Machine learning-based health environmental-clinical risk scores in European children.

Journal

Communications medicine

ISSN: 2730-664X

Titre abrégé: Commun Med (Lond)

Pays: England

ID NLM: 9918250414506676

Informations de publication

Date de publication:
23 May 2024

Historique:

received: 31 03 2023

accepted: 26 04 2024

medline: 24 5 2024

pubmed: 24 5 2024

entrez: 23 5 2024

Statut: epublish

Résumé

Early life environmental stressors play an important role in the development of multiple chronic disorders. Previous studies that used environmental risk scores (ERS) to assess the cumulative impact of environmental exposures on health are limited by the diversity of exposures included, especially for early life determinants. We used machine learning methods to build early life exposome risk scores for three health outcomes using environmental, molecular, and clinical data. In this study, we analyzed data from 1622 mother-child pairs from the HELIX European birth cohorts, using over 300 environmental, 100 child peripheral, and 18 mother-child clinical markers to compute environmental-clinical risk scores (ECRS) for child behavioral difficulties, metabolic syndrome, and lung function. ECRS were computed using LASSO, Random Forest and XGBoost. XGBoost ECRS were selected to extract local feature contributions using Shapley values and derive feature importance and interactions. ECRS captured 13%, 50% and 4% of the variance in mental, cardiometabolic, and respiratory health, respectively. We observed no significant differences in predictive performances between the above-mentioned methods.The most important predictive features were maternal stress, noise, and lifestyle exposures for mental health; proteome (mainly IL1B) and metabolome features for cardiometabolic health; child BMI and urine metabolites for respiratory health. Besides their usefulness for epidemiological research, our risk scores show great potential to capture holistic individual level non-hereditary risk associations that can inform practitioners about actionable factors of high-risk children. As in the post-genetic era personalized prevention medicine will focus more and more on modifiable factors, we believe that such integrative approaches will be instrumental in shaping future healthcare paradigms. Growing up in different environments can greatly affect children’s health later in life. This research looked at how living in cities, being exposed to chemicals, and other experiences before birth and during childhood, work together to influence children’s mental, cardiovascular and respiratory health. We used advanced computer programs to help us understand these effects and estimate health risk scores. These scores are simple numerical measures that help us quantify the likelihood of children developing health issues based on their environmental exposures. Using those scores, the study identified key factors impacting children’s health, in particular psycho-social, perceived environmental and prenatal pollutant exposures for mental health. It also revealed complex patterns and interactions between environmental factors. The results highlighted the potential of such risk scores to support the identification of actionable factors in high-risk children, informing tailored prevention measures in healthcare.

Sections du résumé

BACKGROUND BACKGROUND

METHODS METHODS

In this study, we analyzed data from 1622 mother-child pairs from the HELIX European birth cohorts, using over 300 environmental, 100 child peripheral, and 18 mother-child clinical markers to compute environmental-clinical risk scores (ECRS) for child behavioral difficulties, metabolic syndrome, and lung function. ECRS were computed using LASSO, Random Forest and XGBoost. XGBoost ECRS were selected to extract local feature contributions using Shapley values and derive feature importance and interactions.

RESULTS RESULTS

ECRS captured 13%, 50% and 4% of the variance in mental, cardiometabolic, and respiratory health, respectively. We observed no significant differences in predictive performances between the above-mentioned methods.The most important predictive features were maternal stress, noise, and lifestyle exposures for mental health; proteome (mainly IL1B) and metabolome features for cardiometabolic health; child BMI and urine metabolites for respiratory health.

CONCLUSIONS CONCLUSIONS

Besides their usefulness for epidemiological research, our risk scores show great potential to capture holistic individual level non-hereditary risk associations that can inform practitioners about actionable factors of high-risk children. As in the post-genetic era personalized prevention medicine will focus more and more on modifiable factors, we believe that such integrative approaches will be instrumental in shaping future healthcare paradigms.

Growing up in different environments can greatly affect children’s health later in life. This research looked at how living in cities, being exposed to chemicals, and other experiences before birth and during childhood, work together to influence children’s mental, cardiovascular and respiratory health. We used advanced computer programs to help us understand these effects and estimate health risk scores. These scores are simple numerical measures that help us quantify the likelihood of children developing health issues based on their environmental exposures. Using those scores, the study identified key factors impacting children’s health, in particular psycho-social, perceived environmental and prenatal pollutant exposures for mental health. It also revealed complex patterns and interactions between environmental factors. The results highlighted the potential of such risk scores to support the identification of actionable factors in high-risk children, informing tailored prevention measures in healthcare.

Autres résumés

Type: plain-language-summary (eng)

Identifiants

DOI: 10.1038/s43856-024-00513-y PMID: 38783062

pubmed: 38783062

doi: 10.1038/s43856-024-00513-y

pii: 10.1038/s43856-024-00513-y

doi:

Types de publication

Journal Article

Langues

eng

Pagination

Informations de copyright

Références

Koppe, J. G. et al. Exposure to multiple environmental agents and their effect. Acta Paediatr. 95, 106–113 (2006).

doi: 10.1080/08035320600886646

Rauh, V. A. & Margolis, A. E. Research review: environmental exposures, neurodevelopment, and child mental health – new paradigms for the study of brain and behavioral effects. J. Child Psychol. Psychiatry 57, 775–793 (2016).

pubmed: 26987761 pmcid: 4914412 doi: 10.1111/jcpp.12537

Pryce, C. R. et al. Long-term effects of early-life environmental manipulations in rodents and primates: potential animal models in depression research. Neurosci. Biobehav. Rev. 29, 649–674 (2005).

pubmed: 15925698 doi: 10.1016/j.neubiorev.2005.03.011

Needleman, H. L., Schell, A., Bellinger, D., Leviton, A. & Allred, E. N. The long-term effects of exposure to low doses of lead in childhood. N. Engl. J. Med. 322, 83–88 (1990).

pubmed: 2294437 doi: 10.1056/NEJM199001113220203

Weihrauch-Blüher, S., Schwarz, P. & Klusmann, J.-H. Childhood obesity: increased risk for cardiometabolic disease and cancer in adulthood. Metabolism 92, 147–152 (2019).

pubmed: 30529454 doi: 10.1016/j.metabol.2018.12.001

Maitre, L. et al. Early-life environmental exposure determinants of child behavior in Europe: a longitudinal, population-based study. Environ. Int. 153, 106523 (2021).

pubmed: 33773142 pmcid: 8140407 doi: 10.1016/j.envint.2021.106523

Wild, C. P. Complementing the genome with an ‘exposome’: the outstanding challenge of environmental exposure measurement in molecular epidemiology. Cancer Epidemiol. Biomark. Prev. Publ. Am. Assoc. Cancer Res. Cosponsored Am. Soc. Prev. Oncol 14, 1847–1850 (2005).

doi: 10.1158/1055-9965.EPI-05-0456

Jaffee, S. R. & Price, T. S. Genotype–environment correlations: implications for determining the relationship between environmental exposures and psychiatric illness. Psychiatry 7, 496–499 (2008).

pubmed: 20622930 doi: 10.1016/j.mppsy.2008.10.002

Johns, D. O. et al. Practical advancement of multipollutant scientific and risk assessment approaches for ambient air pollution. Environ. Health Perspect. 120, 1238–1242 (2012).

pubmed: 22645280 pmcid: 3440129 doi: 10.1289/ehp.1204939

D’Agostino, R. B. et al. General cardiovascular risk profile for use in primary care: the framingham heart study. Circulation 117, 743–753 (2008).

pubmed: 18212285 doi: 10.1161/CIRCULATIONAHA.107.699579

Khera, A. V. et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat. Genet. 50, 1219–1224 (2018).

pubmed: 30104762 pmcid: 6128408 doi: 10.1038/s41588-018-0183-z

Park, S. K., Tao, Y., Meeker, J. D., Harlow, S. D. & Mukherjee, B. Environmental risk score as a new tool to examine multi-pollutants in epidemiologic research: an example from the NHANES study using serum lipid levels. PLOS ONE 9, e98632 (2014).

pubmed: 24901996 pmcid: 4047033 doi: 10.1371/journal.pone.0098632

Murray, G. K. et al. Could polygenic risk scores be useful in psychiatry?: a review. JAMA Psychiatry 78, 210–219 (2021).

pubmed: 33052393 doi: 10.1001/jamapsychiatry.2020.3042

Wray, N. R. et al. From basic science to clinical application of polygenic risk scores: a primer. JAMA Psychiatry 78, 101–109 (2021).

pubmed: 32997097 doi: 10.1001/jamapsychiatry.2020.3049

Pries, L.-K., Erzin, G., Rutten, B. P. F., van Os, J. & Guloksuz, S. Estimating aggregate environmental risk score in psychiatry: the exposome score for schizophrenia. Front. Psychiatry. 12, 671334 (2021).

Gao, P. & Snyder, M. Exposome-wide association study for metabolic syndrome. Front. Genet. 12, 783930 (2021).

pubmed: 34950191 pmcid: 8688998 doi: 10.3389/fgene.2021.783930

Le Magueresse-Battistoni, B., Vidal, H. & Naville, D. Environmental pollutants and metabolic disorders: the multi-exposure scenario of life. Front. Endocrinol. 9, 582 (2018).

Vassos, E. et al. The Maudsley environmental risk score for psychosis. Psychol. Med. 50, 1–8 (2019).

Padmanabhan, J. L., Shah, J. L., Tandon, N. & Keshavan, M. S. The ‘polyenviromic risk score’: aggregating environmental risk factors predicts conversion to psychosis in familial high-risk subjects. Schizophr. Res. 181, 17–22 (2017).

pubmed: 28029515 doi: 10.1016/j.schres.2016.10.014

Maitre, L. et al. Human early life exposome (HELIX) study: a european population-based exposome cohort. BMJ Open 8, e021311 (2018).

pubmed: 30206078 pmcid: 6144482 doi: 10.1136/bmjopen-2017-021311

Vrijheid, M. et al. The human early-life exposome (HELIX): project rationale and design. Environ. Health Perspect. 122, 535–544 (2014).

pubmed: 24610234 pmcid: 4048258 doi: 10.1289/ehp.1307204

Wright, J. et al. Cohort profile: the born in bradford multi-ethnic family cohort study. Int. J. Epidemiol. 42, 978–991 (2013).

pubmed: 23064411 doi: 10.1093/ije/dys112

Heude, B. et al. Cohort profile: the EDEN mother-child cohort on the prenatal and early postnatal determinants of child health and development. Int. J. Epidemiol. 45, 353–363 (2016).

pubmed: 26283636 doi: 10.1093/ije/dyv151

Guxens, M. et al. Cohort profile: the INMA—INfancia y medio ambiente—(Environment and childhood) project. Int. J. Epidemiol. 41, 930–940 (2012).

pubmed: 21471022 doi: 10.1093/ije/dyr054

Grazuleviciene, R. et al. Surrounding greenness, proximity to city parks and pregnancy outcomes in kaunas cohort study. Int. J. Hyg. Environ. Health 218, 358–365 (2015).

pubmed: 25757723 pmcid: 4390161 doi: 10.1016/j.ijheh.2015.02.004

Magnus, P. et al. Cohort profile update: the Norwegian mother and child cohort study (MoBa). Int. J. Epidemiol. 45, 382–388 (2016).

pubmed: 27063603 doi: 10.1093/ije/dyw029

Paltiel, L. et al. The biobank of the Norwegian mother and child cohort study – present status. Nor. Epidemiol. 24, 29–35 (2014).

Chatzi, L. et al. Cohort profile: the mother-child cohort in crete, greece (Rhea study). Int. J. Epidemiol. 46, 1392–1393k (2017).

pubmed: 29040580 doi: 10.1093/ije/dyx084

Constantinou, M. P. et al. Changes in general and specific psychopathology factors over a pychosocial intervention. J. Am. Acad. Child Adolesc. Psychiatry 58, 776–786 (2019).

pubmed: 30768397 doi: 10.1016/j.jaac.2018.11.011

Haltigan, J. D. et al. “P” and “DP:” examining symptom-level bifactor models of psychopathology and dysregulation in clinically referred children and adolescents. J. Am. Acad. Child Adolesc. Psychiatry 57, 384–396 (2018).

pubmed: 29859554 doi: 10.1016/j.jaac.2018.03.010

Caspi, A. et al. Longitudinal assessment of mental health disorders and comorbidities across 4 decades among participants in the dunedin birth cohort study. JAMA Netw. Open 3, e203221 (2020).

pubmed: 32315069 pmcid: 7175086 doi: 10.1001/jamanetworkopen.2020.3221

Cervin, M. et al. The p factor consistently predicts long-term psychiatric and functional outcomes in anxiety-disordered youth. J. Am. Acad. Child Adolesc. Psychiatry 60, 902–912.e5 (2021).

pubmed: 32950650 doi: 10.1016/j.jaac.2020.08.440

Rijlaarsdam, J. et al. Genome-wide DNA methylation patterns associated with general psychopathology in children. J. Psychiatr. Res. 140, 214–220 (2021).

pubmed: 34118639 pmcid: 8578013 doi: 10.1016/j.jpsychires.2021.05.029

Rosseel, Y. lavaan: An R package for structural equation modeling. J. Stat. Softw. 48, 1–36 (2012).

doi: 10.18637/jss.v048.i02

Achenbach, T. M. Integrative Guide for the 1991 CBCL/4-18, YSR, and TRF Profiles. (Univ Vermont/Dept Psychiatry, 1991).

Stratakis, N. et al. Association of fish consumption and mercury exposure during pregnancy with metabolic health and inflammatory biomarkers in children. JAMA Netw. Open 3, e201007 (2020).

pubmed: 32176304 pmcid: 7076335 doi: 10.1001/jamanetworkopen.2020.1007

Agier, L. et al. Early-life exposome and lung function in children in Europe: an analysis of data from the longitudinal, population-based HELIX cohort. Lancet Planet. Health 3, e81–e92 (2019).

pubmed: 30737192 doi: 10.1016/S2542-5196(19)30010-5

Quanjer, P. H. et al. Multi-ethnic reference values for spirometry for the 3-95-yr age range: the global lung function 2012 equations. Eur. Respir. J. 40, 1324–1343 (2012).

pubmed: 22743675 pmcid: 3786581 doi: 10.1183/09031936.00080312

Robinson, O. et al. The urban exposome during pregnancy and its socioeconomic determinants. Environ. Health Perspect. 126, 077005 (2018).

pubmed: 30024382 pmcid: 6108870 doi: 10.1289/EHP2862

Lau, C.-H. E. et al. Determinants of the urinary and serum metabolome in children from six European populations. BMC Med. 16, 202 (2018).

pubmed: 30404627 pmcid: 6223046 doi: 10.1186/s12916-018-1190-8

Dieterle, F., Ross, A., Schlotterbeck, G. & Senn, H. Probabilistic quotient normalization as robust method to account for dilution of complex biological mixtures. Application in 1H NMR metabonomics. Anal. Chem. 78, 4281–4290 (2006).

pubmed: 16808434 doi: 10.1021/ac051632c

Cohen, S. Perceived stress in a probability sample of the United States. In The Social Psychology of Health. 31–67 (Sage Publications, Inc, Thousand Oaks, CA, US, 1988).

Sweet, L. H. N-Back Paradigm. In Encyclopedia of Clinical Neuropsychology (eds. Kreutzer, J. S., DeLuca, J. & Caplan, B.) 1718–1719 (Springer, New York, NY, 2011).

Maitre, L. et al. Multi-omics signatures of the human early life exposome. Nat. Commun. 13, 7024 (2022).

pubmed: 36411288 pmcid: 9678903 doi: 10.1038/s41467-022-34422-2

Stekhoven, D. J. & Bühlmann, P. MissForest—non-parametric missing value imputation for mixed-type data. Bioinformatics 28, 112–118 (2012).

pubmed: 22039212 doi: 10.1093/bioinformatics/btr597

Bergstra, J., Bardenet, R., Bengio, Y. & Kégl, B. Algorithms for hyper-parameter optimization. In Advances In Neural Information Processing Systems. 24, 2546–2554 (Curran Associates, Inc., 2011).

Chen, T. & Guestrin, C. XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 785–794 (Association for Computing Machinery, New York, NY, USA, 2016).

Yang, P., Hwa Yang, Y., Zhou, B. B. & Zomaya, Y. A. A review of ensemble methods in bioinformatics. Curr. Bioinforma. 5, 296–308 (2010).

doi: 10.2174/157489310794072508

Lundberg, S. M. et al. From local explanations to global understanding with explainable AI for trees. Nat. Mach. Intell 2, 56–67 (2020).

pubmed: 32607472 pmcid: 7326367 doi: 10.1038/s42256-019-0138-9

Lundberg, S. M., Erion, G. G. & Lee, S.-I. Consistent individualized feature attribution for tree ensembles. arXiv http://arxiv.org/abs/1802.03888 (2019).

Grinsztajn, L., Oyallon, E. & Varoquaux, G. Why do tree-based models still outperform deep learning on tabular data? Adv. Neural. Inf. Process. Syst. 35, 507–520 (2022).

Tamayo-Uria, I. et al. The early-life exposome: description and patterns in six European countries. Environ. Int. 123, 189–200 (2019).

pubmed: 30530161 doi: 10.1016/j.envint.2018.11.067

Lundberg, S. & Lee, S.-I. A Unified approach to interpreting model predictions. Adv. Neural. Inf. Process. Syst. 30 (2017)

Hart, S. Shapley value. in Game Theory (eds. Eatwell, J., Milgate, M. & Newman, P.) 210–216 (Palgrave Macmillan UK, London, 1989).

Balagopal, P. B. et al. Nontraditional risk factors and biomarkers for cardiovascular disease: mechanistic, research, and clinical considerations for youth. Circulation 123, 2749–2769 (2011).

pubmed: 21555711 doi: 10.1161/CIR.0b013e31821c7c64

He, Y. et al. Comparisons of polyexposure, polygenic, and clinical risk scores in risk prediction of type 2 diabetes. Diabetes Care 44, 935–943 (2021).

pubmed: 33563654 pmcid: 7985424 doi: 10.2337/dc20-2049

Hastie, T., Tibshirani, R. & Friedman, J. Overview of supervised learning. In The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Vol. 2 (eds. Hastie, T., Tibshirani, R. & Friedman, J.) 9–41 (Springer, New York, NY, 2009).

Hastie, T., Tibshirani, R. & Friedman, J. Model assessment and selection. In The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Vol. 2 (eds. Hastie, T., Tibshirani, R. & Friedman, J.) 219–259 (Springer, New York, NY, 2009).

Farewell, C. V., Melnick, E. & Leiferman, J. Maternal mental health and early childhood development: Exploring critical periods and unique sources of support. Infant Ment. Health J. 42, 603–615 (2021).

pubmed: 33998003 doi: 10.1002/imhj.21925

Wang, F. & Veugelers, P. J. Self-esteem and cognitive development in the era of the childhood obesity epidemic. Obes. Rev. 9, 615–623 (2008).

pubmed: 18647242 doi: 10.1111/j.1467-789X.2008.00507.x

Lim, J. et al. Negative impact of noise and noise sensitivity on mental health in childhood. Noise Health 20, 199–211 (2018).

pubmed: 30516173 pmcid: 6301087

Esser, N., Legrand-Poels, S., Piette, J., Scheen, A. J. & Paquot, N. Inflammation as a link between obesity, metabolic syndrome and type 2 diabetes. Diabetes Res. Clin. Pract. 105, 141–150 (2014).

pubmed: 24798950 doi: 10.1016/j.diabres.2014.04.006

Wilkins, J. T. et al. Spectrum of apolipoprotein AI and apolipoprotein aII proteoforms and their associations with indices of cardiometabolic health: the CARDIA study. J. Am. Heart Assoc. 10, e019890 (2021).

pubmed: 34472376 pmcid: 8649248 doi: 10.1161/JAHA.120.019890

Tsai, J.-P. The association of serum leptin levels with metabolic diseases. Tzu-Chi Med. J. 29, 192–196 (2017).

pmcid: 5740690 doi: 10.4103/tcmj.tcmj_123_17

Sun, S. et al. Metabolic syndrome and its components are associated with altered amino acid profile in Chinese han population. Front. Endocrinol. 12, 795044 (2022).

doi: 10.3389/fendo.2021.795044

Ding, Y., Wang, S. & Lu, J. Unlocking the potential: amino acids’ role in predicting and exploring therapeutic avenues for type 2 diabetes mellitus. Metabolites 13, 1017 (2023).

pubmed: 37755297 pmcid: 10535527 doi: 10.3390/metabo13091017

Novgorodtseva, T. P. et al. Composition of fatty acids in plasma and erythrocytes and eicosanoids level in patients with metabolic syndrome. Lipids Health Dis 10, 82 (2011).

pubmed: 21595891 pmcid: 3116500 doi: 10.1186/1476-511X-10-82

Sun, Y. et al. BMI is associated with FEV1 decline in chronic obstructive pulmonary disease: a meta-analysis of clinical trials. Respir. Res. 20, 236 (2019).

pubmed: 31665000 pmcid: 6819522 doi: 10.1186/s12931-019-1209-5

Köchli, S. et al. Lung function, obesity and physical fitness in young children: the EXAMIN YOUTH study. Respir. Med. 159, 105813 (2019).

pubmed: 31731085 doi: 10.1016/j.rmed.2019.105813

Agier, L. et al. A systematic comparison of linear regression–based statistical methods to assess exposome-health associations. Environ. Health Perspect. 124, 1848–1856 (2016).

pubmed: 27219331 pmcid: 5132632 doi: 10.1289/EHP172

Vrijheid, M. et al. Advancing tools for human early lifecourse exposome research and translation (ATHLETE). Environ. Epidemiol 5, e166 (2021).

Neufcourt, L. et al. Assessing how social exposures are integrated in exposome research: a scoping review. Environ. Health Perspect. 130, 116001 (2022).

pubmed: 36350665 pmcid: 9645433 doi: 10.1289/EHP11015

Gaye, A. et al. DataSHIELD: taking the analysis to the data, not the data to the analysis. Int. J. Epidemiol. 43, 1929–1944 (2014).

pubmed: 25261970 pmcid: 4276062 doi: 10.1093/ije/dyu188

Guimbaud, J.-B. ML based health ECRS in European children - figure source data. https://doi.org/10.6084/m9.figshare.25625109 .

Guimbaud, J.-B. ML based ECRS for European children - python code. https://doi.org/10.5281/zenodo.10519296 .