Traffic related air pollution and cognitive function in a cohort of older men


Published on

Traffic-Related Air Pollution and Cognitive Function in a Cohort of Older Men

Published in: Health & Medicine, Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Traffic related air pollution and cognitive function in a cohort of older men

  1. 1. ResearchTraffic-Related Air Pollution and Cognitive Function in a Cohort of Older MenMelinda C. Power,1,2 Marc G. Weisskopf,1,2 Stacey E. Alexeeff,1 Brent A. Coull,3 Avron Spiro III,4,5and Joel Schwartz1,21Department of Environmental Health, 2Department of Epidemiology, and 3Department of Biostatistics, Harvard School of Public Health,Boston, Massachusetts, USA; 4Veterans Affairs Boston Healthcare System, Boston, Massachusetts, USA; 5Department of Epidemiology,Boston University School of Public Health, Boston, Massachusetts, USA black carbon (BC), a marker of traffic-related Background: Traffic-related particles induce oxidative stress and may exert adverse effects on air pollution, would be associated with cog- central nervous system function, which could manifest as cognitive impairment. nitive function in a cohort of older men, oBjective: We assessed the association between black carbon (BC), a marker of traffic-related air because traffic-related air pollution is a major pollution, and cognition in older men. source of exposure to ultrafine PM, which Methods: A total of 680 men (mean ± SD, 71 ± 7 years of age) from the U.S. Department of may induce CNS dysfunction. Veterans Affairs Normative Aging Study completed a battery of seven cognitive tests at least once between 1996 and 2007. We assessed long-term exposure to traffic-related air pollution using a vali- Materials and Methods dated spatiotemporal land-use regression model for BC. Study sample. We drew the study sample results: The association between BC and cognition was nonlinear, and we log-transformed BC from participants of the U.S. Department estimates for all analyses [ln(BC)]. In a multivariable-adjusted model, for each doubling in BC on of Veterans Affairs (VA) Normative Aging the natural scale, the odds of having a Mini-Mental State Examination (MMSE) score ≤ 25 was Study (NAS), an ongoing longitudinal cohort 1.3 times higher [95% confidence interval (CI), 1.1 to 1.6]. In a multivariable-adjusted model for study of men established in 1963 (Bell et al. global cognitive function, which combined scores from the remaining six tests, a doubling of BC was associated with a 0.054 SD lower test score (95% CI, –0.103 to –0.006), an effect size similar 1972). Participants in the NAS are invited to that observed with a difference in age of 1.9 years in our data. We found no evidence of heteroge- to undergo an in-person examination every neity by cognitive test. In sensitivity analyses adjusting for past lead exposure, the association with 3 years. In addition to a physical examina- MMSE scores was similar (odds ratio = 1.3; 95% CI, 1.1 to 1.7), but the association with global tion and laboratory tests, participants provide cognition was somewhat attenuated (–0.038 per doubling in BC; 95% CI, –0.089 to 0.012). information on medical history, lifestyle, and conclusions: Ambient traffic-related air pollution was associated with decreased cognitive func- demographic factors at each study visit and, tion in older men. starting in 1993, have been asked to complete key words: aging, black carbon, cognitive dysfunction, epidemiology, particulate matter. Environ a battery of cognitive tests. At each visit, par- Health Perspect 119:682–687 (2011). doi:10.1289/ehp.1002767 [Online 20 December 2010] ticipants provided written informed consent as approved by the VA Boston Healthcare System Institutional Review Board (IRB).Declining or comparatively low cognitive to exhibit white matter hyperintensities in thefunction in older adults is associated with loss prefrontal cortex than were those residing in a Address correspondence to M.C. Power, Harvardof independence in activities of daily living city with lower levels (Calderón-Garcidueñas School of Public Health, EOME, Landmark Center,(Greiner et al. 1996; McGuire et al. 2006), et al. 2008a). Although this ecological evidence 3rd Floor East, P.O. Box 15697, Boston, MA 02215 USA. Telephone: (617) 721-9984. Fax: (617) 384-nursing home admission (Gaugler et al. 2007; is insufficient to conclude that air pollution is 8994. E-mail: mcpower@hsph.harvard.eduJoray et al. 2004), hospitalization (Chodosh the causal agent, other lines of evidence sup- Supplemental Material is available online ( al. 2004), and mortality (Bassuk et al. 2000; port this claim. In animal studies, exposure to 1289/ehp.1002767 via et al. 2003). The burden associated ozone has been linked to neuroinflammation, We thank S. Melly for his help creating Figure 1with deficits in cognitive function is expected lipid peroxidation, and memory deficits (Block and J. Weuve for her contribution to the develop-to grow as the global population ages (Ferri and Calderón-Garcidueñas 2009; Dorado- ment of this work. M.C.P. is supported by a grant from the Nationalet al. 2005), but few interventions to prevent, Martinez et al. 2001), effects that may be medi- Institute of Environmental Health Sciences (NIEHS;delay, or halt the progression of cognitive ated by the induction of systemic inflammation. T32 ES007069). This study was supported by a grantdecline are currently available. The potential Controlled animal experiments demonstrate from the U.S. Environmental Protection Agencyrelationship between exposure to common that ultrafine particulate matter (PM), a com- (RD832416 and RD83479801) and by grants from theenvironmental toxicants and cognition in ponent of traffic-related air pollution, can NIEHS (ES00002, ES015172, ES014663, ES005257,older adults has received relatively little con- translocate to the brain via the olfactory nerve ES012044, and P01 ES009825). The Cognition and Health in Aging Men Project is supported by thesideration compared with nonenvironmental or systemic circulation (Elder et al. 2006; Clinical Science Research and Development Service,risk factors, despite evidence that many toxi- Oberdörster et al. 2004; Peters et al. 2006), U.S. Department of Veterans Affairs, the Nationalcants may be neurotoxic and the modifiable where it may promote CNS dysfunction. In Institutes of Health (grants R01-AG14345, R01-nature of many environmental exposures. animal studies, exposure to fine or ultrafine AG18436, 5-P42-ES05947, and R01-ES05257), the Air pollution may exert adverse effects on PM has been associated with CNS inflamma- U.S. Department of Agriculture, Agricultural Researchcentral nervous system (CNS) function. In tion and lipid peroxidation (Campbell et al. Service (contract 53-K06-510), and Veterans Affairs Merit Review and Research Career Scientist awardsautopsy studies of dogs and people, Calderón- 2005; Kleinman et al. 2008; Zanchi et al. to A.S. The Veterans Affairs Normative Aging StudyGarcidueñas et al. (2002, 2003, 2004, 2008b) 2008), neuronal degeneration (Veronesi et al. is supported by the Cooperative Studies Program/found evidence of increased levels of inflam- 2005), and behavioral changes (Zanchi et al. Epidemiology Research and Information Center, U.S.matory mediators, β-amyloid deposition, and 2008). Alternately, the cardiovascular effects of Department of Veterans Affairs.markers of oxidative damage to DNA, as well air pollution (Brook 2007; Pope and Dockery The views expressed in this article are those of theas evidence of blood–brain barrier disruption, 2006) may lead to CNS dysfunction through authors and do not necessarily represent the views of the U.S. Department of Veterans the brains of those from a city with high lev- promotion of vascular brain pathology. The authors declare they have no actual or potentialels of air pollution compared with those from a One potential manifestation of CNS dys- competing financial with low levels. Children and dogs residing function in older adults is cognitive impair- Received 22 July 2010; accepted 20 Decemberin a highly polluted city were also more likely ment. We hypothesized that exposure to 2010.682 volume 119 | number 5 | May 2011 • Environmental Health Perspectives
  2. 2. Traffic-related air pollution and cognitionThis study has been approved by the IRBs of attention, memory, executive function, lan- random slope for BC effect by cognitive testthe participating institutions. guage, and visuomotor ability. All cognitive score type. If the slope were to be significant, We limited the study sample to partici- tests were drawn from established cognitive this would indicate that the BC effect is dis-pants of the NAS who completed cognitive test batteries, including the Consortium to similar across cognitive test score types andtesting between 1996 and 2007 and who lived Establish a Registry for Alzheimer Disease the global analyses would be inappropriate.within the range of our BC exposure model. (CERAD) battery (Morris et al. 1989), Final models were multivariable adjusted forThese constraints ensured that both cogni- the Wechsler Adult Intelligence Scales for potential confounders or predictors of cog-tive test data and an estimate of BC exposure Adults, Revised (WAIS-R) (Wechsler 1981), nitive function that are not plausible inter-before cognitive assessment were available for the Neurobehavioral Evaluation System 2 mediates for a BC–cognition relationship.each study participant. We also excluded par- (Letz 1991), and the Developmental Test We obtained data on these covariates fromticipants if they experienced a stroke before of Visual-Motor Integration (Beery and information collected at each NAS study start. In addition, we restricted the pres- Buktenica 1967). Final multivariable-adjusted models wereent study population to those self-reporting The present study includes cognitive adjusted for age at cognitive assessment as wellwhite race/ethnicity. data from study visits from 1996 to 2007. as information on several variables at baseline: Exposure assessment. Estimates of BC Therefore, the baseline visit for each study education (< 12, 12–16, > 16 years), alcoholexposure at the residence of each partici- participant is defined as the first cognitive intake (< 2 drinks/day, ≥ 2 drinks/day), physi-pant were used as a surrogate for individual assessment completed on or after 1 January cal activity [< 12, 12 to < 30, ≥ 30 metabolicexposure to traffic-related air pollution. We 1996. This restriction ensures that we have an equivalent hours (MET-hr) per week], diabe-estimated BC exposures using a validated estimate of BC exposure obtained before the tes (yes/no), dark fish consumption (less thanspatio temporal land-use regression model cognitive test data used in our analyses. Most once a week, equal to or greater than oncethat, starting in 1995, provides daily estimates study participants have completed multiple a week), computer experience (yes/no), firstof BC concentrations throughout the greater waves of cognitive testing, and all available language (English/not English), percentage ofBoston, Massachusetts, area, which roughly cognitive data obtained between 1996 and the participant’s census tract that is nonwhite,includes the area within Interstate 495; details 2007 were used in analyses. percentage of the adult residents in the par-of this model have been published previously Statistical analysis. Because we observed ticipant’s census tract with at least a college(Gryparis et al. 2007). Briefly, daily aver- a ceiling effect for MMSE scores, with degree, an indicator for whether the cognitiveage BC estimates from 83 monitoring sites many men achieving the maximum score, data were from the participant’s first cogni-were used to develop a BC prediction model. we created a dichotomous variable for infe- tive assessment (yes/no), and an indicator forPredictors in the final model include infor- rior MMSE performance for use in analyses. whether the participant was a part-time resi-mation on meteorological conditions (e.g., MMSE scores are commonly used to screen dent of the greater Boston area (yes/no). Wewind speed), land use (e.g., traffic density), for dementia (Tombaugh and McIntyre excluded data from study visits for which wedaily BC concentrations at a central monitor, 1992), but few men in our cohort exhibited lacked covariate information from all analysesand other descriptors (e.g., day of the week). scores that would typically trigger further (1.4% of study visits).Based on the training set, the model R2 was evaluation (< 24). We therefore chose to label We log-transformed BC estimates for use0.83, whereas the average correlation between scores ≤ 25 as “low” MMSE scores (18% of in all analyses after examination of restrictedpredicted values and observed BC levels in our observations). Cognitive test scores for cubic splines indicated a log-linear relation-four out-of-sample validation samples was the remaining six tests were z-transformed ship. We additionally explored interactions0.59. For the present study, addresses with based on the mean and standard deviation of between BC and body mass index (BMI),predicted daily BC concentrations outside the the individual test scores at baseline and were hypertension, smoking status, and diabetesrange of the exposure measurements from the considered as continuous variables in analyses, using multiplicative interaction terms andtraining set were excluded. To create a metric with positive z-scores indicating better perfor- conducted sensitivity analyses further adjust-of long-term exposure, we took the average mance for all tests. ing our multivariable models for smokingof the 365 daily estimate at the participant’s We evaluated the association between BC and BMI. We used R (version 2.10.1; Rresidential address prior to the date of the first and the odds of having a low MMSE score Development Core Team 2010) to examinecognitive assessment completed on or after 1 using logistic regression with generalized esti- splines and SAS (version 9.2; SAS InstituteJanuary 1996, the first date for which compu- mating equations and empirical variance esti- Inc., Cary, NC) for all other analyses.tation of a prior 1-year average was possible. mates to account for repeat measures across Sensitivity analyses: potential confounding Cognitive testing. Since the onset of cog- individuals. For the remaining six tests, which by lead. Exposure to lead may confound thenitive testing, the cognitive battery has been yield seven unique scores (the pattern com- relationship between BC and cognitive func-revised, reflecting both the addition and sub- parison task produces two unique scores), we tion. Traffic-related exposures were a primarytraction of individual cognitive tests. Seven treated each cognitive test score as a repeat source of lead exposure in the era of leadedcognitive tests have been administered with measure of global cognitive function in a ran- gasoline, and it is reasonable to expect thatrelative consistency across the span of cogni- dom effects linear mixed model. This model men with high current traffic-related air pol-tive testing and were considered in this study: included a random intercept for individual to lution exposures had high past traffic-relatedthe Mini-Mental State Examination (MMSE), account for repeat measures within individu- lead exposures. In addition, cumulative pastthe digit span backward test, a verbal fluency als and a random intercept for cognitive test lead exposure may have an adverse impact ontask, constructional praxis, immediate recall score to account for use of seven different cognitive function in older adults (Shih et al.of a 10-word list, delayed recall of a 10-word cognitive test score types as repeat measures of 2006), even at low levels (Weisskopf et al.list, and a pattern comparison task. The global cognitive function. This analysis pro- 2007; Wright et al. 2003). Thus, we conductedMMSE assesses multiple cognitive domains vides one hypothesis test for the association sensitivity analyses to determine the influenceand is widely used as a dementia screen- between BC and global cognitive function. of potential confounding by exposure to and research instrument (Tombaugh We evaluated the presence of heterogeneity Tibia bone lead concentrations, a bio-and McIntyre 1992); the six additional of BC effect across cognitive test score type marker for cumulative past lead expo-tests assess a variety of domains, including by adding and testing the significance of a sures (Hu et al. 1998), were available forEnvironmental Health Perspectives • volume 119 | number 5 | May 2011 683
  3. 3. Power et al.363 participants (53.4%). Tibia bone lead Results 1–5). On the natural scale, 1-year averageconcentrations were imputed for 299 partici- A total of 680 men met the inclusion criteria BC exposure estimates ranged from 0.03 topants (44.0%) using a published prediction for this study. Within this group, the number 1.77 µg/m3 (mean ± SD, 0.58 ± 0.28 µg/m3)model developed in the NAS (reduced model) of men completing at least one wave of cogni- and exhibited a skewed distribution. Figure 1(Park et al. 2009). Missing data precluded tive testing varied across the cognitive tests: illustrates the substantial spatial variability ofimputation of bone lead concentrations for MMSE (n = 671), digit span backward test BC over the study area in the year 1995.18 participants (2.6%), and we excluded these (n = 663), verbal fluency task (n = 670), con- Because the association between cogni-participants from the sensitivity analyses. To structional praxis (n = 626), immediate recall tive test scores and BC exposure estimatestest the sensitivity of our results to potential of a 10-word list (n = 669), delayed recall of a appeared log-linear, we log-transformed BCconfounding by exposure to lead, we repeated 10-word list (n = 667), and pattern compari- for use in all analyses and report associationsthe analyses described above additionally son task (n = 645). for a doubling in BC concentration on theadjusting for bone lead concentration and Table 1 summarizes the baseline charac- natural scale, or approximately a 0.69 unitimputation status (yes/no). teristics of the cohort. The mean ± SD age of change in ln(BC). BC exposure was signifi-Table 1. Baseline characteristics of the cohort the study sample at baseline was 71 ± 7 years cantly associated with the risk of having a(n = 680). (range, 51–97 years), and most men had at low MMSE score [multivariable-adjusted least some college or graduate-level educa- odds ratio (OR) = 1.3 for a doubling in BC BC concntrationCharacteristic n (%) (mean ± SD) tion (70.1%). On average, study participants concentration; 95% confidence interval completed 2.14 cognitive assessments (range, (CI), 1.1–1.6; Table 2]. In the analysis ofAge (years)50–59 48 (7.1) 0.550 ± 0.28860–69 284 (41.8) 0.574 ± 0.27770–79 273 (40.1) 0.583 ± 0.26780–89 72 (10.6) 0.605 ± 0.33390–99 3 (0.4) 0.466 ± 0.285Education (years)< 12 203 (29.9) 0.629 ± 0.30412–16 338 (49.7) 0.572 ± 0.276> 16 139 (20.4) 0.521 ± 0.239First languageEnglish 585 (86.0) 0.578 ± 0.282Not English/bilingual 95 (14.0) 0.581 ± 0.266Computer experienceYes 277 (40.7) 0.542 ± 0.255No 403 (59.3) 0.604 ± 0.294Physical activity (MET‑hr/week)< 12 385 (56.6) 0.574 ± 0.28612 to < 30 178 (26.2) 0.602 ± 0.293≥ 30 117 (17.2) 0.558 ± 0.237Alcohol (drinks/day)≥2 166 (24.4) 0.573 ± 0.242<2 514 (75.6) 0.581 ± 0.291DiabetesYes 106 (15.6) 0.567 ± 0.254No 574 (84.4) 0.581 ± 0.285Consumed dark fish (times/week)≥1 100 (14.7) 0.566 ± 0.283<1 580 (85.3) 0.581 ± 0.280Nonwhite (% of census tract) < 5% 268 (39.4) 0.521 ± 0.298 5 to < 10% 188 (27.6) 0.541 ± 0.249 ≥ 10% 224 (32.9) 0.679 ± 0.255≥ 25 years of age with at least a college degree (% ofcensus tract) < 30% 214 (31.5) 0.626 ± 0.301 30 to < 50% 281 (41.3) 0.540 ± 0.277 ≥ 50% 185 (27.2) 0.583 ± 0.250Smoking 0.02–0.07Never 192 (28.2) 0.557 ± 0.255 0.08–0.19 0.20–0.28Former 453 (66.6) 0.584 ± 0.294 0.29–0.35Current 35 (5.1) 0.634 ± 0.208 0.36–0.41BMI (kg/m2) 0.42–0.48< 25 152 (22.4) 0.587 ± 0.262 0.49–0.55≥ 25 528 (77.6) 0.576 ± 0.285 0.56–0.61Hypertension 0.62–0.82 0.83–1.55Yes 454 (66.8) 0.587 ± 0.288No 226 (33.2) 0.562 ± 0.264Tibia bone lead Limited access highwayMeasured 363 (53.4) 0.599 ± 0.286 Town boundaryImputed 299 (44.0) 0.559 ± 0.271Missing 18 (2.6) 0.507 ± 0.289 Figure 1. Predicted 1-year average BC exposures (μg/m3) for the greater Boston area in 1995.684 volume 119 | number 5 | May 2011 • Environmental Health Perspectives
  4. 4. Traffic-related air pollution and cognitionglobal cognitive function, which combined diameter ≤ 10 µm (PM10) and distance to a Several biological mechanisms have beenthe remaining seven cognitive test scores, a major road, and cognitive function in 399 proposed to explain how traffic-related airdoubling in BC was significantly associated women 68–79 years of age who had lived pollution may have an adverse effect onwith a decrement of 0.054 SD in cogni- at the same address for 20 years. Residence CNS function. First, ultrafine particles, suchtive test score in our multivariable-adjusted within 50 m of a major road, but not PM10, as those found in diesel exhaust, are smallmodel (95% CI, –0.103 to –0.006; Table was significantly associated with a decrement enough to pass through the air–blood barrier2), an effect similar to a 1.9-year differ- in the total score of the CERAD-Plus test of the lung, enter the systemic circulation,ence in age in our data. We found no evi- battery, the Stroop test, and a test of olfac- and translocate to other body tissues, includ-dence of heterogeneity of results across tests tory function in multivariable-adjusted mod- ing the brain (Oberdörster et al. 2004; Peters[p-value = 0.25]. See Supplemental Material, els. In addition, residence within 50 m of et al. 2006), and may also translocate directlyTables 1 and 2 (doi:10.1289/ehp.1002767)] a major road was adversely associated with to the brain via the olfactory nerve (Calderón-for the full multivariable-adjusted models. performance on all subtests of the CERAD Garcidueñas et al. 2003; Elder et al. 2006).In our sensitivity analyses, we additionally battery, although few individual associations The presence of such particles in the brain isadjusted for smoking and BMI, and our achieved statistical significance. Our findings associated with neuroinflammation and oxida-results were materially unchanged. are consistent with these, because BC, a traffic tive stress (Calderón-Garcidueñas et al. 2003; We found no evidence of interactions particle, is strongly associated with distance to Elder et al. 2006). Second, traffic-related PMbetween BC and diabetes, smoking, BMI or major road and weakly associated with PM10. may exert indirect effects on CNS functionhypertension on MMSE scores (p-values for A handful of studies have evaluated the through effects on cardiovascular health.all interactions > 0.1). Our exploration of association between air pollution and cogni- Exposure to PM is associated with a variety ofpotential interactions between BC and these tive function in younger populations, and, col- cardiovascular end points (Brook 2007; Popefactors on global cognitive function, in mod- lectively, they also suggest an adverse effect. In and Dockery 2006), including hypertensionels that combined the remaining seven test a study of adults 20–59 years of age, Chen and (Auchincloss et al. 2008; Brook 2007) andscores, suggested that the adverse effects of Schwartz (2009) reported that estimates of atherosclerosis (Künzli et al. 2005, 2010), asBC were concentrated in overweight and exposure to ozone, but not PM10, were asso- well as markers of increased cardiovascularobese individuals (p-value for interaction = ciated with inferior performance on two of risk, including progression of atherosclerotic0.10) and in ever smokers (p-value for inter- three neurobehavioral tests in fully adjusted plaques (Floyd et al. 2009; Künzli et al. 2010;action = 0.07), but we found no evidence for models. Calderón-Garcidueñas et al. (2008a) Sun et al. 2005) and homocysteine (Baccarellieffect modification by hypertension or diabe- found the performance age on several sub- et al. 2007; Park et al. 2008). Several of thesetes (p-values for interactions > 0.1). tests of the Wechsler Intellegence Scale for factors are also associated with worse cognitive Among participants with measured tibia Children, Revised (WISC-R) to be signifi- function (Komulainen et al. 2007; Qiu et al.bone lead concentrations, the mean ± SD bone cantly behind chronological age in children 2005; Sander et al. 2010; Vidal et al. 2008),lead concentration was 19.2 ± 11.9 µg lead/g residing in cities with high levels of air pollu- which likely reflects the link between cogni-bone mineral; among those with imputed tion but not in those residing in cities with low tive impairment and vascular brain pathology.bone lead concentrations, it was 16.1 ± 7.0 µg levels. In a study that compared neurobehav- The present study has several strengths. Thelead/g bone mineral. In the sensitivity analysis ioral test scores across students from two pri- BC exposure model enabled us to report indi-that adjusted for estimates of cumulative lead mary schools in the same city in China, Wang vidual estimates of long-term exposure to traf-exposure, we observed that the multivariable- et al. (2009) found that students who attended fic-related air pollution. In our study, we wereadjusted association between BC and low the school with high levels of traffic-related also able to control for socioeconomic statusMMSE scores was almost identical (OR = 1.3 air pollution were more likely to exhibit poor (SES) using both personal and residence-basedper doubling in BC; 95% CI, 1.1–1.7). The performance than were those who attended characteristics, which should limit residual con-association between BC and global cognitive the school with low levels. Franco Suglia et al. founding by this factor. Furthermore, we werefunction was somewhat attenuated after we (2008) found that higher residence-based able to address the potential for confoundingadjusted for lead in our multivariable-adjusted estimates of average lifetime exposure to BC by past exposure to lead, another environmen-model (–0.038 per doubling in BC; 95% CI, were associated with lower cognitive test scores tal pollutant that may affect cognitive function–0.089 to 0.012). in children 8–11 years of age. Similarly, in in older adults, using a combination of meas- 4- to 5-year-old children from the INMA ured and imputed biomarkers for cumulativeDiscussion (Environment and Childhood) study, Freire past lead exposure. Although correlated withThe results of this study suggest an adverse et al. (2010) observed that higher residence- traffic-related air pollution, exposure to leadeffect of traffic-related air pollution on global based estimates of nitric dioxide were associ- cannot completely account for the observedcognitive function in older men. Confounding ated with lower scores on all subscales of a associations. Finally, our results are based on aby lead cannot fully account for these findings. standardized version of the McCarthy Scales of substantial number of observations, given that,When we explored the potential for effect Children’s Abilities, although only one of these on average, each of our 680 participants com-modification, our results suggest that the effect associations achieved statistical significance. pleted 2.14 cognitive testing sessions.of traffic-related air pollution on cognitionmay be greater in smokers or overweight and Table 2. Adjusted associations for a doubling in BC concentration on the natural scale and cognitive testobese individuals. Because these conditions are score(s).proinflammatory, this is biologically plausible Effect per doubling in BC concentration (95% CI)but requires further confirmation. Model 2: age and Model 3: multivariable To our knowledge, this is only the sec- Analysis Model 1: age adjusted education adjusted adjustedaond study to consider the association between Low MMSE score OR 1.4 (1.1 to 1.6) 1.3 (1.0 to 1.5) 1.3 (1.1 to 1.6)air pollution and cognitive function in older Global analysis estimate –0.073 (–0.122 to –0.023) –0.052 (–0.100 to –0.004) –0.054 (–0.103 to –0.006)adults. Ranft et al. (2009) considered the aAdjusted for age, education, first language, computer experience, physical activity, alcohol consumption, diabetes,association between two markers of air pollu- dark fish consumption, percentage of residential census tract that is nonwhite, percentage of residential census tracttion, a measure of PM with an aerodynamic adults with a college degree, indicator for first cognitive assessment, and indicator for part-time resident.Environmental Health Perspectives • volume 119 | number 5 | May 2011 685
  5. 5. Power et al. Several potential sources of bias must be unlikely to account for the observed associa- RefeRencesconsidered when interpreting our results. The tions. In our study, and generally in studiesuse of exposure estimates based on residential of older adults, participants who are lost to Allen RW, Davies H, Cohen MA, Mallach G, Kaufman JD, Adar SD. 2009. The spatial relationship between traffic-address may misclassify personal exposure lev- follow-up exhibit lower cognitive test scores generated air pollution and noise in 2 U.S. cities. Environels. However, the lack of significant occupa- than do those who complete follow-up (Euser Res 109(3):334–342.tional or commuting exposure in this largely et al. 2008). Because air pollution has been Auchincloss AH, Diez Roux AV, Dvonch JT, Brown PL, Barr RG, Daviglus ML, et al. 2008. Associations between recentretired cohort should mean that residence- related to cardiovascular morbidity and mor- exposure to ambient fine particulate matter and bloodbased exposure estimates are an excellent tality (Brook 2007; Pope and Dockery 2006) pressure in the multi-ethnic study of atherosclerosismeasure of personal exposures over the past and because we observed lower BC estimates (MESA). Environ Health Perspect 116:486–491. Baccarelli A, Zanobetti A, Martinelli I, Grillo P, Hou L, Lanzani G,few years. Because 87% of our participants in participants with three or more follow-up et al. 2007. Air pollution, smoking, and plasma homo-have lived at the same address for at least visits, we expect that a greater loss of those cysteine. Environ Health Perspect 115:176–181.5 years before their baseline cognitive assess- with both lower cognitive test scores and Bassuk SS, Wypij D, Berkman LF. 2000. Cognitive impairment and mortality in the community-dwelling elderly. Am Jment (mean ± SD duration, 18 ± 12 years) higher BC estimates to lead to an underesti- Epidemiol 151(7):676–688.and because people are likely to choose simi- mate of the association. Thus, the true asso- Beery KE, Buktenica NA. 1967. Developmental Test of Visuallar neighborhoods when relocating, misclas- ciation may be stronger than that observed. Motor Integration. Chicago:Follett. Bell B, Rose CL, Damon A. 1972. The Normative Aging Study:sification due to participant relocation is also One additional limitation lies in our an interdisciplinary and longitudinal study of health andexpected to be minimal. In addition, occa- inability to attribute our findings to a particu- aging. Aging Hum Dev 3(1):5–17.sional exposures encountered elsewhere are lar traffic-related exposure. We have estimated Block ML, Calderón-Garcidueñas L. 2009. Air pollution: mecha-unlikely to be correlated with cognitive func- exposure to BC particles for each participant. nisms of neuroinflammation and CNS disease. Trends Neurosci 32(9):506–516.tion and so would result in nondifferential However, traffic-related air pollution is a com- Brook RD. 2007. Is air pollution a cause of cardiovascularmisclassification of exposure, which would plex mixture of gases and particles, and BC is disease? Updated review and controversies. Rev Environbe expected to attenuate our effect estimates correlated with other components of the mix- Health 22(2):115–137. Calderón-Garcidueñas L, Azzarelli B, Acuna H, Garcia R,(Gryparis et al. 2009). Similarly, nondiffer- ture. Although existing evidence suggests that Gambling TM, Osnaya N, et al. 2002. Air pollution and brainential misclassification of cognitive function PM, specifically ultrafine PM, is a likely agent damage. Toxicol Pathol 30(3):373– expected when assessing cognitive func- for adverse effects to the CNS (Calderón- Calderón-Garcidueñas L, Maronpot RR, Torres-Jardon R, Henriquez-Roldán C, Schoonhoven R, Acuna-Ayala H, et al.tion and would make it harder to detect a Garcidueñas et al. 2003; Elder et al. 2006; 2003. DNA damage in nasal and brain tissues of caninestrue effect, although our use of global analy- Oberdörster et al. 2004; Peters et al. 2006), exposed to air pollutants is associated with evidenceses helps to reduce the impact of noise by other components may be important. In of chronic brain inflammation and neuro degeneration. Toxicol Pathol 31(5):524–538.using multiple measures to capture under- addition, measures of traffic-related air pol- Calderón-Garcidueñas L, Mora-Tiscareño A, Ontiveros E,lying global cognitive function. We observed lution are often correlated with other envi- Gómez-Garza G, Barragán-Mejia G, Broadway J, et al.statistically significant associations between ronmental exposures. Although we were able 2008a. Air pollution, cognitive deficits and brain abnor- malities: a pilot study with children and dogs. Brain CognBC and cognitive function, when measured to address the potential for lead to account 68(2):117– the risk of a “low” MMSE score and as for the observed associations, other exposures Calderón-Garcidueñas L, Reed W, Maronpot RR, Henríquez-overall performance on our battery of cogni- must be considered, including noise. Elevated Roldán C, Delgado-Chavez R, Calderón-Garcidueñas A,tive tests, despite the presence of misclassi- noise levels have been linked to metrics of et al. 2004. Brain inflammation and Alzheimer’s-like pathol- ogy in individuals exposed to severe air pollution. Toxicolfication, suggesting that the true association cognitive function in children in a handful Pathol 32(6):650–658.may be even stronger than what we observed. of studies (Clark et al. 2006; Stansfeld et al. Calderón-Garcidueñas L, Solt AC, Henriquez-Roldán C, Torres-Although residual confounding cannot be 2005), and chronic occupational exposure Jardón R, Nuse B, Herritt L, et al. 2008b. Long-term air pollution exposure is associated with neuroinflamma-excluded, we do not believe confounding to noise is associated with neurophysiologi- tion, an altered innate immune response, disruption of theto be a major source of bias in the present cal measures related to attention in adults blood-brain barrier, ultrafine particulate deposition, andstudy given the availability of information (Gomes et al. 1999). Given the high correla- accumulation of amyloid beta-42 and alpha-synuclein in children and young adults. Toxicol Pathol 36(2):289–310.on potential confounders and the use of sev- tion between measures of traffic-related air Campbell A, Oldham M, Becaria A, Bondy SC, Meacher D,eral different, related variables to account for pollution and noise (Allen et al. 2009; Davies Sioutas C, et al. 2005. Particulate matter in polluted airSES. Moreover, we do not believe confound- et al. 2009), it is possible that the associa- may increase biomarkers of inflammation in mouse brain. Neurotoxicology 26(1):133– by regional pollutants, which may have tions observed for one exposure are attribut- Chen J, Schwartz J. 2009. Neurobehavioral effects of ambi-an independent effect on cognition, to be a able to the other. Further research to evaluate ent air pollution on cognitive performance in U.S. adults.major source of bias. Correlation between the relative contribution of particular traffic- Neurotoxicology 30(2):231–239. Chodosh J, Seeman TE, Keeler E, Sewall A, Hirsch SH,BC and regional pollutants may be induced related pollutants or other exposures to the Guralnik JM, et al. 2004. Cognitive decline in high-through similar spatial or temporal patterns. observed association is warranted. functioning older persons is associated with an increasedHowever, traffic-related air pollution exhib- risk of hospitalization. J Am Geriatr Soc 52(9):1456–1462.its significant spatial variation not shared by Conclusions Clark C, Martin R, van Kempen E, Alfred T, Head J, Davies HW, et al. 2006. Exposure-effect relations between aircraft andmore regional pollutants [e.g., ozone or par- Our findings suggest that traffic-related air road traffic noise exposure at school and reading compre-ticulate matter with an aerodynamic diameter pollution may have an adverse effect on cog- hension: the Ranch Project. Am J Epidemiol 163(1):27–37.< 2.5 µm (PM2.5)], and temporal correlation nition in older men. This is the first study to Davies HW, Vlaanderen JJ, Henderson SB, Brauer M. 2009. Correlation between co-exposures to noise and air pollu-is quite weak between our 1-year residence- find an association between traffic-related air tion from traffic sources. Occup Environ Med 66(5):347–350.based BC estimates (which have varying start pollution and cognition in older men, and Dorado-Martinez C, Paredes-Carbajal C, Mascher D, Borgonio-dates depending on the timing of the baseline only the second to consider the relationship Pérez G, Rivas-Arancibia S. 2001. Effects of different ozone doses on memory, motor activity and lipid peroxidationcognitive assessment) and the 1-year average in older adults. Given the ubiquitous nature levels, in rats. Int J Neurosci 108(3–4):149–161.of ozone concentrations at four central moni- of the exposure, if traffic-related air pollution Elder A, Gelein R, Silva V, Feikert T, Opanashuk L, Carter J,tors in the Boston area (r = –0.03) or 1-year is causally related to cognitive impairment in et al. 2006. Translocation of inhaled ultrafine manganese oxide particles to the central nervous system. Environaverage PM2.5 concentrations at a central site older adults, implementation of interventions Health Perspect 114:1172–1178.monitor in Boston (r = 0.16) calculated over to reduce exposure, including establishment Euser SM, Schram MT, Hofman A, Westendorp RG,the same time periods. Incomplete follow-up of more stringent emissions standards, would Breteler MM. 2008. Measuring cognitive function with age: the influence of selection by health and survival.may induce bias in the present study but is be expected to have substantial benefits. Epidemiology 19(3):440–447.686 volume 119 | number 5 | May 2011 • Environmental Health Perspectives
  6. 6. Traffic-related air pollution and cognitionFerri CP, Prince M, Brayne C, Brodaty H, Fratiglioni L, Beckermann B, Gilliland F, et al. 2010. Ambient air pollu- matter impairs cognitive function in the elderly. Environ Ganguli M, et al. 2005. Global prevalence of dementia: a tion and the progression of atherosclerosis in adults. PLoS Res 109(8):1004–1011. Delphi consensus study. Lancet 366(9503):2112–2117. One 5(2):e9096; doi:10.1371/journal.pone.0009096 [Online Sander K, Bickel H, Förstl H, Etgen T, Briesenick C, Poppert H,Floyd HS, Chen LC, Vallanat B, Dreher K. 2009. Fine ambient 8 February 2010]. et al. 2010. Carotid-intima media thickness is independently air particulate matter exposure induces molecular altera- Künzli N, Jerrett M, Mack WJ, Beckerman B, LaBree L, associated with cognitive decline. The Invade Study. Int J tions associated with vascular disease progression within Gilliland F, et al. 2005. Ambient air pollution and atheroscle- Geriatr Psychiatry 25(4):389–394. plaques of atherosclerotic susceptible mice. Inhal Toxicol rosis in Los Angeles. Environ Health Perspect 113:201–206. Shih RA, Glass TA, Bandeen-Roche K, Carlson MC, Bolla KI, 21(5):394–403. Letz R. 1991. NES2 User’s Manual (version 4.4). Winchester, Todd AC, et al. 2006. Environmental lead exposure andFranco Suglia S, Gryparis A, Schwartz J, Wright RJ. 2008. MA:Neurobehavioral Systems, Inc. cognitive function in community-dwelling older adults. Association between traffic-related black carbon expo- McGuire LC, Ford ES, Ajani UA. 2006. Cognitive functioning as a Neurology 67(9):1556–1562. sure and lung function among urban women. Environ predictor of functional disability in later life. Am J Geriatr Stansfeld SA, Berglund B, Clark C, Lopez-Barrio I, Fischer P, Health Perspect 116:1333–1337. Psychiatry 14(1):36–42. Ohrström E, et al. 2005. Aircraft and road traffic noise andFreire C, Ramos R, Puertas R, Lopez-Espinosa MJ, Julvez J, Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, children’s cognition and health: a cross-national study. Aguilera I, et al. 2010. Association of traffic-related air pol- Fillenbaum G, et al. 1989. The Consortium to Establish a Lancet 365(9475):1942–1949. lution with cognitive development in children. J Epidemiol Registry for Alzheimer’s Disease (CERAD). Part I. Clinical Sun Q, Wang A, Jin X, Natanzon A, Duquaine D, Brook RD, Community Health 64(3):223–228. and neuropsychological assessment of Alzheimer’s disease. et al. 2005. Long-term air pollution exposure and accel-Gaugler JE, Duval S, Anderson KA, Kane RL. 2007. Predicting Neurology 39(9):1159–1165. eration of atherosclerosis and vascular inflammation in an nursing home admission in the U.S.: a meta-analysis. BMC Nguyen HT, Black SA, Ray LA, Espino DV, Markides KS. 2003. animal model. JAMA 294(23):3003–3010. Geriatr 7:13; doi:10.1186/1471-2318-7-13 [Online 19 June Cognitive impairment and mortality in older Mexican Tombaugh TN, McIntyre NJ. 1992. The Mini-Mental State 2007]. Americans. J Am Geriatr Soc 51(2):178–183. Examination: a comprehensive review. J Am Geriatr SocGomes LM, Martinho Pimenta AJ, Castelo Branco NA. 1999. Oberdörster G, Sharp Z, Atudorei V, Elder A, Gelein R, 40(9):922–935. Effects of occupational exposure to low frequency noise on Kreyling W, et al. 2004. Translocation of inhaled ultrafine Veronesi B, Makwana O, Pooler M, Chen LC. 2005. Effects of cognition. Aviat Space Environ Med 70(3 pt 2):A115–A118. particles to the brain. Inhal Toxicol 16(6–7):437–445. subchronic exposures to concentrated ambient particles.Greiner PA, Snowdon DA, Schmitt FA. 1996. The loss of indepen- Park SK, Mukherjee B, Xia X, Sparrow D, Weisskopf MG, Nie H, VII. Degeneration of dopaminergic neurons in Apo E -/- dence in activities of daily living: the role of low normal cogni- et al. 2009. Bone lead level prediction models and their mice. Inhal Toxicol 17(4–5):235–241. tive function in elderly nuns. Am J Public Health 86(1):62–66. application to examine the relationship of lead exposure Vidal JS, Dufouil C, Ducros V, Tzourio C. 2008. Homocysteine,Gryparis A, Coull BA, Schwartz J, Suh HH. 2007. Semiparametric and hypertension in the third National Health and Nutrition folate and cognition in a large community-based sample latent variable regression models for spatiotemporal mod- Examination Survey. J Occup Environ Med 51(12):1422–1436. of elderly people—the 3C Dijon study. Neuroepidemiology elling of mobile source particles in the greater Boston Park SK, O’Neill MS, Vokonas PS, Sparrow D, Spiro A III, 30(4):207–214. area. J R Stat Soc Ser C Appl Stat 56(2):183–209. Tucker KL, et al. 2008. Traffic-related particles are associ- Wang S, Zhang J, Zeng X, Zeng Y, Chen S. 2009. Association ofGryparis A, Paciorek CJ, Zeka A, Schwartz J, Coull BA. 2009. ated with elevated homocysteine: the VA Normative Aging traffic-related air pollution with children’s neurobehavioral Measurement error caused by spatial misalignment in Study. Am J Respir Crit Care Med 178(3):283–289. functions in Quanzhou, China. Environ Health Perspect environmental epidemiology. Biostatistics 10(2):258–274. Peters A, Veronesi B, Calderón-Garcidueñas L, Gehr P, 117:1612–1618.Hu H, Rabinowitz M, Smith D. 1998. Bone lead as a biological Chen LC, Geiser M, et al. 2006. Translocation and potential Wechsler D. 1981. Wechsler Adult Intelligence Scale-Revised marker in epidemiologic studies of chronic toxicity: con- neurological effects of fine and ultrafine particles a critical (WAIS-R). San Antonio, TX:Psychological Corp. ceptual paradigms. Environ Health Perspect 106:1–8. update. Part Fibre Toxicol 3:13; doi:10.1186/1743-8977-3-13 Weisskopf MG, Proctor SP, Wright RO, Schwartz J,Joray S, Wietlisbach V, Bula CJ. 2004. Cognitive impairment in [Online 8 September 2006]. Spiro A III, Sparrow D, et al. 2007. Cumulative lead expo- elderly medical inpatients: detection and associated six- Pope CA III, Dockery DW. 2006. Health effects of fine particu- sure and cognitive performance among elderly men. month outcomes. Am J Geriatr Psychiatry 12(6):639–647. late air pollution: lines that connect. J Air Waste Manag Epidemiology 18(1):59–66.Kleinman MT, Araujo JA, Nel A, Sioutas C, Campbell A, Assoc 56(6):709–742. Wright RO, Tsaih SW, Schwartz J, Spiro A III, McDonald K, Cong PQ, et al. 2008. Inhaled ultrafine particulate matter Qiu C, Winblad B, Fratiglioni L. 2005. The age-dependent rela- Weiss ST, et al. 2003. Lead exposure biomarkers and Mini- affects CNS inflammatory processes and may act via MAP tion of blood pressure to cognitive function and dementia. Mental Status Exam scores in older men. Epidemiology kinase signaling pathways. Toxicol Lett 178(2):127–130. Lancet Neurol 4(8):487–499. 14(6):713–718.Komulainen P, Kivipelto M, Lakka TA, Hassinen M, Helkala EL, R Development Core Team 2010. R: A Language and Zanchi AC, Venturini CD, Saiki M, Nascimento Saldiva PH, Patja K, et al. 2007. Carotid intima-media thickness and Environment for Statistical Computing. Vienna, Austria: R Tannhauser Barros HM, Rhoden CR. 2008. Chronic nasal cognitive function in elderly women: a population-based Foundation for Statistical Computing. instillation of residual-oil fly ash (ROFA) induces brain lipid study. Neuroepidemiology 28(4):207–213. Ranft U, Schikowski T, Sugiri D, Krutmann J, Kramer U. peroxidation and behavioral changes in rats. Inhal ToxicolKünzli N, Jerrett M, Garcia-Esteban R, Basagana X, 2009. Long-term exposure to traffic-related particulate 20(9):795–800.Environmental Health Perspectives • volume 119 | number 5 | May 2011 687