and smoking-related cancer mortality but not for CHD mortality (15). The most likely explanation for our results is that the men with diabetes diagnosed by IPH or the men with IGT and non-IFG had a high alcohol intake and subsequently died from diseases associated with excessive alcohol consumption rather than from coronary causes. The frequency of men with a higher mean corpuscular volume in these groups supports this hypothesis. However, surprisingly, although the men with IGT and non-IFG had higher alcohol-related cancer death rates than the men with IGT and IFG, death rates were similar in these two groups for all alcohol-related diseases. Alcohol has been associated with diabetes in other studies. In the Rancho-Bernardo Study, during a 14-year follow-up, the subjects in whom diabetes developed (FPG ≥7.8 mmol/l or 2-h plasma glucose ≥11.1 mmol/l) had a mean daily alcohol intake of 31 g/day compared with 19 g/day in the other subjects (16). In the Paris Prospective Study, we have already reported that diabetic subjects (FPG ≥7.8 mmol/l or 2-h plasma glucose ≥11.1 mmol/l) had a high death rate from alcohol-related causes (17). Based on a dietary survey, the average daily alcohol intake was 49 g of pure alcohol (equivalent to 0.6 l of wine per day) in a subgroup of 446 of these men; 3% of the men with an alcohol intake ≥20 g/day had diabetes (FPG ≥7.8 mmol/l or 2-h plasma glucose ≥11.1 mmol/l), whereas none of the 82 men with a lower alcohol intake had diabetes (18). In contrast, among the subjects from Mauritius, alcohol intake was not associated significantly with diabetes (2-h plasma glucose ≥11.1 mmol/l) in either men or women, even though the alcohol intake was not low: 39% of the men had three or more drinks per day (19). A recent study in the U.S. showed that the incidence of diabetes (FPG ≥7.0 mmol/l) increased above the median intake of 17 g/day of alcohol (20). From these various studies, we have no information about the alcohol consumption in subjects with diabetes diagnosed by IPH or subjects with IGT and non-IFG, and the dietary survey in the Paris Prospective Study comprised too few subjects.

Studies of causes of death in subjects with diabetes diagnosed by IPH and subjects with IGT and non-IFG are required in other populations to determine whether deaths caused by cancer and deaths caused by CHD are common and whether the subjects in whom diabetes was diagnosed based on isolated elevated 2-h postload glucose concentrations are similar to those in whom diabetes was diagnosed by fasting glucose concentrations. If subjects with high 2-h plasma glucose concentrations die principally because of cancer, screening for diabetes and subsequent treatment for hyperglycemia would not seem to be of any benefit for increasing the longevity of these individuals.

Figure 1—

Death rates over 20 years for all causes of death (A); deaths caused by CHD (B); and deaths caused by cancer (C), according to FPG <7.0 and ≥7.0 mmol/l and 2-h plasma glucose values <11.1 and ≥11.1 mmol/l for men aged 44–55 years in the Paris Prospective Study.

Figure 1—

Death rates over 20 years for all causes of death (A); deaths caused by CHD (B); and deaths caused by cancer (C), according to FPG <7.0 and ≥7.0 mmol/l and 2-h plasma glucose values <11.1 and ≥11.1 mmol/l for men aged 44–55 years in the Paris Prospective Study.

Close modal
Table 1—

Characteristics at baseline for the 6,881 men studied in the Paris Prospective Study

IGT
Diabetes
Normal/IFG
IGT and non-IFG
P*IGT and IFG
Diagnosed by IPH
PDiagnosed by FPG
P
FPG (mmol/l)<7.0<6.1≥6.1 and <7.0<7.0≥7.0
2-h glucose (mmol/l)<7.8≥7.8 and <11.1≥7.8 and <11.1≥11.1
n 5,957 364  263 68  229  
Age (years) 48 (2) 48 (2)  49 (2) 49 (2)  49 (2)  
Family history of diabetes (%) 7.1 9.4  7.2 6.3 0.05 15.5  
Smokers (%) 42 58  51 32  43  
BMI (kg/m225.7 (3.1) 25.8 (3.6) 0.001 27.7 (3.5) 24.9 (3.4) 0.0001 28.6 (3.8) 0.05 
Systolic blood pressure (mmHg) 143 (20) 151 (23) 0.01 157 (24) 162 (23)  163 (29) 0.0001 
Diastolic blood pressure (mmHg) 80 (12) 85 (13)  88 (13) 86 (13)  90 (16) 0.001 
Fasting glucose (mmol/l) 5.5 (0.5) 5.6 (0.4) 0.001 6.4 (0.2) 5.8 (0.6) 0.0001 7.8 (1.0) 0.01 
2-h glucose (mmol/l) 5.2 (1.2) 8.8 (0.9)  8.8 (0.8) 12.3 (1.2) 0.0001 10.2 (3.8) 0.0001 
Fasting insulin (pmol/l) 68 (20–235) 78 (20–208) 0.001 101 (30–351) 69 (15–330) 0.0001 129 (36–461)  
2-h insulin (pmol/l) 209 (46–952) 438 (110–1,738) 0.001 582 (168–2,012) 407 (88–1,890)  397 (104–1,510) 0.0001 
Cholesterol (mmol/l) 5.6 (1.1) 5.7 (1.1)  5.8 (1.1) 5.7 (1.2)  6.0 (1.2)  
Triglycerides (mmol/l) 1.2 (0.5–3.3) 1.4 (0.4–4.2)  1.5 (0.5–4.3) 1.2 (0.5–3.2) 0.0001 1.7 (0.6–5.1)  
Nonesterified fatty acids (mmol/l) 77 (34–172) 107 (45–252)  103 (44–240) 143 (58–351) 0.0001 105 (46–240) 0.0001 
Mean corpuscular volume (≥ 100 fl) (%) 29 39 0.05 28 58 0.001 36 0.001 
IGT
Diabetes
Normal/IFG
IGT and non-IFG
P*IGT and IFG
Diagnosed by IPH
PDiagnosed by FPG
P
FPG (mmol/l)<7.0<6.1≥6.1 and <7.0<7.0≥7.0
2-h glucose (mmol/l)<7.8≥7.8 and <11.1≥7.8 and <11.1≥11.1
n 5,957 364  263 68  229  
Age (years) 48 (2) 48 (2)  49 (2) 49 (2)  49 (2)  
Family history of diabetes (%) 7.1 9.4  7.2 6.3 0.05 15.5  
Smokers (%) 42 58  51 32  43  
BMI (kg/m225.7 (3.1) 25.8 (3.6) 0.001 27.7 (3.5) 24.9 (3.4) 0.0001 28.6 (3.8) 0.05 
Systolic blood pressure (mmHg) 143 (20) 151 (23) 0.01 157 (24) 162 (23)  163 (29) 0.0001 
Diastolic blood pressure (mmHg) 80 (12) 85 (13)  88 (13) 86 (13)  90 (16) 0.001 
Fasting glucose (mmol/l) 5.5 (0.5) 5.6 (0.4) 0.001 6.4 (0.2) 5.8 (0.6) 0.0001 7.8 (1.0) 0.01 
2-h glucose (mmol/l) 5.2 (1.2) 8.8 (0.9)  8.8 (0.8) 12.3 (1.2) 0.0001 10.2 (3.8) 0.0001 
Fasting insulin (pmol/l) 68 (20–235) 78 (20–208) 0.001 101 (30–351) 69 (15–330) 0.0001 129 (36–461)  
2-h insulin (pmol/l) 209 (46–952) 438 (110–1,738) 0.001 582 (168–2,012) 407 (88–1,890)  397 (104–1,510) 0.0001 
Cholesterol (mmol/l) 5.6 (1.1) 5.7 (1.1)  5.8 (1.1) 5.7 (1.2)  6.0 (1.2)  
Triglycerides (mmol/l) 1.2 (0.5–3.3) 1.4 (0.4–4.2)  1.5 (0.5–4.3) 1.2 (0.5–3.2) 0.0001 1.7 (0.6–5.1)  
Nonesterified fatty acids (mmol/l) 77 (34–172) 107 (45–252)  103 (44–240) 143 (58–351) 0.0001 105 (46–240) 0.0001 
Mean corpuscular volume (≥ 100 fl) (%) 29 39 0.05 28 58 0.001 36 0.001 

Data are means (SD), geometric means (95% CI), or %.

*

IGT diagnosed by non-IFG versus IFG;

diabetes diagnosed by IPH versus FPG;

diabetes diagnosed by IPH versus normal/IFG.

Table 2—

Death rate at 10 and 20 years of follow-up by cause, according to diabetic status, in the 6,881 men followed for vital status and causes of death in the Paris Prospective Study

IGT
Diabetes
Normal/IFG
IGT and non-IFG
P*IGT and IFG
Diagnosed by IPH
PDiagnosed by FPG
P
FPG (mmol/l)<7.0<6.1≥6.1 and <7.0<7.0≥7.0
2-h glucose (mmol/l)<7.8≥7.8 and <11.1≥7.8 and <11.1≥11.1
n 5,957 364  263 68  229  
10-year follow-up         
 All causes 7.0 (403) 14.8 (52) 0.01 7.9 (20) 18.1 (12)  14.3 (32) 0.0001 
 CHD 1.4 (79) 2.7 (9)  1.6 (4) 1.5 (1)  5.1 (11)  
 All cancers 3.0 (165) 7.1 (24) 0.03 2.8 (7) 8.3 (5) 0.001 1.5 (3) 0.05 
 Alcohol-related cancers 0.6 (34) 1.8 (6)  0.4 (1) 3.4 (2) 0.07 0.5 (1) 0.009 
 All alcohol-related causes of death 0.8 (43) 2.1 (7)  1.7 (4) 4.9 (3) 0.1 1.5 (3) 0.0003 
20-year follow-up         
 All causes 21.1 (1,188) 36.3 (127) 0.02 27.4 (69) 54.6 (36) 0.1 44.1 (98) 0.0001 
 CHD 4.2 (216) 8.1 (24)  6.7 (15) 5.8 (3)  11.1 (21)  
 All cancers 9.9 (519) 19.1 (59) 0.03 12.0 (27) 31.3 (17) 0.01 17.0 (29) 0.0001 
 Alcohol-related cancers 1.9 (96) 5.6 (16)  3.8 (8) 11.5 (6) 0.06 5.3 (9) 0.0001 
 All alcohol-related causes of death 2.4 (123) 7.3 (21)  5.9 (13) 17.3 (9) 0.03 7.8 (14) 0.0001 
IGT
Diabetes
Normal/IFG
IGT and non-IFG
P*IGT and IFG
Diagnosed by IPH
PDiagnosed by FPG
P
FPG (mmol/l)<7.0<6.1≥6.1 and <7.0<7.0≥7.0
2-h glucose (mmol/l)<7.8≥7.8 and <11.1≥7.8 and <11.1≥11.1
n 5,957 364  263 68  229  
10-year follow-up         
 All causes 7.0 (403) 14.8 (52) 0.01 7.9 (20) 18.1 (12)  14.3 (32) 0.0001 
 CHD 1.4 (79) 2.7 (9)  1.6 (4) 1.5 (1)  5.1 (11)  
 All cancers 3.0 (165) 7.1 (24) 0.03 2.8 (7) 8.3 (5) 0.001 1.5 (3) 0.05 
 Alcohol-related cancers 0.6 (34) 1.8 (6)  0.4 (1) 3.4 (2) 0.07 0.5 (1) 0.009 
 All alcohol-related causes of death 0.8 (43) 2.1 (7)  1.7 (4) 4.9 (3) 0.1 1.5 (3) 0.0003 
20-year follow-up         
 All causes 21.1 (1,188) 36.3 (127) 0.02 27.4 (69) 54.6 (36) 0.1 44.1 (98) 0.0001 
 CHD 4.2 (216) 8.1 (24)  6.7 (15) 5.8 (3)  11.1 (21)  
 All cancers 9.9 (519) 19.1 (59) 0.03 12.0 (27) 31.3 (17) 0.01 17.0 (29) 0.0001 
 Alcohol-related cancers 1.9 (96) 5.6 (16)  3.8 (8) 11.5 (6) 0.06 5.3 (9) 0.0001 
 All alcohol-related causes of death 2.4 (123) 7.3 (21)  5.9 (13) 17.3 (9) 0.03 7.8 (14) 0.0001 

Data are % (n).

*

IGT diagnosed by non-IFG versus IFG;

diabetes diagnosed by IPH versus FPG;

diabetes diagnosed by IPH versus normal/IFG.

1
The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus: Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus.
Diabetes Care
20
:
1183
–1197,
1997
2
World Health Organization:
Definition, Diagnosis and Classification of Diabetes Mellitus and Its Complications. Part 1: Diagnosis and Classification of Diabetes Mellitus: Report of a WHO Consultation
. Geneva, World Health Org., 1999 (WHO/NCD/NCS/99.2)
3
World Health Organization: Diabetes Mellitus: Report of a WHO Study Group. Geneva, World Health Org., 1985 (Tech. Rep. Ser., no. 727)
4
CNAMTS: Exploitation des Données de Codage des Actes de Biologie Médicale, Année 1998. Ed L’Assurance Maladie, Paris, 1999, p. 87
5
Papoz L, Warnet JM, Péquignot G, Eschwège E, Claude JR, Schwartz D: Alcohol consumption in a healthy population: relationship to γ-glutamyl transferase activity and mean corpuscular volume.
JAMA
245
:
1748
–1751,
1981
6
World Health Organization: International Classification of Diseases, Injuries and Causes of Death, 9th Revision. Geneva, World Health Org., 1977
7
Ducimetiere P, Eschwege E, Papoz L, Richard JL, Claude JR, Rosselin G: Relationship of plasma insulin levels to the incidence of myocardial infarction and coronary heart disease mortality in a middle-aged population.
Diabetologia
19
:
205
–210,
1980
8
Barrett-Connor E, Ferrara A: Isolated postchallenge hyperglycemia and the risk of fatal cardiovascular disease in older women and men.
Diabetes Care
21
:
1236
–1239,
1998
9
The DECODE Study Group on Behalf of the European Diabetes Epidemiology Group: Is fasting glucose sufficient to define diabetes? Epidemiological data from 20 European Studies.
Diabetologia
42
:
647
–654,
1999
10
Balkau B for the DECODE Study Group on Behalf of the EDEG: New diagnostic criteria for diabetes and mortality in older adults.
Lancet
353
:
68
–69,
1999
11
Wahl PW, Savage PJ, Psaty BM, Orchard TJ, Robins JA, Tracy RP: Diabetes in older adults: comparison of 1997 American Diabetes Association classification of diabetes mellitus with the 1985 WHO classification.
Lancet
352
:
1012
–1015,
1998
12
Shaw JE, Hodge AM, de Courten M, Chitson P, Zimmet PZ: Isolated post-challenge hyperglycaemia confirmed as a risk factor for mortality.
Diabetologia
42
:
1050
–1054,
1999
13
Resnick HE, Harris MI, Brock DB, Harris TB: American Diabetes Association diabetes diagnostic criteria, advancing age, and cardiovascular disease risk profiles.
Diabetes Care
23
:
176
–180,
2000
14
Argiles JM, Lopez-Soriano FJ: Insulin and cancer.
Int J Oncol
18
:
683
–687,
2001
15
Charles MA, Fontbonne A, Thibult N, Claude J-R, Warnet J-M, Rosselin G, Ducimetière P, Eschwège E: High plasma nonesterified fatty acids are predictive of cancer mortality but not of coronary heart disease mortality: results from the Paris Prospective Study.
Am J Epidemiol
153
:
292
–298,
2001
16
Holbrook TL, Barrett-Connor E, Wingard DL: A prospective population-based study of alcohol use and non-insulin-dependent diabetes.
Am J Epidemiol
132
:
902
–909,
1990
17
Balkau B, Eschwège E, Fontbonne A, Claude J-R, Warnet J-M: Cardiovascular and alcohol-related deaths in abnormal glucose tolerant and diabetic subjects.
Diabetologia
35
:
39
–44,
1992
18
Balkau B, Randrianjohany A, Papoz L, Eschwège E: A prospective population-based study of alcohol use and non-insulin-dependent diabetes.
Am J Epidemiol
134
:
1469
–1470,
1991
19
Hodge AM, Dowse GK, Collins VR, Zimmet PZ: Abnormal glucose tolerance and alcohol consumption in three populations at high risk of non-insulin-dependent diabetes mellitus.
Am J Epidemiol
137
:
178
–189,
1993
20
Wei M, Gibbons LW, Mitchell TL, Kampert JB, Blair SN: Alcohol intake and incidence of type 2 diabetes in men.
Diabetes Care
23
:
18
–22,
2000

Address correspondence and reprint requests to Eveline Eschwège, MD, INSERM Unit 258, 16 Ave. Paul Vaillant Couturier, 94807 Villejuif, France. E-mail: eschwege@vjf.inserm.fr.

Received for publication 2 November 2000 and accepted in revised form 5 July 2001.

A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.