Table S1. The distribution of IgE (kU/L), 8-OHdG (ng/mL), and 1-OHP levels (ng/mL) (n=453)

> LOD (%) / Minimum / Medium / Maximum
IgE / 97.47 / 1.90 / 113.00 / 2366.00
8-OHdG / 99.77 / 0.32 / 19.37 / 2644.13
1-OHP / 88.74 / 0.53 / 1.94 / 438.90

Abbreviations: LOD: limit of detection
Table S2 Baseline characteristics of the study population and the geometric means (S.D.) of urine 1-OHP levels in different subgroups (n=453)

Characteristics / GM (S.D.)
1-OHP levels (ng/mL) / p-value
Mother
Maternal age 34 years / 3.65(3.97) / 0.646
³ 34 years / 3.69(4.01)
Maternal education < College / 3.73(4.04) / 0.554
³ College / 3.61(3.79)
Maternal nationality Taiwan / 3.67(3.93) / 0.460
Foreign countries / 4.28(4.31)
Maternal history of atopy No / 3.22(3.73) / 0.269
Yes / 4.39(4.16)
Children
Sex Male / 3.72(4.00) / 0.340
Female / 3.45(3.54)
Birth weight < 2500 gm / 6.69(4.27) / 0.402
³ 2500 gm / 3.61(3.92)
Gestational age <37 weeks / 3.68(4.02) / 0.415
³37 weeks / 3.85(3.19)
Environmental factors
Breast feeding No / 4.04(3.90) / 0.742
Yesa / 3.65(3.97)
Number of older siblings
< 2 / 3.38(3.74) / 0.121
³ 2 / 5.41(4.81)
Current incensing at home No / 3.29(3.66) / 0.307
Yes / 4.07(4.17)
Current ETS exposure No / 3.39(3.77) / 0.179
Yes / 3.99(4.07)
Family income per year
600,000 NT dollars / 3.69(3.94) / 0.525
600,000-1,500,000 NT dollars / 3.50(4.00)
> 1,500,000 NT dollars / 4.01(3.92)

Abbreviations: ETS, environmental tobacco smoke; NT, New Taiwan dollars

aBreast feeding “Yes” is any breast feeding at all.
Figure S1. Steps of mediation causal analysis showing the effect of PAH exposure on asthma was mediated by oxidative stress.

Step 1 Show that the independent variable (X) is correlated with the mediator (M) (Model 1)

Step 2 Show that the independent variable (X) is correlated with the outcome (Y) (Model 2)

Step 3 Show that the mediator affects the outcome variable (Model 3)

Variables in our article are as follows:

Y: asthma (1 = No; 2=Yes)

X: Ln1-OHP ((Ln)-transformed urine 1-OHP levels )

M: Ln8-OHdG ((Ln)-transformed urine 8-OHdG levels )

Step 1 Mediator Model (Ln1OHP --> Ln8OHdG)

Step 2 Outcome Model (Ln1OHP --> Asthma)

Step 3 Mediation Analysis

The direct mediation effect was 65%.

35% risk of asthma in relation to Ln1-OHP was mediated by changes in 8-OHdG.

Hence, other pathways were likely to contribute to the mediation as well.

If was to be interpreted as a mediation proportion (95% CI), then 8-OHdG mediated 35% (16-0.61%) of the total effect of Ln1-OHP on asthma.

Mediation analysis needs some strict assumptions and has also been criticized. For the âb̂ estimator of the mediated effect, the model assumes that the residuals in Equations 2 and 3 are independent and that M and the residual in Equation 2 are independent (McDonald 1997). It is also assumed that there is not an XM interaction in Equation 3, although this can and should be routinely tested. The assumptions of a correctly specified model requires no misspecification of causal order (e.g., Y → M → X rather than X → M → Y), no misspecification of causal direction (e.g., there is reciprocal causation between the mediator and the dependent variable), no misspecification due to unmeasured variables that cause variables in the mediation analysis, and no misspecification due to imperfect measurement (McDonald 1997). A more realistic approach is to incorporate additional information from prior research, including randomized experimental studies, theory, and qualitative methods to bolster the tentative conclusion that a mediation relation exists. (MacKinnon et al., 2007)

Methods based on the observed regression approach to estimating mediation have been criticized based on causal analysis of the relations among variables. One of these critique addressed above is the equivalent model criticism. If X, M, and Y are measured simultaneously, there are other models that would explain the data equally well (e.g., X is the mediator of the M to Y relationship or M and Y both cause X), and in many situations it is not possible to distinguish these alternatives without more information (Spirtes et al. 1993). The case in which X represents random assignment to conditions improves causal interpretation of mediating variables because X precedes M and Y (Robins & Greenland 1992).

Reference

MacKinnon DP,Fairchild AJ,Fritz MS. Mediation analysis. Annu Rev Psychol.2007;58:593-614.

McDonald RP. Haldane’s lungs: a case study in path analysis. Multivar Behav Res 1997;32:1–38.

Spirtes, P.; Glymour, C.; Scheines, R. Causation, Prediction, and Search. New York: Springer-Verlag; 1993.

Robins JM, Greenland S. Identifiability and exchangeability for direct and indirect effects. Epidemiology 1992;3:143–55.


Figure S2. Steps of mediation causal analysis showing the effect of PAH exposure on IgE was not medicated by oxidative stress.

Step 1 Show that the independent variable (X) is correlated with the outcome (Y) (Model 1)

Step 2 Show that the independent variable (X) is correlated with the mediator (M) (Model 2)

Step 3 Show that the mediator affects the outcome variable (Model 3)

Variables in our article are as follows:

Y: IgE

X: Ln1-OHP ((Ln)-transformed urine 1-OHP levels )

M: Ln8OHdG ((Ln)-transformed urine 8-OHdG levels )

Step 1 IgE→Ln1-OHP

Model 1: IgE=225.86+59.48* Ln1-OHP

→ c is 59.48 and p-value <0.027.

Step 2 Ln8OHdG→Ln1-OHP

Model 2: Ln8OHdG = 3.918 + 0.327* Ln1-OHP

→ a is 0.327 and p-value0.001.

Step 3 IgE→Ln1-OHP & Ln8OhdG

Model 3: IgE =210.15+59.46*Ln1-OHP +2.71*Ln8OHdG

→b is 2.71 and p-value 0.858.

→ is 59.46 and p-value=0.033.

It’s a none mediation because the b is not significant.


Figure S3. Possible mechanisms

PAH exposure may increase IgE levels and the risk of asthma.

The effect of PAH exposure on IgE was not mediated by oxidative stress.

The effect of PAH exposure on asthma was partially mediated by oxidative stress.