Applying the IDF definition in a large Norwegian population aged 20–89 years classified 29.6% as having the metabolic syndrome. The prevalence was highly age-dependent. This was evident especially in women, with a sevenfold increase in prevalence from age group 20–29 years to age group 80–89 years.
The IDF emphasizes central obesity as an essential criterion for the metabolic syndrome, due to the evidence linking ethnicity-specific waist circumference to cardiovascular disease and to the other components of the metabolic syndrome . Applying the recommended cut-offs for white people of European origin classified a large portion of this Norwegian population as having central obesity. In women aged 60 years or more, three out of four were defined as obese. As central obesity is regarded as a likely early step in the development of full metabolic syndrome [6, 10], this definition puts a very large number of individuals belonging to one of the longest living and healthiest populations in the world  at increased risk for cardiovascular disease and type 2 diabetes.
The IDF definition classified the majority of our population as hypertensive. Mean blood pressure in the HUNT 2 population increased markedly with age, especially among women. As the IDF definition, like other definitions of the metabolic syndrome, has not included any adjustment for age, 90% of those aged 60 years or more would be defined as being hypertensive.
The prevalence of the metabolic syndrome by the IDF criteria was higher than by the 2005 ATP III criteria in all sex-stratified age groups except in men 20–29 years old. As the two definitions are based on much of the same components, the difference in prevalence was mainly related to different waist circumference and to the focus on central obesity as an obligatory component in the IDF definition in contrast to being one out of five equally weighted components in the 2005 ATP III definition.
Several population studies have reported an increase in the prevalence of the metabolic syndrome with age regardless of definition [12–31], although some have reported a peak in the seventh decade and then a decline in both sexes [12, 19, 23, 31] or only in men [13, 14, 18, 22, 26, 30]. However, this has been incompletely studied in European populations, where only one study from Greece has assessed age-specific prevalence using the IDF definition . In addition, to our knowledge, only one study worldwide has previously assessed the prevalence separately for those aged 80 years and above . In contrast to several other studies [12–14, 18, 19, 22, 23, 26, 31], we found that the prevalence of the metabolic syndrome, by both definitions, increased linearly with age beyond the seventh decade in both sexes. To our knowledge, our study is the first to show that this increase continued into the ninth decade.
Strengths and limitations
The HUNT 2 study covered the whole adult population (homogenous, <3% non-Caucasian) in Nord-Trøndelag County, which is considered fairly representative of Norway. The county is mostly rural; the largest of six small towns has a population of 21,000. The main objectives of the HUNT study concerned large public health issues like diabetes, cardiovascular disease, obstructive lung disease, osteoporosis and mental health .
The participation rate was age dependent, with the highest participation rate (85.6%) in the age group 60–69 years. A study of non-participants  showed that the main reasons for not attending were of a practical nature in young people and poor health in elderly people. There is a potential selection bias from non-participation, although the participation rate in the HUNT study was similar to or higher than in comparable population studies .
A limitation of the HUNT study is that it was impractical to request the whole population to be fasting since examinations were spread out during daytime. Blood sampling was done whenever participants attended. However, our estimates of prevalence of the metabolic syndrome based on the sub-sample of participants reporting ≥4 hours fasting should be quite representative for the entire HUNT 2 population as the statistically significant differences between fasting and non-fasting participants were all minor (except triglycerides and glucose). Further, using the proxies of IDF and ATP metabolic syndrome, we were unable to reveal any statistically significant difference between fasting and non-fasting participants. Our adjustment of the levels of plasma glucose and triglycerides among the included individuals reporting 4–8 hours fasting should further increase the validity of our findings. Other referenced prevalence studies have used 6 hours fasting without adjustments [15, 31, 32].
Management of metabolic syndrome
The IDF Epidemiology Task Force Consensus Group  states, "Researchers and clinicians should use the new criteria for the identification of high-risk individuals and for research studies. Preventive measures are obviously needed in people identified". According to the AHA/NHLBI Scientific Statement 2005, the prime emphasis in individuals with metabolic syndrome is to mitigate the underlying risk factors (obesity, physical inactivity, and atherogenic diet) through lifestyle changes . In addition, 10-year risk assessment (e.g. age, gender, total cholesterol, smoking status) for cardiovascular disease should be carried out with algorithms such as the Framingham risk score [1, 7, 33].
All components of the IDF criteria can easily be measured in primary care and used as tools for counselling. However, when the majority of elderly individuals are defined as being in need of risk management, this becomes a challenge both for clinicians and researchers. It seems ethical questionable to identify such a large part of the population as being at risk, particularly since there still is doubt regarding the value of the metabolic syndrome as a risk marker. On the one hand, data indicate that the metabolic syndrome carries increased long-term risk both for cardiovascular disease and diabetes as well as a short-term risk . On the other hand, the meaning of the metabolic syndrome as a cardiovascular risk factor independent of other conventional risk factors (e.g., smoking, family history) has been questioned . In addition, few studies have examined whether this risk factor is as strong in elderly individuals as it is in younger or middle-aged individuals. A recent study among Swedish men followed for a maximum of 33 years found that the metabolic syndrome was an independent risk factor in middle age but not consistently in elderly men for cardiovascular and total mortality, when established cardiovascular risk factors were taken into account . Other studies among elderly people have not been conclusive: in British women, the syndrome was found only modestly associated with risk of coronary heart events in un-adjusted analyses, and not associated when adjusted for other risk factors , and in American men and women, the syndrome was found associated with increased cardiovascular events but not with total mortality .
Thus, implementation of the recent IDF and AHA/NHLBI guidelines for clinical management of the metabolic syndrome in elderly individuals may seem somewhat premature. First, further data are needed to assess the risk for cardiovascular and total mortality. Secondly, as the metabolic syndrome is not a reliable tool for overall risk assessment in the short term but more carries increased long-term risk , this may be of lower value among elderly than among younger people. Thirdly, our findings should give increased attention to the practical, ethical, and economic aspects of classifying a very high portion of asymptomatic elderly individuals as in need of counselling, overall risk assessment for cardiovascular disease, and long-term follow-up.