Machine learning model to determine associations between metabolic syndrome and lactation

In a recent study published in Scientific Reports, researchers developed and validated machine learning models to investigate the relationship of lactation with metabolic syndrome risk, comparing its clinical importance to other known risk factors for the condition.

Study: Machine learning analysis for the association between breast feeding and metabolic syndrome in women. Image Credit: evso/Shutterstock.comStudy: Machine learning analysis for the association between breast feeding and metabolic syndrome in women. Image Credit: evso/Shutterstock.com

Background

Metabolic syndrome, which includes hypertension, dyslipidemia, insulin resistance, and central obesity, is associated with type 2 diabetes and cardiovascular disease. Researchers aim to lower this risk, especially during pregnancy-related events like delivery and nursing.

While nursing protects against pregnancy-induced metabolic alterations, several studies find no link. Understanding this relationship is critical for devising preventative methods.

About the study

In the present study, researchers examined the relationship between obstetric features such as metabolic syndrome and lactation and the occurrence of cardiovascular disease (CVD) among Asian women.

The team used artificial intelligence to develop a metabolic syndrome prediction model considering 86 variables, such as general obstetric features, demographics, medical history, dietary choices, lifestyle habits, and socioeconomic aspects.

The study included 30,204 female Korean National Health and Nutrition Examination Survey 2010-2109 (KNHANES) participants aged ≥20 years.

Metabolic syndrome was the dependent study variable, and the 86 variables of the independent type comprised demographic and socioeconomic factors and medical and obstetric data, including cardiovascular disease and lactation duration. The team excluded individuals with incomplete metabolic syndrome or cardiovascular disease data.

The team used questionnaires to measure sociodemographic data such as age at recruitment, gender, body mass index (BMI), family income, residence, educational level, economic activities, marital status, and professions.

The surveys also revealed general obstetric factors such as parity, gravidity, lactation, abortion history, menarche age, and menstrual status.  
The researchers conducted interviews to determine the prevalence rates of diseases such as hypertension, myocardial infarction, angina, stroke, osteoarthritis, rheumatoid arthritis, pulmonary tuberculosis, asthma, thyroid disorders, major depressive disorders, kidney failure, hepatitis B, hepatitis C, liver cirrhosis, cancers, and atopic dermatitis.

The surveys also included questions on family history of hyperlipidemia, hypertension, diabetes mellitus, ischemic heart disease, and stroke.

The team used the European Quality of Life-5 Dimensions (EQ-5D) scale to measure quality of life, and the nutrition survey determined daily intakes of calories, carbohydrates, protein, fat, salt, water, calcium, phosphorus, iron, vitamin C, and potassium. CVD diagnosis requires the presence of hypertension, angina, or myocardial infarction.

The researchers predicted metabolic syndrome risk using several algorithms, including artificial neural networks, decision trees, naïve Bayes, logistic regression, support vector machines, and random forest classifiers. They split 70% and 30% of data for model training and validation, respectively.

They used the accuracy and area under the curve (AUC) curve values to validate the model and random forest relevance to investigate the primary metabolic syndrome predictors.

Results

The study examined data from 30,204 patients (mean age, 51 years) with a 28% prevalence of metabolic syndrome. Random forest classifiers showed the highest AUC, with 91% accuracy for all participants, 88% for those diagnosed with cardiovascular disease, and 83% for those not diagnosed.

The primary metabolic syndrome estimators were BMI, antihypertensive medication use, hypertension, CVD, age at enrolment, leukocyte count, low-density-type lipoprotein-cholesterol (LDL) levels, menstrual status, lipid-lowering agent use, erythrocyte count, total cholesterol, subjective-type body image, education level, daily fat consumption, hematocrit levels, and lactation duration.

The relevance rankings of numerous main predictors changed dramatically in subgroup analysis, specifically between individuals with or without cardiovascular disease.

For example, hypertension drugs and diagnosis estimators attained the second and third positions overall but declined to the 23rd rank or below in both subgroups.

The team rated lactation duration 16th as an estimator for all individuals, somewhat higher at the 14th position for individuals without cardiovascular disease, and significantly lower at the 26th position for CVD patients.

The adjusted odds ratio (aOR) of 1.0 indicated that lactation duration was related to a lower metabolic syndrome risk. Extending breastfeeding duration by one month and one year reduced metabolic syndrome risk by 0.2% and 2.4%, respectively.

The impact of lactation duration on the risk of metabolic syndrome seems minimal in one month but becomes considerable after one or two years.

The OR was non-significant at the 5.0% level but provided helpful information for machine learning models. Logistic regression findings would augment the relevance of random forest variables.

Conclusion

Overall, the study findings showed that breastfeeding duration with body mass index, hypertension, cardiovascular illness, and age is a primary estimator of metabolic syndrome in women. Pregnancy induces metabolic alterations that enhance insulin resistance and blood cholesterol levels.

Breastfeeding speeds up the recovery of postpartum metabolic alterations in mothers and provides long-term benefits for maternal glucose levels, lipid metabolism, and obesity.

Women without cardiovascular disease scored better for age, breastfeeding duration, and gravidity. Nutrient consumption, particularly fat intake, was strongly associated with metabolic syndrome.

Journal reference:
Pooja Toshniwal Paharia

Written by

Pooja Toshniwal Paharia

Pooja Toshniwal Paharia is an oral and maxillofacial physician and radiologist based in Pune, India. Her academic background is in Oral Medicine and Radiology. She has extensive experience in research and evidence-based clinical-radiological diagnosis and management of oral lesions and conditions and associated maxillofacial disorders.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Toshniwal Paharia, Pooja Toshniwal Paharia. (2024, February 22). Machine learning model to determine associations between metabolic syndrome and lactation. News-Medical. Retrieved on October 31, 2024 from https://www.news-medical.net/news/20240222/Machine-learning-model-to-determine-associations-between-metabolic-syndrome-and-lactation.aspx.

  • MLA

    Toshniwal Paharia, Pooja Toshniwal Paharia. "Machine learning model to determine associations between metabolic syndrome and lactation". News-Medical. 31 October 2024. <https://www.news-medical.net/news/20240222/Machine-learning-model-to-determine-associations-between-metabolic-syndrome-and-lactation.aspx>.

  • Chicago

    Toshniwal Paharia, Pooja Toshniwal Paharia. "Machine learning model to determine associations between metabolic syndrome and lactation". News-Medical. https://www.news-medical.net/news/20240222/Machine-learning-model-to-determine-associations-between-metabolic-syndrome-and-lactation.aspx. (accessed October 31, 2024).

  • Harvard

    Toshniwal Paharia, Pooja Toshniwal Paharia. 2024. Machine learning model to determine associations between metabolic syndrome and lactation. News-Medical, viewed 31 October 2024, https://www.news-medical.net/news/20240222/Machine-learning-model-to-determine-associations-between-metabolic-syndrome-and-lactation.aspx.

Comments

The opinions expressed here are the views of the writer and do not necessarily reflect the views and opinions of News Medical.
Post a new comment
Post

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.

You might also like...
Machine learning reveals sleep quality and anxiety as major predictors of depression