Schrijvers EMC, Koudstaal PJ, Hofman A, Breteler MMB. Plasma Clusterin and the Risk of Alzheimer Disease. JAMA. 2011;305(13):1322-1326. doi:10.1001/jama.2011.381
Author Affiliations: Departments of Epidemiology (Drs Schrijvers, Hofman, and Breteler) and Neurology (Drs Schrijvers and Koudstaal), Erasmus MC University Medical Center, Rotterdam, the Netherlands; and German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany (Dr Breteler).
Context Variants in the clusterin gene are associated with the risk of Alzheimer disease (AD) and clusterin levels have been found to be increased in brain and cerebrospinal fluid of patients with AD. Plasma clusterin was reported to be associated with brain atrophy, baseline disease severity, and rapid clinical progression in patients with AD.
Objective To evaluate the potential of plasma clusterin as a biomarker of the presence, severity, and risk of AD.
Design, Setting, and Participants A case-cohort study nested within the Rotterdam Study, a prospective population-based cohort study conducted in Rotterdam, the Netherlands. Plasma levels of clusterin were measured at baseline (1997-1999) in 60 individuals with prevalent AD, a random subcohort of 926 participants, and an additional 156 participants diagnosed with AD during follow-up until January 1, 2007 (mean [SD], 7.2 [2.3] years).
Main Outcome Measures Prevalent AD, severity of AD measured by the Mini-Mental State Examination (MMSE) score, and the risk of developing AD during follow-up.
Results The likelihood of prevalent AD increased with increasing plasma levels of clusterin (odds ratio [OR] per SD increase of plasma clusterin level, 1.63; 95% confidence interval [CI], 1.21-2.20; adjusted for age, sex, education level, apolipoprotein E status, diabetes, smoking, coronary heart disease, and hypertension). Among patients with AD, higher clusterin levels were associated with more severe disease (adjusted difference in MMSE score per SD increase in clusterin levels, −1.36; 95% CI, −2.70 to −0.02; P = .047). Plasma clusterin levels were not related to the risk of incident AD during total follow-up (adjusted HR, 1.00; 95% CI, 0.85-1.17; P for trend = .77) or within 3 years of baseline (adjusted HR, 1.09; 95% CI, 0.84-1.42; P for trend = .65).
Conclusion Plasma clusterin levels were significantly associated with baseline prevalence and severity of AD, but not with incidence of AD.
Several genome-wide association studies have identified the CLU gene, which encodes for clusterin, as a genetic locus involved in Alzheimer disease (AD).1- 3 The protein clusterin, also known as apolipoprotein J, has been suggested to be involved in the pathogenesis of AD.4,5 Clusterin has been found in the frontal cortex and hippocampus of postmortem AD brains6 and is increased in the cerebrospinal fluid of patients with AD.7 Plasma clusterin was reported to be associated with brain atrophy, baseline disease severity, and rapid clinical progression in AD, suggesting its possible use as a biomarker of AD.8
Data from a large population-based cohort study was used to examine the associations between plasma levels of clusterin and the prevalence, severity, and risk of AD in participants.
This study was based on participants of the Rotterdam Study, a large prospective population-based cohort study that is conducted among all inhabitants aged 55 years or older of Ommoord, a district of Rotterdam, the Netherlands.9 Baseline examinations were conducted between 1990 and 1993, with follow-up examinations conducted in 1993-1994, 1997-1999, and 2002-2004. At each survey, the diagnosis of dementia was made following a 3-step protocol.10 Two brief tests of cognition (Mini-Mental State Examination [MMSE]11 and Geriatric Mental State Schedule organic level12) were used to screen all participants. Participants who screened positive (an MMSE score of <26 or Geriatric Mental State Schedule organic level of >0) underwent the Cambridge Examination for Mental Disorders of the Elderly.13 Participants who were suspected of having dementia were examined by a neuropsychologist, if necessary. In addition, the total cohort was continuously monitored for incident dementia through computerized linkage between the study database and digitized medical records from general practitioners and the Regional Institute for Outpatient Mental Health Care.10 The diagnoses of dementia and AD were made in accordance with internationally accepted criteria for dementia (Diagnostic and Statistical Manual of Mental Disorders [Third Edition Revised]),14 Alzheimer disease (NINCDS-ADRDA),15 and vascular dementia (NINDS-AIREN)16 by a panel of a neurologist, neuropsychologist, and research physician. Follow-up for incident dementia was virtually complete (>98%) through January 1, 2007. The medical ethics committee at Erasmus University of Rotterdam, Rotterdam, the Netherlands, approved the study, and written informed consent was obtained from all participants.
We used a case-cohort study design, which is an established method that increases efficiency, especially when costly measurements are required.17 In this study design, a random subcohort is drawn from the total cohort at risk. Participants from the total cohort who develop the disease outside the subcohort are added to the analyses; however, only persons from the subcohort contribute follow-up time.
At the third survey, fasting blood samples were obtained at the research center. Citrate plasma (5 mL) was collected and stored at −80°C. In July 2008, 200 μL of citrate plasma from each participant was sent to Rules-Based Medicine, Austin,
Texas (http://www.rulesbasedmedicine.com), where clusterin levels were analyzed via multiplex immunoassay on a human multi-analyte profile. The least detectable dose was 1.3 μg/mL. The intra-assay variability was less than 4% and the interassay variability was less than 13%.
Educational level was assessed during the first interview, which took place between 1990 and 1993, and was dichotomized into primary education (with or without an unfinished higher education) vs lower vocational to university education. APOE (apolipoprotein E) genotype (RefSeqNG_007084) was assessed on coded DNA samples using polymerase chain reaction without knowledge of the dementia diagnosis. APOE ε4 status was defined as carriership of 1 or 2 ε4 alleles. If APOE genotype was missing (n = 42, 4.3%), APOE ε4 status was imputed as 0.28 (the proportion with an APOE ε4 allele in the total population with APOE genotyping). The MMSE score11 was assessed at the research center during the third survey. In addition, a dedicated neuropsychological test battery was used to assess executive function, attention, and information processing speed. The test battery included the Letter-Digit Substitution Task,18 the Word Fluency Test,19 and the abbreviated Stroop test.20
Hypertension was defined as a blood pressure of at least 140/90 mm Hg or use of antihypertensive medication, prescribed for the indication of hypertension. Coronary heart disease was defined as a previous myocardial infarction, percutaneous transluminal coronary angiography, or coronary artery bypass graft surgery. Smoking habits were assessed at the home interview. Diabetes was defined as a self-reported history of diabetes, registration by a general practitioner as having diabetes, or a fasting glucose level of at least 126 mg/dL (to convert to millimoles per liter, multiply by 0.055). Missing values in covariates (<5%) were imputed as the mean.
We used linear regression analyses to investigate the associations between the baseline characteristics and plasma clusterin levels. Analyses were adjusted for age and sex when applicable. All analyses were performed using SPSS statistical package 15.0 (SPSS Inc, Chicago, Illinois) or SAS version 9.2 (SAS Institute Inc, Cary, North Carolina). A priori level of significance was set at P ≤ .05 for all analyses.
First, we investigated the cross-sectional association between plasma levels of clusterin and prevalent AD and dementia using logistic regression models. After establishing that clusterin followed a normal distribution, clusterin was entered continuously per SD increase into the models and per quartile of its distribution. All analyses were adjusted for age and sex, and additional adjustments were made for educational level, APOE ε4 status, and vascular risk factors.
Second, to test whether plasma clusterin levels are associated with severity of AD within individuals with prevalent AD, we performed linear regression analyses of clusterin levels with the MMSE score and other cognitive test scores as the dependent variable.
Third, we investigated the association between plasma clusterin and the risk of developing incident AD during follow-up using Cox proportional hazards regression models with modification of the standard errors based on robust variance estimates. We used the method according to Barlow in which the random subcohort is weighted by the inverse of the sampling fraction from the total cohort at risk.17 All analyses were adjusted for age (used as the time scale) and sex, and additional adjustments were made for the abovementioned covariates.
In addition, to see whether plasma clusterin levels might have changed due to subclinical AD, subsequent analyses were performed on incident cases identified within and after 3 years of follow-up.
Of the 5990 individuals alive at the time of the third Rotterdam survey (1997-1999), 4797 participated in the survey. A total of 3795 participants had fasting blood samples drawn that could be used for clusterin assessment. Among these participants, 79 were diagnosed with prevalent dementia and 7 did not undergo dementia screening, resulting in a cohort of 3709 participants at risk for incident dementia.
From this cohort, we drew a random subcohort of 952 participants in 2008, of whom 926 had sufficient plasma remaining for clusterin measurement. As of follow-up through January 1, 2007, we identified 61 participants who developed dementia in this subcohort (of whom 52 were diagnosed with AD) with clusterin measurement and 178 participants who developed dementia in the rest of the cohort (of whom 156 were diagnosed with AD), resulting in 237 incident dementia cases in the analysis (2 incident dementia cases did not have enough plasma available for measurement of clusterin). Because we wanted to investigate the associations of plasma clusterin levels with both prevalent and incident dementia and AD, we also measured clusterin levels in the 77 patients with prevalent dementia with sufficient plasma for measurement.
Baseline characteristics of the source population, the subcohort, and the prevalent AD cases are shown in Table 1. The random subcohort with plasma clusterin measurements did not differ from the total cohort at risk. Mean follow-up time was 7.2 years (SD, 2.3 years; range, 0.1-9.7 years). Of the 77 prevalent dementia cases, 60 were diagnosed with AD, 9 with vascular dementia, and 8 with other types of dementia. Of the 237 incident dementia cases, 208 were diagnosed with AD (of whom 76 were diagnosed within 3 years of baseline), 20 with vascular dementia, and 9 with other types of dementia. Associations of the baseline characteristics with plasma clusterin levels are shown in Table 2.
Table 3 shows the associations of plasma clusterin levels with prevalent AD and with the risk of incident AD during follow-up. The odds that a participant had prevalent AD significantly increased by 49% for every SD increase in clusterin levels. This association became even stronger after further adjustments for educational level, APOE ε4 status, and vascular factors. There was no statistically significant association of plasma clusterin levels with incident AD during total follow-up or with incident AD within or after 3 years of baseline. Results for all-cause dementia and vascular dementia were similar and are shown in Table 4.
After adjusting for age and sex, clusterin levels were associated with the MMSE score in patients with prevalent AD (difference in MMSE score per SD increase in clusterin levels, −1.34; 95% confidence interval [CI], −2.54 to −0.13; P = .03), but not in controls without dementia (difference in MMSE score per SD increase in clusterin levels, −0.004; 95% CI, −0.128 to 0.120; P = .95). Adjusting for education level, APOE ε4 status, smoking, diabetes, coronary heart disease, and hypertension did not change the results (difference in MMSE score per SD increase in clusterin levels, −1.36; 95% CI, −2.70 to −0.02; P = .047 for patients with prevalent AD and −0.005; 95% CI, −0.126 to 0.116; P = .93 for controls without dementia). A smaller subset underwent additional cognitive tests (Letter-Digit Substitution Task, Word Fluency Test, and Stroop test), which largely showed the same pattern but did not reach statistical significance (Table 5).
In our population-based cohort study, plasma levels of clusterin were associated with the prevalence and severity of AD, but not with the development of incident AD during follow-up. Major strengths of our study were the population-based design and the long and virtually complete follow-up for incident dementia. Therefore, we were able not only to investigate the associations of plasma clusterin with the presence and severity of AD, but also to investigate whether plasma clusterin might be a preclinical marker of AD. However, magnetic resonance imaging was not routinely performed in the third survey; therefore, we were not able to investigate the relationship of plasma clusterin with brain or hippocampal atrophy. The relationship between plasma clusterin and progression of AD was also not investigated. We did explore the associations of clusterin with vascular dementia and all-cause dementia, which were similar to the associations with AD, suggesting that clusterin cannot be used to distinguish AD from vascular dementia. Other subtypes of dementia could not be investigated because of small numbers.
Our finding that plasma clusterin was associated with MMSE in patients with prevalent AD was similar to that of Thambisetty et al8; however, unlike their study, our patients with AD had significantly higher levels of plasma clusterin than controls. In addition, our data do not support the suggestion that clusterin is increased, possibly as an etiopathological event, before the development of AD,8 but fits the hypothesis that the increased expression of clusterin in AD reflects a neuroprotective response.5 Several protective effects of clusterin on the brain that may play a role in AD have been described in in vitro or in vivo studies, including inhibition of amyloid formation21 through binding amyloid-beta or enhancing its clearance over the blood-brain barrier,22 clearance by endocytosis of amyloid-beta aggregates and cell debris to brain phagocytes, and inhibition of complement activation.5 The neurodegenerative changes that occur in AD may trigger an increased expression of clusterin.5 This is in line with our finding that plasma clusterin was associated with prevalent AD and severity of AD, but not with the risk of developing incident AD during follow-up. Clusterin was also associated with prevalent all-cause dementia and vascular dementia, supporting a reactive rather than a causative role of clusterin and suggesting that clusterin will not be useful in the differential diagnosis of AD vs other subtypes of dementia.
In conclusion, our data from the general population show that increased plasma clusterin levels are associated with prevalent AD and are higher in more severe cases of AD. However, increased levels of clusterin do not precede development of AD and therefore are not a potential early marker of subclinical disease.
Corresponding Author: Monique M. B. Breteler, MD, PhD, Department of Epidemiology, Erasmus MC University Medical Center, PO Box 2040, 3000 CA Rotterdam, the Netherlands (email@example.com).
Author Contributions: Drs Schrijvers and Breteler had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Schrijvers, Breteler.
Acquisition of data: Schrijvers, Koudstaal, Hofman, Breteler.
Analysis and interpretation of data: Schrijvers, Breteler.
Drafting of the manuscript: Schrijvers, Breteler.
Critical revision of the manuscript for important intellectual content: Schrijvers, Koudstaal, Hofman, Breteler.
Statistical analysis: Schrijvers.
Obtained funding: Hofman, Breteler.
Administrative, technical, or material support: Schrijvers, Koudstaal, Hofman, Breteler.
Study supervision: Breteler.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Breteler reported receiving an unrestricted research grant from Pfizer. No other authors reported any disclosures.
Funding/Support: The Rotterdam Study is supported by the Erasmus MC University Medical Center and Erasmus University Rotterdam; the Netherlands Organization for Scientific Research (NWO); the Netherlands Organization for Health Research and Development (ZonMw); the Research Institute for Diseases in the Elderly (RIDE); the Ministry of Education, Culture, and Science; the Ministry of Health, Welfare, and Sports; the European Commission (DG XII); and the Municipality of Rotterdam. This study was financially supported by grant IIRG-06-27261 from the Alzheimer's Association and by an unrestricted research grant from Pfizer (Dr Breteler).
Role of the Sponsor: The funding organizations and sponsors had no role in the design and conduct of the study, in the collection, management, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript.