- Email: [email protected]
Epidemiology of air pollution and diabetes
Review
Epidemiology of air pollution and diabetes Elisabeth Thiering1,2 and Joachim Heinrich1,3 1
Institute of Epidemiology I, Helmholtz Zentrum Mu¨nchen – German Research Center for Environmental Health, Neuherberg, Germany 2 Division of Metabolic and Nutritional Medicine, Dr von Hauner Children’s Hospital, University of Munich Medical Center, Munich, Germany 3 Institute and Outpatient Clinic for Occupational, Social and Environmental Medicine, Inner City Clinic, University Hospital Munich, Ludwig Maximilian University of Munich, Munich, Germany
Air pollution affects a large proportion of the global population. Air pollutants are hypothesized to exert their effects via impaired endothelial function, elevated systemic inflammation, mitochondrial dysfunction, and oxidative stress, all of which are hallmarks of type 2 diabetes (T2D). Here we review epidemiological studies aimed at answering whether diabetes patients are more vulnerable to ambient (outdoor) air pollution exposure and whether air pollution is associated with diabetes development or other predisposing conditions for T2D. Current evidence suggests an association between air pollution exposure and T2D, but more critical analysis is warranted. Understanding the associations between air pollution exposure and the development of T2D is critical in our efforts to control sources of air pollution and their impact on the disease. The diabetes epidemic and ubiquitous air pollution exposure Air pollution is a ubiquitous exposure that affects large proportions of the global population [1]. Air pollutants are emitted from many sources, such as industrial facilities, cars, trucks, ships, and airplanes, but also from household combustion devices or during forest fires and volcanic eruptions. Ambient or outdoor air pollution (referred to as air pollution hereafter) comprises particulate matter (PM) (see Glossary) of various sizes [<10 mm in diameter, PM10; <2.5 mm, PM2.5; <100 nm, ultrafine particles (UFPs)], chemicals such as persistent organic pollutants (POPs), and gaseous compounds such as nitric oxides (NOx), carbon monoxide (CO), ozone (O3), and sulfur oxides (SOx). These compounds differ in their dispersion, reactivity, and toxicity. Air pollution exposure has been linked to a reduced life expectancy, mainly attributable to cardiovascular and respiratory diseases such as heart disease and lung cancer [2–4], even among individuals exposed to annual average concentrations below the current air quality standards in Europe [5,6]. Corresponding author: Thiering, E. ([email protected]). Keywords: air pollution; type 2 diabetes; gestational diabetes; mortality. 1043-2760/ ß 2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tem.2015.05.002
384
Trends in Endocrinology and Metabolism, July 2015, Vol. 26, No. 7
In 2010, more than 3.2 million deaths worldwide were attributed to outdoor PM exposure and 3.5 million deaths to indoor household air pollution from solid fuels [7]. Air pollution exposure also accounts for a large number of disease-adjusted life-years lost (29.4 million for PM pollution and 47.9 million for household air pollution from solid fuels, which are equivalent to 3% and 4.3% of the worldwide burden of disease, respectively). Newer data from 2012 revealed 3.7 million deaths attributable to ambient air pollution and 4.3 million to household air pollution [Global Health Observatory Data Repository (http://apps.
Glossary Black carbon (BC): a fraction of PM2.5 that is mainly emitted during open biomass burning and from diesel cars and trucks without soot particle filters. Confidence interval (CI): an interval estimation of a parameter to show the reliability of the estimates. Typically, 95% CIs are reported, which represent an interval that contains the true value with a 95% probability. Hazard ratio (HR): a measure of mortality risk between different groups corresponding to a certain period of time. It is defined as the ratio between two hazard rates and is typically estimated via proportional hazard models. Interquartile range (IQR): the distance between the 75% percentile (third quartile) and 25% percentile (first quartile) of a distribution. Being a measure of dispersion it is often used to calculate effect sizes per increase in IQR to make them comparable between studies. Mortality rate ratio (MRR): defines the ratio between two mortality rates corresponding to two subgroups or to different causes of death. Nitric oxide (NO): a free radical and a byproduct of combustion in automobile engines and fossil-fuel power plants. In humans it is an important cellular signaling molecule and it is involved in many physiological processes. In the air it is rapidly oxidized to nitrogen dioxide (NO2). Ozone (O3): also known as trioxygen; an inorganic molecule. At ground level O3 is primarily formed by the action of sunlight on air containing hydrocarbons and NOx from combustion processes. It is not emitted directly from car engines or industrial operations. Particulate matter (PM): a complex mixture of small particles and liquid droplets. Particle pollution comprises numerous components, including acids (such as nitrates and sulfates), organic chemicals, metals, and soil or dust particles. These compounds can be directly emitted or formed through reactions with other compounds. The size of a particle is directly linked to its potential to penetrate the lung and cause health problems. Therefore, particles are divided into those of diameter <10 mm (PM10), which can pass into the throat and nose and enter the lung, and those of diameter <2.5 mm (PM2.5), which can additionally reach the alveolar region of the lung. Sulfur (S): an abundant, multivalent, non-metal element. It is oxidized to sulfur dioxide (SO2) during combustion processes. Major sources of SO2 emissions are volcanic eruptions and fossil fuel combustion at power plants and industrial facilities. In the atmosphere SO2 is converted to sulfate (SO4). Ultrafine particles (UFPs): defined as particles with aerodynamic diameter <0.1 mm and best characterized using particle number concentrations instead of particle mass.
Review who.int/gho/data/node.main.122?lang=en)]. While household indoor air pollution from solid fuels plays a major role in Asia, Latin America, and Africa [8], outdoor air pollution from PM is of global importance. Although there is strong evidence that air pollution adversely influences a broad spectrum of health indicators [2–6,9,10], links between air pollution exposure and T2D development have been only recently revealed. T2D affects approximately 9% of the global population [11] and accounts for 2.7% of global deaths (WHO, 2012). The number of affected people is expected to increase to 592 million by 2035 [12], with the steepest increases predicted in low- and middle-income countries. Changes in lifestyle factors due to urbanization, such as reduced physical activity, overnutrition, and obesity [13], were previously advanced as explanations for this trend. Recently, there is also interest in, and some evidence for, an association between air pollution exposure and T2D. Several biological mechanisms have been proposed to explain such a link. For example, PM exposure has been associated with impaired endothelial function, elevated systemic inflammation and oxidative stress, endoplasmic reticulum stress, cardiac autonomic nervous system dysfunction, and mitochondrial dysfunction [10,14–16]. Additionally, epigenetic changes leading to activation of key signaling pathways, or changes in markers of coagulation, inflammation, and endothelial function, have been described after exposure to air pollutants [17,18]. An ecological study by Alan Lockwood [19] was the first to link diabetes prevalence in adults to total air releases from all industries. More recently, several epidemiological studies varying in design, exposure, and outcome that are discussed below have explored the connection between air quality and diabetes. In this review we focus on epidemiological studies that examined associations between outdoor air pollution exposure and T2D. We also consider mortality due to diabetes, diabetes-related traits, and gestational diabetes and comprehensively summarize the results of five previous metaanalyses [20–24]. We address three fundamental questions: are diabetes patients more vulnerable to air pollution exposure, is exposure to air pollution associated with diabetes development, and/or is it associated with other conditions that predispose to diabetes? While most T2D determinants, such as inactivity or overweight, are individual factors, environmental hazards have a mostly involuntary nature. Therefore, policies that modify environmental risk factors can protect the general population. However, to fully evaluate potential public health impacts, better understanding of who is at risk [e.g., only predisposed subjects (for example, those having T2D) or also healthy individuals] is required. Although the currently reported risk increases are relatively small, given the large proportion of subjects exposed to air pollution and the high prevalence of T2D establishing and modifying a link between these two factors could have a large impact on healthy living. Although the body of evidence for adverse health effects of air pollution, especially of PM, is increasing, results specific to diabetes-related outcomes could be used in the ongoing debate on air quality guidelines setting. Thus, evaluating
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
whether air pollution is related to T2D is highly important and very timely. Are diabetes patients more vulnerable to ambient air pollution exposure? Mainly case-crossover and time-series study designs have been used to compare the effects of short-term exposure to air pollutants on all-cause mortality rates among the general population and diabetic patients. A total of seven studies have been conducted in American, Italian, Canadian, and European populations [25–32] involving between 12 978 [25] and 2935 647 [30] subjects with diabetes (for an overview see Figure 1A, for details see Table 1). Diabetes was not the original primary focus of these investigations with the exception of the study by Zanobetti et al. [30], in which 2935 647 subjects with a history of diabetes-related emergency hospital admission were followed. In other studies, death rates after days of high air pollution exposure among diabetic subjects were compared with those in the general population as a subgroup analysis. Among 65 180 elderly people who had a history of hospitalization for lung or heart disease, increased PM10 levels were positively associated with mortality risk for participants with a prior diagnosis of diabetes and this risk was greater than that observed among those without a prior diagnosis of diabetes [25]. Similar PM10 effects on mortality were seen in a study including nine Italian cities [26]. A national case-crossover analysis conducted in 121 cities in the USA observed associations between short-term PM2.5 exposure and increased risk for diabetes-related hospitalization and all-cause mortality in subjects with a previous diabetesrelated hospital admission [27]. Again, these PM2.5-associated risks were greater than those observed among individuals without diabetes. Increased nonaccidental mortality among subjects previously diagnosed with diabetes was observed in association with higher PM2.5 and SO4 concentrations in a mortality time-series study in Montreal, Canada [28]. However, no increased risk was observed among subjects who had diabetes but not cancer, cardiovascular disease, or an airways disease. An extension of this study published in 2013, which included deaths in Montreal during a later time period, revealed similar but slightly decreased effect estimates [29]. Associations between mortality among diabetes patients and O3 levels were also investigated. Higher average summer concentrations O3 were positively associated with mortality risk among 2935 647 patients with a history of diabetes-related emergency hospital admission living in 105 cities in the USA [30]. A nonsignificantly increased mortality risk following higher short-term O3 exposure was also seen among 14 350 subjects with diabetes in Italy [31]. Finally, a study on short-term NO2 exposure effects reported an increase in all-cause mortality risk among subjects with diabetes that was greater than that reported among those without diabetes [32] (Figure 1A). In summary, all seven studies investigating the effects of air pollution exposure on all-cause mortality observed a tendency for higher mortality rates among those with diabetes compared with those without. However, statistical significance (P < 0.05) was reached in only two of the 385
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Are diabetes paents more vulnerable to ambient air polluon exposure? Ambient air polluon exposure
General populaon
Diabetes paents
?
?
(B) Diabetes mortality
(A) All-cause mortality Bateson, 2004 Forasere, 2008 Zanobe, 2014 Goldberg, 2013 Zanobe, 2011 Staffogia, 2010 Chiusolo, 2011
PM10 PM10 PM2.5 PM2.5, NO2 O3 O3 NO2
(+) + (+) (+) + (+) (+)
Ostro, 2006 Samoli, 2014 Goldberg, 2006 Kan, 2004 Maynard, 2007 Ren, 2010
PM2.5 PM2.5, PM10 PM2.5, SO4, SO2, NO2, CO PM10, SO2, NO2 BC, SO4 O3
Long-term exposure Brook, 2013 Raaschou-Nielsen, 2013
PM2.5 NO2
+ (+) + + (+) + + +
TRENDS in Endocrinology & Metabolism
Figure 1. Overview of studies examining the vulnerability of diabetes patients to ambient air pollutants. (A) Studies that compare all-cause mortality rates in subjects with diabetes to those in the general population (B). Studies on diabetes-related mortality risk in the general population. The direction of the effect is denoted by for negative association, + for positive association, or o for no association; parentheses represent nonsignificant association or association restricted to subgroups. The summary shows that air pollution exposure leads to a higher mortality risk in diabetic subjects and to increased mortality due to diabetes in the general population in all of the studies despite nonsignificance in most of them.
studies. Compared with subjects without diabetes, the observed up-to-twofold increased mortality risks suggest that patients with diabetes are more susceptible to the immediate effects of short-term air pollution. Diabetes-related mortality in the general population Eight studies have investigated how air pollution exposure affects diabetes-related mortality in the general population (Figure 1B). Although the study design is different from the abovementioned studies, the underlying research question is similar, as it is always difficult to disentangle underlying and contributing causes of death in complex diseases. Five studies suggested a positive change in daily diabetes-related mortality after days of high PM or black carbon (BC) exposure [28,33–36]. These studies were conducted in the Montreal area [28], Boston [37], nine Californian counties [33], ten Mediterranean cities [34], and Shanghai [35] and involved between 431 and 15 184 diabetes-related deaths. However, statistical significance was reached in only three of the studies. These studies demonstrate that short-term exposure to air pollutants can potentially trigger diabetes-related mortality in the general population and among subjects with other medical conditions. Two large cohorts in Canada and Europe have studied long-term air pollution exposure effects in relation to diabetes mortality. Using a cohort of 2.1 million adults from the 1991 Canadian census mortality follow-up study, higher long-term PM2.5 exposure was associated with increased diabetes-related mortality [37]. In the Danish 386
Diet, Cancer, and Health cohort (DCH), long-term NO2 exposure was positively associated with diabetes-related mortality in 52 061 adults [38]. Although many of the estimates in the above-described studies were not statistically significant, the cumulative evidence appears to suggest that short-term ambient air pollution exposure increases the daily all-cause mortality risk in diabetic subjects and that both short-term and longterm exposure might be associated with diabetes-related mortality risk in the general population (Figure 1). Nevertheless, more large studies investigating long-term air pollution exposure in relation to mortality in diabetic patients as well as diabetes-related mortality in the general population are needed to establish causality and draw final conclusions. Is ambient air pollution exposure associated with diabetes development? Prospective cohort studies As discussed above, diabetic patients might be more susceptible to air pollution effects. Prospective studies can elucidate whether air pollution exposure only leads to accelerated progression of the disease or whether it directly contributes to its development. Prospective studies follow participants without diabetes for several years and evaluate potential associations between air pollution exposure at baseline or average air pollution exposure during the whole follow-up period with newly diagnosed T2D cases (Table 2). Figure 2 summarizes the results of the
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 1. Characteristics and results of case-crossover, time-series, and cohort studies on diabetes mortality Study design
Outcome
Population
Case crossover
Mortality
Case crossover
Mortality
Case crossover
Mortality
Mortality data of US citizens 65 years admitted to hospital with a diagnosis of heart disease or lung disease 1988– 1991 (n = 65 180, with prior T2D diagnosis n = 12 978) Mortality data of residents of nine Italian cities 35 years with natural cause of death obtained from regional registers of causes of death 1997–2004 (n = 321 014, 30 173 with hospital admission due to T2D within 2 years before death) Mortality data of residents of Montreal 65 years 1993–2003 (n = 158 350, n = 38 883 with T2D)
Case crossover
Mortality
Case crossover
Mortality
Case crossover
Diabetes-related mortality
Case crossover
Diabetes-related mortality
Time-series
Diabetes-related mortality
Time series
Diabetes-related mortality
Time series
Diabetes-related mortality
Time series
Diabetes-related mortality
Cohort (follow-up 5.6 years)
Mortality
Cohort (follow-up 10 years)
Diabetes-related mortality
Cohort (follow-up 13 years)
Diabetes-related mortality
Mortality data of residents of ten Italian cities 35 years with natural cause of death 2001–2005 between April and September (n = 126 647, with prior T2D hospital admission 14 350) Mortality data of residents of ten Italian cities 35 years with natural cause of death 2001–2005 (n = 276 205, with prior T2D 30 620) Mortality data from the Boston metropolitan area 1995–2002 (n = 107 925, n = 2694 due to T2D) Mortality data of residents from three counties in eastern Massachusetts 35 years 1995–2002 (n = 157 197, n = 3845 due to T2D) Deaths due to diabetes for all residents in nine Californian counties 1999–2002 (n = 15 184) Deaths due to diabetes in nine metropolitan areas in the European Mediterranean region (n = 9832) Mortality data of residents of Montreal 65 years 1984–1993 (n = 2922 deaths due to T2D)
Mortality data of residents of Zhabei district of Shanghai 2001–2002 (n = 431 due to T2D) Mortality data of US residents 65 years after emergency admission to hospital with a primary diagnosis of diabetes 1985–2005 (n = 2935 647)
1991 Canadian census mortality follow-up, subjects >25 years (n = 2145 400, n = 5200 deaths due to diabetes) Danish Diet, Cancer, and Health cohort, mortality in the nationwide Register of Causes of Death 1993– 2009 (n = 52 061)
Exposure duration and assessment Short term; daily mean from monitoring stations
Pollutants (increments) and estimates a % Change in mortality PM10 (10 mg/m3): b: 1.5 ( 0.1;3.1)
Short term; daily mean from monitoring stations
% Change in mortality PM10 (10 mg/m3): b: 1.0 (0.3;1.8)
[26]
Monitoring stations
% Change in mortality PM2.5 (IQR 6.9 mg/m3): b: 1.8 ( 0.5;4.3) NO2 (IQR 17.6 mg/m3): b: 3.5 (1.3;5.7) % Change in mortality O3 (10 mg/m3): b: 1.5 ( 0.6;3.7)
[29]
Short term; hourly data from monitoring stations
Refs [25]
[31]
Short term; daily average monitoring stations
% Change in mortality NO2 (10 mg/m3): b: 3.6 (1.7;5.5)
[32]
Short term; spatial– temporal LUR
% Change in mortality BC (IQR 0.2 g/m): b: 5.7 ( 1.7;13.7)
[36]
Short term; maximal 8-h concentration monitoring stations
% Change in mortality O3 (20 mg/m3): b: 8.3 (0.7,16.5)
[67]
Short term; monitoring stations
% Change in mortality PM2.5 (10 mg/m3): b: 2.4 (0.6;4.2)
[33]
Short term; monitoring stations
% Change in mortality PM2.5 (10 mg/m3): b: 2.0 ( 1.5;5.7) PM10 (10 mg/m3): b: 0.9 ( 1.9;3.8) % Change in mortality PM2.5 (IQR 9.5 mg/m3): b: 8.4 (1.8;15.4) NO2 (IQR 19.3 mg/m3): b: 11.6 (4.2;19.6) SO2 (IQR 2.5 mg/m3): b: 5.1 (0.7;9.7) % Change in mortality PM10 (10 mg/m3): b: 0.6 (0.0;1.2) NO2 (10 mg/m3): b: 1.3 (0.0;2.6) Summer O3 (10 mg/m3): HR: 1.07 (1.05–1.10) Winter O3 (10 mg/m3): HR: 1.03 (1.00– 1.06)
[34]
Short term; monitoring stations
Short term; one monitoring station Medium term; average during summer and winter from 8-h means from monitoring stations Long term; satellite derived
Long term; LUR and dispersion model
[28]
[35]
[27]
PM2.5 (10 mg/m3): HR:1.49 (1.37–1.62)
[37]
NO2 (third quartile 15.5–19.4 mg/m3): HR: 1.65 (0.91–2.98) NO2 (fourth quartile >19.4 mg/m3): HR: 2.15 (1.21–3.83) Reference: first quartile <13.6 mg/m3 NO2 (10 mg/m3): HR: 1.31 (0.98–1.76)
[38]
a
If estimates for different lags were provided, we extracted the lag nearest to 5 days due to availability of studies.
387
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 2. Characteristics and results of cohort, cross-sectional, and ecological studies on diabetes prevalence and incidence Study design
Outcome
Population
Cohort (follow-up T2D incidence 9.7 years)
Cohort (follow-up T2D incidence 16 years) Cohort (follow-up T2D incidence 10 years) Cohort (follow-up T2D incidence 13 years)
Cohort (follow-up T2D incidence 8 years) T2D prevalence Cross-sectional
Cross-sectional
T2D prevalence
Cross-sectional
T2D prevalence
Ecological study
Danish Diet, Cancer, and Health cohort (n = 51 818)
SALIA (n = 1775 women aged 54–55 years at baseline) Black Women’s Health Study (BWHS) (n = 3992 females) NHS (n = 74 412 females) and HPFS (n = 15 048 males)
Participants in five populationbased health surveys (n = 62 012) Patients attending two respiratory diseases clinics in Hamilton (n = 5228) and Toronto (n = 2406) with a median age of 60 years Diabetes screening study in 50–75-year olds (n = 8018)
Swiss Cohort Study on Air Pollution and Lung and Heart Disease in Adults (SAPALDIA) (n = 6392) County-level T2D Prevalence data based on a prevalence monthly telephone survey of the adult US population
Exposure duration and assessment a Long term; LUR and dispersion model
Long term; LUR models Long term; Kriging model Average exposure 12 months before T2D diagnosis; LUR models Long term; satellite derived Medium term; LUR model
Medium term; LUR model
Long term; dispersion model (PM10), dispersion + LUR model (NO2) Medium term; annual mean level obtained from monitoring stations
Pollutants (increments) and estimates
Refs
NO2 (IQR 4.9 mg/m3) All cases: HR: 1.00 (0.97–1.04) Confirmed cases: HR: 1.04 (1.01–1.07) Nonsmoker: HR: 1.12 (1.05–1.20) Physically active: HR: 1.10 (1.03–1.16) NO2 (IQR 15 mg/m3): HR: 1.42 (1.16–1.73) PM/soot (IQR 0.39 3 10 5 m): HR: 1.27 (1.09–1.48) PM2.5 (10 mg/m3): HR: 1.63 (0.78–3.44) NOx (IQR 23 mg/m3): HR: 1.25 (1.07–1.46) PM2.5 (IQR 4 mg/m3): HR: 1.00 (0.93–1.08) PM10 (IQR 7 mg/m3): HR: 1.04 (0.98–1.10)
[43]
[39]
[41] [42]
PM2.5 (10 mg/m3): HR: 1.11 (1.02–1.21)
[40]
NO2 (1.9 mg/m3): OR: 1.02 (0.98–1.049)
[44]
NO2 (third quartile 15.2–16.5 mg/m3): OR: [45] 1.25 (0.99–1.56) NO2 (fourth quartile 16.5–36.0 mg/m3): OR: 0.80 (0.63–1.02) Reference: first quartile 8.8–14.2 mg/m 3 PM10 (10 mg/m3): OR: 1.40 (1.17–1.67) [46] NO2 (10 mg/m3): OR: 1.19 (1.03–1.38)
% Increase in diabetes prevalence in 2004 [47] PM2.5 (10 mg/m3): b: 0.77 (0.39;1.25) in 2004 PM2.5 (10 mg/m3): b: 0.81 (0.48;1.07)
a
If no date for the exposure assessment was provided, medium-term exposure was assumed.
Is ambient air polluon exposure associated with diabetes development? Ambient air polluon exposure
General populaon ?
?
(A) Diabetes incidence
(B) Diabetes prevalence
(high impact) Andersen, 2012
NO2
Krämer, 2010
NO2, PM
Coogan, 2012
PM2.5 NOx
Pue, 2011 Chen, 2013
(+)
Brook, 2008
NO2
(+)
Dijkema, 2011
NO2
ο
(+) +
Eze, 2014
PM10, NO2
+
PM2.5, PM10
ο
Pearson, 2010
PM2.5
+
PM2.5
+
+
TRENDS in Endocrinology & Metabolism
Figure 2. Ambient air pollution exposure and its association with diabetes development. The studies investigate associations between ambient air pollution exposure and (A) diabetes incidence and (B) diabetes prevalence. The direction of the effect is denoted by for negative association, + for positive association, or o for no association; parentheses represent nonsignificant association or association restricted to subgroups. Most of the studies reported statistically significantly increased risk for incident and prevalent diabetes with increased exposure to ambient air pollution.
388
Review studies for diabetes incidence (Figure 2A) and prevalence (Figure 2B). Two cohort studies observed significant associations between long-term PM exposure and incident diabetes during the follow-up period in 1775 inhabitants of the highly industrialized Ruhr district in Germany and 62 012 adults living in Ontario, Canada [39,40]. Nonsignificant adverse effects were seen in 3992 African American women living in Los Angeles, USA [41]. An analysis combining data from the Nurses’ Health Study (NHS) (74 412 females) and the Health Professionals Follow-Up Study (HPFS) (15 048 males) in the USA did not yield strong evidence for an association between PM2.5 and PM10 exposure in the 12 months before T2D diagnosis and incident cases, although the study was sufficiently powered [42]. However, only recent air pollution exposure was considered in this analysis. Two of these previous studies additionally reported significant associations between NO2/NOx exposure and incident T2D [39,41]. In the 57 053 participants of the DCH cohort, no significant association between NO2 exposure and all incident diabetes cases was found but a borderline statistically significant association for confirmed diabetes cases was observed [43]. While some of the studies were conducted solely in women [39,41], Chen et al. [40] reported higher effect estimates in women compared with men. It is often hypothesized that the degree of exposure misclassification is lower in women compared with men as women tend to spend more time at home (especially at advanced ages) and air pollution exposures are generally assessed at the residential address, as was done in all studies mentioned here. Other effect modifications were also studied, revealing that air pollution effects on T2D were stronger in participants with pre-existing chronic obstructive pulmonary disease (COPD) [40], increased subclinical inflammation [39], or low education levels [38,39,43], those who undertook increased physical activity, and never-smokers [43]. Cross-sectional studies Cross-sectional studies can help assess associations between air pollution exposure and T2D (Figure 2B and Table 2). However, in these studies only prevalent diabetes cases can be analyzed. Therefore, effect estimates are more likely to be influenced by reverse causation and residual confounding. The first epidemiological evidence for a cross-sectional association was published in 2008 using data on 7634 patients attending two respiratory clinics in Hamilton and Toronto, Canada [44]. While NO2 exposure was positively associated with T2D diagnosis among women, no significant association was observed among men or in the total study population. Also, no association between NO2 exposure and diabetes prevalence was seen among 8018 individuals participating in a diabetes screening study in the semirural area of Westfriesland in The Netherlands [45]. By contrast, a recent study conducted in a Swiss cohort of 6392 participants observed a positive association between both NO2 and PM10 exposure in the 10 years preceding the survey and a higher number of prevalent T2D cases [46]. In an ecological US study, country-level diabetes prevalence
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
was greater in countries with higher annual mean PM25 concentrations [47]. A limitation of this study is the lack of individual-level data, as is the case for all ecological studies. These aforementioned studies have been partly analyzed in five recent meta-analyses [20–24] (Table 3). The meta-analysis by Balti et al. [22] estimated the combined effects of NO2 and NOx in three prospective cohort studies [39,41,43] and of PM2.5 in five prospective cohort studies [39–42]. A significantly increased risk for incident T2D was seen for both NO2/NOx and PM2.5. Park and Wang [24] meta-analyzed effects for only PM2.5 in relation to incident diabetes using data from four cohorts also included in Balti et al. [40–42]. Eze et al. [23] considered similar studies for PM2.5 but meta-analyzed pooled estimates for males and females separately, resulting in slightly different risk estimation. In Eze et al. [23], NO2 estimates derived from nonlinear estimates from Dijkema et al. [45] were included in addition. The meta-analysis by Wang et al. [20] included two published studies on prevalent diabetes [44,46] and an additional study on incident diabetes published as a congress abstract [48] as well as all of the abovementioned studies on incident diabetes [39–43]. The combined risk estimate for NO2/NOx was similar to that reported by Balti et al. For PM2.5, the combined effect estimate was higher than reported by Balti et al., although similar cohorts were included. This difference is due to the fact that Wang et al. extracted risk estimates from sensitivity analyses in which: (i) the analysis was restricted to the last 2 years of follow-up only [42]; and (ii) multi-pollutant models were considered [41]. Thus, the results of the meta-analysis by Wang et al. may overstate the adverse effects of air pollution in the general population as the estimates extracted from these sensitivity analyses were higher than those reported in the general population. Wang et al. [20] also reported a combined effect for PM10 using data from the four cohorts with available estimates for this pollutant [39,42,46]. In the meta-analysis by Janghorbani et al. [22], increased risks were observed for NO2, O3, PM10, and SO4 exposure. However, in this analysis diabetes prevalence and incidence were meta-analyzed together with diabetes mortality. Furthermore, the estimates were included in the meta-analysis without standardization, although the reported increments differed considerably. For example, effect estimates per 20mg/m3 increase in pollutant concentration, per 10-mg/m3 increase in pollutant concentration, and when comparing the upper and lower concentration quartiles were all metaanalyzed together. Following this strategy makes the resultant combined effect estimates difficult to interpret. Overall, cross-sectional studies appear to yield estimates with considerable heterogeneity whereas the results from prospective studies more consistently support an association between PM2.5, PM10, and NO2/NOx exposure and T2D development. All of the five existing meta-analyses concluded that there is evidence for an association between long-term exposure to air pollutants and T2D. However, it should be noted that the conducted studies have limitations, especially due to uncontrolled confounding by other factors such as lifestyle (diet, physical activity, socioeconomic status), other exposures (indoor, occupational, smoking), and comorbidities [23]. 389
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Is ambient air pollution exposure associated with predisposing conditions for diabetes? Studies on diabetes-related traits, which might be altered in subjects who have a higher diabetes risk, can provide further insight into the association between air pollution exposure and diabetes as well as the potential biological mechanisms underlying any association. The results of such studies are summarized in Figure 3 (for study details see Table 4). HbA1c is a marker of the average blood glucose levels in an individual during the previous 30–120 days. In 9102 newly diagnosed T2D patients in Germany, higher regional levels of PM10 were associated with higher HbA1c concentrations [49]. Adverse effects of several pollutants on HbA1c and fasting glucose concentrations were also observed in an elderly population in Taiwan [50]. In 363 women participating in the Study on the Influence of Air Pollution on Lung, Inflammation, and Aging (SALIA) cohort in Germany, a tendency for impaired glucose metabolism was observed in participants with higher air pollution exposure, although none of the associations remained statistically significant after adjusting for multiple testing [51]. Furthermore, no significant associations were observed for the 14 pro- and anti-inflammatory immune mediators investigated in this study. The authors
suggested that low statistical power due to their small sample size might explain their null findings. Insulin resistance, another potential precursor of T2D, can lead to elevated fasting glucose in combination with elevated insulin levels due to impaired insulin action. Among 560 elderly individuals living in Korea, short-term NO2 exposure was associated with elevated homeostatic model assessment of insulin resistance (HOMA-IR) in participants without a history of diabetes. For participants with diabetes, NO2, PM10, and O3 were all associated with increased HOMA-IR [52]. Furthermore, increased susceptibility was observed for participants carrying risk genotypes in the oxidative stress-related human glutathione Stransferase genes (GSTM1, GSTT1, and GSTP1). In an experimental study in the USA, 25 healthy adults living in rural Michigan had reduced insulin sensitivity after they were exposed to air pollution in an urban location for 4–5 h each day for five consecutive days [53]. Air pollution exposure effects on insulin resistance have also been investigated in healthy children and adolescents. Long-term NO2 and PM10 exposures were associated with increased HOMA-IR in 397 German children aged 10 years [54]. Additionally, short-term exposure PM10 and CO exposures led to elevated HOMA-IR in 374 Iranian children aged 10–18 years [55]. PM10, CO, and NO2 exposures were
Is ambient air polluon exposure associated with other predisposing condions for diabetes? Ambient air polluon exposure
General populaon ?
? (D) Gestaonal diabetes
(C) Diabetes related traits HbA1c Tamayo, 2014
PM10
+
Chuang, 2011
PM10, PM2.5, NO2, O3
+
HOMA-IR Kim, 2012
NO2
+
Brook, 2013
PM2.5
+
Thiering, 2013
NO2, PM10
+
Kelishadi, 2009
PM10, CO
+
NO2, NOx
(+)
Fleisch, 2014
PM2.5
o
Robledo, 2015
NOx SO2 O3
+
Malmqvist, 2013
NOx
(–) +
IGT Teichert, 2013
TRENDS in Endocrinology & Metabolism
Figure 3. Ambient air pollution exposure and its association with other predisposing conditions for diabetes. Overview of studies on ambient air pollution exposure in relation to (C) diabetes-related traits and (D) gestational diabetes. The direction of the effect is denoted by for negative association, + for positive association, or o for no association; parentheses represent nonsignificant association or association restricted to subgroups. While all reviewed studies report adverse effect of air pollutants on diabetes-related traits, there is less evidence for an association with gestational diabetes due to the heterogeneity of the results.
390
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 3. Evidence synthesis: summary of published meta-analyses on ambient air pollution exposure and T2D (X indicates that a study was included in the respective meta-analysis) Balti et al. 2014 Prospective NO2 and NOx Andersen, 2012 Kra¨mer, 2010 Coogan, 2012 Brook, 2008 Eze, 2014 Weinmayr, 2012 Dijkema, 2011 Combined estimate PM2.5 and PM10 Chen, 2013 Coogan, 2012 Kra¨mer, 2010 Puett, 2011 Eze, 2014 Brook, 2013 Combined estimate
Wang et al. 2014
X X X X X
1.13 (1.04–1.22)
1.16 (1.00–1.25)
X X X X
X X Xa X X X
Eze et al. 2015
X X X X
1.11 (1.07–1.16) PM2.5 X Xa Xa
X 1.11 (1.03–1.20)
Park and Wang 2014
Cross-sectional
X 1.39 (1.14–1.68)
PM2.5 X X
X 1.08 (1.00–1.17) PM2.5 X X
X Xa X
X
X
1.34 (1.22–1.47)
1.11 (1.03–1.19)
1.10 (1.02–1.18)
PM10
a
Estimates from sensitivity analyses in Puett et al., 2011 were used in the meta-analyses by Wang et al., which explains the high combined effect estimates observed for PM in Wang et al. For Coogan et al., 2012 effect estimates were extracted from the multi-pollutant model.
also associated with increased C-reactive protein and oxidized low-density lipoprotein in this study. Women who experience gestational diabetes are at higher risk for subsequent T2D [56]. To date, three studies have investigated associations between air pollution exposure and gestational diabetes. In 2093 pregnant women without previous type 1 diabetes or T2D living in the area of Boston, USA, PM2.5 exposure was associated with impaired glucose tolerance [57]. However, no association with gestational diabetes was observed. A large retrospective cohort in the USA that included 219 952 singleton deliveries observed higher relative risks for gestational diabetes in women with higher NOx and SO2 exposures during the 3 months preconception but a reduced risk for O3 exposure [58]. When considering only air pollution exposures that occurred during the first trimester, similar results were observed for NOx and SO2 but there was no association for O3. In a study from Taiwan, exposure to O3 led to more preterm births in women with gestational diabetes [59]. Finally, a Swedish study linked registry data from 81 110 pregnancies with individual NOx exposure and found a dose–response relationship [60]. These studies demonstrate that air pollutants might have the potential to affect diabetes development and may also be associated with gestational diabetes. However, additional studies are needed to draw firmer conclusions. Concluding remarks and future perspectives There appears to be evidence indicative of an association between long-term exposure to main air pollutants and the development of T2D and diabetes-related mortality in adults. The results from studies that have investigated precursors of diabetes in healthy individuals strengthen the support for a potential association. However, a firm conclusion regarding the association between ambient air pollution exposure and diabetes-related traits and gestational diabetes cannot be presently drawn due to the limited data available. For short(er)-term air pollution
exposures, the results are more heterogeneous and the evidence too weak to demonstrate a causal relationship. Most of the studies reviewed here were conducted in developed countries in North America and Europe. Studies in other parts of the world are needed, where air pollution concentrations are higher and the diabetes burden is greater, especially to determine potential dose–response relationships. Furthermore, an unanswered question exists regarding the latency period of air pollution exposure: does high air pollution exposure lead to priming effects that manifest their risk even after a certain time of low exposure? It also remains unclear whether exposure improvement – for example, by moving to a cleaner environment – could reduce the risk for T2D or intermediate phenotypes such as insulin resistance or glycemic control. The experimental study by Brook et al. [53] only studied effects after subacute air pollution exposure in individuals living in a rural environment. Exposure reduction effects among individuals living in polluted areas have not been assessed so far. Furthermore, air pollution exposure sources are often located indoors. Household indoor air pollution concentrations often exceed outdoor concentrations by a factor of two to five [61,62], and even occasionally up to a factor of 100. Air pollution concentrations in the work environment are also often higher than the mean concentration outdoors [63,64]. Air pollution concentrations indoors and in work environments have not yet been studied in relation to T2D. Another important gap is that the role of potential environmental confounders, such as noise and reduced greenness, that are correlated with air pollution have not yet been extensively evaluated [65,66]. In conclusion, the available evidence supporting adverse effects of air pollution on diabetes and the related high public health impact of such an association justify the need for further investigations. Future studies are also needed to establish how and to what extent air pollution control measures may reduce the global diabetes-related burden of disease. 391
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 4. Characteristics and results of cross-sectional and experimental studies on diabetes-related traits Study design
Outcome
Population
Exposure duration and assessment a Medium term; chemical transport model with interpolation, area level of 100 five-digit postcode areas
Cross-sectional
HbA1c
Newly diagnosed T2D patients 18 years in the DPV database (n = 9102)
Cross-sectional
HbA1c, fasting glucose
Participants in the Social Environment and Biomarkers of Aging study 54 years (n = 1023)
1 year; 1-year mean of monitoring stations at area level
Experimental
Insulin resistance
Short term; monitoring stations
Cross-sectional
Insulin resistance
Cross-sectional
Insulin resistance
Cross-sectional
Insulin resistance
Cross-sectional
Impaired glucose tolerance
Healthy adults living in rural Michigan were transported to an urban location for five consecutive days (n = 25) Participants in the Korean Elderly Environmental Panel (KEEP) 50 years (n = 560) 10-year-old participants of the birth cohorts GINIplus and LISAplus (n = 397) Population-based study of adolescents aged 10–18 years (n = 374) Participants in the SALIA cohort (n = 363 females)
Cross-sectional
Gestational diabetes
Project Viva cohort (n = 2093 pregnant women)
Cross-sectional
Gestational diabetes
Consortium on Safe Labor retrospective cohort in 19 hospitals (n = 219 952 women with singleton pregnancy)
Cross-sectional
Gestational diabetes
Singleton deliveries during 1999–2005 from the Swedish Medical Birth Registry (n = 81 110)
Pollutants (increments) and estimates
Refs
Increase in HbA1c PM10 (third quartile 18.1–21.1 mg/m3): b: 0.21 (0.09;0.32) PM10 (fourth quartile >21.1 mg/m3): b: 0.36 (0.22;0.49) Reference: first quartile <16.4 mg/m 3 For % in HbA1c PM10 (IQR 48 mg/m3): b: 1.40 (1.11;0.32) PM2.5 (IQR 20.4 mg/m3): b: 2.24 (1.47;3.00) NO2 (IQR 24 mg/m3): b: 1.08 (0.84;1.33) For mg/dl in fasting glucose PM10 (IQR 48 mg/m3): b: 22.9 (14.9;30.8) PM2.5 (IQR 20.4 mg/m3): b: 36.6 (19.2;53.9) NO2 (IQR 24 mg/m3): b: 17.0 (10.4;23.7) Increase in HOMA-IR PM2.5 (10 mg/m3): b: 0.7 (0.1;1.3)
[49]
[50]
[53]
Short term; daily concentrations measured by monitoring stations
PM10 (IQR 20.8 mg/m3): b: 0.14 ( 0.003;0.29) NO2 (IQR 10.8 mg/m3): b: 0.28 (0.13;0.42)
[52]
Long term; LUR model
% Increase in HOMA-IR NO2 (2SD 10.6 mg/m3): b: 17.0 (5.0;30.3) PM10 (2SD 6.0 mg/m3): b: 18.7 (2.9;36.9) PM2.5 (2SD 3.7 mg/m3): b: 17.0 ( 0.9;51.5) PM10 (unknown): b: 1.1 (0.5;1.7) NO2 (unknown): b: 0.8 (0.6;1.0)
[54]
NO2 (IQR 14.7 mg/m3): OR: 1.47 (1.05–2.05) PM2.5 (IQR unknown): OR: 1.15 (0.81–1.63) PM10 (IQR unknown): OR 1.20 (0.87–1.68)
[51]
Impaired glucose tolerance PM2.5 (IQR 2 mg/m3): OR: 1.64 (1.11–2.42) Gestational diabetes PM2.5 (IQR 2 mg/m3): OR 0.94 (0.67–1.34) PM2.5 (IQR 5.54 mg/m3): RR: 0.97 (0.93–1.02) PM10 (IQR 6.3 mg/m3): RR: 0.99 (0.96–1.02) NOx (IQR 15.1 mg/m3): RR: 1.09 (1.04–1.13) O3 (IQR: 6.1 mg/m3): RR: 0.93 (0.90–0.96)
[57]
Short term; daily concentrations measured by monitoring stations Long term; local monitoring stations or LUR model, backextrapolated Second trimester; spatiotemporal model and LUR model 91 days before last menstrual period, first trimester; Community Multi-scale Air Quality Model: average exposure level of the hospital First trimester; dispersion model
NOx (third quartile 14.2–22.6 mg/m3): OR: 1.52 (1.28–1.82) NOx (fourth quartile >22.7 mg/m3): OR: 1.69 (1.41–2.03) Reference: first quartile (2.5–8.9 mg/m3)
[55]
[58]
[60]
a
If no date for the exposure assessment was provided, medium-term exposure was assumed.
Acknowledgments The authors thank Dr Elaine Fuertes (Institute of Epidemiology I, Helmholtz Zentrum Mu¨nchen, Germany) for editorial assistance in the preparation of the manuscript.
References 1 van Donkelaar, A. et al. (2015) Use of satellite observations for longterm exposure assessment of global concentrations of fine particulate matter. Environ. Health Perspect. 123, 135–143 2 Turner, M.C. et al. (2011) Long-term ambient fine particulate matter air pollution and lung cancer in a large cohort of never-smokers. Am. J. Respir. Crit. Care Med. 184, 1374–1381 392
3 Chen, H. et al. (2013) Long-term exposure to traffic-related air pollution and cardiovascular mortality. Epidemiology 24, 35–43 4 Shah, A.S. et al. (2013) Global association of air pollution and heart failure: a systematic review and meta-analysis. Lancet 382, 1039–1048 5 Cesaroni, G. et al. (2014) Long term exposure to ambient air pollution and incidence of acute coronary events: prospective cohort study and meta-analysis in 11 European cohorts from the ESCAPE Project. BMJ 21, 348 6 Raaschou-Nielsen, O. et al. (2013) Air pollution and lung cancer incidence in 17 European cohorts: prospective analyses from the European Study of Cohorts for Air Pollution Effects (ESCAPE). Lancet Oncol. 14, 813–822
Review 7 Lim, S.S. et al. (2012) A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 380, 2224–2260 8 Gordon, S.B. et al. (2014) Respiratory risks from household air pollution in low and middle income countries. Lancet Respir. Med. 2, 823–860 9 Kampa, M. and Castanas, E. (2008) Human health effects of air pollution. Environ. Pollut. 151, 362–367 10 Brook, R.D. et al. (2010) Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the American Heart Association. Circulation 121, 2331–2378 11 WHO (2012) Global Status Report on Noncommunicable Diseases 2014, World Health Organization 12 Guariguata, L. et al. (2014) Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res. Clin. Pract. 103, 137–149 13 Finucane, M.M. et al. (2011) National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet 377, 557–567 14 Rajagopalan, S. and Brook, R.D. (2012) Air pollution and type 2 diabetes: mechanistic insights. Diabetes 61, 3037–3045 15 Ru¨ckerl, R. et al. (2014) Associations between ambient air pollution and blood markers of inflammation and coagulation/fibrinolysis in susceptible populations. Environ. Int. 70, 32–49 16 Liu, C. et al. (2013) Epidemiological and experimental links between air pollution and type 2 diabetes. Toxicol. Pathol. 41, 361–373 17 Bind, M.A. et al. (2012) Air pollution and markers of coagulation, inflammation, and endothelial function: associations and epigene– environment interactions in an elderly cohort. Epidemiology 23, 332–340 18 Carmona, J.J. et al. (2014) Short-term airborne particulate matter exposure alters the epigenetic landscape of human genes associated with the mitogen-activated protein kinase network: a cross-sectional study. Environ. Health 13, 94 19 Lockwood, A.H. (2002) Diabetes and air pollution. Diabetes Care 25, 1487–1488 20 Wang, B. et al. (2014) Effect of long-term exposure to air pollution on type 2 diabetes mellitus risk: a systemic review and meta-analysis of cohort studies. Eur. J. Endocrinol. 171, R173–R182 21 Balti, E.V. et al. (2014) Air pollution and risk of type 2 diabetes mellitus: a systematic review and meta-analysis. Diabetes Res. Clin. Pract. 106, 161–172 22 Janghorbani, M. et al. (2014) Systematic review and metaanalysis of air pollution exposure and risk of diabetes. Eur. J. Epidemiol. 29, 231– 242 23 Eze, I.C. et al. (2015) Association between ambient air pollution and diabetes mellitus in Europe and North America: systematic review and meta-analysis. Environ. Health Perspect. 123, 381–389 24 Park, S.K. and Wang, W. (2014) Ambient air pollution and type 2 diabetes: a systematic review of epidemiologic research. Curr. Environ. Health Rep. 1, 275–286 25 Bateson, T.F. and Schwartz, J. (2004) Who is sensitive to the effects of particulate air pollution on mortality? A case-crossover analysis of effect modifiers. Epidemiology 15, 143–149 26 Forastiere, F. et al. (2008) Particulate matter and daily mortality: a case-crossover analysis of individual effect modifiers. Epidemiology 19, 571–580 27 Zanobetti, A. et al. (2014) A national case-crossover analysis of the short-term effect of PM2.5 on hospitalizations and mortality in subjects with diabetes and neurological disorders. Environ. Health 13, 38 28 Goldberg, M.S. et al. (2006) Associations between ambient air pollution and daily mortality among persons with diabetes and cardiovascular disease. Environ. Res. 100, 255–267 29 Goldberg, M.S. et al. (2013) Associations between ambient air pollution and daily mortality among elderly persons in Montreal. Quebec. Sci. Total Environ. 463–464, 931–942 30 Zanobetti, A. and Schwartz, J. (2011) Ozone and survival in four cohorts with potentially predisposing diseases. Am. J. Respir. Crit. Care Med. 184, 836–841 31 Stafoggia, M. et al. (2010) Susceptibility factors to ozone-related mortality: a population-based case-crossover analysis. Am. J. Respir. Crit. Care Med. 182, 376–384
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
32 Chiusolo, M. et al. (2011) Short-term effects of nitrogen dioxide on mortality and susceptibility factors in 10 Italian cities: the EpiAir Study. Environ. Health Perspect. 119, 1233–1238 33 Ostro, B. et al. (2006) Fine particulate air pollution and mortality in nine California counties: results from CALFINE. Environ. Health Perspect. 114, 29–33 34 Samoli, E. et al. (2014) Which specific causes of death are associated with short term exposure to fine and coarse particles in Southern Europe? Results from the MED-PARTICLES project. Environ. Int. 67, 54–61 35 Kan, H. et al. (2004) he association of daily diabetes mortality and outdoor air pollution in Shanghai, China. J. Environ. Health 67, 21–26 36 Maynard, D. (2007) Mortality risk associated with short-term exposure to traffic particles and sulfates. Environ. Health Perspect. 115, 751–755 37 Brook, R.D. et al. (2013) Long-term fine particulate matter exposure and mortality from diabetes in Canada. Diabetes Care 36, 3313–3320 38 Raaschou-Nielsen, O. et al. (2013) Long-term exposure to trafficrelated air pollution and diabetes-associated mortality: a cohort study. Diabetologia 56, 36–46 39 Kra¨mer, U. et al. (2010) Traffic-related air pollution and incident type 2 diabetes: results from the SALIA cohort study. Environ. Health Perspect. 118, 1273–1279 40 Chen, H. et al. (2013) Risk of incident diabetes in relation to long-term exposure to fine particulate matter in Ontario, Canada. Environ. Health Perspect. 121, 804–810 41 Coogan, P.F. et al. (2012) Air pollution and incidence of hypertension and diabetes mellitus in black women living in Los Angeles. Circulation 125, 767–772 42 Puett, R.C. et al. (2011) Are particulate matter exposures associated with risk of type 2 diabetes? Environ. Health Perspect. 119, 384–389 43 Andersen, Z.J. et al. (2012) Diabetes incidence and long-term exposure to air pollution: a cohort study. Diabetes Care 35, 92–98 44 Brook, R.D. et al. (2008) The relationship between diabetes mellitus and traffic-related air pollution. J. Occup. Environ. Med. 50, 32–38 45 Dijkema, M.B. et al. (2011) Long-term exposure to traffic-related air pollution and type 2 diabetes prevalence in a cross-sectional screeningstudy in The Netherlands. Environ. Health 10, 76 46 Eze, I.C. et al. (2014) Long-term air pollution exposure and diabetes in a population-based Swiss cohort. Environ. Int. 70, 95–105 47 Pearson, J.F. et al. (2010) Association between fine particulate matter and diabetes prevalence in the US. Diabetes Care 33, 2196–2201 48 Weinmayr, G. et al. (2014) Industry and traffic related air pollution and diabetes type two incidence: results from a German cohort study. Eur. J. Epidemiol. 27, S12 49 Tamayo, T. et al. (2014) Is particle pollution in outdoor air associated with metabolic control in type 2 diabetes? PLoS ONE 9, e91639 50 Chuang, K.J. et al. (2011) Long-term air pollution exposure and risk factors for cardiovascular diseases among the elderly in Taiwan. Occup. Environ. Med. 68, 64–68 51 Teichert, T. et al. (2013) Association between traffic-related air pollution, subclinical inflammation and impaired glucose metabolism: results from the SALIA study. PLoS ONE 8, e83042 52 Kim, J.H. and Hong, Y.C. (2012) GSTM1, GSTT1, and GSTP1 polymorphisms and associations between air pollutants and markers of insulin resistance in elderly Koreans. Environ. Health Perspect. 120, 1378–1384 53 Brook, R.D. et al. (2013) Reduced metabolic insulin sensitivity following sub-acute exposures to low levels of ambient fine particulate matter air pollution. Sci. Total Environ. 448, 66–71 54 Thiering, E. et al. (2013) Long-term exposure to traffic-related air pollution and insulin resistance in children: results from the GINIplus and LISAplus birth cohorts. Diabetologia 56, 1696–1704 55 Kelishadi, R. et al. (2009) Lifestyle and environmental factors associated with inflammation, oxidative stress and insulin resistance in children. Atherosclerosis 203, 311–319 56 Kwak, S.H. et al. (2013) Clinical and genetic risk factors for type 2 diabetes at early or late postpartum after gestational diabetes mellitus. J. Clin. Endocrinol. Metab. 98, E744–E752 57 Fleisch, A.F. (2014) Air pollution exposure and abnormal glucose tolerance during pregnancy: the Project Viva cohort. Environ. Health Perspect. 122, 378–383
393
Review 58 Robledo, C.A. et al. (2015) Preconception and early pregnancy air pollution exposures and risk of gestational diabetes mellitus. Environ. Res. 137C, 316–322 59 Lin, Y.T. et al. (2015) Associations between ozone and preterm birth in women who develop gestational diabetes. Am. J. Epidemiol. 181, 280– 287 60 Malmqvist, E. et al. (2013) Gestational diabetes and preeclampsia in association with air pollution at levels below current air quality guidelines. Environ. Health Perspect. 121, 488–493 61 Bordada, J.C. et al. (2012) Exposure to airborne ultrafine particles from cooking in Portuguese homes. J. Air Waste Manag. Assoc. 62, 1116– 1126 62 Nadasena, S. et al. (2012) Indoor fine particle (PM2.5) pollution exposure due to secondhand smoke in selected public places of Sri Lanka. Am. J. Ind. Med. 55, 1129–1136
394
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
63 Mitchell, D.C. et al. (2013) Particulate matter, endotoxin, and worker respiratory health on large Californian dairies. J. Occup. Environ. Med. 57, 79–87 64 Dioni, L. et al. (2011) Effects of short-term exposure to inhalable particulate matter on telomere length, telomerase expression, and telomerase methylation in steel workers. Environ. Health Perspect. 119, 622–627 65 Te´treault, L.F. et al. (2013) Cardiovascular health, traffic-related air pollution and noise: are associations mutually confounded? A systematic review. Int. J. Public Health 58, 649–666 66 Dadvand, P. et al. (2012) Surrounding greenness and exposure to air pollution during pregnancy: an analysis of personal monitoring data. Environ. Health Perspect. 120, 1286–1290 67 Ren, C. et al. (2010) Modifiers of short-term effects of ozone on mortality in eastern Massachusetts – a case-crossover analysis at individual level. Environ. Health 9, 3
Epidemiology of air pollution and diabetes Elisabeth Thiering1,2 and Joachim Heinrich1,3 1
Institute of Epidemiology I, Helmholtz Zentrum Mu¨nchen – German Research Center for Environmental Health, Neuherberg, Germany 2 Division of Metabolic and Nutritional Medicine, Dr von Hauner Children’s Hospital, University of Munich Medical Center, Munich, Germany 3 Institute and Outpatient Clinic for Occupational, Social and Environmental Medicine, Inner City Clinic, University Hospital Munich, Ludwig Maximilian University of Munich, Munich, Germany
Air pollution affects a large proportion of the global population. Air pollutants are hypothesized to exert their effects via impaired endothelial function, elevated systemic inflammation, mitochondrial dysfunction, and oxidative stress, all of which are hallmarks of type 2 diabetes (T2D). Here we review epidemiological studies aimed at answering whether diabetes patients are more vulnerable to ambient (outdoor) air pollution exposure and whether air pollution is associated with diabetes development or other predisposing conditions for T2D. Current evidence suggests an association between air pollution exposure and T2D, but more critical analysis is warranted. Understanding the associations between air pollution exposure and the development of T2D is critical in our efforts to control sources of air pollution and their impact on the disease. The diabetes epidemic and ubiquitous air pollution exposure Air pollution is a ubiquitous exposure that affects large proportions of the global population [1]. Air pollutants are emitted from many sources, such as industrial facilities, cars, trucks, ships, and airplanes, but also from household combustion devices or during forest fires and volcanic eruptions. Ambient or outdoor air pollution (referred to as air pollution hereafter) comprises particulate matter (PM) (see Glossary) of various sizes [<10 mm in diameter, PM10; <2.5 mm, PM2.5; <100 nm, ultrafine particles (UFPs)], chemicals such as persistent organic pollutants (POPs), and gaseous compounds such as nitric oxides (NOx), carbon monoxide (CO), ozone (O3), and sulfur oxides (SOx). These compounds differ in their dispersion, reactivity, and toxicity. Air pollution exposure has been linked to a reduced life expectancy, mainly attributable to cardiovascular and respiratory diseases such as heart disease and lung cancer [2–4], even among individuals exposed to annual average concentrations below the current air quality standards in Europe [5,6]. Corresponding author: Thiering, E. ([email protected]). Keywords: air pollution; type 2 diabetes; gestational diabetes; mortality. 1043-2760/ ß 2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tem.2015.05.002
384
Trends in Endocrinology and Metabolism, July 2015, Vol. 26, No. 7
In 2010, more than 3.2 million deaths worldwide were attributed to outdoor PM exposure and 3.5 million deaths to indoor household air pollution from solid fuels [7]. Air pollution exposure also accounts for a large number of disease-adjusted life-years lost (29.4 million for PM pollution and 47.9 million for household air pollution from solid fuels, which are equivalent to 3% and 4.3% of the worldwide burden of disease, respectively). Newer data from 2012 revealed 3.7 million deaths attributable to ambient air pollution and 4.3 million to household air pollution [Global Health Observatory Data Repository (http://apps.
Glossary Black carbon (BC): a fraction of PM2.5 that is mainly emitted during open biomass burning and from diesel cars and trucks without soot particle filters. Confidence interval (CI): an interval estimation of a parameter to show the reliability of the estimates. Typically, 95% CIs are reported, which represent an interval that contains the true value with a 95% probability. Hazard ratio (HR): a measure of mortality risk between different groups corresponding to a certain period of time. It is defined as the ratio between two hazard rates and is typically estimated via proportional hazard models. Interquartile range (IQR): the distance between the 75% percentile (third quartile) and 25% percentile (first quartile) of a distribution. Being a measure of dispersion it is often used to calculate effect sizes per increase in IQR to make them comparable between studies. Mortality rate ratio (MRR): defines the ratio between two mortality rates corresponding to two subgroups or to different causes of death. Nitric oxide (NO): a free radical and a byproduct of combustion in automobile engines and fossil-fuel power plants. In humans it is an important cellular signaling molecule and it is involved in many physiological processes. In the air it is rapidly oxidized to nitrogen dioxide (NO2). Ozone (O3): also known as trioxygen; an inorganic molecule. At ground level O3 is primarily formed by the action of sunlight on air containing hydrocarbons and NOx from combustion processes. It is not emitted directly from car engines or industrial operations. Particulate matter (PM): a complex mixture of small particles and liquid droplets. Particle pollution comprises numerous components, including acids (such as nitrates and sulfates), organic chemicals, metals, and soil or dust particles. These compounds can be directly emitted or formed through reactions with other compounds. The size of a particle is directly linked to its potential to penetrate the lung and cause health problems. Therefore, particles are divided into those of diameter <10 mm (PM10), which can pass into the throat and nose and enter the lung, and those of diameter <2.5 mm (PM2.5), which can additionally reach the alveolar region of the lung. Sulfur (S): an abundant, multivalent, non-metal element. It is oxidized to sulfur dioxide (SO2) during combustion processes. Major sources of SO2 emissions are volcanic eruptions and fossil fuel combustion at power plants and industrial facilities. In the atmosphere SO2 is converted to sulfate (SO4). Ultrafine particles (UFPs): defined as particles with aerodynamic diameter <0.1 mm and best characterized using particle number concentrations instead of particle mass.
Review who.int/gho/data/node.main.122?lang=en)]. While household indoor air pollution from solid fuels plays a major role in Asia, Latin America, and Africa [8], outdoor air pollution from PM is of global importance. Although there is strong evidence that air pollution adversely influences a broad spectrum of health indicators [2–6,9,10], links between air pollution exposure and T2D development have been only recently revealed. T2D affects approximately 9% of the global population [11] and accounts for 2.7% of global deaths (WHO, 2012). The number of affected people is expected to increase to 592 million by 2035 [12], with the steepest increases predicted in low- and middle-income countries. Changes in lifestyle factors due to urbanization, such as reduced physical activity, overnutrition, and obesity [13], were previously advanced as explanations for this trend. Recently, there is also interest in, and some evidence for, an association between air pollution exposure and T2D. Several biological mechanisms have been proposed to explain such a link. For example, PM exposure has been associated with impaired endothelial function, elevated systemic inflammation and oxidative stress, endoplasmic reticulum stress, cardiac autonomic nervous system dysfunction, and mitochondrial dysfunction [10,14–16]. Additionally, epigenetic changes leading to activation of key signaling pathways, or changes in markers of coagulation, inflammation, and endothelial function, have been described after exposure to air pollutants [17,18]. An ecological study by Alan Lockwood [19] was the first to link diabetes prevalence in adults to total air releases from all industries. More recently, several epidemiological studies varying in design, exposure, and outcome that are discussed below have explored the connection between air quality and diabetes. In this review we focus on epidemiological studies that examined associations between outdoor air pollution exposure and T2D. We also consider mortality due to diabetes, diabetes-related traits, and gestational diabetes and comprehensively summarize the results of five previous metaanalyses [20–24]. We address three fundamental questions: are diabetes patients more vulnerable to air pollution exposure, is exposure to air pollution associated with diabetes development, and/or is it associated with other conditions that predispose to diabetes? While most T2D determinants, such as inactivity or overweight, are individual factors, environmental hazards have a mostly involuntary nature. Therefore, policies that modify environmental risk factors can protect the general population. However, to fully evaluate potential public health impacts, better understanding of who is at risk [e.g., only predisposed subjects (for example, those having T2D) or also healthy individuals] is required. Although the currently reported risk increases are relatively small, given the large proportion of subjects exposed to air pollution and the high prevalence of T2D establishing and modifying a link between these two factors could have a large impact on healthy living. Although the body of evidence for adverse health effects of air pollution, especially of PM, is increasing, results specific to diabetes-related outcomes could be used in the ongoing debate on air quality guidelines setting. Thus, evaluating
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
whether air pollution is related to T2D is highly important and very timely. Are diabetes patients more vulnerable to ambient air pollution exposure? Mainly case-crossover and time-series study designs have been used to compare the effects of short-term exposure to air pollutants on all-cause mortality rates among the general population and diabetic patients. A total of seven studies have been conducted in American, Italian, Canadian, and European populations [25–32] involving between 12 978 [25] and 2935 647 [30] subjects with diabetes (for an overview see Figure 1A, for details see Table 1). Diabetes was not the original primary focus of these investigations with the exception of the study by Zanobetti et al. [30], in which 2935 647 subjects with a history of diabetes-related emergency hospital admission were followed. In other studies, death rates after days of high air pollution exposure among diabetic subjects were compared with those in the general population as a subgroup analysis. Among 65 180 elderly people who had a history of hospitalization for lung or heart disease, increased PM10 levels were positively associated with mortality risk for participants with a prior diagnosis of diabetes and this risk was greater than that observed among those without a prior diagnosis of diabetes [25]. Similar PM10 effects on mortality were seen in a study including nine Italian cities [26]. A national case-crossover analysis conducted in 121 cities in the USA observed associations between short-term PM2.5 exposure and increased risk for diabetes-related hospitalization and all-cause mortality in subjects with a previous diabetesrelated hospital admission [27]. Again, these PM2.5-associated risks were greater than those observed among individuals without diabetes. Increased nonaccidental mortality among subjects previously diagnosed with diabetes was observed in association with higher PM2.5 and SO4 concentrations in a mortality time-series study in Montreal, Canada [28]. However, no increased risk was observed among subjects who had diabetes but not cancer, cardiovascular disease, or an airways disease. An extension of this study published in 2013, which included deaths in Montreal during a later time period, revealed similar but slightly decreased effect estimates [29]. Associations between mortality among diabetes patients and O3 levels were also investigated. Higher average summer concentrations O3 were positively associated with mortality risk among 2935 647 patients with a history of diabetes-related emergency hospital admission living in 105 cities in the USA [30]. A nonsignificantly increased mortality risk following higher short-term O3 exposure was also seen among 14 350 subjects with diabetes in Italy [31]. Finally, a study on short-term NO2 exposure effects reported an increase in all-cause mortality risk among subjects with diabetes that was greater than that reported among those without diabetes [32] (Figure 1A). In summary, all seven studies investigating the effects of air pollution exposure on all-cause mortality observed a tendency for higher mortality rates among those with diabetes compared with those without. However, statistical significance (P < 0.05) was reached in only two of the 385
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Are diabetes paents more vulnerable to ambient air polluon exposure? Ambient air polluon exposure
General populaon
Diabetes paents
?
?
(B) Diabetes mortality
(A) All-cause mortality Bateson, 2004 Forasere, 2008 Zanobe, 2014 Goldberg, 2013 Zanobe, 2011 Staffogia, 2010 Chiusolo, 2011
PM10 PM10 PM2.5 PM2.5, NO2 O3 O3 NO2
(+) + (+) (+) + (+) (+)
Ostro, 2006 Samoli, 2014 Goldberg, 2006 Kan, 2004 Maynard, 2007 Ren, 2010
PM2.5 PM2.5, PM10 PM2.5, SO4, SO2, NO2, CO PM10, SO2, NO2 BC, SO4 O3
Long-term exposure Brook, 2013 Raaschou-Nielsen, 2013
PM2.5 NO2
+ (+) + + (+) + + +
TRENDS in Endocrinology & Metabolism
Figure 1. Overview of studies examining the vulnerability of diabetes patients to ambient air pollutants. (A) Studies that compare all-cause mortality rates in subjects with diabetes to those in the general population (B). Studies on diabetes-related mortality risk in the general population. The direction of the effect is denoted by for negative association, + for positive association, or o for no association; parentheses represent nonsignificant association or association restricted to subgroups. The summary shows that air pollution exposure leads to a higher mortality risk in diabetic subjects and to increased mortality due to diabetes in the general population in all of the studies despite nonsignificance in most of them.
studies. Compared with subjects without diabetes, the observed up-to-twofold increased mortality risks suggest that patients with diabetes are more susceptible to the immediate effects of short-term air pollution. Diabetes-related mortality in the general population Eight studies have investigated how air pollution exposure affects diabetes-related mortality in the general population (Figure 1B). Although the study design is different from the abovementioned studies, the underlying research question is similar, as it is always difficult to disentangle underlying and contributing causes of death in complex diseases. Five studies suggested a positive change in daily diabetes-related mortality after days of high PM or black carbon (BC) exposure [28,33–36]. These studies were conducted in the Montreal area [28], Boston [37], nine Californian counties [33], ten Mediterranean cities [34], and Shanghai [35] and involved between 431 and 15 184 diabetes-related deaths. However, statistical significance was reached in only three of the studies. These studies demonstrate that short-term exposure to air pollutants can potentially trigger diabetes-related mortality in the general population and among subjects with other medical conditions. Two large cohorts in Canada and Europe have studied long-term air pollution exposure effects in relation to diabetes mortality. Using a cohort of 2.1 million adults from the 1991 Canadian census mortality follow-up study, higher long-term PM2.5 exposure was associated with increased diabetes-related mortality [37]. In the Danish 386
Diet, Cancer, and Health cohort (DCH), long-term NO2 exposure was positively associated with diabetes-related mortality in 52 061 adults [38]. Although many of the estimates in the above-described studies were not statistically significant, the cumulative evidence appears to suggest that short-term ambient air pollution exposure increases the daily all-cause mortality risk in diabetic subjects and that both short-term and longterm exposure might be associated with diabetes-related mortality risk in the general population (Figure 1). Nevertheless, more large studies investigating long-term air pollution exposure in relation to mortality in diabetic patients as well as diabetes-related mortality in the general population are needed to establish causality and draw final conclusions. Is ambient air pollution exposure associated with diabetes development? Prospective cohort studies As discussed above, diabetic patients might be more susceptible to air pollution effects. Prospective studies can elucidate whether air pollution exposure only leads to accelerated progression of the disease or whether it directly contributes to its development. Prospective studies follow participants without diabetes for several years and evaluate potential associations between air pollution exposure at baseline or average air pollution exposure during the whole follow-up period with newly diagnosed T2D cases (Table 2). Figure 2 summarizes the results of the
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 1. Characteristics and results of case-crossover, time-series, and cohort studies on diabetes mortality Study design
Outcome
Population
Case crossover
Mortality
Case crossover
Mortality
Case crossover
Mortality
Mortality data of US citizens 65 years admitted to hospital with a diagnosis of heart disease or lung disease 1988– 1991 (n = 65 180, with prior T2D diagnosis n = 12 978) Mortality data of residents of nine Italian cities 35 years with natural cause of death obtained from regional registers of causes of death 1997–2004 (n = 321 014, 30 173 with hospital admission due to T2D within 2 years before death) Mortality data of residents of Montreal 65 years 1993–2003 (n = 158 350, n = 38 883 with T2D)
Case crossover
Mortality
Case crossover
Mortality
Case crossover
Diabetes-related mortality
Case crossover
Diabetes-related mortality
Time-series
Diabetes-related mortality
Time series
Diabetes-related mortality
Time series
Diabetes-related mortality
Time series
Diabetes-related mortality
Cohort (follow-up 5.6 years)
Mortality
Cohort (follow-up 10 years)
Diabetes-related mortality
Cohort (follow-up 13 years)
Diabetes-related mortality
Mortality data of residents of ten Italian cities 35 years with natural cause of death 2001–2005 between April and September (n = 126 647, with prior T2D hospital admission 14 350) Mortality data of residents of ten Italian cities 35 years with natural cause of death 2001–2005 (n = 276 205, with prior T2D 30 620) Mortality data from the Boston metropolitan area 1995–2002 (n = 107 925, n = 2694 due to T2D) Mortality data of residents from three counties in eastern Massachusetts 35 years 1995–2002 (n = 157 197, n = 3845 due to T2D) Deaths due to diabetes for all residents in nine Californian counties 1999–2002 (n = 15 184) Deaths due to diabetes in nine metropolitan areas in the European Mediterranean region (n = 9832) Mortality data of residents of Montreal 65 years 1984–1993 (n = 2922 deaths due to T2D)
Mortality data of residents of Zhabei district of Shanghai 2001–2002 (n = 431 due to T2D) Mortality data of US residents 65 years after emergency admission to hospital with a primary diagnosis of diabetes 1985–2005 (n = 2935 647)
1991 Canadian census mortality follow-up, subjects >25 years (n = 2145 400, n = 5200 deaths due to diabetes) Danish Diet, Cancer, and Health cohort, mortality in the nationwide Register of Causes of Death 1993– 2009 (n = 52 061)
Exposure duration and assessment Short term; daily mean from monitoring stations
Pollutants (increments) and estimates a % Change in mortality PM10 (10 mg/m3): b: 1.5 ( 0.1;3.1)
Short term; daily mean from monitoring stations
% Change in mortality PM10 (10 mg/m3): b: 1.0 (0.3;1.8)
[26]
Monitoring stations
% Change in mortality PM2.5 (IQR 6.9 mg/m3): b: 1.8 ( 0.5;4.3) NO2 (IQR 17.6 mg/m3): b: 3.5 (1.3;5.7) % Change in mortality O3 (10 mg/m3): b: 1.5 ( 0.6;3.7)
[29]
Short term; hourly data from monitoring stations
Refs [25]
[31]
Short term; daily average monitoring stations
% Change in mortality NO2 (10 mg/m3): b: 3.6 (1.7;5.5)
[32]
Short term; spatial– temporal LUR
% Change in mortality BC (IQR 0.2 g/m): b: 5.7 ( 1.7;13.7)
[36]
Short term; maximal 8-h concentration monitoring stations
% Change in mortality O3 (20 mg/m3): b: 8.3 (0.7,16.5)
[67]
Short term; monitoring stations
% Change in mortality PM2.5 (10 mg/m3): b: 2.4 (0.6;4.2)
[33]
Short term; monitoring stations
% Change in mortality PM2.5 (10 mg/m3): b: 2.0 ( 1.5;5.7) PM10 (10 mg/m3): b: 0.9 ( 1.9;3.8) % Change in mortality PM2.5 (IQR 9.5 mg/m3): b: 8.4 (1.8;15.4) NO2 (IQR 19.3 mg/m3): b: 11.6 (4.2;19.6) SO2 (IQR 2.5 mg/m3): b: 5.1 (0.7;9.7) % Change in mortality PM10 (10 mg/m3): b: 0.6 (0.0;1.2) NO2 (10 mg/m3): b: 1.3 (0.0;2.6) Summer O3 (10 mg/m3): HR: 1.07 (1.05–1.10) Winter O3 (10 mg/m3): HR: 1.03 (1.00– 1.06)
[34]
Short term; monitoring stations
Short term; one monitoring station Medium term; average during summer and winter from 8-h means from monitoring stations Long term; satellite derived
Long term; LUR and dispersion model
[28]
[35]
[27]
PM2.5 (10 mg/m3): HR:1.49 (1.37–1.62)
[37]
NO2 (third quartile 15.5–19.4 mg/m3): HR: 1.65 (0.91–2.98) NO2 (fourth quartile >19.4 mg/m3): HR: 2.15 (1.21–3.83) Reference: first quartile <13.6 mg/m3 NO2 (10 mg/m3): HR: 1.31 (0.98–1.76)
[38]
a
If estimates for different lags were provided, we extracted the lag nearest to 5 days due to availability of studies.
387
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 2. Characteristics and results of cohort, cross-sectional, and ecological studies on diabetes prevalence and incidence Study design
Outcome
Population
Cohort (follow-up T2D incidence 9.7 years)
Cohort (follow-up T2D incidence 16 years) Cohort (follow-up T2D incidence 10 years) Cohort (follow-up T2D incidence 13 years)
Cohort (follow-up T2D incidence 8 years) T2D prevalence Cross-sectional
Cross-sectional
T2D prevalence
Cross-sectional
T2D prevalence
Ecological study
Danish Diet, Cancer, and Health cohort (n = 51 818)
SALIA (n = 1775 women aged 54–55 years at baseline) Black Women’s Health Study (BWHS) (n = 3992 females) NHS (n = 74 412 females) and HPFS (n = 15 048 males)
Participants in five populationbased health surveys (n = 62 012) Patients attending two respiratory diseases clinics in Hamilton (n = 5228) and Toronto (n = 2406) with a median age of 60 years Diabetes screening study in 50–75-year olds (n = 8018)
Swiss Cohort Study on Air Pollution and Lung and Heart Disease in Adults (SAPALDIA) (n = 6392) County-level T2D Prevalence data based on a prevalence monthly telephone survey of the adult US population
Exposure duration and assessment a Long term; LUR and dispersion model
Long term; LUR models Long term; Kriging model Average exposure 12 months before T2D diagnosis; LUR models Long term; satellite derived Medium term; LUR model
Medium term; LUR model
Long term; dispersion model (PM10), dispersion + LUR model (NO2) Medium term; annual mean level obtained from monitoring stations
Pollutants (increments) and estimates
Refs
NO2 (IQR 4.9 mg/m3) All cases: HR: 1.00 (0.97–1.04) Confirmed cases: HR: 1.04 (1.01–1.07) Nonsmoker: HR: 1.12 (1.05–1.20) Physically active: HR: 1.10 (1.03–1.16) NO2 (IQR 15 mg/m3): HR: 1.42 (1.16–1.73) PM/soot (IQR 0.39 3 10 5 m): HR: 1.27 (1.09–1.48) PM2.5 (10 mg/m3): HR: 1.63 (0.78–3.44) NOx (IQR 23 mg/m3): HR: 1.25 (1.07–1.46) PM2.5 (IQR 4 mg/m3): HR: 1.00 (0.93–1.08) PM10 (IQR 7 mg/m3): HR: 1.04 (0.98–1.10)
[43]
[39]
[41] [42]
PM2.5 (10 mg/m3): HR: 1.11 (1.02–1.21)
[40]
NO2 (1.9 mg/m3): OR: 1.02 (0.98–1.049)
[44]
NO2 (third quartile 15.2–16.5 mg/m3): OR: [45] 1.25 (0.99–1.56) NO2 (fourth quartile 16.5–36.0 mg/m3): OR: 0.80 (0.63–1.02) Reference: first quartile 8.8–14.2 mg/m 3 PM10 (10 mg/m3): OR: 1.40 (1.17–1.67) [46] NO2 (10 mg/m3): OR: 1.19 (1.03–1.38)
% Increase in diabetes prevalence in 2004 [47] PM2.5 (10 mg/m3): b: 0.77 (0.39;1.25) in 2004 PM2.5 (10 mg/m3): b: 0.81 (0.48;1.07)
a
If no date for the exposure assessment was provided, medium-term exposure was assumed.
Is ambient air polluon exposure associated with diabetes development? Ambient air polluon exposure
General populaon ?
?
(A) Diabetes incidence
(B) Diabetes prevalence
(high impact) Andersen, 2012
NO2
Krämer, 2010
NO2, PM
Coogan, 2012
PM2.5 NOx
Pue, 2011 Chen, 2013
(+)
Brook, 2008
NO2
(+)
Dijkema, 2011
NO2
ο
(+) +
Eze, 2014
PM10, NO2
+
PM2.5, PM10
ο
Pearson, 2010
PM2.5
+
PM2.5
+
+
TRENDS in Endocrinology & Metabolism
Figure 2. Ambient air pollution exposure and its association with diabetes development. The studies investigate associations between ambient air pollution exposure and (A) diabetes incidence and (B) diabetes prevalence. The direction of the effect is denoted by for negative association, + for positive association, or o for no association; parentheses represent nonsignificant association or association restricted to subgroups. Most of the studies reported statistically significantly increased risk for incident and prevalent diabetes with increased exposure to ambient air pollution.
388
Review studies for diabetes incidence (Figure 2A) and prevalence (Figure 2B). Two cohort studies observed significant associations between long-term PM exposure and incident diabetes during the follow-up period in 1775 inhabitants of the highly industrialized Ruhr district in Germany and 62 012 adults living in Ontario, Canada [39,40]. Nonsignificant adverse effects were seen in 3992 African American women living in Los Angeles, USA [41]. An analysis combining data from the Nurses’ Health Study (NHS) (74 412 females) and the Health Professionals Follow-Up Study (HPFS) (15 048 males) in the USA did not yield strong evidence for an association between PM2.5 and PM10 exposure in the 12 months before T2D diagnosis and incident cases, although the study was sufficiently powered [42]. However, only recent air pollution exposure was considered in this analysis. Two of these previous studies additionally reported significant associations between NO2/NOx exposure and incident T2D [39,41]. In the 57 053 participants of the DCH cohort, no significant association between NO2 exposure and all incident diabetes cases was found but a borderline statistically significant association for confirmed diabetes cases was observed [43]. While some of the studies were conducted solely in women [39,41], Chen et al. [40] reported higher effect estimates in women compared with men. It is often hypothesized that the degree of exposure misclassification is lower in women compared with men as women tend to spend more time at home (especially at advanced ages) and air pollution exposures are generally assessed at the residential address, as was done in all studies mentioned here. Other effect modifications were also studied, revealing that air pollution effects on T2D were stronger in participants with pre-existing chronic obstructive pulmonary disease (COPD) [40], increased subclinical inflammation [39], or low education levels [38,39,43], those who undertook increased physical activity, and never-smokers [43]. Cross-sectional studies Cross-sectional studies can help assess associations between air pollution exposure and T2D (Figure 2B and Table 2). However, in these studies only prevalent diabetes cases can be analyzed. Therefore, effect estimates are more likely to be influenced by reverse causation and residual confounding. The first epidemiological evidence for a cross-sectional association was published in 2008 using data on 7634 patients attending two respiratory clinics in Hamilton and Toronto, Canada [44]. While NO2 exposure was positively associated with T2D diagnosis among women, no significant association was observed among men or in the total study population. Also, no association between NO2 exposure and diabetes prevalence was seen among 8018 individuals participating in a diabetes screening study in the semirural area of Westfriesland in The Netherlands [45]. By contrast, a recent study conducted in a Swiss cohort of 6392 participants observed a positive association between both NO2 and PM10 exposure in the 10 years preceding the survey and a higher number of prevalent T2D cases [46]. In an ecological US study, country-level diabetes prevalence
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
was greater in countries with higher annual mean PM25 concentrations [47]. A limitation of this study is the lack of individual-level data, as is the case for all ecological studies. These aforementioned studies have been partly analyzed in five recent meta-analyses [20–24] (Table 3). The meta-analysis by Balti et al. [22] estimated the combined effects of NO2 and NOx in three prospective cohort studies [39,41,43] and of PM2.5 in five prospective cohort studies [39–42]. A significantly increased risk for incident T2D was seen for both NO2/NOx and PM2.5. Park and Wang [24] meta-analyzed effects for only PM2.5 in relation to incident diabetes using data from four cohorts also included in Balti et al. [40–42]. Eze et al. [23] considered similar studies for PM2.5 but meta-analyzed pooled estimates for males and females separately, resulting in slightly different risk estimation. In Eze et al. [23], NO2 estimates derived from nonlinear estimates from Dijkema et al. [45] were included in addition. The meta-analysis by Wang et al. [20] included two published studies on prevalent diabetes [44,46] and an additional study on incident diabetes published as a congress abstract [48] as well as all of the abovementioned studies on incident diabetes [39–43]. The combined risk estimate for NO2/NOx was similar to that reported by Balti et al. For PM2.5, the combined effect estimate was higher than reported by Balti et al., although similar cohorts were included. This difference is due to the fact that Wang et al. extracted risk estimates from sensitivity analyses in which: (i) the analysis was restricted to the last 2 years of follow-up only [42]; and (ii) multi-pollutant models were considered [41]. Thus, the results of the meta-analysis by Wang et al. may overstate the adverse effects of air pollution in the general population as the estimates extracted from these sensitivity analyses were higher than those reported in the general population. Wang et al. [20] also reported a combined effect for PM10 using data from the four cohorts with available estimates for this pollutant [39,42,46]. In the meta-analysis by Janghorbani et al. [22], increased risks were observed for NO2, O3, PM10, and SO4 exposure. However, in this analysis diabetes prevalence and incidence were meta-analyzed together with diabetes mortality. Furthermore, the estimates were included in the meta-analysis without standardization, although the reported increments differed considerably. For example, effect estimates per 20mg/m3 increase in pollutant concentration, per 10-mg/m3 increase in pollutant concentration, and when comparing the upper and lower concentration quartiles were all metaanalyzed together. Following this strategy makes the resultant combined effect estimates difficult to interpret. Overall, cross-sectional studies appear to yield estimates with considerable heterogeneity whereas the results from prospective studies more consistently support an association between PM2.5, PM10, and NO2/NOx exposure and T2D development. All of the five existing meta-analyses concluded that there is evidence for an association between long-term exposure to air pollutants and T2D. However, it should be noted that the conducted studies have limitations, especially due to uncontrolled confounding by other factors such as lifestyle (diet, physical activity, socioeconomic status), other exposures (indoor, occupational, smoking), and comorbidities [23]. 389
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Is ambient air pollution exposure associated with predisposing conditions for diabetes? Studies on diabetes-related traits, which might be altered in subjects who have a higher diabetes risk, can provide further insight into the association between air pollution exposure and diabetes as well as the potential biological mechanisms underlying any association. The results of such studies are summarized in Figure 3 (for study details see Table 4). HbA1c is a marker of the average blood glucose levels in an individual during the previous 30–120 days. In 9102 newly diagnosed T2D patients in Germany, higher regional levels of PM10 were associated with higher HbA1c concentrations [49]. Adverse effects of several pollutants on HbA1c and fasting glucose concentrations were also observed in an elderly population in Taiwan [50]. In 363 women participating in the Study on the Influence of Air Pollution on Lung, Inflammation, and Aging (SALIA) cohort in Germany, a tendency for impaired glucose metabolism was observed in participants with higher air pollution exposure, although none of the associations remained statistically significant after adjusting for multiple testing [51]. Furthermore, no significant associations were observed for the 14 pro- and anti-inflammatory immune mediators investigated in this study. The authors
suggested that low statistical power due to their small sample size might explain their null findings. Insulin resistance, another potential precursor of T2D, can lead to elevated fasting glucose in combination with elevated insulin levels due to impaired insulin action. Among 560 elderly individuals living in Korea, short-term NO2 exposure was associated with elevated homeostatic model assessment of insulin resistance (HOMA-IR) in participants without a history of diabetes. For participants with diabetes, NO2, PM10, and O3 were all associated with increased HOMA-IR [52]. Furthermore, increased susceptibility was observed for participants carrying risk genotypes in the oxidative stress-related human glutathione Stransferase genes (GSTM1, GSTT1, and GSTP1). In an experimental study in the USA, 25 healthy adults living in rural Michigan had reduced insulin sensitivity after they were exposed to air pollution in an urban location for 4–5 h each day for five consecutive days [53]. Air pollution exposure effects on insulin resistance have also been investigated in healthy children and adolescents. Long-term NO2 and PM10 exposures were associated with increased HOMA-IR in 397 German children aged 10 years [54]. Additionally, short-term exposure PM10 and CO exposures led to elevated HOMA-IR in 374 Iranian children aged 10–18 years [55]. PM10, CO, and NO2 exposures were
Is ambient air polluon exposure associated with other predisposing condions for diabetes? Ambient air polluon exposure
General populaon ?
? (D) Gestaonal diabetes
(C) Diabetes related traits HbA1c Tamayo, 2014
PM10
+
Chuang, 2011
PM10, PM2.5, NO2, O3
+
HOMA-IR Kim, 2012
NO2
+
Brook, 2013
PM2.5
+
Thiering, 2013
NO2, PM10
+
Kelishadi, 2009
PM10, CO
+
NO2, NOx
(+)
Fleisch, 2014
PM2.5
o
Robledo, 2015
NOx SO2 O3
+
Malmqvist, 2013
NOx
(–) +
IGT Teichert, 2013
TRENDS in Endocrinology & Metabolism
Figure 3. Ambient air pollution exposure and its association with other predisposing conditions for diabetes. Overview of studies on ambient air pollution exposure in relation to (C) diabetes-related traits and (D) gestational diabetes. The direction of the effect is denoted by for negative association, + for positive association, or o for no association; parentheses represent nonsignificant association or association restricted to subgroups. While all reviewed studies report adverse effect of air pollutants on diabetes-related traits, there is less evidence for an association with gestational diabetes due to the heterogeneity of the results.
390
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 3. Evidence synthesis: summary of published meta-analyses on ambient air pollution exposure and T2D (X indicates that a study was included in the respective meta-analysis) Balti et al. 2014 Prospective NO2 and NOx Andersen, 2012 Kra¨mer, 2010 Coogan, 2012 Brook, 2008 Eze, 2014 Weinmayr, 2012 Dijkema, 2011 Combined estimate PM2.5 and PM10 Chen, 2013 Coogan, 2012 Kra¨mer, 2010 Puett, 2011 Eze, 2014 Brook, 2013 Combined estimate
Wang et al. 2014
X X X X X
1.13 (1.04–1.22)
1.16 (1.00–1.25)
X X X X
X X Xa X X X
Eze et al. 2015
X X X X
1.11 (1.07–1.16) PM2.5 X Xa Xa
X 1.11 (1.03–1.20)
Park and Wang 2014
Cross-sectional
X 1.39 (1.14–1.68)
PM2.5 X X
X 1.08 (1.00–1.17) PM2.5 X X
X Xa X
X
X
1.34 (1.22–1.47)
1.11 (1.03–1.19)
1.10 (1.02–1.18)
PM10
a
Estimates from sensitivity analyses in Puett et al., 2011 were used in the meta-analyses by Wang et al., which explains the high combined effect estimates observed for PM in Wang et al. For Coogan et al., 2012 effect estimates were extracted from the multi-pollutant model.
also associated with increased C-reactive protein and oxidized low-density lipoprotein in this study. Women who experience gestational diabetes are at higher risk for subsequent T2D [56]. To date, three studies have investigated associations between air pollution exposure and gestational diabetes. In 2093 pregnant women without previous type 1 diabetes or T2D living in the area of Boston, USA, PM2.5 exposure was associated with impaired glucose tolerance [57]. However, no association with gestational diabetes was observed. A large retrospective cohort in the USA that included 219 952 singleton deliveries observed higher relative risks for gestational diabetes in women with higher NOx and SO2 exposures during the 3 months preconception but a reduced risk for O3 exposure [58]. When considering only air pollution exposures that occurred during the first trimester, similar results were observed for NOx and SO2 but there was no association for O3. In a study from Taiwan, exposure to O3 led to more preterm births in women with gestational diabetes [59]. Finally, a Swedish study linked registry data from 81 110 pregnancies with individual NOx exposure and found a dose–response relationship [60]. These studies demonstrate that air pollutants might have the potential to affect diabetes development and may also be associated with gestational diabetes. However, additional studies are needed to draw firmer conclusions. Concluding remarks and future perspectives There appears to be evidence indicative of an association between long-term exposure to main air pollutants and the development of T2D and diabetes-related mortality in adults. The results from studies that have investigated precursors of diabetes in healthy individuals strengthen the support for a potential association. However, a firm conclusion regarding the association between ambient air pollution exposure and diabetes-related traits and gestational diabetes cannot be presently drawn due to the limited data available. For short(er)-term air pollution
exposures, the results are more heterogeneous and the evidence too weak to demonstrate a causal relationship. Most of the studies reviewed here were conducted in developed countries in North America and Europe. Studies in other parts of the world are needed, where air pollution concentrations are higher and the diabetes burden is greater, especially to determine potential dose–response relationships. Furthermore, an unanswered question exists regarding the latency period of air pollution exposure: does high air pollution exposure lead to priming effects that manifest their risk even after a certain time of low exposure? It also remains unclear whether exposure improvement – for example, by moving to a cleaner environment – could reduce the risk for T2D or intermediate phenotypes such as insulin resistance or glycemic control. The experimental study by Brook et al. [53] only studied effects after subacute air pollution exposure in individuals living in a rural environment. Exposure reduction effects among individuals living in polluted areas have not been assessed so far. Furthermore, air pollution exposure sources are often located indoors. Household indoor air pollution concentrations often exceed outdoor concentrations by a factor of two to five [61,62], and even occasionally up to a factor of 100. Air pollution concentrations in the work environment are also often higher than the mean concentration outdoors [63,64]. Air pollution concentrations indoors and in work environments have not yet been studied in relation to T2D. Another important gap is that the role of potential environmental confounders, such as noise and reduced greenness, that are correlated with air pollution have not yet been extensively evaluated [65,66]. In conclusion, the available evidence supporting adverse effects of air pollution on diabetes and the related high public health impact of such an association justify the need for further investigations. Future studies are also needed to establish how and to what extent air pollution control measures may reduce the global diabetes-related burden of disease. 391
Review
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
Table 4. Characteristics and results of cross-sectional and experimental studies on diabetes-related traits Study design
Outcome
Population
Exposure duration and assessment a Medium term; chemical transport model with interpolation, area level of 100 five-digit postcode areas
Cross-sectional
HbA1c
Newly diagnosed T2D patients 18 years in the DPV database (n = 9102)
Cross-sectional
HbA1c, fasting glucose
Participants in the Social Environment and Biomarkers of Aging study 54 years (n = 1023)
1 year; 1-year mean of monitoring stations at area level
Experimental
Insulin resistance
Short term; monitoring stations
Cross-sectional
Insulin resistance
Cross-sectional
Insulin resistance
Cross-sectional
Insulin resistance
Cross-sectional
Impaired glucose tolerance
Healthy adults living in rural Michigan were transported to an urban location for five consecutive days (n = 25) Participants in the Korean Elderly Environmental Panel (KEEP) 50 years (n = 560) 10-year-old participants of the birth cohorts GINIplus and LISAplus (n = 397) Population-based study of adolescents aged 10–18 years (n = 374) Participants in the SALIA cohort (n = 363 females)
Cross-sectional
Gestational diabetes
Project Viva cohort (n = 2093 pregnant women)
Cross-sectional
Gestational diabetes
Consortium on Safe Labor retrospective cohort in 19 hospitals (n = 219 952 women with singleton pregnancy)
Cross-sectional
Gestational diabetes
Singleton deliveries during 1999–2005 from the Swedish Medical Birth Registry (n = 81 110)
Pollutants (increments) and estimates
Refs
Increase in HbA1c PM10 (third quartile 18.1–21.1 mg/m3): b: 0.21 (0.09;0.32) PM10 (fourth quartile >21.1 mg/m3): b: 0.36 (0.22;0.49) Reference: first quartile <16.4 mg/m 3 For % in HbA1c PM10 (IQR 48 mg/m3): b: 1.40 (1.11;0.32) PM2.5 (IQR 20.4 mg/m3): b: 2.24 (1.47;3.00) NO2 (IQR 24 mg/m3): b: 1.08 (0.84;1.33) For mg/dl in fasting glucose PM10 (IQR 48 mg/m3): b: 22.9 (14.9;30.8) PM2.5 (IQR 20.4 mg/m3): b: 36.6 (19.2;53.9) NO2 (IQR 24 mg/m3): b: 17.0 (10.4;23.7) Increase in HOMA-IR PM2.5 (10 mg/m3): b: 0.7 (0.1;1.3)
[49]
[50]
[53]
Short term; daily concentrations measured by monitoring stations
PM10 (IQR 20.8 mg/m3): b: 0.14 ( 0.003;0.29) NO2 (IQR 10.8 mg/m3): b: 0.28 (0.13;0.42)
[52]
Long term; LUR model
% Increase in HOMA-IR NO2 (2SD 10.6 mg/m3): b: 17.0 (5.0;30.3) PM10 (2SD 6.0 mg/m3): b: 18.7 (2.9;36.9) PM2.5 (2SD 3.7 mg/m3): b: 17.0 ( 0.9;51.5) PM10 (unknown): b: 1.1 (0.5;1.7) NO2 (unknown): b: 0.8 (0.6;1.0)
[54]
NO2 (IQR 14.7 mg/m3): OR: 1.47 (1.05–2.05) PM2.5 (IQR unknown): OR: 1.15 (0.81–1.63) PM10 (IQR unknown): OR 1.20 (0.87–1.68)
[51]
Impaired glucose tolerance PM2.5 (IQR 2 mg/m3): OR: 1.64 (1.11–2.42) Gestational diabetes PM2.5 (IQR 2 mg/m3): OR 0.94 (0.67–1.34) PM2.5 (IQR 5.54 mg/m3): RR: 0.97 (0.93–1.02) PM10 (IQR 6.3 mg/m3): RR: 0.99 (0.96–1.02) NOx (IQR 15.1 mg/m3): RR: 1.09 (1.04–1.13) O3 (IQR: 6.1 mg/m3): RR: 0.93 (0.90–0.96)
[57]
Short term; daily concentrations measured by monitoring stations Long term; local monitoring stations or LUR model, backextrapolated Second trimester; spatiotemporal model and LUR model 91 days before last menstrual period, first trimester; Community Multi-scale Air Quality Model: average exposure level of the hospital First trimester; dispersion model
NOx (third quartile 14.2–22.6 mg/m3): OR: 1.52 (1.28–1.82) NOx (fourth quartile >22.7 mg/m3): OR: 1.69 (1.41–2.03) Reference: first quartile (2.5–8.9 mg/m3)
[55]
[58]
[60]
a
If no date for the exposure assessment was provided, medium-term exposure was assumed.
Acknowledgments The authors thank Dr Elaine Fuertes (Institute of Epidemiology I, Helmholtz Zentrum Mu¨nchen, Germany) for editorial assistance in the preparation of the manuscript.
References 1 van Donkelaar, A. et al. (2015) Use of satellite observations for longterm exposure assessment of global concentrations of fine particulate matter. Environ. Health Perspect. 123, 135–143 2 Turner, M.C. et al. (2011) Long-term ambient fine particulate matter air pollution and lung cancer in a large cohort of never-smokers. Am. J. Respir. Crit. Care Med. 184, 1374–1381 392
3 Chen, H. et al. (2013) Long-term exposure to traffic-related air pollution and cardiovascular mortality. Epidemiology 24, 35–43 4 Shah, A.S. et al. (2013) Global association of air pollution and heart failure: a systematic review and meta-analysis. Lancet 382, 1039–1048 5 Cesaroni, G. et al. (2014) Long term exposure to ambient air pollution and incidence of acute coronary events: prospective cohort study and meta-analysis in 11 European cohorts from the ESCAPE Project. BMJ 21, 348 6 Raaschou-Nielsen, O. et al. (2013) Air pollution and lung cancer incidence in 17 European cohorts: prospective analyses from the European Study of Cohorts for Air Pollution Effects (ESCAPE). Lancet Oncol. 14, 813–822
Review 7 Lim, S.S. et al. (2012) A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 380, 2224–2260 8 Gordon, S.B. et al. (2014) Respiratory risks from household air pollution in low and middle income countries. Lancet Respir. Med. 2, 823–860 9 Kampa, M. and Castanas, E. (2008) Human health effects of air pollution. Environ. Pollut. 151, 362–367 10 Brook, R.D. et al. (2010) Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the American Heart Association. Circulation 121, 2331–2378 11 WHO (2012) Global Status Report on Noncommunicable Diseases 2014, World Health Organization 12 Guariguata, L. et al. (2014) Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res. Clin. Pract. 103, 137–149 13 Finucane, M.M. et al. (2011) National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet 377, 557–567 14 Rajagopalan, S. and Brook, R.D. (2012) Air pollution and type 2 diabetes: mechanistic insights. Diabetes 61, 3037–3045 15 Ru¨ckerl, R. et al. (2014) Associations between ambient air pollution and blood markers of inflammation and coagulation/fibrinolysis in susceptible populations. Environ. Int. 70, 32–49 16 Liu, C. et al. (2013) Epidemiological and experimental links between air pollution and type 2 diabetes. Toxicol. Pathol. 41, 361–373 17 Bind, M.A. et al. (2012) Air pollution and markers of coagulation, inflammation, and endothelial function: associations and epigene– environment interactions in an elderly cohort. Epidemiology 23, 332–340 18 Carmona, J.J. et al. (2014) Short-term airborne particulate matter exposure alters the epigenetic landscape of human genes associated with the mitogen-activated protein kinase network: a cross-sectional study. Environ. Health 13, 94 19 Lockwood, A.H. (2002) Diabetes and air pollution. Diabetes Care 25, 1487–1488 20 Wang, B. et al. (2014) Effect of long-term exposure to air pollution on type 2 diabetes mellitus risk: a systemic review and meta-analysis of cohort studies. Eur. J. Endocrinol. 171, R173–R182 21 Balti, E.V. et al. (2014) Air pollution and risk of type 2 diabetes mellitus: a systematic review and meta-analysis. Diabetes Res. Clin. Pract. 106, 161–172 22 Janghorbani, M. et al. (2014) Systematic review and metaanalysis of air pollution exposure and risk of diabetes. Eur. J. Epidemiol. 29, 231– 242 23 Eze, I.C. et al. (2015) Association between ambient air pollution and diabetes mellitus in Europe and North America: systematic review and meta-analysis. Environ. Health Perspect. 123, 381–389 24 Park, S.K. and Wang, W. (2014) Ambient air pollution and type 2 diabetes: a systematic review of epidemiologic research. Curr. Environ. Health Rep. 1, 275–286 25 Bateson, T.F. and Schwartz, J. (2004) Who is sensitive to the effects of particulate air pollution on mortality? A case-crossover analysis of effect modifiers. Epidemiology 15, 143–149 26 Forastiere, F. et al. (2008) Particulate matter and daily mortality: a case-crossover analysis of individual effect modifiers. Epidemiology 19, 571–580 27 Zanobetti, A. et al. (2014) A national case-crossover analysis of the short-term effect of PM2.5 on hospitalizations and mortality in subjects with diabetes and neurological disorders. Environ. Health 13, 38 28 Goldberg, M.S. et al. (2006) Associations between ambient air pollution and daily mortality among persons with diabetes and cardiovascular disease. Environ. Res. 100, 255–267 29 Goldberg, M.S. et al. (2013) Associations between ambient air pollution and daily mortality among elderly persons in Montreal. Quebec. Sci. Total Environ. 463–464, 931–942 30 Zanobetti, A. and Schwartz, J. (2011) Ozone and survival in four cohorts with potentially predisposing diseases. Am. J. Respir. Crit. Care Med. 184, 836–841 31 Stafoggia, M. et al. (2010) Susceptibility factors to ozone-related mortality: a population-based case-crossover analysis. Am. J. Respir. Crit. Care Med. 182, 376–384
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
32 Chiusolo, M. et al. (2011) Short-term effects of nitrogen dioxide on mortality and susceptibility factors in 10 Italian cities: the EpiAir Study. Environ. Health Perspect. 119, 1233–1238 33 Ostro, B. et al. (2006) Fine particulate air pollution and mortality in nine California counties: results from CALFINE. Environ. Health Perspect. 114, 29–33 34 Samoli, E. et al. (2014) Which specific causes of death are associated with short term exposure to fine and coarse particles in Southern Europe? Results from the MED-PARTICLES project. Environ. Int. 67, 54–61 35 Kan, H. et al. (2004) he association of daily diabetes mortality and outdoor air pollution in Shanghai, China. J. Environ. Health 67, 21–26 36 Maynard, D. (2007) Mortality risk associated with short-term exposure to traffic particles and sulfates. Environ. Health Perspect. 115, 751–755 37 Brook, R.D. et al. (2013) Long-term fine particulate matter exposure and mortality from diabetes in Canada. Diabetes Care 36, 3313–3320 38 Raaschou-Nielsen, O. et al. (2013) Long-term exposure to trafficrelated air pollution and diabetes-associated mortality: a cohort study. Diabetologia 56, 36–46 39 Kra¨mer, U. et al. (2010) Traffic-related air pollution and incident type 2 diabetes: results from the SALIA cohort study. Environ. Health Perspect. 118, 1273–1279 40 Chen, H. et al. (2013) Risk of incident diabetes in relation to long-term exposure to fine particulate matter in Ontario, Canada. Environ. Health Perspect. 121, 804–810 41 Coogan, P.F. et al. (2012) Air pollution and incidence of hypertension and diabetes mellitus in black women living in Los Angeles. Circulation 125, 767–772 42 Puett, R.C. et al. (2011) Are particulate matter exposures associated with risk of type 2 diabetes? Environ. Health Perspect. 119, 384–389 43 Andersen, Z.J. et al. (2012) Diabetes incidence and long-term exposure to air pollution: a cohort study. Diabetes Care 35, 92–98 44 Brook, R.D. et al. (2008) The relationship between diabetes mellitus and traffic-related air pollution. J. Occup. Environ. Med. 50, 32–38 45 Dijkema, M.B. et al. (2011) Long-term exposure to traffic-related air pollution and type 2 diabetes prevalence in a cross-sectional screeningstudy in The Netherlands. Environ. Health 10, 76 46 Eze, I.C. et al. (2014) Long-term air pollution exposure and diabetes in a population-based Swiss cohort. Environ. Int. 70, 95–105 47 Pearson, J.F. et al. (2010) Association between fine particulate matter and diabetes prevalence in the US. Diabetes Care 33, 2196–2201 48 Weinmayr, G. et al. (2014) Industry and traffic related air pollution and diabetes type two incidence: results from a German cohort study. Eur. J. Epidemiol. 27, S12 49 Tamayo, T. et al. (2014) Is particle pollution in outdoor air associated with metabolic control in type 2 diabetes? PLoS ONE 9, e91639 50 Chuang, K.J. et al. (2011) Long-term air pollution exposure and risk factors for cardiovascular diseases among the elderly in Taiwan. Occup. Environ. Med. 68, 64–68 51 Teichert, T. et al. (2013) Association between traffic-related air pollution, subclinical inflammation and impaired glucose metabolism: results from the SALIA study. PLoS ONE 8, e83042 52 Kim, J.H. and Hong, Y.C. (2012) GSTM1, GSTT1, and GSTP1 polymorphisms and associations between air pollutants and markers of insulin resistance in elderly Koreans. Environ. Health Perspect. 120, 1378–1384 53 Brook, R.D. et al. (2013) Reduced metabolic insulin sensitivity following sub-acute exposures to low levels of ambient fine particulate matter air pollution. Sci. Total Environ. 448, 66–71 54 Thiering, E. et al. (2013) Long-term exposure to traffic-related air pollution and insulin resistance in children: results from the GINIplus and LISAplus birth cohorts. Diabetologia 56, 1696–1704 55 Kelishadi, R. et al. (2009) Lifestyle and environmental factors associated with inflammation, oxidative stress and insulin resistance in children. Atherosclerosis 203, 311–319 56 Kwak, S.H. et al. (2013) Clinical and genetic risk factors for type 2 diabetes at early or late postpartum after gestational diabetes mellitus. J. Clin. Endocrinol. Metab. 98, E744–E752 57 Fleisch, A.F. (2014) Air pollution exposure and abnormal glucose tolerance during pregnancy: the Project Viva cohort. Environ. Health Perspect. 122, 378–383
393
Review 58 Robledo, C.A. et al. (2015) Preconception and early pregnancy air pollution exposures and risk of gestational diabetes mellitus. Environ. Res. 137C, 316–322 59 Lin, Y.T. et al. (2015) Associations between ozone and preterm birth in women who develop gestational diabetes. Am. J. Epidemiol. 181, 280– 287 60 Malmqvist, E. et al. (2013) Gestational diabetes and preeclampsia in association with air pollution at levels below current air quality guidelines. Environ. Health Perspect. 121, 488–493 61 Bordada, J.C. et al. (2012) Exposure to airborne ultrafine particles from cooking in Portuguese homes. J. Air Waste Manag. Assoc. 62, 1116– 1126 62 Nadasena, S. et al. (2012) Indoor fine particle (PM2.5) pollution exposure due to secondhand smoke in selected public places of Sri Lanka. Am. J. Ind. Med. 55, 1129–1136
394
Trends in Endocrinology and Metabolism July 2015, Vol. 26, No. 7
63 Mitchell, D.C. et al. (2013) Particulate matter, endotoxin, and worker respiratory health on large Californian dairies. J. Occup. Environ. Med. 57, 79–87 64 Dioni, L. et al. (2011) Effects of short-term exposure to inhalable particulate matter on telomere length, telomerase expression, and telomerase methylation in steel workers. Environ. Health Perspect. 119, 622–627 65 Te´treault, L.F. et al. (2013) Cardiovascular health, traffic-related air pollution and noise: are associations mutually confounded? A systematic review. Int. J. Public Health 58, 649–666 66 Dadvand, P. et al. (2012) Surrounding greenness and exposure to air pollution during pregnancy: an analysis of personal monitoring data. Environ. Health Perspect. 120, 1286–1290 67 Ren, C. et al. (2010) Modifiers of short-term effects of ozone on mortality in eastern Massachusetts – a case-crossover analysis at individual level. Environ. Health 9, 3