Studies and principal investigators participating in this analysis

• Denmark, Glostrup 1936 cohort: Thomas Drivsholm (Centre of Preventive Medicine, Glostrup University Hospital, Glostrup), Knut Borch-Johnsen (Steno Diabetes Centre, Gentofte);

• England, Ely Study: Nick Wareham (Cambridge University, Cambridge);

• England, Goodinge Study: John S Yudkin (University College, London), Richard Morris (Royal Free and University College Medical School; London)

• France, D.E.S.I.R.: Beverley Balkau (INSERM U258, Villejuif);

• Italy, Barilla Study and Barilla follow-up: Ivana Zavaroni (University of Parma, Parma);

• The Netherlands, MORGEN: Rob van Dam, Edith Feskens (National Institute for Public Health and the Environment, Bilthoven);

• Spain, VIVA: Rafael Gabriel (Hospital Universitario de La Princesa, Madrid);

• Sweden, Malmö Diet and Cancer Study: Peter Nilsson, Bo Hedblad (University Hospital, Malmö).

The National Cholesterol Education Program Expert Panel (NCEP) recently gave a clinical definition of the metabolic syndrome which was based on the WHO definition and required that three abnormalities be present, among abdominal obesity, hypertriglyceridaemia, low LDL-cholesterol, high blood pressure, high fasting glucose. In comparison to the WHO and EGIR definitions, neither insulin resistance nor a surrogate measure of insulin resistance was included. Also, the thresholds for triglycerides and blood pressure were lower, for HDL-cholesterol higher, for glucose intolerance only the fasting plasma glucose level was used and for obesity, as for the EGIR syndrome, the waist circumference was used by the NCEP [4].

As there is very little published data which enables the comparison of the frequency of the syndrome in different populations using common definitions and methods of analysis, we present here data on the frequency of the WHO and the EGIR syndromes in some European cohorts. Thus, the frequency of the metabolic abnormalities and the importance of these syndromes can be evaluated in the European context.

The term “impaired glucose regulation” in the WHO metabolic syndrome was taken to include both impaired glucose tolerance and impaired fasting glucose, as well as diabetes (Fig. 1).

We have used a wider definition of microalbuminuria than that defined by the WHO consultation, and included measures of casual microalbuminuria, and not just the albumin excretion rate and the albumin creatinine ratio (Fig. 2). Only the Glostrup, Goodinge, D.E.S.I.R. and the Barilla follow-up studies had some measure of microalbuminuria available, and the measures and definitions used were: the albumin to creatinine ratio >= 30 mg/g (Glostrup), albumin excretion rate >= 20 mg/min, 2-hours after an oral glucose tolerance test (Goodinge), a spot albuminuria >= 20 mg/l or dip-stick proteinuria (D.E.S.I.R.), a timed albumin secretion >= 20 mg/min (Barilla follow-up study).

The data was analysed in the various centres by sex and age class, according to a written protocol, and a synthesis made of the combined data. As this is a post-hoc study, there was no standardisation of any of the measures used, apart from the common clinical standards that exist. As the studies are shown by sex and age class, we have only used those cohorts and those age-sex groups where there were more than 40 subjects.

Overall percentages, means and standard deviations have been calculated, using weights proportional to the reciprocal of the variance for the percentages and for the means; the D.E.S.I.R. study is included twice in the tables, to show the effect of not-including and including microalbuminuria as a factor, however, only the former has been included in the calculated overall average figures; for the Barilla Study, the 1981 cohort and the sub-group of subjects who were followed up are both included, however, in this case, very few subjects would appear in the same sex-age groups.

For the WHO syndrome, for any given population, the frequency of each abnormality increased with age, and generally was a little higher in men than in women (Table II). Impaired glucose regulation was the least frequent abnormality, central obesity the most frequent. Raised blood pressure was also frequent and overall involved more than half the men and women 55 years and over. The overall prevalence of the syndrome was 14% in men and 4% in women under 40 years, 23% and 13% respectively for 40 to 55 years, and 41% and 26% respectively over 55 years of age.

For comparison with the EGIR syndrome, the analysis of the WHO syndrome was then restricted to non-diabetic subjects, (not already not known as diabetic, fasting plasma glucose < 7.0 mmol/l (venous whole blood < 6.1 mmol/l) and a 2-hour plasma glucose < 11.1 mmol/l (venous whole blood < 10.0 mmol/l)). By removing the diabetic subjects, who were mainly in the older age classes, the frequency of impaired glucose regulation decreased markedly, but there was only a marginal decrease for the other abnormalities (Tables IIandIII).

For the WHO and the EGIR syndromes in non-diabetic subjects (Tables IIIandIV), the frequencies of the insulin abnormality remained identical, by definition. For the glucose abnormality, the two syndromes had slightly differing frequencies, depending on whether an oral glucose tolerance test (OGTT) had been carried out. The treatment of hypertension was included in the EGIR syndrome, and thus the frequency of high blood pressure increased over that in the WHO syndrome, by 4% in men and 5% in women over the age of 55 years, indicating the frequency of subjects with blood pressures controlled by anti-hypertensive medication. For the lipid abnormalities, the frequencies were similar with both definitions of the syndrome: the EGIR syndrome used slightly higher thresholds but included individuals on treatment. Central obesity was much more frequent for men with the WHO than the EGIR syndrome as the former definition included obesity (BMI > 30 kg/m²); this was most evident in the oldest age group, > 55 years, where the overall frequencies were, for the WHO and EGIR definitions, 85% and 49%, respectively. In contrast, in women, the EGIR definition gave higher frequencies for central obesity than the WHO definition, with 44% and 27% respectively in the oldest age class.

For the overall syndrome frequencies, in men, the WHO metabolic syndrome was close to 50% more frequent than the EGIR insulin resistance syndrome, corresponding to, on average, more abnormalities. For women, while the WHO syndrome was more frequent the difference was less than for men (Fig. 3).

We have defined insulin resistance in terms of fasting insulin concentrations, so that we could provide descriptive data on the frequency of the WHO metabolic syndrome, as direct measures of insulin resistance are rarely available in epidemiological studies. Subjects in the top 25% of the fasting insulin distribution for non-diabetic subjects have been taken to be insulin resistant. It was not clear what the WHO consultation intended by the term “background population” and perhaps the cut-point for the definition of insulin resistance should have included the diabetic patients; alternately, perhaps the definition should have been more restrictive using only the subjects without impaired glucose regulation.

As we commented earlier [2], it is not evident that microalbuminuria should be included in the definition of the syndrome. For the D.E.S.I.R. Study, inclusion of the microalbuminuria criteria in the WHO syndrome definition changed little the frequency of the syndrome, suggesting either a minor contribution of this factor or that microalbuminuria tends to occur with the other conditions, even if its relation with insulin is tenuous [16][17][18].

The difference between the WHO and EGIR syndromes lies firstly in the inclusion of diabetic subjects in the WHO syndrome. Comparing the two syndrome definitions in the non-diabetic population, the main differences are in the inclusion of treatment for hypertension and dyslipidemia in the EGIR definition, in the definition of obesity. Both the waist hip ratio and the body mass index in the WHO definition, while the EGIR definition uses only one measurement, the waist circumference. There are minor differences in the cut-points to define dyslipidemia.

Overall, in men, the frequency of the WHO syndrome is 50% higher than the EGIR syndrome, although there is variation between studies. This higher frequency is mainly due to the differing definitions of central obesity which results in a higher frequency of central obesity with the WHO definition. In women, the frequencies are closer, even though, in contrast to men, central obesity is more common with the EGIR definition. The thresholds of the central obesity measures require refinement and justification, and a given measure may have different implications in different populations. In the NCEP definition of the metabolic syndrome, the waist circumference only is used, and the thresholds are 102 cm for men and 88 cm for women, somewhat higher than the 94 and 80 cm used for the EGIR syndrome definition.

Overall the frequency of hyperinsulinaemia changes little with age class, for example it is 27% in the non-diabetic men under 40 years, 33% in those over 55 years. In contrast, the syndrome frequency more than doubles in men over the same age classes: from 13% to 33% for the WHO syndrome and from 10% to 22% for the EGIR syndrome.

Surprisingly, there has been little use of this definition of the metabolic syndrome proposed by the WHO consultation; instead studies appear to have used the proposed definitions of the abnormalities, but have not combined them to give the frequency of the syndrome.

A very recent article has been published on the US prevalence of the metabolic syndrome, as defined by the National Cholesterol Education Program [4], using data from the third National Health and Nutrition Examination [19]. For the US population aged over 20 years, the prevalence was very similar in men and women, 24.0% and 23.4% respectively. The prevalence increased with age, from 6.7% for the youngest age group, 20 to 29 years, to 42.0% for those over 79 years. In contrast, in our study, men had the syndrome much more frequently than women (with either the WHO or the EGIR definition) for all age classes.

Only prospective studies can evaluate the usefulness of identifying and treating this syndrome in individuals and any definitions await refinement after elucidation of the physio-pathological mechanisms involved.