Environment International 69 (2014) 141–147

Contents lists available at ScienceDirect

Environment International journal homepage: www.elsevier.com/locate/envint

Plasma polychlorinated biphenyl and organochlorine pesticide concentrations in dementia: The Canadian Study of Health and Aging Thierry Comlan Marc Medehouenou a,b, Pierre Ayotte c,d,e, Pierre-Hugues Carmichael b, Edeltraut Kröger a,b, René Verreault b,c, Joan Lindsay c,f, Éric Dewailly c,d,e, Suzanne L. Tyas g,h, Alexandre Bureau c,i, Danielle Laurin a,b,⁎ a

Faculté de pharmacie, Université Laval, Quebec City, Québec, Canada Centre d'excellence sur le vieillissement de Québec, Santé des populations et pratiques optimales en santé, Centre de recherche du CHU de Québec, Quebec City, Québec, Canada c Faculté de médecine, Département de médecine sociale et préventive, Université Laval, Quebec City, Québec, Canada d Axe Santé des populations et pratiques optimales en santé, Centre de recherche du CHU de Québec, Quebec City, Québec, Canada e Laboratoire de toxicologie, Institut national de santé publique du Québec, Quebec City, Québec, Canada f Department of Epidemiology and Community Medicine, Faculty of Medicine, University of Ottawa, Ottawa, Ontario, Canada g School of Public Health and Health Systems, University of Waterloo, Waterloo, Ontario, Canada h Department of Psychology, University of Waterloo, Waterloo, Ontario, Canada i Centre de recherche de l'Institut universitaire en santé mentale de Québec, Quebec City, Québec, Canada b

a r t i c l e

i n f o

Article history: Received 12 November 2013 Accepted 22 April 2014 Available online 20 May 2014 Keywords: Alzheimer disease Dementia Exposure Organochlorine pesticide Polychlorinated biphenyl

a b s t r a c t Background: Even though polychlorinated biphenyls (PCBs) and organochlorine (OC) pesticides are recognized as neurotoxicants, few studies have investigated their associations with dementia. Here, we assess associations of plasma PCB and OC pesticide concentrations with all-cause dementia and Alzheimer's disease (AD). Methods: Analyses are based on data from the Canadian Study of Health and Aging, a population-based study of men and women aged 65+ years at baseline. PCB and OC pesticide concentrations were measured in 2023 participants who had complete clinical evaluations and blood samples; 574 had dementia, including 399 cases of AD. Concentrations were log-transformed and used as continuous variables in logistic regression models to assess their individual associations with dementia and AD. Results: After adjustment for blood collection period, total plasma lipids, age, sex, education, apolipoprotein E e4 allele (ApoE4), tobacco and alcohol use, rural/urban residence, and comorbidities, elevated plasma PCB concentrations were not associated with increased prevalence of dementia and AD. Elevated concentrations of some OC pesticides and metabolites such as hexachlorobenzene, cis-nonachlor and 1,1,1-trichloro-2,2bis(p-chlorophenyl)ethane were significantly associated with a reduced prevalence of dementia. A significant reduced prevalence of AD was also observed with elevated hexachlorobenzene concentrations. Other OC pesticides and metabolites were not associated with the prevalence of dementia or AD. No effect modification by sex and ApoE4 was observed for either dementia or AD. Conclusions: Elevated plasma PCB and OC pesticide concentrations were not associated with higher prevalence of all-cause dementia and AD. The possibility of modest reductions in prevalence with specific OC pesticides remains to be further investigated given the cross-sectional design of this study. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction Organochlorines (OCs), including polychlorinated biphenyls (PCBs) and OC pesticides, are ubiquitous and persistent environmental contaminants that accumulate in lipid tissues of living organisms (Li et al., 2006). Experimental studies have shown that PCBs and OC pesticides cause adverse effects on the central nervous system, suggesting potential ⁎ Corresponding author at: Centre d'excellence sur le vieillissement de Québec, Hôpital du Saint-Sacrement, 1050 chemin Sainte-Foy, Local L2-30, Quebec City, Québec G1S 4L8, Canada. Tel.: +1 418 682 7511x4832; fax: +1 418 682 7998. E-mail address: [email protected] (D. Laurin).

http://dx.doi.org/10.1016/j.envint.2014.04.016 0160-4120/© 2014 Elsevier Ltd. All rights reserved.

neurotoxic effects on humans (Mariussen and Fonnum, 2006). Few cross-sectional (Fitzgerald et al., 2008; Haase et al., 2009; Schantz et al., 2001; van Wendel de Joode et al., 2001) and prospective (Lin et al., 2008) epidemiologic studies have linked PCBs and OC pesticides to neurocognitive and neurobehavioral impairment in older adults. Dementia and its most common subtype, Alzheimer's disease (AD), are characterized by cognitive deficits. The etiology of AD is still unknown although several genetic and environmental risk factors have been identified (Chouliaras et al., 2010; Maloney et al., 2012; Marques et al., 2011). Even though PCBs and OC pesticides are recognized as neurotoxicants and thus potential environmental contributors to AD, little is known about their effects on the risk

142

T.C.M. Medehouenou et al. / Environment International 69 (2014) 141–147

of AD. Over the last decades, few studies have investigated their associations with dementia or AD. One case report suggested a relationship between occupational exposure to a PCB mixture (Aroclor 1260) and the development of dementia with features similar to AD (Troster et al., 1991). Another study from the National Institute for Occupational Safety and Health databases showed an increased standardized mortality ratio for dementia in women highly exposed to PCBs, but not in men (Steenland et al., 2006). A cohort study in an agricultural community suggested significant increased risk estimates for dementia and AD among pesticide-exposed individuals; however, the increased risk estimates were non-significant among those exposed to OC pesticides (Hayden et al., 2010). In the latter two studies, the use of exposure assessments based on job–exposure matrices or detailed questionnaires in person, which are prone to bias, limits the validity of the results. To date, no epidemiologic study has examined the association of plasma concentrations of PCBs and OC pesticides with clinically assessed cases of dementia and AD. The use of such circulating concentrations yields more valid and reliable information and allows identification of specific PCB congeners and OC pesticides. The aim of the present study was to assess the associations of plasma PCB and OC pesticide concentrations with the prevalence of all-cause dementia and AD using data from the Canadian Study of Health and Aging (CSHA). 2. Materials and methods 2.1. Study population and dementia assessment The CSHA is a national cohort study of older Canadians designed to examine prevalence, incidence and risk factors for dementia. Eighteen research centers across the country and one coordinating center were involved. Methodological details have been described elsewhere (Lindsay et al., 2004; The Canadian Study of Health and Aging Working Group, 1994). The baseline examination was carried out in 1991–1992 (CSHA-1) with two follow-up examinations performed five and ten years later. The three phases received approval from institutional ethics committees in all participating centers; participants and/or family representatives gave written consent at each phase. In CSHA-1, a random sample of 10,263 men and women, representative of the Canadian population aged 65 and over, was drawn from Enumeration Composite Records in Ontario and from Medicare lists in the other provinces for 36 urban and surrounding rural areas where roughly 60% of Canadians were living. Institutionalized participants were randomly selected from residents in stratified random samples of institutions in each region. The study excluded the Yukon and the Northwest Territories, Indian reserves and military units. Of the study participants, 9008 were living in the community and 1255 in institutions. Community-living participants were screened for dementia using a cut-off of 77/78 on the 100-point Modified Mini-Mental State (3MS) examination (The Canadian Study of Health and Aging Working Group, 1994). Participants who screened positive (3MS score b78), a random sample of those who screened negative (3MS score N 77) and all participants living in institutions were invited to attend an extensive standardized three-stage clinical evaluation. A nurse readministered the 3MS, collected information on medication use, and obtained the participant's medical and family histories from a relative. A physician solicited information on the participant's medical history and performed a standardized clinical and neurological examination. Non-fasting blood samples were drawn for laboratory tests (required if dementia or delirium was presumed) or collected for future analyses (optional) at the end of the examination. Finally, a psychometrist administered a neuropsychological test battery to participants with a 3MS score of 50 and over; results were interpreted by a neuropsychologist. Preliminary diagnoses of all-cause dementia according to the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition, revised criteria (DSM-III-R) were made independently by the physician and the neuropsychologist, who subsequently arrived at a final diagnosis in a consensus conference.

Final diagnoses also included AD according to the National Institute of Neurological and Communicative Disorders and Stroke — Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) criteria. Vascular dementia (VaD), other specific and unclassifiable dementias, no cognitive impairment, and cognitive impairment with no dementia (CIND) were diagnosed according to DSM-III-R and the International Classification of Diseases, 10th revision criteria. Information on risk factors was collected at baseline (CSHA-1) with a self-administered risk factor questionnaire (RFQ) covering sociodemographic and lifestyle characteristics as well as family and medical histories. Participants with no dementia were invited to complete the RFQ at home and return it by mail. As part of a first case–control study with prevalent cases of AD, the same RFQ was administered to proxies of prevalent cases of participants with dementia, and to proxies of controls (The Canadian Study of Health and Aging, 1994). All participants initially evaluated were contacted in 1996–1997 (CSHA-2) to assess changes in health status and functioning. CSHA-2 participants underwent a similar diagnostic process (screening and clinical evaluation) as in CSHA-1. If participants had been clinically examined at CSHA-1, they were automatically invited to CSHA-2 clinical examination. CSHA-2 diagnoses were made blinded to CSHA-1 diagnoses. To account for the modification of diagnostic criteria over the interval between CSHA-1 and CSHA-2, final diagnoses for dementia and AD were made for all participants according to the criteria used in CSHA-1 as well as revised according to the more recent criteria (Lindsay et al., 2004). All participants who were clinically assessed were asked to provide blood samples, except for those who previously gave blood in CSHA-1. The last phase of CSHA (CSHA-3) took place in 2001–2002. The cutoff for screening with the 3MS was increased to 89/90 to focus on more complete identification of participants with cognitive impairment. Participants unable to complete the neuropsychological evaluation or who had received a diagnosis of CIND or dementia from the neuropsychologist were asked to attend the clinical examination. The final diagnosis was made according to the same processes and diagnostic criteria as those used in CSHA-2 (Lindsay et al., 2004). All participants who were clinically assessed in CSHA-3 were asked to provide blood samples, except for those who previously gave blood either in CSHA-1 or CSHA-2. 2.2. Blood sampling Nine out of the 18 study centers in CSHA-1 and all study centers in CSHA-2 and CSHA-3 participated in the creation of a blood bank for future analyses. Plasma fractions were divided into aliquots and stored at −20 °C at the National Microbiology Laboratory, initially part of Health Canada in Ottawa, and more recently part of the Public Health Agency of Canada in Winnipeg, Canada. In total, blood samples were available for 2119 participants: 422 out of the 2914 clinically assessed in CSHA-1, 1312 out of the 2305 in CSHA-2, and 385 out of the 1386 in CSHA-3. Of the 2119 participants, 96 had to be excluded because of lack of consensus diagnosis or stored plasma, leaving 2023 participants for chemical and statistical analyses (Fig. 1). The measurement of OCs received approval from the research ethics review board of the CHU de Québec (Hôpital de l'Enfant-Jésus), Quebec City, Canada. 2.3. Laboratory analyses Plasma OC concentrations were determined as described previously (Medehouenou et al., 2011) at the Laboratoire de toxicologie of the Institut national de santé publique du Québec, which is accredited under ISO 17025 by the Standards Council of Canada, and participates regularly in many international quality control programs (Raaschou-Nielsen et al., 2005). Briefly, 15 PCB congeners (nos. 28, 52, 99, 101, 105, 118, 128, 138, 153, 156, 163, 170, 180, 183, 187) and 11 OC pesticides or their metabolites [aldrin, mirex, α-chlordane, γ-chlordane, oxychlordane, cis-nonachlor, trans-nonachlor, β-hexachlorocyclohexane (β-HCH), hexachlorobenzene (HCB), 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane

T.C.M. Medehouenou et al. / Environment International 69 (2014) 141–147

143

CSHA (n = 10,263)

CSHA-1, 1991-92

Clinical examination (n= 2914)

Blood collection (n = 422)

No blood collection

Clinical examination (n = 2305)

CSHA-2, 1996-97

Blood collection (n = 1312)

No blood collection

Clinical examination (n = 1386)

CSHA-3, 2001-02

Blood collection (n = 385)

No blood collection

Study sample (n = 2119) Exhausted plasma (n=40) No consensus diagnosis (n=56) Analytic sample (n = 2023)

With dementia With no dementia

CSHA-1 200 212

CSHA-2 273 977

CSHA-3 101 260

Fig. 1. Flowchart of the analytic sample.

(p,p′-DDT), 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (p,p′-DDE)] were identified and quantified by gas chromatography–mass spectrometry using negative chemical ionization. Limits of detection (LOD) ranged from 0.01 to 0.70 μg/L for PCB congeners, and from 0.005 to 0.20 μg/L for OC pesticides/metabolites. Between-day coefficients of variation were ≤11% for PCB congeners except for PCB 28 (13.3%) and PCB 128 (20.8%), and ≤ 12% for OC pesticides and metabolites except for oxychlordane (14.7%), p,p′-DDT (16.9%) and aldrin (29.7%). Samples with measures below LOD were imputed with a value of half of the LOD (Hornung and Reed, 1990). Cholesterol and triglyceride concentrations were determined by enzymatic tests on Roche automated clinical chemistry analyzers at the Laboratoire de biochimie du Centre de recherche du CHU de Québec, Quebec City. Total plasma lipids were calculated using the two-component formula proposed by Phillips and colleagues (Phillips et al., 1989). ApoE genotypes were determined by the modified method of Zivelin (McLeod et al., 1998). 2.4. Covariates and potential confounders Information on covariates and potential confounders based on the literature on AD were extracted from the RFQ at baseline or from the clinical examination at blood collection. These included blood collection

period, sex, presence of an apolipoprotein E e4 allele (ApoE4) (yes/no) and residence area (rural/urban). Smoking and drinking status were dichotomized as having ever been a regular smoker (almost every day or not) or a regular drinker (at least one alcoholic drink a week or not). A summary score (0–3) for vascular risk factors was based on the presence of hypertension (defined as a supine blood pressure N160 mm Hg systolic or 95 mm Hg diastolic, a physician's diagnosis, or the use of medication for hypertension), history of myocardial infarction or stroke, and history of diabetes mellitus. Total plasma lipids (g/L), age (years), education (years), body mass index (BMI, weight (kg)/(height (m2)) and the vascular score were treated as continuous variables. 2.5. Statistical analysis Analyses were performed individually for OCs where at least 60% of samples had concentrations greater than the LOD. These OCs were 10 PCBs: congeners 105 (% N LOD, 85.7%), 118 (99.8%), 138 (99.0%), 153 (100%), 156 (99.8%), 163 (99.8%), 170 (100%), 180 (100%), 183 (94.5%) and 187 (99.8%); and 7 OC pesticides/metabolites: β-HCH (99.2%), HCB (91.1%), oxychlordane (99.8%), trans-nonachlor (99.8%), cis-nonachlor (95.6%), p,p′-DDE (98.4%), and p,p′-DDT (62.5%). Given the cross-sectional nature of the study design and the three time points at which blood samples were collected, diagnosis at the

144

T.C.M. Medehouenou et al. / Environment International 69 (2014) 141–147

time of blood collection was based on the CSHA-1 criteria. Characteristics of individuals with dementia and those with no dementia were compared using chi-square test for categorical variables and Student's t test or non-parametric Wilcoxon rank-sum test (also called the Mann–Whitney U test) for continuous variables. For multivariate analyses, multiple imputation (MI) was performed to take into account observations with missing covariates. MI allows unbiased estimation under the assumption that information is missing at random, conditional on measured covariates (van Buuren, 2007). Five imputed datasets were generated and analyzed separately using multivariate logistic regression. The estimate odds ratios (ORs) from imputed datasets were then pooled to generate a single OR with 95% confidence interval (CI). Generalized additive models (Hastie and Tibshirani, 1990), which make no prior assumptions on the shape of the link function, were used to assess the assumption of linearity in logits for each continuous variable. Total plasma lipids were introduced in the model on a cubic scale and education on a piecewise linear scale, while individual logtransformed wet-weight (crude) OC concentrations, age, BMI and the vascular score on the usual linear scale. To account for the three sampling times at which blood samples were collected and pooling together for this cross-sectional study, and for the potential time elapsed after environmental peak emissions and human metabolic and environmental degradation rates (Quinn and Wania, 2012) in this study, the blood collection period (i.e., CSHA-1, CSHA-2 and CSHA-3) was included in all models as a covariate. Since lipid-standardized OC concentrations (i.e., OC concentrations divided by total plasma lipid concentrations) are highly prone to bias (Schisterman et al., 2005), total plasma lipid concentrations were rather included in all models as covariate. Therefore, the first model was adjusted for blood collection period, total plasma lipids, age, sex, education, and ApoE4. The second model was additionally adjusted for BMI, smoking, alcohol, rural/urban residence, and vascular score. Effect modification by sex and ApoE4 was tested by entering interaction terms in fully adjusted models for the associations of four indicator PCBs (congeners 118, 138, 153, 180) and three sentinel OC pesticides or metabolites (β-HCH, trans-nonachlor, p,p′-DDE) with dementia and AD. These seven selected OCs were characterized by both high prevalence (i.e., N 98% of participants had concentrations of these specific OCs above LODs) and high concentrations (Medehouenou et al., 2011). We also conducted a sensitivity analysis using individual log-transformed plasma lipid-standardized OC concentrations instead of individual logtransformed wet-weight OC concentrations in the models. Statistical analyses were performed using SAS software (version 9.2 SAS Institute Inc., Cary, NC, USA). 3. Results Baseline characteristics of participants from the analytic sample were similar to those from the CSHA baseline cohort in age, sex and education. Of the 2023 participants in the analytic sample, 574 were diagnosed with all-cause dementia; this group included 399 participants with AD, 113 with VaD, 32 with other specific dementias and 30 with unclassifiable dementia. Compared with individuals with no dementia, those with dementia were older (85.8 vs. 82.7 years), more likely to be women (66.6 vs. 58.5%), had fewer years of education (9.1 vs. 9.6 years), showed a higher proportion of ApoE4 carriers (34.8 vs. 20.6%), had a lower mean BMI (23.7 vs. 25.2 kg/m2), a lower mean concentration of total plasma lipids (5.64 vs. 5.92 g/L), and were less likely to be regular smokers (41.0 vs. 50.2%) or alcohol drinkers (29.9 vs. 38.1%) (Table 1). No difference between individuals with no dementia and those with dementia groups for rural/urban residence or vascular score was observed. Results were similar for AD cases except that they showed a lower vascular score than individuals with no dementia (1.04 vs. 1.24). Plasma cis-nonachlor and trans-nonachlor concentrations were modestly lower in those with dementia, including AD cases, than in

Table 1 Baseline characteristics of individuals with no dementia and those with dementia including AD. Baseline characteristics

Individuals with no dementia (n = 1449)

Individuals with dementia (n = 574)

Individuals with AD (n = 399)

Age, y Sex, female, n (%) Education, y Body mass index, kg/m2 Total plasma lipids, g/L ApoE4 carrier, n (%) Smoking, n (%)a No Yes Alcohol drinking, n (%)b No Yes Residence area, n (%) Rural Urban Vascular scorec

82.7 ± 6.5 848 (58.5) 9.6 ± 4.0 25.2 ± 4.8 5.92 ± 1.63 298 (20.6)

85.8 ± 6.8⁎⁎⁎ 382 (66.6)⁎⁎⁎ 9.1 ± 3.6⁎ 23.7 ± 5.0⁎⁎⁎ 5.64 ± 1.75⁎⁎ 199 (34.8)⁎⁎⁎

86.8 ± 6.7⁎⁎⁎ 283 (70.9)⁎⁎⁎ 9.1 ± 3.4⁎ 23.7 ± 5.2⁎⁎⁎ 5.68 ± 1.75⁎ 139 (34.9)⁎⁎⁎

628 (49.8) 633 (50.2)

280 (59.0)⁎⁎⁎ 195 (41.0)⁎⁎⁎

206 (62.6)⁎⁎⁎ 123 (37.4)⁎⁎⁎

789 (61.9) 486 (38.1)

330 (70.1)⁎⁎⁎ 141 (29.9)⁎⁎⁎

240 (73.4)⁎⁎⁎ 87 (26.6)⁎⁎⁎

183 (12.7) 1257 (87.3) 1.24 ± 0.88

79 (14.1) 482 (85.9) 1.21 ± 0.92

45 (11.5) 345 (88.5) 1.04 ± 0.88⁎⁎⁎

AD, Alzheimer's disease; ApoE4, apolipoprotein E e4 allele. Values are represented as mean ± standard deviation unless stated otherwise. Information was missing for 276 individuals on body mass index (131 with no dementia vs. 145 with dementia including 104 with AD); 7 individuals on ApoE4 status (5 with no dementia vs. 2 with dementia including 1 with AD); 22 individuals on residence area (9 with no dementia vs. 13 with dementia including 9 with AD); 287 individuals on smoking (188 with no dementia vs. 99 with dementia 70 with AD); 277 individuals on alcohol drinking (174 with no dementia vs. 103 with dementia including 72 with AD); and 44 individuals on total plasma lipids (38 with no dementia vs. 6 with dementia including 5 with AD). P-values were obtained using χ2 test for dichotomous variables and t-tests for continuous variables. ⁎ P b 0.05 (referent group includes individuals with no dementia). ⁎⁎ P b 0.01 (referent group includes individuals with no dementia). ⁎⁎⁎ P b 0.001 (referent group includes individuals with no dementia). a Ever been smoking regularly (nearly everyday). b Ever been drinking regularly (at least once a week). c A summary score for vascular risk factors including hypertension, history of myocardial infarction or stroke, and history of diabetes mellitus.

individuals with no dementia. Modestly higher plasma PCB 163 and p,p′-DDE concentrations were observed in AD cases compared to individuals with no dementia (Table 2). No differences for other plasma OC concentrations between groups were noted. Table 3 summarizes the association of prevalent dementia and AD with a 1-unit increase in log-transformed plasma PCB and OC pesticide concentrations. Elevations of PCB congeners 105 and 118 in model 1 were associated with a reduced prevalence of dementia (PCB 105: OR = 0.87, 95% confidence interval (CI) 0.77–0.99; PCB 118: OR = 0.86, 95% CI 0.74–0.99), but not in the fully adjusted models. Other PCBs were not associated with dementia. No associations were found between PCB congeners and the prevalence of AD. Analyses of OC pesticides and their metabolites showed that elevated concentrations of HCB in both partially and fully adjusted models were associated with a reduced prevalence of dementia (model 1: OR = 0.85, 95% CI 0.75–0.97; model 2: OR = 0.87, 95% CI 0.77–0.99) and AD (model 1: OR = 0.83, 95% CI 0.71–0.96; model 2: OR = 0.86, 95% CI 0.73–1.00, p = 0.044). Elevations of cis-nonachlor were associated with a reduced prevalence of dementia (model 1: OR = 0.83, 95% CI 0.72–0.97; model 2: OR = 0.85, 95% CI 0.73–0.99); no association was observed with the prevalence of AD. Elevations of p,p′-DDT were associated with a reduced prevalence of dementia in both models (model 1: OR = 0.82, 95% CI 0.72–0.93; model 2: OR = 0.84, 95% CI 0.74–0.96). Higher p,p′-DDT concentrations were also associated with a reduced prevalence of AD in the partially (OR = 0.86, 95% CI 0.75–1.00), but not the fully adjusted model. No associations with dementia or AD were found with β-HCH, oxychlordane, trans-nonachlor or p,p′-DDE. The association of dementia or AD with the four indicator PCBs and the three sentinel OC pesticides was not modified by sex or ApoE4

T.C.M. Medehouenou et al. / Environment International 69 (2014) 141–147 Table 2 Plasma OC concentrations (μg/L) in individuals with no dementia and those with dementia including AD. OC concentrations

PCB congeners PCB 105 PCB 118 PCB 138 PCB 153 PCB 156 PCB 163 PCB 170 PCB 180 PCB 183 PCB 187 OC pesticides β-HCH HCB Oxychlordane cis-Nonachlor trans-Nonachlor p, p′-DDT p,p′-DDE

Individuals with no dementia (n = 1449)

Individuals with dementia (n = 574)

Individuals with AD (n = 399)

0.02 (0.01–0.04) 0.13 (0.08–0.22) 0.24 (0.16–0.35) 0.42 (0.29–0.60) 0.06 (0.04–0.08) 0.079 (0.05–0.12) 0.10 (0.07–0.14) 0.34 (0.24–0.49) 0.03 (0.02–0.05) 0.09 (0.06–0.15)

0.02 (0.01–0.04) 0.14 (0.08–0.22) 0.25 (0.15–0.36) 0.42 (0.28–0.63) 0.06 (0.04–0.08) 0.08 (0.06–0.12) 0.10 (0.07–0.14) 0.36 (0.25–0.49) 0.03 (0.02–0.05) 0.10 (0.06–0.15)

0.02 (0.01–0.04) 0.15 (0.09–0.23) 0.26 (0.16–0.37) 0.45 (0.30–0.64) 0.06 (0.04–0.09) 0.083 (0.06–0.13)⁎⁎ 0.11 (0.07–0.15) 0.37 (0.26–0.50) 0.04 (0.02–0.05) 0.10 (0.06–0.15)

0.12 (0.08–0.19) 0.16 (0.09–0.27) 0.11 (0.07–0.15) 0.02 (0.01–0.03) 0.14 (0.10–0.21) 0.07 (0.03–0.12) 3.90 (1.80–7.80)

0.12 (0.08–0.18) 0.15 (0.08–0.28) 0.10 (0.07–0.14) 0.01 (0.01–0.02)⁎⁎ 0.13 (0.09–0.20)⁎ 0.06 (0.03–0.12) 4.40 (2.20–7.70)

0.12 (0.08–0.18) 0.16 (0.08–0.29) 0.10 (0.07–0.14) 0.01 (0.01–0.02)⁎⁎ 0.13 (0.09–0.20)⁎ 0.07 (0.02–0.13) 4.70 (2.40–8.10)⁎

Values are represented as median (interquartile range). AD, Alzheimer's disease; PCB, polychlorinated biphenyl; β-HCH, beta hexachlorocyclohexane; HCB, hexachlorobenzene; OC, organochlorine; p,p′-DDT, 1,1,1-trichloro-2,2bis(p-chlorophenyl)ethane; p,p′-DDE, 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene. P-values were obtained using non-parametric Wilcoxon rank-sum tests. ⁎ P b 0.05 (referent group includes individuals with no dementia). ⁎⁎ P b 0.01 (referent group includes individuals with no dementia).

(data not shown). Similar results were found in sensitivity analyses using plasma lipid-standardized OC concentrations instead of wetweight OC concentrations (data not shown). 4. Discussion Although some results were indicative of marginal protective effects, the present study suggests that overall there is no association between

145

plasma concentrations of PCBs and OC pesticides or their metabolites and the prevalence of all-cause dementia and AD. These results are based on a large representative sample of older Canadians, clinically assessed according to a standardized procedure, and plasma concentrations of PCBs and OC pesticides, which are regarded as the best estimates of cumulative exposure to PCBs and OC pesticides and their metabolites across the lifespan. These plasma OC concentrations were well within the ranges observed in other exposure studies (Medehouenou et al., 2011). Few studies have examined the effects of PCBs and OC pesticides on dementia in older populations. A retrospective study of 16,906 PCB-exposed US plant workers showed an excess of death due to noncerebrovascular dementia mortality among highly exposed women (Steenland et al., 2006). Limitations of this study included the use of modified job–exposure matrices, case assessment from death certificates, and a small number of deaths related to dementia. Other studies have explored the relationship between exposure to unspecific pesticides and dementia/AD (Zaganas et al., 2013). Using a question from the RFQ on occupational exposure to pesticides and fertilizers, the CSHA previously reported a significant two-fold increase in risk for AD in a case–control study (The Canadian Study of Health and Aging, 1994). Environmental exposure to pesticides, based on residential and agriculture census histories, was not related to the risk of AD in age- and sex-paired case–control study (68 pairs of Canadians aged N70 years) (Gauthier et al., 2001). In a cohort study of 1507 older French adults, an increased risk of AD was noted in men, but not in women, with occupational exposure to pesticides assessed by a jobexposure matrix (Baldi et al., 2003). An ecological study conducted in Spain showed that prevalence rates and the risk of having AD were significantly higher in districts with greater pesticide use than in those with lower use (Parrón et al., 2011). While previous studies failed to provide information on the type of pesticides, the study of 3084 older residents of an agricultural community in Utah reported a significantly increased risk for AD, but not dementia, among individuals exposed to organophosphate pesticides, and a nearly significant increased risk of AD among those exposed to OC pesticides (technical grade DDT) (Hayden et al., 2010). The comparability of our results to those of other studies is limited by differences in exposure ascertainment, and

Table 3 Associations of prevalent dementia and AD with a 1-unit increase in log-transformed plasma OC concentrations (μg/L). OC concentrations

PCB congeners PCB 105 PCB 118 PCB 138 PCB 153 PCB 156 PCB 163 PCB 170 PCB 180 PCB 183 PCB 187 OC pesticides β-HCH HCB Oxychlordane cis-Nonachlor trans-Nonachlor p,p′-DDT p,p′-DDE

All-cause dementia (n = 2023)

Alzheimer's disease (n = 1848)

Model 1

Model 2

Model 1

Model 2

OR (95% CI)

OR (95% CI)

OR (95% CI)

OR (95% CI)

0.87 (0.77–0.99) 0.86 (0.74–0.99) 0.93 (0.80–1.07) 0.95 (0.80–1.12) 1.14 (0.94–1.38) 1.04 (0.87–1.24) 1.07 (0.91–1.27) 1.06 (0.90–1.25) 0.98 (0.85–1.13) 1.03 (0.89–1.19)

0.90 (0.79–1.02) 0.88 (0.76–1.02) 0.93 (0.80–1.07) 0.93 (0.79–1.10) 1.09 (0.90–1.33) 1.02 (0.85–1.22) 1.04 (0.87–1.24) 1.02 (0.86–1.21) 0.98 (0.85–1.13) 1.02 (0.88–1.18)

0.95 (0.82–1.10) 0.94 (0.80–1.12) 1.06 (0.89–1.26) 1.11 (0.91–1.36) 1.23 (0.99–1.54) 1.15 (0.93–1.42) 1.17 (0.96–1.43) 1.14 (0.94–1.38) 1.09 (0.92–1.29) 1.12 (0.95–1.33)

0.99 (0.85–1.15) 0.99 (0.83–1.17) 1.07 (0.89–1.28) 1.11 (0.91–1.36) 1.23 (0.97–1.54) 1.15 (0.93–1.43) 1.16 (0.95–1.42) 1.12 (0.92–1.37) 1.09 (0.92–1.29) 1.12 (0.94–1.34)

0.90 (0.78–1.05) 0.85 (0.75–0.97) 0.96 (0.79–1.18) 0.83 (0.72–0.97) 0.89 (0.74–1.06) 0.82 (0.72–0.93) 0.93 (0.85–1.03)

0.92 (0.79–1.07) 0.87 (0.77–0.99) 0.97 (0.79–1.19) 0.85 (0.73–0.99) 0.90 (0.75–1.08) 0.84 (0.74–0.96) 0.94 (0.86–1.04)

0.90 (0.75–1.07) 0.83 (0.71–0.96) 0.91 (0.72–1.16) 0.85 (0.72–1.02) 0.91 (0.73–1.12) 0.86 (0.75–1.00) 1.01 (0.96–1.21)

0.92 (0.77–1.10) 0.86 (0.73–1.00) 0.93 (0.73–1.19) 0.89 (0.74–1.06) 0.93 (0.75–1.16) 0.89 (0.77–1.03) 1.02 (0.90–1.15)

Model 1 was adjusted for period of blood collection, total plasma lipid concentrations, age, sex, education, and ApoE4. Model 2 was additionally adjusted for body mass index, smoking status, alcohol drinking status, rural/urban residence, and vascular score. 95% CI, 95% confidence interval; AD, Alzheimer's disease; β-HCH, beta hexachlorocyclohexane; HCB, hexachlorobenzene; OC, organochlorine; OR, odds ratio; PCB, polychlorinated biphenyl; p,p′-DDT, 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane; p,p′-DDE, 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene. The p value for this specific result is b0.05.

146

T.C.M. Medehouenou et al. / Environment International 69 (2014) 141–147

by the fact that we studied subjects in the general population whereas other studies focussed on highly exposed workers. In our study, analyses were performed with individual PCB congeners (nos. 105, 118, 138, 153, 156, 163, 170, 180, 183, and 187) and with individual OC pesticides/metabolites (β-HCH, HCB, oxychlordane, cis-nonachlor, trans-nonachlor, p,p′-DDT, and p,p′-DDE). As results, none of these PCB congeners were associated with all-cause dementia or AD, although exposure to PCB leads to their bioaccumulation in fatty tissues including brain tissue where they can cause several adverse effects such as oxidative and neuroinflammatory responses, alterations of Ca2 + homeostasis and impairment in neurotransmitter functions (Kodavanti, 2005; Mariussen and Fonnum, 2006). These adverse effects are mainly involved in dementia due to AD characterized pathologically by a progressive loss of neurons and synapses, and by extracellular deposits of amyloid-β (Aβ) as senile plaques, Aβ deposits in the cerebral blood vessels, and intracellular inclusions of hyperphosphorylated tau as neurofibrillary tangles (Querfurth and LaFerla, 2010). The relatively low plasma PCB concentrations in addition to the possibility that those concentrations could not reflect exposure levels at the relevant window of exposure, although the half-lives of PCBs are approximately on the order of decades (Ritter et al., 2011; Seegal et al., 2011), might explain the lack of association found in our study. Among OC pesticides/metabolites, elevated concentrations of HCB, cis-nonachlor and p,p′-DDT were associated with a reduced prevalence of all-cause dementia; HCB concentrations were also associated with a reduced prevalence of AD. The meaning of these apparent inverse associations is unclear. It is reasonable to think that these associations are rather fallacious than compelling since p,p′-DDE, the long-lasting metabolite of p,p′-DDT, failed to show any significant association. Our findings contrast from those reported in a recent US case–control study including 79 control and 86 AD cases, where elevated serum DDE concentrations were associated with an increased risk for AD, with ApoE4 carriers being more susceptible (Richardson et al., 2014). Overall, there is no evidence of association between prevalence of dementia/AD and concentrations of plasma PCBs and OC pesticides/ metabolites in this cross-sectional design. In view of the limited research on the relationship between plasma PCB concentrations and dementia, additional studies using cohort designs should be conducted to confirm and extend the present results. The major strength of this study compared with the previous is that it is a population-based study with a relatively large sample size, and with dementia/AD diagnoses according to the international criteria. A second advantage is the use of plasma concentrations of PCBs and OC pesticides which are more informative than questionnaires and job–exposure matrices. A third advantage is the measurement and control for important confounders. Inasmuch as dementia may affect total lipids, plasma concentrations of total lipids were used as a covariate in all models to perform lipid correction of crude plasma OC concentrations in order to avoid systematic bias rather than simply dividing crude plasma OC values by the plasma total lipid values. The major limitation to our study is its cross-sectional nature: causal inference cannot be drawn from these results, and longitudinal data are needed. In addition, as it is frequently the case in population-based studies, not all participants were clinically evaluated and thus had the opportunity to provide a blood sample, which may introduce a selection bias. Moreover, the cut-off for the screening test determining eligibility for clinical assessment was lower in the first two phases of CSHA, which may also affect the representativeness of our sample. Compared to the initial CSHA cohort, our results showed, however, that our sample was similar in baseline characteristics (age, sex, education). Misclassification is an additional limitation, since some of the information on smoking and alcohol drinking were based on proxy reports. However, comparability between self and proxy reports has been examined in a subsample of the CSHA cohort: kappa statistics between the RFQ completed by non-cognitively impaired participants and their proxies were 0.85 for smoking status, and 0.70, 0.61 and 0.47 for beer, spirits and

wine consumption, respectively (The Canadian Study of Health and Aging, 1994). We think that this potential information bias has not substantially influenced our results; nevertheless, residual confounding is possible. Lastly, we were not able to account for potential confounders including heavy metals such as lead and mercury. These heavy metals have been associated with brain development impairment at a very early age, but the evidence regarding their putative role on neurodegeneration at older age is less compelling and still elusive (Mutter et al., 2010; Sanders et al., 2009). These variables were not available in our analytic sample. Moreover, we were not able to account for weight change from the blood collection period. However, we do not think that weight change within old age could significantly affect our results, although weight loss is a concern of both dementia/AD and OC. It will be interesting to account for the weight change from middle or late adulthood to old age. 5. Conclusion This study is one of the first to assess the association of PCB and OC pesticide/metabolite concentrations with the prevalence of all-cause dementia and AD using data from a large population-based study of older Canadians. Elevated plasma PCB and OC pesticide/metabolite concentrations were not related to an increased prevalence of all-cause dementia and AD. Further assessment of these associations in a longitudinal study is warranted given the cross-sectional design of the study. Acknowledgments The core funding for phases 1 and 2 of the Canadian Study of Health and Aging was provided by the Seniors' Independence Research Program, through the Health Canada's National Health Research and Development Program (NHRDP). Other funding was provided by Pfizer Canada Incorporated, by Bayer Incorporated and by the British Columbia Health Research Foundation. The core funding for phase 3 was provided by the Canadian Institutes of Health Research (CIHR). Other funding was provided by Merck-Frosst and Janssen-Ortho. The present study was financially supported by grants from the CIHR. Dr. Medehouenou is a recipient of a PhD scholarship from the Alzheimer Society of Canada (ASC) and the Canadian Dementia Knowledge Translation Network (CDKTN). Drs. Bureau and Laurin are supported by scientist awards from the Fonds de Recherche du Québec — Santé (FRQS). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. References Baldi I, Lebailly P, Mohammed-Brahim B, Letenneur L, Dartigues JF, Brochard P. Neurodegenerative diseases and exposure to pesticides in the elderly. Am J Epidemiol 2003; 157:409–14. Chouliaras L, Sierksma AS, Kenis G, Prickaerts J, Lemmens MA, Brasnjevic I, et al. Gene– environment interaction research and transgenic mouse models of Alzheimer's disease. Int J Alzheimers Dis 2010. http://dx.doi.org/10.4061/2010/859101. [pii: 859101]. Fitzgerald EF, Belanger EE, Gomez MI, Cayo M, McCaffrey RJ, Seegal RF, et al. Polychlorinated biphenyl exposure and neuropsychological status among older residents of upper Hudson River communities. Environ Health Perspect 2008;116:209–15. Gauthier E, Fortier I, Courchesne F, Pepin P, Mortimer J, Gauvreau D. Environmental pesticide exposure as a risk factor for Alzheimer's disease: a case–control study. Environ Res 2001;86:37–45. Haase RF, McCaffrey RJ, Santiago-Rivera AL, Morse GS, Tarbell A. Evidence of an agerelated threshold effect of polychlorinated biphenyls (PCBs) on neuropsychological functioning in a Native American population. Environ Res 2009;109:73–85. Hastie TJ, Tibshirani RJ. Generalized Additive Models. New York: Chapman and Hall; 1990. Hayden KM, Norton MC, Darcey D, Ostbye T, Zandi PP, Breitner JC, et al. Occupational exposure to pesticides increases the risk of incident AD: the Cache County study. Neurology 2010;74:1524–30. Hornung RW, Reed LD. Estimation of average concentration in the presence of nondetectable values. Appl Occup Environ Hyg 1990;5:48–51. Kodavanti PR. Neurotoxicity of persistent organic pollutants: possible mode(s) of action and further considerations. Dose Response 2005;3:273–305. Li QQ, Loganath A, Chong YS, Tan J, Obbard JP. Persistent organic pollutants and adverse health effects in humans. J Toxicol Environ Health A 2006;69:1987–2005.

T.C.M. Medehouenou et al. / Environment International 69 (2014) 141–147 Lin KC, Guo NW, Tsai PC, Yang CY, Guo YL. Neurocognitive changes among elderly exposed to PCBs/PCDFs in Taiwan. Environ Health Perspect 2008;116:184–9. Lindsay J, Sykes E, McDowell I, Verreault R, Laurin D. More than the epidemiology of Alzheimer's disease: contributions of the Canadian Study of Health and Aging. Can J Psychiatry 2004;49:83–91. Maloney B, Sambamurti K, Zawia N, Lahiri DK. Applying epigenetics to Alzheimer's disease via the latent early-life associated regulation (LEARn) model. Curr Alzheimer Res 2012;9:589–99. Mariussen E, Fonnum F. Neurochemical targets and behavioral effects of organohalogen compounds: an update. Crit Rev Toxicol 2006;36:253–89. Marques SC, Oliveira CR, Pereira CM, Outeiro TF. Epigenetics in neurodegeneration: a new layer of complexity. Prog Neuropsychopharmacol Biol Psychiatry 2011;35:348–55. McLeod D, Arnott B, Gaudreault N, Boudreau S, Sevigny P. A comparison of two methods for routine, accurate determination of apolipoprotein E genotypes. Alzheimers Rep 1998;1:211–5. Medehouenou TCM, Ayotte P, Carmichael PH, Kroger E, Verreault R, Lindsay J, et al. Polychlorinated biphenyls and organochlorine pesticides in plasma of older Canadians. Environ Res 2011;111:1313–20. Mutter J, Curth A, Naumann J, Deth R, Walach H. Does inorganic mercury play a role in Alzheimer's disease? A systematic review and an integrated molecular mechanism. J Alzheimers Dis 2010;22:357–74. Parrón T, Requena M, Hernández AF, Alarcon R. Association between environmental exposure to pesticides and neurodegenerative diseases. Toxicol Appl Pharmacol 2011; 256:379–85. Phillips DL, Pirkle JL, Burse VW, Bernert Jr JT, Henderson LO, Needham LL. Chlorinated hydrocarbon levels in human serum: effects of fasting and feeding. Arch Environ Contam Toxicol 1989;18:495–500. Querfurth HW, LaFerla FM. Alzheimer's disease. N Engl J Med 2010;362:329–44. Quinn CL, Wania F. Understanding differences in the body burden–age relationships of bioaccumulating contaminants based on population cross sections versus individuals. Environ Health Perspect 2012;120:554–9. Raaschou-Nielsen O, Pavuk M, Leblanc A, Dumas P, Philippe Weber J, Olsen A, et al. Adipose organochlorine concentrations and risk of breast cancer among postmenopausal Danish women. Cancer Epidemiol Biomarkers Prev 2005;14:67–74. Richardson JR, Roy A, Shalat SL, von Stein RT, Hossain MM, Buckley B, et al. Elevated serum pesticide levels and risk for Alzheimer disease. JAMA Neurol 2014;71:284–90.

147

Ritter R, Scheringer M, MacLeod M, Moeckel C, Jones KC, Hungerbuhler K. Intrinsic human elimination half-lives of polychlorinated biphenyls derived from the temporal evolution of cross-sectional biomonitoring data from the United Kingdom. Environ Health Perspect 2011;119:225–31. Sanders T, Liu Y, Buchner V, Tchounwou PB. Neurotoxic effects and biomarkers of lead exposure: a review. Rev Environ Health 2009;24:15–45. Schantz SL, Gasior DM, Polverejan E, McCaffrey RJ, Sweeney AM, Humphrey HE, et al. Impairments of memory and learning in older adults exposed to polychlorinated biphenyls via consumption of Great Lakes fish. Environ Health Perspect 2001;109: 605–11. Schisterman EF, Whitcomb BW, Louis GM, Louis TA. Lipid adjustment in the analysis of environmental contaminants and human health risks. Environ Health Perspect 2005;113:853–7. Seegal RF, Fitzgerald EF, Hills EA, Wolff MS, Haase RF, Todd AC, et al. Estimating the halflives of PCB congeners in former capacitor workers measured over a 28-year interval. J Expo Sci Environ Epidemiol 2011;21:234–46. Steenland K, Hein MJ, Cassinelli II RT, Prince MM, Nilsen NB, Whelan EA, et al. Polychlorinated biphenyls and neurodegenerative disease mortality in an occupational cohort. Epidemiology 2006;17:8–13. The Canadian Study of Health and Aging. The Canadian Study of Health and Aging: risk factors for Alzheimer's disease in Canada. Neurology 1994;44:2073–80. The Canadian Study of Health and Aging Working Group. Canadian study of health and aging: study methods and prevalence of dementia. CMAJ 1994;150:899–913. Troster AI, Ruff RM, Watson DP. Dementia as a neuropsychological consequence of chronic occupational exposure to polychlorinated biphenyls (PCBs). Arch Clin Neuropsychol 1991;6:301–18. van Buuren S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat Methods Med Res 2007;16:219–42. van Wendel de Joode B, Wesseling C, Kromhout H, Monge P, Garcia M, Mergler D. Chronic nervous-system effects of long-term occupational exposure to DDT. Lancet 2001;357: 1014–6. Zaganas I, Kapetanaki S, Mastorodemos V, Kanavouras K, Colosio C, Wilks MF, et al. Linking pesticide exposure and dementia: what is the evidence? Toxicology 2013; 307:3–11.