Indoor air quality of 5,000 households and its determinants. Part B: Volatile organic compounds and inorganic gaseous pollutants in the Japan Environment and Children's study

Development and evaluation of machine learning models for predicting large-for-gestational-age newborns in women exposed to radiation prior to pregnancy

INTRODUCTION

The correlation between radiation exposure before pregnancy and abnormal birth weight has been previously proven. However, for large-for-gestational-age (LGA) babies in women exposed to radiation before becoming pregnant, there is no prediction model yet.

MATERIAL AND METHODS

The data were collected from the National Free Preconception Health Examination Project in China. A sum of 455 neonates (42 SGA births and 423 non-LGA births) were included. A training set (n = 319) and a test set (n = 136) were created from the dataset at random. To develop prediction models for LGA neonates, conventional logistic regression (LR) method and six machine learning methods were used in this study. Recursive feature elimination approach was performed by choosing 10 features which made a big contribution to the prediction models. And the Shapley Additive Explanation model was applied to interpret the most important characteristics that affected forecast outputs.

RESULTS

The random forest (RF) model had the highest average area under the receiver-operating-characteristic curve (AUC) for predicting LGA in the test set (0.843, 95% confidence interval [CI]: 0.714-0.974). Except for the logistic regression model (AUC: 0.603, 95%CI: 0.440-0.767), other models' AUCs displayed well. Thereinto, the RF algorithm's final prediction model using 10 characteristics achieved an average AUC of 0.821 (95% CI: 0.693-0.949).

CONCLUSION

The prediction model based on machine learning might be a promising tool for the prenatal prediction of LGA births in women with radiation exposure before pregnancy.

The impact of formaldehyde exposure on lung inflammatory disorders: Insights into asthma, bronchitis, and pulmonary fibrosis

Lung inflammatory disorders are a major global health burden, impacting millions of people and raising rates of morbidity and death across many demographic groups. An industrial chemical and common environmental contaminant, formaldehyde (FA) presents serious health concerns to the respiratory system, including the onset and aggravation of lung inflammatory disorders. Epidemiological studies have shown significant associations between FA exposure levels and the incidence and severity of several respiratory diseases. FA causes inflammation in the respiratory tract via immunological activation, oxidative stress, and airway remodelling, aggravating pre-existing pulmonary inflammation and compromising lung function. Additionally, FA functions as a respiratory sensitizer, causing allergic responses and hypersensitivity pneumonitis in sensitive people. Understanding the complicated processes behind formaldehyde-induced lung inflammation is critical for directing targeted strategies aimed at minimizing environmental exposures and alleviating the burden of formaldehyde-related lung illnesses on global respiratory health. This abstract explores the intricate relationship between FA exposure and lung inflammatory diseases, including asthma, bronchitis, allergic inflammation, lung injury and pulmonary fibrosis.

The association between clustering based on composition of volatile organic compound in indoor air and building-related symptoms

Volatile organic compounds (VOCs) have been suspected to cause building-related symptoms (BRSs). Although some studies investigated the association between BRSs and VOCs in indoor air, those results were inconsistent. This study investigated the contamination status of VOCs in the indoor air of 154 houses in Japan. Additionally, these samples were grouped by hierarchical clustering analysis based on the VOC composition, and the relationship between a VOC cluster and the BRSs was investigated. The median concentration of the sum of VOCs (ΣVOCs) was 140 μg m-3 (range: 18-3500 μg m-3). The levels of acetaldehyde in four samples and p-dichlorobenzene in one sample exceeded the guideline value. As a result of the hierarchical clustering analysis, the samples in this study were divided into six characteristic clusters based on the VOC composition. The ΣVOCs in cluster 1 were significantly lower than those in other clusters. In cluster 2, acyclic and aromatic hydrocarbons were dominant. Cluster 3 had a relatively high proportion of limonene. In cluster 4, the concentrations and composition ratios of α-pinene were higher than those of other clusters. In cluster 5, p-dichlorobenzene accounted for 42 %-72 % of the total VOCs. Cluster 6 had a relatively high proportion of decamethyl cyclopentasiloxane. This clustering likely depended on the construction of houses and lifestyles. As a result of logistic regression analysis, cluster 5 was associated with the cough symptoms of the BRSs. The results of the present study suggest that investigating the association between VOCs and BRSs is necessary to consider not only total concentrations such as TVOC and ΣVOCs but also VOC composition.

The prediction of in-hospital mortality in elderly patients with sepsis-associated acute kidney injury utilizing machine learning models

Background

Sepsis-associated acute kidney injury (SA-AKI) is a severe complication associated with poorer prognosis and increased mortality, particularly in elderly patients. Currently, there is a lack of accurate mortality risk prediction models for these patients in clinic.

Objectives

This study aimed to develop and validate machine learning models for predicting in-hospital mortality risk in elderly patients with SA-AKI.

Methods

Machine learning models were developed and validated using the public, high-quality Medical Information Mart for Intensive Care (MIMIC)-IV critically ill database. The recursive feature elimination (RFE) algorithm was employed for key feature selection. Eleven predictive models were compared, with the best one selected for further validation. Shapley Additive Explanations (SHAP) values were used for visualization and interpretation, making the machine learning models clinically interpretable.

Results

There were 16,154 patients with SA-AKI in the MIMIC-IV database, and 8426 SA-AKI patients were included in this study (median age: 77.0 years; female: 45%). 7728 patients excluded based on these criteria. They were randomly divided into a training cohort (5,934, 70%) and a validation cohort (2,492, 30%). Nine key features were selected by the RFE algorithm. The CatBoost model achieved the best performance, with an AUC of 0.844 in the training cohort and 0.804 in the validation cohort. SHAP values revealed that AKI stage, PaO2, and lactate were the top three most important features contributing to the CatBoost model.

Conclusion

We developed a model capable of predicting the risk of in-hospital mortality in elderly patients with SA-AKI.

Daily Exposure to Environmental Volatile Organic Compounds Triggers Oxidative Damage: Evidence from a Large-Scale Survey in China

Volatile organic compounds (VOCs) are ubiquitous environmental pollutants and have been implicated in adverse health outcomes. In this study, concentrations of 11 VOC metabolites (mVOCs) and three oxidative stress biomarkers (8-oxo-7,8-dihydro-2'-deoxyguanosine, 8-oxo-7,8-dihydro-guanosine, and dityrosine) were determined in 205 urine samples collected from 12 cities across mainland China. Urinary ∑11mVOC concentrations ranged from 498 to 1660 ng/mL, with a geometric mean (GM) value of 1070 ng/mL. The factorial analysis revealed that cooking, solvents, and vehicle emissions were the three primary sources of VOC exposure. A significant regional variation was clearly found in ∑11mVOC concentrations across four regions in China, with high urine VOC concentrations found in North and South China (GM: 1450 and 1340 ng/mL). The multiple linear regression model revealed that most mVOCs were significantly positively correlated with three oxidative stress markers (β range: 0.06-0.22). Mixture effect regression showed that isoprene, crotonaldehyde, acrolein, and benzene were the strongest contributors to oxidative stress. Approximately 80% of the participants have HQ values greater than 1.0 for 1,3-butadiene and benzene, suggesting that their exposure doses were close to potential adverse health effects. Our findings provide comprehensive information on human exposure and potential health risks of VOCs in China.

Association of air pollution exposure during pregnancy and early childhood with children's cognitive performance and behavior at age six

BACKGROUND

The impact of air pollution on neurodevelopment in children has attracted much attention in recent times. We aim to clarify the association between prenatal and postnatal air pollutant exposure and children's cognitive performance and behavior at age six.

METHODS

This study was conducted based on a birth cohort study in Japan. Children's intelligence quotient (IQ) was assessed using the Wechsler Intelligence Scale for Children and a score <85 was deemed as low intelligence. A score ≥60 on the Child Behavior Checklist indicated behavioral problems. Exposure to outdoor fine particulate matter (PM2.5) during pregnancy and early childhood was estimated using a spatiotemporal model, while indoor concentrations of air pollutants inside subjects' homes were measured for a week when the child was of ages 1.5 and 3. The associations of exposure to air pollution during pregnancy and after childbirth with cognitive performance and behavior were analyzed using logistic regression models.

RESULTS

The estimated exposure to outdoor PM2.5 during pregnancy and early childhood was not associated with decreased cognitive performance. However, exposure during the first trimester, 0-1 and 3-5 years of age was associated with children's externalizing problems (odds ratios (ORs) were 2.77 [95% confidence interval (CI): 1.05-7.29], 1.66 [95%CI: 1.05-2.62], and 1.80 [95%CI: 1.19-2.74] per interquartile range (IQR) increase, respectively). Exposure to indoor PM2.5 and coarse particles after childbirth was associated with lower full scale IQ (ORs were 1.46 [95%CI: 1.03-2.08] and 1.85 [95%CI: 1.12-3.07] per IQR increase, respectively). However, some inverse associations were also observed.

CONCLUSIONS

These results suggest associations between prenatal and postnatal exposure to outdoor air pollution and behavioral problems, and between indoor air pollution after childbirth and cognitive performance at age six. However, the effects of exposure to outdoor PM2.5 during pregnancy on cognitive performance were not observed.

A hybrid model for estimating the number concentration of ultrafine particles based on machine learning algorithms in central Taiwan

Modeling is a cost-effective measure to estimate ultrafine particle (UFP) levels. Previous UFP estimates generally relied on land-use regression with insufficient temporal resolution. We carried out in-situ measurements for UFP in central Taiwan and developed a model incorporating satellite-based measurements, meteorological variables, and land-use data to estimate daily UFP levels at a 1-km resolution. Two sampling campaigns were conducted for measuring hourly UFP concentrations at six sites between 2008-2010 and 2017-2021, respectively, using scanning mobility particle sizers. Three machine learning algorithms, namely random forest, eXtreme gradient boosting (XGBoost), and deep neural network, were used to develop UFP estimation models. The performances were evaluated with a 10-fold cross-validation, temporal, and spatial validation. A total of 1,022 effective sampling days were conducted. The XGBoost model had the best performance with a training coefficient of determination (R²) of 0.99 [normalized root mean square error (nRMSE): 6.52%] and a cross-validation R² of 0.78 (nRMSE: 31.0%). The ten most important variables were surface pressure, distance to the nearest road, temperature, calendar year, day of the year, NO₂, meridional wind, the total length of roads, PM_2.5, and zonal wind. The UFP levels were elevated along the main roads across different seasons, suggesting that traffic emission is an important contributor to UFP. This hybrid model outperformed prior land use regression models and thus can provide more accurate estimates of UFP for epidemiological studies.

A review of critical residential buildings parameters and activities when investigating indoor air quality and pollutants

Indoor air in residential dwellings can contain a variety of chemicals, sometimes present at concentrations or in combinations which can have a negative impact on human health. Indoor Air Quality (IAQ) surveys are often required to characterize human exposure or to investigate IAQ concerns and complaints. Such surveys should include sufficient contextual information to elucidate sources, pathways, and the magnitude of exposures. The aim of this review was to investigate and describe the parameters that affect IAQ in residential dwellings: building location, layout, and ventilation, finishing materials, occupant activities, and occupant demography. About 180 peer-reviewed articles, published from 01/2013 to 09/2021 (plus some important earlier publications), were reviewed. The importance of the building parameters largely depends on the study objectives and whether the focus is on a specific pollutant or to assess health risk. When considering classical pollutants such as particulate matter (PM) or volatile organic compounds (VOCs), the building parameters can have a significant impact on IAQ, and detailed information of these parameters needs to be reported in each study. Research gaps and suggestions for the future studies together with recommendation of where measurements should be done are also provided.

The relationship of residential formaldehyde pollution in 11 Chinese cities to schoolchildren pneumonia prevalence in actual living condition

Residential formaldehyde pollution is one of the leading residential harmful pollutants with a large production and consumption globally and remains much uncertainty in Chinese families with huge health burden for children worldwide. A multi-center observation study from 11 cities was conducted to investigate residential formaldehyde pollution levels measured by phenol reagent spectrophotometry. Data on household characteristics and schoolchildren's health were collected by questionnaire. The median concentration of residential formaldehyde was 0.025 (0.002-0.281) mg/m³ among 11 cities with the total exceeding standard rate of 7.40% according to the reference value of 0.10 mg/m³ (1-h average). Residential formaldehyde pollution in warm season, bedrooms and northern cities was more serious than that in cold season, living rooms and southern cities, respectively. The potential influencing factors of residential formaldehyde included household characteristics (distance from a traffic road, building history, residence duration, window glass layers, decoration and furniture) and use of air conditioner. The positive regulation effect of temperature on residential formaldehyde was explored with the approximately turning-point temperature of 28.9 °C for peak concentration. Long-term exposure to residential formaldehyde of low concentrations (0.010-0.090 mg/m³) would increase the prevalence of childhood pneumonia and a more stringent criteria value for residential formaldehyde should be discussed cautiously.

Predictive models for small-for-gestational-age births in women exposed to pesticides before pregnancy based on multiple machine learning algorithms

Background

The association between prenatal pesticide exposures and a higher incidence of small-for-gestational-age (SGA) births has been reported. No prediction model has been developed for SGA neonates in pregnant women exposed to pesticides prior to pregnancy.

Methods

A retrospective cohort study was conducted using information from the National Free Preconception Health Examination Project between 2010 and 2012. A development set (n = 606) and a validation set (n = 151) of the dataset were split at random. Traditional logistic regression (LR) method and six machine learning classifiers were used to develop prediction models for SGA neonates. The Shapley Additive Explanation (SHAP) model was applied to determine the most influential variables that contributed to the outcome of the prediction.

Results

757 neonates in total were analyzed. SGA occurred in 12.9% (n = 98) of cases overall. With an area under the receiver-operating-characteristic curve (AUC) of 0.855 [95% confidence interval (CI): 0.752–0.959], the model based on category boosting (CatBoost) algorithm obtained the best performance in the validation set. With the exception of the LR model (AUC: 0.691, 95% CI: 0.554–0.828), all models had good AUCs. Using recursive feature elimination (RFE) approach to perform the feature selection, we included 15 variables in the final model based on CatBoost classifier, achieving the AUC of 0.811 (95% CI: 0.675–0.947).

Conclusions

Machine learning algorithms can develop satisfactory tools for SGA prediction in mothers exposed to pesticides prior to pregnancy, which might become a tool to predict SGA neonates in the high-risk population.

Baseline Complete Blood Count and Chemistry Panel Profile from the Japan Environment and Children's Study (JECS)

BACKGROUND

The Japan Environment and Children's Study (JECS) is a nationwide birth cohort study of environmental factors affecting children's health and development. We summarize the complete blood count and chemistry panel statistical data from pregnant women enrolled in JECS.

METHODS

Statistical data of up to 58,056 mother's sample in their first (gestational age ≤ 13 weeks) and second trimester (22-27 weeks) were calculated.

RESULTS

Mean (SD) values in the first trimester were: white blood cell count, 7866 (1906)/μL; red blood cell count, 417 (33) 10⁴/μL; hemoglobin, 12.4 (1.0) g/dL; hematocrit, 37.1 (2.6)%; mean corpuscular volume, 89.2 (4.7) fL; mean corpuscular hemoglobin, 29.8 (1.9) pg; mean corpuscular hemoglobin concentration, 33.3 (0.9)%; platelet count, 24.8 (5.2) 10⁴/μL; HbA1c, 5.26 (0.26)%; total cholesterol, 181 (28) mg/dL; low density lipoprotein cholesterol, 95 (23) mg/dL; high density lipoprotein cholesterol, 73 (13) mg/dL; free cholesterol, 41 (7) mg/dL; triglycerides, 109 (47) mg/dL; total protein, 6.9 (0.4) g/dL; and albumin, 4.1 (0.2) g/dL. Mean values (SD) in the second trimester were: total cholesterol, 246 (38) mg/dL; free cholesterol, 61 (9) mg/dL; triglycerides, 183 (70) mg/dL; total protein, 6.5 (0.4) g/dL; and albumin, 3.6 (0.2) g/dL.

CONCLUSIONS

These data will be useful for future JECS studies.

Paving the Way for a Green Transition in the Design of Sensors and Biosensors for the Detection of Volatile Organic Compounds (VOCs)

The efficient and selective detection of volatile organic compounds (VOCs) provides key information for various purposes ranging from the toxicological analysis of indoor/outdoor environments to the diagnosis of diseases or to the investigation of biological processes. In the last decade, different sensors and biosensors providing reliable, rapid, and economic responses in the detection of VOCs have been successfully conceived and applied in numerous practical cases; however, the global necessity of a sustainable development, has driven the design of devices for the detection of VOCs to greener methods. In this review, the most recent and innovative VOC sensors and biosensors with sustainable features are presented. The sensors are grouped into three of the main industrial sectors of daily life, including environmental analysis, highly important for toxicity issues, food packaging tools, especially aimed at avoiding the spoilage of meat and fish, and the diagnosis of diseases, crucial for the early detection of relevant pathological conditions such as cancer and diabetes. The research outcomes presented in the review underly the necessity of preparing sensors with higher efficiency, lower detection limits, improved selectivity, and enhanced sustainable characteristics to fully address the sustainable manufacturing of VOC sensors and biosensors.

Early life exposure to indoor air pollutants and the risk of neurodevelopmental delays: The Japan Environment and Children's Study

Air pollution has been associated with childhood neurodevelopment. However, the role of indoor air pollution, especially volatile organic compounds (VOCs), on childhood neurodevelopment has been poorly explored to date. We investigated the association between indoor air pollutants and childhood neurodevelopment in 5,017 randomly selected children from the Japan Environment and Children's Study. When the participants reached 1.5 and 3 years of age, they were followed up with home visits and neurodevelopmental tests using the Ages and Stages Questionnaire (ASQ). At both ages, we collected indoor air samples for 1 week and measured 13 indoor air pollutants: particulate matter with an aerodynamic diameter of ≤2.5 μm, ozone, nitrogen dioxide, sulfur dioxide, and nine VOCs. The associations between air pollutants and ASQ scores were estimated using linear mixed effects models and weighted quantile sum regressions (WQS) at each age separately. Stratified analysis by sex was conducted. Exposure to m,p-xylene at the age of 3 was associated with lower communication, fine motor, and overall ASQ scores (coefficients: -0.18 [99% confidence intervals (CI): -0.35, -0.02], -0.23 [99 %CI: -0.43, -0.03], and - 0.72 [99 %CI: -1.41, -0.04] per 1 µg/m³ increase, respectively). Exposure to o-xylene at the age of 3 was associated with lower communication, gross motor, fine motor, and overall ASQ scores (coefficients: -0.48 [99 %CI: -0.90, -0.07], -0.45 [99 %CI: -0.78, -0.13], -0.65 [99 %CI: -1.14, -0.16], and -2.15 [99 %CI: -3.83, -0.47] per 1 µg/m³ increase, respectively). The WQS index was associated with lower gross motor ASQ scores at the age of 3 (coefficient: -0.27 [95 %CI: -0.51, -0.03] for one-unit WQS index increases), which was attributed to benzene (33.96%), toluene (26.02%), o-xylene (13.62%), and ethylbenzene (9.83%). Stratified analysis showed similar results. Although further investigations are required, our results suggest an association of neurodevelopmental delays with indoor low-level exposure to m,p-xylene and o-xylene in early life.

Urinary Metabolites of Organophosphate Pesticides among Pregnant Women Participating in the Japan Environment and Children's Study (JECS)

Organophosphate pesticides (OPPs) exhibit neurodevelopmental toxicity. To evaluate the effect of prenatal exposure to OPPs in the Japan Environment and Children's Study, a nationally representative birth cohort study, 4575 maternal urine samples were analysed for six OPP metabolites, i.e., dialkylphosphates (DAPs). This study aimed to investigate predictors of urinary DAPs using machine learning approaches and to assess the cumulative risk based on relative potency factors among Japanese pregnant women. The median creatinine-normalised urinary concentrations (interquartile ranges) of dimethylphosphate, dimethylthiophosphate and diethylphosphate, which had a detection rate of 50% or higher, were 3.53 (1.91-6.78), 4.09 (1.66-10.8) and 3.28 (1.88-5.98) µg/g-creatinine, respectively. Possible predictors of urinary DAP concentrations were the month of urine sampling, consumption of apple and maternal body mass index. When fenitrothion was used as an index chemical for cumulative risk assessment, 0.36% of participants exceeded the lower 95% confidence limit of the benchmark dose₁₀.