629
Journal of Alzheimer’s Disease 47 (2015) 629–638 DOI 10.3233/JAD-150193 IOS Press
Aluminum Levels in Brain, Serum, and Cerebrospinal Fluid are Higher in Alzheimer’s Disease Cases than in Controls: A Series of Meta-Analyses Sohaib A. Virk and Guy D. Eslick∗ The Whiteley-Martin Research Centre, Discipline of Surgery, The University of Sydney, Nepean Hospital, Penrith, New South Wales, Australia
Accepted 5 May 2015
Abstract. Background: Aluminum is the most studied environmental agent linked with Alzheimer’s disease (AD). However, it remains unclear whether levels are significantly elevated in AD sufferers. Objective: To systematically assess levels of aluminum in brain, serum, and cerebrospinal fluid (CSF) of AD cases and controls. Methods: Electronic searches of Medline, Embase, PubMed, and Cochrane Library were conducted up to June 2015. Studies reporting brain, serum, or CSF aluminum levels in individuals with AD and non-demented controls were included. Meta-analyses were performed using random-effects models and the pooled standardized mean difference (SMD) reported with 95% confidence intervals (CI). Results: Overall, 34 studies involving 1,208 participants and 613 AD cases met the criteria for inclusion. Aluminum was measured in brain tissue in 20 studies (n = 386), serum in 12 studies (n = 698), and CSF in 4 studies (n = 124). Compared to control subjects, AD sufferers had significantly higher levels of brain (SMD 0.88; 95% CI, 0.25–1.51), serum (SMD 0.28; 95% CI, 0.03–0.54), and CSF (SMD 0.48; 95% CI, 0.03–0.93) aluminum. Sensitivity analyses excluding studies without age-matched controls did not impact upon these results. Conclusions: The findings of the present meta-analyses demonstrate that aluminum levels are significantly elevated in brain, serum, and CSF of patients with AD. These findings suggest that elevated aluminum levels, particularly in serum, may serve as an early marker of AD and/or play a role in the development of the disease. These results substantially clarify the existing evidence examining the link between chronic aluminum exposure and the development of AD. Keywords: Alzheimer disease, aluminum, meta-analysis, systematic review
INTRODUCTION Alzheimer’s disease (AD) is the most common neurodegenerative condition, accounting for 60–80% of dementia cases [1]. In 2011, the global prevalence ∗ Correspondence
to: Guy D. Eslick, The Whiteley-Martin Research Center, Discipline of Surgery, The University of Sydney, Nepean Hospital, Level 3, Clinical Building, P.O. Box 63, Penrith, NSW 2751, Australia. Tel.: +61 247341373; Fax: +61 247343432; E-mail:
[email protected].
of AD was approximately 33.9 million and this figure is projected to more than triple by 2050 [1, 2]. Pathologically, AD is characterized by cortical atrophy and several pathological hallmarks, including the aggregation of amyloid- peptide in the form of neuritic plaques and the polymerization of hyperphosphorylated tau proteins as neurofibrillary tangles [3]. Although AD causality is known to have a genetic component, epidemiological studies have demonstrated discordance rates of up to 80% for monozygotic twins,
ISSN 1387-2877/15/$35.00 © 2015 – IOS Press and the authors. All rights reserved
630
S.A. Virk and G.D. Eslick / Aluminum Levels in Alzheimer’s Disease
indicating a significant etiological role for one or more environmental factors [4]. Aluminum is the most prominent of the environmental agents postulated to play a causal role in the development of AD. It is a potent neurotoxin and the most abundant metal in the Earth’s crust. Although fresh foods and most non-alum-treated drinking water contain relatively small amounts of aluminum, a diet rich in processed foods and alum-treated drinking water can contain 95 milligrams or more per day in the form of aluminum additives [5]. Aluminum additives tend to be more readily absorbed and thus have greater bioavailability than naturally occurring aluminum compounds [6]. Hence, over a lifetime, individuals can routinely be exposed to significant levels of aluminum through the chronic intake of small dietarysized doses. The evidence cited in support of a role for aluminum in AD can be summarized as follows: i) in vivo, aluminum demonstrates numerous adverse effects relevant to the pathophysiology of AD, including the depletion of microtubules and consequent cortical atrophy [7, 8]; ii) hippocampal tissue from humans with AD and brain tissue in transgenic and non-transgenic animal models demonstrate aluminum involvement in the formation of various AD hallmarks: neurofibrillary tangles [9], amyloid deposits [10–12], and granulovaculor degeneration [13, 14]; iii) aluminum has been implicated in the Alzheimer-type dementia that sometimes accompanies Parkinson’s disease [15] and in amyotrophic lateral sclerosis with parkinsonism dementia of Guam [16]; iv) an increased incidence of AD has been observed in populations with high levels of aluminum in drinking water supplies [17–20]; and v) aluminum/iron chelation is the most promising pharmacological treatment available for AD [21]. It is still unclear whether aluminum levels are significantly elevated in AD sufferers [22]. Numerous studies have been conducted to answer this question, but the small sample sizes of most individual studies and their conflicting findings have prevented a clear answer. Thus, the present series of meta-analyses were conducted to comprehensively and systematically quantify the association between aluminum levels and AD. METHODS Search strategy and study selection Electronic databases, including PubMed, Medline, Embase, and the Cochrane Library, were searched from their dates of inception to June 2015 for relevant
studies. To maximize sensitivity, the search terms “Alzheimer’s OR Alzheimer” AND “aluminum OR aluminium” were combined as both key words and MeSH terms without any language restrictions. This was supplemented by hand-searching the reference lists of review articles and all potentially relevant studies. The two authors (S.V. and G.E.) independently screened the title and abstract of records identified in the search. Full-text publications were subsequently reviewed separately if either reviewer considered the manuscript to be potentially eligible. Disagreements regarding final study inclusion were resolved by discussion and consensus. Eligibility criteria Eligible studies were those in which aluminum levels were reported in the brain, serum, or cerebrospinal fluid (CSF) of individuals with AD and non-demented control participants. Studies involving populations with multiple dementia types were excluded unless aluminum levels were specifically reported for the AD cohort. To be eligible for inclusion, studies were required to report sufficient data to compute an effect size with 95% confidence intervals (CI). To account for the risk of sample contamination when measuring aluminum levels, studies in which control serum values exceeded 25 g/L were excluded. Abstracts, case reports, conference presentations, editorials, and expert opinions were excluded. When institutions published duplicated studies with accumulating numbers of patients or increased lengths of follow-up, only the most complete reports were included for assessment. Data extraction Data extraction was performed using a standardized form, collecting information on the publication year, country, study design, number of cases and controls, mean age of participants, sampling source, the diagnostic criteria used to establish diagnosis of AD, the method used for analysis of aluminum levels, and data used to compute effect sizes and CI. All data were independently extracted from text, tables, and figures by both authors (S.V. and G.E.). When multiple aluminum samples were presented for a single participant, a combined mean and standard deviation was computed using the formula provided by the Cochrane Collaboration [23].
S.A. Virk and G.D. Eslick / Aluminum Levels in Alzheimer’s Disease
631
Statistical analysis
Effect sizes
The standardized mean difference (SMD) was used as the summary statistic. Meta-analyses were performed using random-effects models to take into account anticipated methodological diversity between studies. The I2 statistic was used to estimate the percentage of total variation across studies due to heterogeneity rather than chance, with values exceeding 50% indicative of considerable heterogeneity. In such cases, the possible clinical and methodological reasons for heterogeneity were explored qualitatively. Sensitivity analysis was conducted excluding studies in which there was a significant difference in the mean age of AD and control participants. Publication bias was assessed using funnel plots comparing effect sizes with their standard error. Egger’s regression test was used to detect funnel plot asymmetry and the Trim-and-Fill method was used to explore the impact of studies potentially missing due to publication bias [24, 25]. Meta-analyses were performed using Review Manager Version 5.2.1 and publication bias assessed using Comprehensive Metaanalysis v2.2 (Biostat Inc., Englewood, NJ, US). All p-values were two-sided, and values
