Metabolic syndrome and the risk of vascular dementia: the Italian Longitudinal Study on Ageing
- Vincenzo Solfrizzi1,
- Emanuele Scafato2,
- Cristiano Capurso3,
- Alessia D'Introno1,
- Anna Maria Colacicco1,
- Vincenza Frisardi1,
- Gianluigi Vendemiale3,4,
- Marzia Baldereschi5,
- Gaetano Crepaldi6,
- Antonio Di Carlo5,
- Lucia Galluzzo2,
- Claudia Gandin2,
- Domenico Inzitari7,
- Stefania Maggi6,
- Antonio Capurso1,
- Francesco Panza1
- for the Italian Longitudinal Study on Ageing Working Group*
- 1Department of Geriatrics, Center for Ageing Brain, Memory Unit, University of Bari, Bari, Italy;
- 2Population Health and Health Determinants Unit, National Centre for Epidemiology, Surveillance and Health Promotion (CNESPS), Istituto Superiore di Sanità (ISS), Rome, Italy;
- 3Department of Geriatrics, University of Foggia, Foggia, Italy;
- 4Internal Medicine Unit, IRCSS Casa Sollievo dalla Sofferenza, San Giovanni Rotondo, Foggia, Italy;
- 5Institute of Neuroscience, Italian National Research Council (CNR), Florence, Italy;
- 6Italian National Research Council (CNR), Ageing Section, Padua, Italy;
- 7Department of Neurological and Psychiatric Sciences, University of Firenze, Florence, Italy
- Correspondence to Dr Vincenzo Solfrizzi, Department of Geriatrics, Center for Ageing Brain, Memory Unit, University of Bari, Policlinico, Piazza Giulio Cesare, 11-70124 Bari, Italy;
Contributors VS and FP contributed to the concept, interpretation, and manuscript preparation. ES, CC, AD'I, AMC, VF, GV, MB, GC, ADC, LG, CG, DI, SM and AC contributed to interpretation and manuscript preparation.
- Received 27 April 2009
- Revised 14 September 2009
- Accepted 14 October 2009
- Published Online First 3 December 2009
Objective The authors investigated the relationship of metabolic syndrome (MetS) and its individual components with incident dementia in a prospective population-based study with a 3.5-year follow-up.
Methods A total of 2097 participants from a sample of 5632 subjects (65–84 years old) from the Italian Longitudinal Study on Ageing were evaluated. MetS was defined according to the Third Adults Treatment Panel of the National Cholesterol Education Program criteria. Dementia, Alzheimer disease (AD) and vascular dementia (VaD) were classified using current published criteria.
Results MetS subjects (N=918) compared with those without MetS (N=1179) had an increased risk for VaD (1.63% vs 0.85%, adjusted hazard ratio (HR) 3.71, 95% CI 1.40 to 9.83). After excluding 338 subjects with baseline undernutrition, MetS subjects compared with those without MetS had an elevated risk of VaD (adjusted HR, 3.82; 95% CI 1.32 to 11.06). Moreover, those with MetS and high inflammation had a still further higher risk of VaD (multivariate adjusted HR, 9.55; 95% CI 1.17 to 78.17) compared with those without MetS and high inflammation. On the other hand, those with MetS and low inflammation compared with those without MetS and low inflammation did not exhibit a significant increased risk of VaD (adjusted HR, 3.31, 95% CI 0.91 to 12.14). Finally, a synergistic MetS effect versus its individual component effects was verified on the risk of VaD.
Conclusion In our population, MetS subjects had an elevated risk of VaD that increased after excluding patients with baseline undernutrition and selecting MetS subjects with high inflammation.
Metabolic syndrome (MetS) is defined as a cluster of abdominal obesity, impaired fasting glucose, hypertension, low high-density lipoprotein (HDL) and/or high triglycerides.1 Most of the MetS components have been shown to be independent risk factors for coronary artery disease (CAD) and stroke, and MetS itself was already evidenced to be an independent risk factor for CAD, CAD mortality, and fatal and non-fatal stroke.2–4 In particular, MetS was associated with an increased risk of both clinical stroke and silent brain infarction by two- to fourfold.3 4
Several individual components of MetS have been linked to a risk of developing dementia and mild cognitive impairment (MCI)5. However, few studies have looked at the components of MetS as a whole. MetS has been previously shown to increase the risk of cognitive decline in different ethnic groups,6–10 overall dementia,11 Alzheimer disease (AD) in both population-based and case-control studies,12 13 frontal-subcortical geriatric syndrome14 and vascular dementia (VaD).15 While a cross-sectional study found a higher prevalence of dementia in women with MetS,8 a prospective study showed no association between MetS and incident dementia.16 Furthermore, one study suggested that the association between MetS and accelerated cognitive decline disappeared at the age of 85 and older.17
Some investigators have suggested that inflammation should be added as a component in the definition of MetS because it is such an important part of the pathophysiology.2 Moreover, in older, frail subjects, there is an increased prevalence of poorer health status, which is often the result of undernutrition, a condition that could also influence MetS.18 In the present study, we examined the association between MetS and its individual components with dementia (AD, VaD and other dementias), accounting for inflammation and undernutrition, factors that reflect frailty, in a large prospective epidemiological study conducted in community-dwelling older subjects.
Subjects and setting
The subjects of this study took part in a larger study, the Italian Longitudinal Study on Ageing (ILSA), promoted by the Italian National Research Council, in which a sample of 5632 independent subjects 65–84 years old was randomly selected from demographic rolls of eight municipalities, after stratification for age and gender. Data were obtained from the first prevalence survey study between March 1992 and June 1993, and from the follow-up survey study between September 1995 and October 1996. The ILSA was described in detail elsewhere.5 The study population consisted of 2097 free-living elderly subjects, representing the 37.2% of randomly selected ILSA participants in 1992–1993 (n=5632), who underwent both a clinical examination and laboratory assessment (figure 1). The study project was approved by the Institutional Review Board of the eight municipalities. Voluntary informed consent was obtained from each subject and/or their relatives before enrolment.
Laboratory analyses and clinical examination
Blood samples were obtained early in the morning after a 13 h overnight fast. Serum was removed after centrifugation at 1500×g for 20 min and was rapidly frozen and stored at –80°C, except for the volume needed for lipid determination analysed in the same day. Serum total cholesterol, triglycerides, HDL cholesterol, low-density lipoprotein (LDL) cholesterol, apolipoproteins B (apo B) and A-I (apo A-I), plasma glucose concentrations and fibrinogen were measured as detailed elsewhere.19 Non-HDL cholesterol was calculated, subtracting HDL cholesterol from serum total cholesterol. Serum albumin was measured by electrophoresis, as detailed elsewhere.20
The case-finding strategy for the diagnosis of dementia consisted of a two-phase procedure as reported in detail elsewhere.5 In phase 1, each subject was administered a screening questionnaire, a series of brief screening tests to identify suspect cases for further investigation, and clinical evaluation; in phase 2, suspected cases were confirmed with a standardised clinical examination by a certified neurologist or geriatrician. The main screening criteria for cognitive impairment or dementia were the MMSE with a cut-off score of 23, or a previous diagnosis reported by the respondent proxy. The diagnosis was based on the Diagnostic and Statistical Manual of Mental Disorders, third edition revised (DSM-III-R) criteria for dementia syndrome,21 the National Institute of Neurological and Communicative Disorders and Stroke—Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) criteria for possible and probable AD,22 and the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10) criteria for VaD and other dementing diseases.23
Cases of CAD (myocardial infarction or angina pectoris), type 2 diabetes mellitus, hypertension and stroke were identified with a two-phase procedure and using clinical criteria, as detailed elsewhere.5 MetS was diagnosed according to the Third Adults Treatment Panel of the National Cholesterol Education Program (ATP-III-NCEP) criteria.1 Based on this definition, subjects with MetS were identified by any combination of three or more of the following components: abdominal obesity (waist circumference >102 cm for men and >88 cm for women); elevated plasma triglycerides (≥150 mg/dl); low HDL cholesterol (<40 mg/dl for men and <50 mg/dl for women); high blood pressure (≥130/≥85 mm Hg) or being in hypertensive treatment; high fasting plasma glucose (≥110 mg/dl) or being in oral antidiabetic treatment. The inflammatory marker was fibrinogen.24 High inflammation status was defined as serum levels of fibrinogen ≥338 mg/dl (median value).
Based on self-reports, smoking habits were categorised as ‘ever’ or ‘never,’ and were assessed asking the study participants the amount of cigarettes smoked and the ages when they started and stopped smoking; from these data the variable ‘pack-years’ (years smoked×usual number of cigarettes smoked/20 cigarettes per pack) was derived. We collected information on alcohol consumption by food-frequency questionnaires completed in 1992. These data were transformed in drink as reported elsewhere.25 Body weight was measured on a balance beam platform scale (Salus, Milan, Italy) to the nearest 0.1 kg. Height was taken by a stadiometer (Salus Milan, Italy) at head level to the nearest centimetre with the subject standing barefoot, with feet together. Body mass index (BMI) was calculated as weight/height2 (kg/m2). The abdominal circumference was measured using a flexible steel tape at the end of expiration, by wrapping the tape at the level of the umbilicus to the nearest centimetre, with the subject standing.26 Undernutrition status was defined as BMI<22,27 and/or albumin serum level <3.2 mg/dl.28
All analyses have been performed using SAS statistical software (SAS/STAT user's guide, V.9.1 Cary, North Carolina: SAS Institute, 2004). Except for pack-years cigarettes and drink per day (Mann–Whitney test), we have used parametric usual statistical tests. Incidence rates of dementia (all causes) and types of dementia (AD, VaD and other dementias) for MetS and its individual components were calculated as the number of events per 1000 person-years at risk. Time to new events of dementia, AD, VaD and other dementias were modelled separately using Cox proportional hazards. The proportional hazard assumption over time for the covariates of interest was checked including in the Cox model each covariate by time as a predictor variable. The synergistic effect of MetS versus its individual component effect on risk of dementia (all causes), AD, VaD and other dementias was estimated using the Rothman synergy test.29 To assess this interaction as a departure from joint additive effects of individual components of MetS, we computed an approximation to the Rothman synergy index and its 95% CI based on model coefficients from the above Cox model. The difference between –2 log-likelihood of two multivariate Cox models for patients with and without the overall MetS parameter estimating the risk of dementia, AD, VaD and other dementias was calculated by the likelihood ratio test (LR test). These two models for each pathology of interest included all individual components of MetS and traditional risk factors of dementia. We used a model-building strategy involving three stages: variable specification, interaction assessment and confounding assessment by consideration of precision. Covariates that have been modelled as continuous were examined by quartile analysis to obtain the correct scale in the log hazard of dementia and types of dementia, using the lowest quartile as the reference group. The evidence of nonlinearity suggested that a categorical model be calculated. In order to check the proportional hazard assumption over time for the covariates for each type of event, we included in the Cox model each covariate by time as a predictor variable. The statistical significance threshold was set at 5% in the Cox model involving the whole sample, while the p value threshold was adjusted according to Bonferroni inequality in the subsample analyses. In particular, in the Cox model involving participants with high inflammation status at baseline and participants excluding undernutrition at baseline, the p value was set at 0.025 (0.05/2=2.5%); in the Cox model involving participants with high inflammation status excluding undernutrition at baseline, the p value was set at 0.0125 (0.05/4=1.3%).
The sociodemographic and clinical characteristics (mean±SD or %) of subjects with or without MetS are shown in table 1. The figure 1 shows the sample size and attrition in assessing the role of MetS on incident dementia. The incident cases of dementia were 88 (47 AD, 25 VaD and 16 other dementias). Differences in age (72.9±5.6 vs 75.0±5.7, p≤0.01 evaluated by separate variance t test) and sex (Pearson χ2=10.05, p≤0.01) were observed between participants and non-participants (2097 participants: 1106 (39.3%) men and 991 (35.2%) women; 3535 non-participants: 1.710 (60.7%) men and 1825 (64.8%) women). The mean (SD) age of the participants at baseline was 72.94 (5.56) years, 47.26% were women, and 50.30% had a high marker of inflammation (fibrinogen ≥338 mg/dl). Compared with participants without MetS (n=1179), those with MetS (n=918) were more likely to be women (63.94%), have a higher BMI and a history of vascular disease, use drugs affecting lipoprotein metabolism (statins and fibric acids) and have an elevated inflammation status (table 1). Among those with MetS, 16.09% met four, and 5.29% met five components of the ATP-III-NCEP criteria. Moreover, among those with MetS, the most common ATP-III-NCEP criterion was hypertension (93.57%), followed by abdominal obesity (78.72%), hypertriglyceridaemia (64.71%), low HDL cholesterol (61.22%) and high fasting glucose/antidiabetic medication use (48.15%). The prevalence of MetS in the study sample was 43.78%. The prevalence of the individual components of MetS in the whole population was 50.15% for abdominal obesity, 80.11% for hypertension, 36.93% for high triglycerides, 34.62% for low HDL cholesterol and 29.14% for hyperglycaemia.
Incidence rates of dementia for MetS and its individual components in whole sample
In individuals affected by MetS, the incidence rates (the number of events per 1000 person-years at risk) of dementia and AD were roughly the same for each component of MetS, while the incidence rate of VaD in affected individuals was slightly higher for hypertension than for other MetS components (table 2). On the other hand, in non-affected individuals, the incidence rates of dementia and AD were higher for low HDL cholesterol than for other MetS components (table 2).
Synergistic effect of MetS versus its individual component effects on the risk of dementia in a whole sample
No individual MetS component was statistically significantly associated with the change in risk of dementia (all causes), AD, VaD and other dementias in the whole sample. The difference in –2 loglikelihood between the two multivariate Cox models of patients with and without the overall MetS parameter estimating the risk of VaD was statistically significant (LR test: χ2=9.87, df=1, p<0.01) (table 3). No significant differences in −2loglikelihood of two multivariate Cox models for dementia (all causes), AD and other dementias with and without overall MetS parameter and each individual MetS component, were found. Finally, a synergistic effect due to overall MetS (ie, ≥3 MetS components) versus individual MetS component effects was verified on the risk of VaD (table 3). Moreover, the risk of VaD due to MetS was about four and a half times higher than the additive risk of its individual components (Rothman synergy index 4.66, 95% CI 1.05 to 28.67), indicating statistically significant interaction on an additive scale (table 3).
Risk of dementia according to MetS in the whole sample
In the first unadjusted model, the risk of VaD did not change significantly in MetS subjects in comparison with subjects without MetS (table 4, a1). On the other hand, in the second adjusted model compared with the subjects without MetS, those with MetS were more likely to have VaD (1.63% vs 0.85%, hazard ratio (HR) 3.71, 95% CI 1.40 to 9.83 in a multivariate adjusted model) (table 4, a2). The HRs of dementia, AD or other types of dementia in the whole study population did not change significantly in MetS subjects in comparison with subjects without MetS (table 4, a1 and a2).
Adjustment for undernutrition and inflammation (stratification)
In analysing possible confounders in the relationship between MetS and dementia, we evaluated the interaction of MetS with undernutrition (for dementia: HR 0.61 95% CI 0.14 to 2.61; for AD: HR 0.65, 95% CI 0.08 to 5.01; for VaD: HR 2.53 95% CI 0.65 to 9.85; and for other dementias: 5.99 95% CI 0.47 to 76.32.) and inflammation status (for dementia: HR 0.77 95% CI 0.36 to 1.66; for AD: HR 1.05 95% CI 0.37 to 2.97; for VaD: HR 1.12 95% CI 0.27 to 4.64; and for other dementias: HR 0.28 CI 95% 0.02 to 3.34). Then, we performed the analysis reported in the previous section excluding subjects with undernutrition at baseline (BMI<22 and/or albumin serum level <3.2 mg/dl). Among the excluded undernourished subjects, those with MetS have a borderline lower risk of dementia (HR 0.19, 95% CI 0.032 to 1.13 for multivariate model) compared with those without MetS (data no shown). There were 17 cases of dementia occurring in the undernourished subjects group (nine AD, four VaD and four other dementias; only two subjects will be demented during the follow-up in the undernourished subjects with MetS). However, no statistical difference in age (72.98±5.47 vs 74.10±5.97), education (8.61±5.57 vs 7.89±5.57), and MMSE score (26.24±4.54 vs 26.33±4.08) at baseline between undernourished subjects with and without MetS has been observed. The adjustment for MMSE score at baseline significantly reduces the rate of dementia in subjects affected by MetS in comparison with non-affected individuals (HR 0.10, 95% CI 0.01 to 0.77). We did not repeat the analysis for AD, VaD and other dementias for the small number of new events in these subsamples. Moreover, excluding subjects with undernutrition at baseline, those with MetS have an elevated risk of VaD (1.79% vs 0.65%, HR, 5.49; 95% CI 1.62 to 18.55 for multivariate model) compared with those without MetS (table 4, c2). On the other hand, excluding subjects with baseline undernutrition, MetS subjects with a high inflammation status had a still further higher risk of VaD (1.66% vs 0.44%, HR, 9.55; 95% CI 1.17 to 78.17 for multivariate model) compared with those without MetS and high inflammation status (table 4, d2). Moreover, those with MetS and low inflammation (No=357) compared with those without MetS and low inflammation (No=464) did not exhibit a significant increased risk of VaD (2.00% vs 0.87%, fully adjusted HR, 3.31, 95% CI 0.91 to 12.14) (data not shown). Finally, the hazard ratios of dementia, AD or other dementias in subjects with high inflammatory status, in subjects excluding undernutrition, and in subjects with high inflammatory status excluding undernutrition (table 4, c1, c2, d1, and d2) or low inflammatory status excluding undernutrition (data not shown) did not change significantly in MetS subjects in comparison with subjects without MetS. When we repeated the analysis, subdividing the whole sample on the basis of subjects with a high inflammation status at baseline, we did not found any significantly elevated risk of dementia, AD, VaD and other dementias in MetS subjects compared with those without MetS (table 4, b1 and b2).
In the present study with a 3.5-year follow-up, MetS subjects had a threefold increased risk of VaD, but not of AD or other dementias, compared with those without MetS only after adjustment for traditional risk factors. Moreover, after excluding subjects with undernutrition at baseline, MetS subjects had an approximately 5.5-fold increased risk of VaD, and those with MetS and a high inflammation status still showed a 9.5-fold increased risk of VaD. On the other hand, those with MetS and low inflammation status did not show any significant change. Among the five components identifying MetS, hypertension had the higher incidence rate for VaD and low HDL cholesterol for AD in the whole study sample. Finally, a synergistic effect due to MetS versus its individual component effects was verified on the risk of VaD.
The main finding of the present study is that MetS appears to be a risk factor of VaD, and not AD. In the Honolulu–Asia Ageing Study (HAAS) a clustering of seven metabolic cardiovascular risk factors increased the risk of dementia, and in particular of VaD, but not AD.11 However, this clustering of cardiovascular risk factors did not identify a MetS diagnosis according to current criteria. More recently, in the Brazilian community-dwelling elderly study, frontal-subcortical syndrome was strongly associated with MetS diagnosed by an adapted ATP III criteria, independently of age, gender or presence of stroke.14 In this recent study, MetS was also significantly associated with lower cognitive, executive and neuromotor functions, and depressive symptoms.14 However, at present, there is no clear consensus on the diagnosis of frontal-subcortical geriatric syndrome,30 and while white-matter lesions (WMLs) and lacunar infarcts strongly correlate and predict aspects of the frontal-subcortical geriatric syndrome,30 this syndrome appears to be clinically different by VaD diagnosed with current criteria. Therefore, whereas MetS has been transversally linked to a syndrome that seems to be equivalent to an earlier stage of VaD,14 to date only a study has evaluated if MetS is a risk factor for VaD itself, moreover utilising for this purpose a proper longitudinal design. In fact, very recently, in the Three-City Study, on 7087 French non-institutionalised elderly subjects aged 65 years, the presence of MetS diagnosed by ATP III criteria at baseline was associated with an increased risk of incident VaD but not all-cause dementia and AD over 4 years, independently of sociodemographic characteristics and the apolipoprotein E ɛ4 allele.15
Indicators of health status are not usually included in the set of traditional risk factors for dementia. In the present study, the association between MetS and VaD was strengthened by almost 30% when subjects with baseline undernutrition were excluded, and the risk increased by another 50% when those with high inflammation status were selected in multivariate model. In fact, recent evidence suggested that being underweight in midlife may increase the risk of dementia (AD and VaD) in late life,31 so by eliminating patients with baseline undernutrition, we improved the homogeneity of our populations sample. Moreover, in this study, among the excluded undernourished subjects, those with MetS have a borderline lower risk of dementia compared with those without MetS, although this observation could be explained by the presence of subclinical dementia at baseline. Then, we also hypothesised that in an elderly population, the usual adjustment for the traditional risk factors, cardiovascular and cerebrovascular diseases (stroke or small-vessels disease) does not account for possible biological changes associated with frailty, which can affect the relationship between MetS and VaD. Therefore, adjustment for factors reflecting frailty, such as undernutrition and inflammatory status,32 could allow us to demonstrate with increased validity a direct and significant relation between MetS and dementia disorders in older persons.
Furthermore, high levels of inflammation increase the risk of the development of diabetes and atherosclerosis33 and have been associated with an increased risk of developing dementia and cognitive decline.34 Inflammation and MetS are probably related in a circular process, with inflammation leading to MetS, and MetS increasing inflammation.6–8 Different studies on different ethnic groups have consistently found that MetS subjects with elevated inflammation have a greater risk of cognitive decline or cognitive impairment than those with MetS and without elevated inflammation status.6–8 In this picture, MetS contributes to accelerated atherosclerosis that is associated with an inflammatory response, and in turn either the atherosclerosis or inflammation or both contribute to cognitive decline.35 On the other hand, proinflammatory cytokines have been implicated as mediators of undernutrition and cachexia in a wide variety of chronic diseases.32 These cytokines can affect nutritional status by induction of anorexia or appetite suppression and by direct metabolic changes such as induction of an increased catabolic state.36
In the present study, a synergistic effect due to MetS versus its individual component effects was verified on the risk of VaD. This evidence reinforces the relationship of cause and effect between MetS and VaD in comparison with the individual components that constitute the syndrome. Moreover, the risk of VaD due to MetS is about four times higher than the additive risk of the individual MetS components. Our present findings confirmed the results from the HAAS and the Three-City Study in which no association between MetS and AD was found,11 15 but our results did not support those from population-based and case-control studies in which MetS appeared to increase the risk of AD.12 13 In fact, in a population-based study of 980 randomly selected Finnish elderly subjects, MetS increased the risk for AD, but only in women; however, this study was only cross-sectional and, consequently, cannot address the question of whether MetS predicts AD.12 A more recent case-control study demonstrated a link between MetS and AD, but important limitations were the relatively small sample and the design of the study.13 In fact, it was not possible to distinguish between those factors that may precede the development of AD and have a causal role, and those factors that may have been altered as a consequence of the disease. There was also a selection bias because AD patients were selected from memory clinics; therefore, there could be a spurious identification of risk factors associated with clinic attendance rather than the disease itself.13
The mechanism by which MetS could increase the risk of VaD is, at present, unknown. However, most previous studies assessing the risk factors for VaD focused on stroke and hypertension,18 and MetS increases the risk of both clinical stroke and silent brain infarction by two- to fourfold.3 4 Moreover, recent studies have identified other risk factors for VaD, including diabetes mellitus, obesity and triglyceride levels,18 all components of MetS. Finally, in a very recent Japanese study, the presence of MetS tripled the risk of having any leucoaraiosis.37 Leucoaraiosis, which belongs to the microvascular extreme on the spectrum of cerebrovascular diseases, is an insidious and often progressive process that is associated with brain atrophy and cognitive decline.38 On the basis of these findings, we can speculate that MetS contributes to VaD by causing microvascular damage with subsequent white-matter injury and disrupted cortical connectivity.
The strengths of the present study were its prospective design, the population-based setting and its large number of subjects, although the number of the new cases was relatively small (overall 88 dementia patients). Moreover, we identified a more specific elderly population in which to study the role of MetS in the risk of VaD. Nonetheless, some limitations in this study have to be considered. Although only in the whole sample, undernourished subjects showed a non-significant lower risk of dementia (all causes) in MetS subjects in comparison with subjects without MetS, the highly skewed CI suggested that this effect, while not statistically significant, may be important. It should be considered that undernourished people with MetS might have a milder form of MetS due to lower peripheral insulin resistance. Inflammation status was defined only by increased fibrinogen levels, while in other studies on this topic, baseline serum levels of interleukin-6, α1-antichymotrypsin or C reactive protein (CRP) were also measured.7 9 However, fibrinogen is an acute-phase protein, and high levels serve as non-specific markers for inflammatory disease. In the Rotterdam Study, higher levels of fibrinogen but not of CRP were associated with an increased risk of both VaD and AD, and this association was independent of cardiovascular risk factors, clinical stroke and other inflammatory markers.39 In the ILSA, among factors that are potential risk factors for dementia and might be associated with some MetS components, thus acting as potential confounders, we did not have any information on the apolipoprotein E (APOE) ε4 allele status. However, few studies on the relationship between MetS and dementia included the APOE ε4 status, with no relevant effect found.12 15 16 Finally, ICD-10 criteria for VaD diagnosis have a high sensitivity but a rather low specificity, but are quite inclusive, as compared for example with NINDS-AIREN criteria.40 Future studies will need to address whether preventing MetS or lowering inflammatory status may prevent VaD onset in healthy elderly individuals.
Funding This work was supported by the Italian Longitudinal Study on Ageing (ILSA) (Italian National Research Council—CNR-Targeted Project on Ageing—Grants 9400419PF40 and 95973PF40) (VS, ES, CC, AD'I, AMC, MB, GC, ADC, LG, CG, DI, SM, AC and FP).
Competing interests None.
Ethics approval Ethics approval was provided by the regional medical school ethics committees.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
The ILSA Working Group E Scafato (Scientific Coordinator), G Farchi, L Galluzzo, C Gandin, Istituto Superiore di Sanità, Rome; A Capurso, F Panza, V Solfrizzi, V Lepore, P Livrea, University of Bari; L Motta, G Carnazzo, M Motta, P Bentivegna, University of Catania; S Bonaiuto, G Cruciani, D Postacchini, Italian National Research Centre on Ageing (INRCA), Fermo; D Inzitari, L Amaducci, University of Florence; A Di Carlo, M Baldereschi, Italian National Research Council (CNR), Firenze; C Gandolfo, M Conti, University of Genoa; N Canal, M Franceschi, San Raffaele Institute, Milan; G Scarlato, L Candelise, E Scapini, University of Milan; F Rengo, P Abete, F Cacciatore, University of Naples; G Enzi, L Battistin, G Sergi, G Crepaldi, University of Padua; S Maggi, N Minicucci, M Noale, Italian National Research Council (CNR), Ageing Section, Padua; F Grigoletto, E Perissinotto, Institute of Hygiene, University of Padua; P Carbonin, Università Cattolica del Sacro Cuore, Rome.