Data-driven endocrine–metabolic phenotypes in young women with polycystic ovary syndrome and associations with cardiometabolic risk markers
4 affiliations
- 1Department of Applied Computer Science, Faculty of Computer Science and Telecommunications, Wroclaw University of Science and Technology, Wrocław, Poland
- 2Department of Medical Anthropology, Faculty of Medical Sciences, Medical University of Silesia, Katowice, 40-752 Poland
- 3Department of Endocrinological Gynecology, Faculty of Medical Sciences, Medical University of Silesia, Katowice, 40-752 Poland
- 4Department of Toxicology, Faculty of Pharmacy, Wroclaw Medical University, Wrocław, Poland
Abstract
Polycystic ovary syndrome (PCOS) is a heterogeneous endocrine disorder associated with reproductive dysfunction and increased cardiometabolic risk. Conventional diagnostic phenotypes may not adequately capture endocrine-metabolic variability, particularly in young women. We aimed to identify data-driven endocrine–metabolic phenotypes in young women with PCOS and assess their associations with selected cardiometabolic risk markers. This cross-sectional study included 1032 women aged 16–25 years with PCOS diagnosed according to the Rotterdam criteria, selected from 1300 consecutive patients evaluated at a tertiary gynecological endocrinology center in Poland between 2018 and 2025. Principal component analysis (PCA) followed by Gaussian mixture model clustering was used to identify latent endocrine–metabolic subgroups. Cardiometabolic outcomes included impaired glucose tolerance (2-hour plasma glucose ≥ 140 mg/dL), elevated triglyceride/high-density lipoprotein cholesterol (TG/HDL) ratio, elevated non–high-density lipoprotein cholesterol (≥ 130 mg/dL), and a composite endpoint. PCA retained 10 components explaining 81.9% of total variance. The optimal clustering solution identified two subgroups: a predominant mixed endocrine–metabolic subgroup (n = 954; 92.4%) and a smaller thyroid/autoimmune-enriched subgroup (n = 78; 7.6%) characterized by higher anti-thyroid peroxidase antibody and thyroid-stimulating hormone levels. Cluster assignment certainty within the selected model was high, although sensitivity analyses suggested that the latent structure was partially driven by thyroid/autoimmune-related variables. The thyroid/autoimmune-enriched subgroup showed a higher prevalence of an elevated TG/HDL ratio, whereas differences in glucose tolerance and non–high-density lipoprotein cholesterol were less pronounced. These findings suggest the presence of exploratory endocrine–metabolic subgroup structure in young women with PCOS, including a thyroid/autoimmune-enriched subgroup associated with modest differences in lipid-related cardiometabolic risk markers. The observed heterogeneity extends beyond conventional diagnostic classifications, although external validation and longitudinal studies are required to determine clinical relevance and reproducibility.
Clinical Trial Registration: This study was not registered as a clinical trial, as it was an observational cross-sectional analysis.
Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s40618-026-02998-x.
Related
- Aerobic and Aerobic-Oriented Exercise for Anthropometric, Androgen-Related, and Reproductive Endocrine Outcomes in Polyendocrine Metabolic Ovarian Syndrome (Polycystic Ovary Syndrome): A Systematic Review and Meta-Analysis
- Endocrine and metabolic profiling of PMOS rat models in relation to clinical obese and non-obese phenotypes
- Physical activity phenotypes identified by latent class analysis and their associations with metabolic syndrome among Korean adults
- Beyond Comorbidity: A Three-Level Clinical Framework for Endocrine–Metabolic Associations in Rosacea
- Circulating low- and medium-molecular-weight adiponectin are associated with metabolic phenotypes in adults undergoing cardiometabolic screening
BibTeX
@article{piorkowska2026datadriven,
title = {{Data-driven endocrine–metabolic phenotypes in young women with polycystic ovary syndrome and associations with cardiometabolic risk markers}},
author = {Natalia Piórkowska and Kinga Nicifur and Marcin Leśniewski and Grzegorz Franik and Anna Bizoń},
journal = {Journal of Endocrinological Investigation},
year = {2026},
month = jul,
doi = {10.1007/s40618-026-02998-x},
url = {https://doi.org/10.1007/s40618-026-02998-x},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Data-driven endocrine–metabolic phenotypes in young women with polycystic ovary syndrome and associations with cardiometabolic risk markers
AU - Natalia Piórkowska
AU - Kinga Nicifur
AU - Marcin Leśniewski
AU - Grzegorz Franik
AU - Anna Bizoń
PY - 2026
DA - 2026/07/28
JO - Journal of Endocrinological Investigation
DO - 10.1007/s40618-026-02998-x
UR - https://doi.org/10.1007/s40618-026-02998-x
AB - Polycystic ovary syndrome (PCOS) is a heterogeneous endocrine disorder associated with reproductive dysfunction and increased cardiometabolic risk. Conventional diagnostic phenotypes may not adequately capture endocrine-metabolic variability, particularly in young women. We aimed to identify data-driven endocrine–metabolic phenotypes in young women with PCOS and assess their associations with selected cardiometabolic risk markers. This cross-sectional study included 1032 women aged 16–25 years with PCOS diagnosed according to the Rotterdam criteria, selected from 1300 consecutive patients evaluated at a tertiary gynecological endocrinology center in Poland between 2018 and 2025. Principal component analysis (PCA) followed by Gaussian mixture model clustering was used to identify latent endocrine–metabolic subgroups. Cardiometabolic outcomes included impaired glucose tolerance (2-hour plasma glucose ≥ 140 mg/dL), elevated triglyceride/high-density lipoprotein cholesterol (TG/HDL) ratio, elevated non–high-density lipoprotein cholesterol (≥ 130 mg/dL), and a composite endpoint. PCA retained 10 components explaining 81.9% of total variance. The optimal clustering solution identified two subgroups: a predominant mixed endocrine–metabolic subgroup (n = 954; 92.4%) and a smaller thyroid/autoimmune-enriched subgroup (n = 78; 7.6%) characterized by higher anti-thyroid peroxidase antibody and thyroid-stimulating hormone levels. Cluster assignment certainty within the selected model was high, although sensitivity analyses suggested that the latent structure was partially driven by thyroid/autoimmune-related variables. The thyroid/autoimmune-enriched subgroup showed a higher prevalence of an elevated TG/HDL ratio, whereas differences in glucose tolerance and non–high-density lipoprotein cholesterol were less pronounced. These findings suggest the presence of exploratory endocrine–metabolic subgroup structure in young women with PCOS, including a thyroid/autoimmune-enriched subgroup associated with modest differences in lipid-related cardiometabolic risk markers. The observed heterogeneity extends beyond conventional diagnostic classifications, although external validation and longitudinal studies are required to determine clinical relevance and reproducibility. Clinical Trial Registration: This study was not registered as a clinical trial, as it was an observational cross-sectional analysis. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s40618-026-02998-x.
SN - 0391-4097
SN - 1720-8386
ER -