Introduction

Recent evidence from epidemiological studies has shown that air pollution is associated with adverse health effects such as an increase in cardiovascular and pulmonary disease [1–5]. In particular, increased risk of hospitalization for cardiovascular and pulmonary diseases has also been reported to be associated with particulate air pollution [6, 7]. However, most epidemiological studies have used outdoor central monitoring data as the exposure estimates, and evidence supporting relevant biological mechanisms of how ambient particulate matter causes adverse health effects is insufficient.

Although the pathophysiological mechanisms linking inhalation of air pollution to acute exacerbation of cardiovascular or pulmonary disease are not completely understood in humans, inflammation has been reported to be induced by particulate air pollution in several human-exposure studies, but it is uncertain whether oxidative stress is mediated by particle-induced inflammation [8, 9]. Although results from cell-culture and animal experiments suggest that oxidative stress is involved in its pathogenesis, evidence of the relationship between human exposure to ambient particulate matter and oxidative stress is still lacking [10, 11]. This lack of clear understanding arises in part from the difficulty of measuring human exposure to ambient concentrations of pollutants and evaluating the resulting health effects. Monitoring personal exposure instead of using outdoor central monitoring data and utilizing relevant biomarkers instead of disease or mortality record could be one of the more precise methods for characterizing individual exposure and early effects, which are expected to explain their mechanisms.

Because people spend most of their time indoors, using outdoor environmental data to represent a subject’s exposure may not be an accurate method of evaluating the effects of air pollution exposure [12]. In addition, different exposures are expected individually according to considerable spatial variations, various indoor sources and composition of ambient particulate pollutants, and different personal activity levels [13, 14]. Therefore, measuring personal exposure and its biological effects on an individual basis is an important issue in air pollution epidemiological studies.

Biomarkers can be used to relate exposure to biological effects such as oxidative stress, which is thought to be involved in the mechanisms of adverse health effects of particulate pollutants. We measured urinary malondialdehyde (MDA) concentrations in spot urine samples as a relevant biomarker representing oxidative stress, considering times between exposure and its biological effects. As for human biomarkers of oxidative stress, MDA, one of the aldehyde products of lipid peroxidation, is known to be an end-product of free radical reactions and elevated MDA levels are found in many diseases including cardiovascular disease and cancer [15, 16].

As for human exposure and oxidative damage, some studies regarding outdoor central monitoring data as exposure estimates in young subjects found a significant association between particulate air pollution and oxidative stress, but others found no association [17–19]. Recently, as measurement of personal exposure and issue of monitoring devices have progressed, researchers have shown associations between individual exposure to particulate pollutants and biomarkers of oxidative damage [20, 21]. Further, because populations susceptible to the effects of particulate matter could be identified by personal monitoring, it helps to determine target populations for air pollution control [22].

The primary objective of this study was to examine the link between exposure to fine particulate matter and oxidative stress, with the hypothesis that not all groups of people respond to particulate pollutant exposure in the same way. In particular, young children and the elderly were recruited for this study with the assumption that they have different susceptibility to oxidative stress induced by particulate air pollution.

Methods

Participants

A total of 38 elderly volunteers at two day-care facilities and 51 children in one kindergarten class in Seoul participated in this study. General information about sociodemographic factors, past medical history, and family history were acquired by means of interviews, and anthropometric measurements were obtained during the first visit. We also obtained smoking history (no smoking/past smoking/current smoking) and alcohol consumption (no drinking/less than 2–3 times per month/1–2 times per week/3–4 times per week/almost daily) from the elderly participants. The ethical committee of the Seoul National University Hospital approved this study, and informed consent was obtained from each participant.

PM2.5 sampling and analysis

We measured personal exposure to PM2.5 of participating subjects at two day-care facilities for elderly and at a kindergarten for children. To account for temporal variation, the measurements were repeated three times in August, October, and December 2005 for the elderly participants and in October and December 2006 and February 2007 for the children.

The particles were sampled with a PM2.5 cyclon (Personal Environmental Monitor, SKC, USA) containing a pump (Mine Safety Appliances, USA), and a battery for 12 h operation. The air flow rate (2 l/min) of the pump was adjusted before each sampling. The PM2.5 cyclon was attached with a pin between the chest and head, and the pump was placed in a small bag that the subjects carried during measurement.

The suspended fine particles were sampled over 6 h for each subject. The 37 mm, 2.0 μm Teflon filters (PTFE, USA) were weighed before and after sampling using a microbalance and the concentration of ambient particulate matter was calculated considering air flow and measurement time.

Measurement of urinary MDA

For each participant, 30 ml urine was collected at each visit between 4:00 and 5:00 pm; these samples were kept frozen at −70°C until analysis. Urinary MDA was determined by measurement of MDA-thiobarbituric acid (TBA) adducts. Aliquots of the upper layers of centrifuged urine samples were mixed with phosphoric acid and TBA reagent in an iced methanol glass tube, and the mixtures were boiled for 60 min. After the mixture had been boiled it was cooled in ice water for 5 min and centrifuged after addition of methanol. The absorbance was measured by high-performance liquid chromatography (HPLC), using a system equipped with an SP930D solvent-delivery pump and a UV730D absorbance detector (Youngin, Korea). Compounds were separated on a Nova-Pak C18 column (150 × 3.9 nm) with 50 mM KH2PO4 (pH 6.8)–methanol 58:42 (v/v) as mobile phase. Detection was at 532 nm.

Statistical analysis

Associations between exposure to PM2.5 and urinary MDA concentrations were assessed. A linear mixed-effects model was used to estimate the effects of pollutants on urinary MDA, controlling for age, sex, BMI, passive smoking, and measurement session. For the elderly, day-care facility site, alcohol consumption, cigarette smoking, and medical history (heart disease, hypertension, and bronchial asthma) were also included in the model. Because the distribution of urinary MDA concentrations was skewed, log-transformed data were used for these measurements in the linear-mixed models. Age, sex, height, weight, alcohol consumption, and smoking exposure were treated as fixed effects, whereas each subject was treated as a random effect in the models. Statistical analyses were conducted with SAS (version 9.1). All tests of statistical significance were two-sided (α-error 0.05).

Results

General characteristics of the study participants are shown in Table 1. Mean subject age was 5.9 ± 0.2 for the children and 78.6 ± 6.0 for the elderly. In this study population 72.5% of the preschoolers and 42.1% of the elderly were male; 39.5% of the elderly subjects were current smokers. In addition, 42.1% of the elderly reported the presence of hypertension. The proportion of subjects that had the heart disease and bronchial asthma were 7.9 and 5.3% (not shown in Table 1).

Table 1 General characteristics of preschoolers and the elderly
Full size table

Personal exposure to PM2.5 and concentrations of urinary MDA are shown in Table 2. The mean personal PM2.5 was 80.5 ± 29.9 μg/m3 in the children and 20.7 ± 12.7 μg/m3 in the elderly. However, the mean urinary MDA concentrations were not significantly different between two groups, with means of 3.6 ± 1.9 μmol/g cr for the children and 4.0 ± 1.6 μmol/g cr for the elderly.

Table 2 Mean concentrations of personal PM2.5 and urinary MDA for the preschoolers in October and December 2006 and February 2007, and for the elderly in August, October, and December 2005
Full size table

The relationship between personal exposure to PM2.5 and urinary MDA concentrations is shown in Fig. 1. The horizontal axis is the level of personal exposure to PM2.5 assessed over 6 h and the vertical axis is the log-transformed urinary concentration of MDA adjusted by urinary creatinine (cr) level from spot urine collected in three sessions from 51 preschoolers and 38 elderly subjects.

Fig. 1
figure 1

Relationship between personal exposure to PM2.5 assessed over 6 h and urinary MDA concentrations measured three times for 51 preschool children (a) and 38 elderly subjects (b)

Full size image

The mixed-effects model of log MDA concentration for personal PM2.5 exposure showed a statistically significant increase of 0.3% log MDA concentration per 1 μg/m3 increase of personal PM2.5 exposure in the elderly. This study showed that estimates of change for urinary MDA concentrations attributable to personal PM2.5 exposure were quite different for children and the elderly. In children, the mixed-effects model of log MDA concentration for personal PM2.5 exposure showed no association between them, controlling for age, sex, BMI, passive smoking, and measurement session.

Discussion

This study found a significant association between fine particulate pollution and urinary MDA concentration in the elderly, but not in children, demonstrating that the elderly are more susceptible than young children to oxidative stress induced by ambient fine particulates.

We used personal exposure to PM2.5 as exposure variable. The mean personal PM2.5 was 80.5 ± 29.9 μg/m3 for the children and 20.7 ± 12.7 μg/m3 for the elderly subjects. Air monitoring data for ambient air pollutants in Seoul showed that annual mean values for PM10 were 58 and 60 μg/m3 in 2005 and 2006 [23]. Ambient PM2.5 concentrations were not monitored regularly in Seoul, but a report showed that 24-h means of outdoor and indoor PM2.5 were 56.0 and 43.2 μg/m3 [24].

Personal concentrations of PM2.5 may reflect the level of indoor PM2.5 rather than outdoor levels [14, 25]. However, despite the higher PM2.5 level in children, there is little difference in the oxidative stress concentrations between the elderly and the preschoolers. Urinary MDA concentrations in the elderly (4.0 μmol/g cr) and in children (3.6 μmol/g cr) in this study were comparable with the levels of MDA in asthmatic children under stable conditions (3.0 μmol/g cr) and during asthma attacks (4.4 μmol/g cr) [26].

Mechanisms by which the particulate matter induces oxidative stress have been investigated in cell and animal experiments. Soluble transition metals and organic molecules adsorbed on to the surface of particles are capable of generating reactive oxygen intermediates in lung cells and activating redox-sensitive transcription factors to enhance inflammatory reactions [27]. Moreover, soluble constituents of PM2.5 that can cross the pulmonary epithelium into the circulation may directly affect the cardiovascular system via imbalance of the autonomic nervous system, systemic oxidative stress, and inflammatory response [28].

The production of reactive oxygen species and subsequent biomolecular damage can be repaired or removed by defense systems in normal metabolism. Therefore, the balance between production and degradation of reactive oxygen metabolites determines oxidative stress and health effects [29, 30]. However, the functional capacity of antioxidant defense may be reduced with aging; aging is, therefore, a contributing factor for induction of oxidative stress [31, 32]. This mechanism was supported by the results of our study, which showed that exposure to particulate matter significantly produced metabolites of oxidative stress in the aged group, regarded as more vulnerable to oxidative stress, than in the young children group.

Biological markers are useful in explaining pathogenic mechanisms and recognizing early effects before development of disease. Therefore, through personal monitoring of exposure to particulate matter and use of biomarkers, it is possible to characterize individual exposure and explore the mechanisms leading to relevant biological effects [33]. Sørensen et al. [34] measured personal exposure to PM2.5 and 8-oxodG in the blood and urine of students in Copenhagen, and found that personal PM2.5 exposure was associated with 8-oxodG concentrations in lymphocyte DNA, but not in urine.

MDA used as biomarker of oxidative stress in this study is known to be associated with many diseases including cardiovascular disease, cancer, Alzheimer’s disease, liver disease, and adverse pregnancy outcomes [35–40]. There are numerous factors that are known to affect the formation of lipid peroxides. The level of MDA may be associated with age, body mass index and dietary factors. The consumption of some vegetables may have an inverse association with MDA concentration [41, 42]. On the other hand, alcohol and meat consumption and smoking behavior may increase the level [42]. In statistical analysis, we considered these factors and determined the final model including alcohol consumption, BMI, smoking behavior, passive smoking, and medical history (heart disease, hypertension, and bronchial asthma).

Elderly people spend most of their time indoors and personal exposure to PM2.5 does not correlate strongly with outdoor central site PM2.5 concentrations [14]. Sources of indoor ambient particles are not only from outdoor contributions, but also from indoor sources such as smoking, cooking, combustion, indoor resuspension of dusts by human activities, and transformation processes [43, 44]. To reduce errors in the measurement and bias of the exposure-effect relationship, we repeatedly measured personal exposure instead of using outdoor or indoor area monitoring data as an exposure estimate.

In this study, personal exposure to PM2.5 was four times higher for the children than for the elderly. Personal exposure to PM2.5 has been reported to be associated with activities such as walking or cycling and proximity to microenvironments [45]. Therefore, assuming the indoor and outdoor environments were not so different between two subgroups, but the elderly at day-care facilities are relatively inactive compared with the preschoolers in kindergarten, the reason for the differences in personal exposure level to PM2.5 may be resuspension of particulate matter and more particle collection in the filter because of the greater physical activity of the preschool children.

We found that the elderly are at a higher risk of oxidative stress associated with fine particulate pollutant exposure compared with children. Ambient concentrations of particles may be sufficient to cause oxidative stress in elderly subjects with compromised defense mechanisms against the attack of oxygen-containing free radicals, whereas young children will be able to tolerate higher exposure levels. However, it is unclear whether elderly people are more vulnerable to the effects of all components in the air pollution mixture. When studying associations between particulate matter and oxidative stress other considerations must be included, for example concomitant personal exposure to other pollutants, e.g. ozone, nitrous oxides, polycyclic aromatic hydrocarbons, heavy metals, and volatile organic chemicals.

In conclusion, the association between personal exposure to fine particulates and oxidative stress markers suggests that oxidative stress may be involved in the disease processes induced by particulate air pollution. This result explains the link between toxicological mechanisms in in-vitro experiments on particulate exposure and epidemiological evidence of particulate air pollution. This finding also suggests the need for air pollution-management policies to focus on susceptible populations, particularly the elderly.