Diagnostic Performance of Sex-Specific Modified Metabolite Patterns in Urine for Screening of Prediabetes

Zaifang Li, Yanhui Zhang, Miriam Hoene, Louise Fritsche, Sijia Zheng, Andreas L. Birkenfeld, Andreas Fritsche, Andreas Peter, Xinyu Liu, Xinjie Zhao, Lina Zhou, Ping Luo, Cora Weigert, Xiaohui Lin, Guowang Xu, Rainer LehmannFrontiers in Endocrinology 2022 · DOI

This research paper, published in Frontiers in Endocrinology, presents a novel approach to screening for prediabetes by analyzing "modified" metabolites in urine. By moving away from standard blood tests and focusing on sex-specific metabolic signatures, the authors achieved significantly higher diagnostic accuracy than traditional methods like fasting glucose or HbA1c.

TL;DR

Researchers have developed a non-invasive urine-based screening tool for prediabetes that identifies sex-specific patterns of modified metabolites. This approach significantly outperforms traditional blood-based markers, achieving an area under the curve (AUC) of up to 0.977 in men when combined with standard tests.

What Problem Does This Paper Address?

Prediabetes is a global pandemic, affecting over 860 million people, yet it often goes undetected. The current "gold standard" for diagnosis is the Oral Glucose Tolerance Test (OGTT), but it is too time-consuming and labor-intensive for large-scale population screening.

Commonly used alternatives, such as Fasting Plasma Glucose (FPG) and Hemoglobin A1c (HbA1c), are convenient but notoriously insensitive—missing approximately 50% of prediabetic individuals. Furthermore, most existing metabolomics research focuses on blood samples and "one-size-fits-all" biomarkers, often ignoring the distinct biological differences between sexes and the diagnostic potential of modified metabolites (metabolites that have undergone chemical changes like sulfation or glucuronidation).

Methods

The study utilized non-targeted, modification group-assisted metabolomics.

  • Instrumentation: Liquid chromatography–mass spectrometry (LC-MS) was used to analyze second morning urine samples.
  • The "Modified" Focus: Unlike standard studies that look for primary metabolites (like simple amino acids), this team specifically targeted 15 types of modifications. These modifications often represent the body’s attempt to detoxify or process substances, providing a more nuanced view of metabolic stress.
  • Feature Selection: The researchers used the ReliefF algorithm to sift through hundreds of metabolic features to identify the most predictive "diagnostic patterns" for men and women separately.
  • Workflow: The study followed a rigorous discovery-and-validation pipeline, comparing the new urinary patterns against conventional markers (FPG, HbA1c, and fasting insulin).

Datasets

The study drew from the TULIP prediabetes cohort in Tübingen, Germany:

  • Total Participants: 340 individuals (208 with normal glucose tolerance [NGT] and 132 with impaired glucose tolerance [IGT]).
  • Discovery Cohort: 119 individuals (used to identify the patterns).
  • Validation Cohort: 221 individuals (used to test the patterns' accuracy on "unseen" data).
  • Groups were carefully matched for age and Body Mass Index (BMI) to ensure results weren't skewed by these factors.

Key Insights and Ideas

  • Sex Specificity is Crucial: Only three biomarkers were identical between the male and female diagnostic patterns. This suggests that prediabetes manifests differently at a molecular level in men and women, and "unisex" screening tools are likely suboptimal.
  • Urine as a Diagnostic Powerhouse: While blood is the clinical standard, urine is easier to collect and contains concentrated metabolic end-products (like advanced glycation end products and gut microbiota-associated compounds) that serve as early warning signs of metabolic dysfunction.
  • Modified Metabolites Matter: By focusing on modified compounds (e.g., glucuronides and sulfates), the researchers tapped into a "hidden" layer of the metabolome that is often overlooked in standard databases but carries high diagnostic value.

Key Results

The urinary patterns consistently outperformed traditional blood tests across the board:

  • In Men: The male-specific pattern achieved an AUC of 0.889 in the validation cohort with a diagnostic sensitivity of 92.6%. When combined with HbA1c, the AUC rose to a near-perfect 0.977.
  • In Women: The female-specific pattern achieved an AUC of 0.722. While lower than the male result, it still significantly exceeded FPG (0.595) and HbA1c (0.604).
  • Comparison: For context, the AUCs for FPG, HbA1c, and insulin alone in the same cohort were mediocre, ranging from roughly 0.57 to 0.67.
  • Identified Markers: Key biomarkers included compounds related to gut health (indoxyl sulfate) and protein damage (pentosidine-glucuronide), linking prediabetes to broader systemic changes.

Limitations and Future Work

  • Female Variability: The lower performance in the female cohort is a known challenge in metabolomics, likely due to hormonal fluctuations (menstrual cycle or menopause) that add "noise" to the data. Future models may need to account for hormonal status.
  • Sample Size: While robust, the male validation cohort was relatively small (n=73). Larger, more ethnically diverse cohorts are needed to confirm these patterns globally.
  • Clinical Integration: While LC-MS is a sophisticated laboratory technique, the authors note that the recent development of fully automated LC-MS systems could eventually make this type of sex-specific urinary screening a reality in routine clinical practice.

Summary updated

Abstract

Our results demonstrate that the sex-specific search for diagnostic metabolite biomarkers can be superior to common metabolomics strategies. The diagnostic performance for IGT detection was significantly better than routinely applied blood parameters. Together with recently developed fully automatic LC-MS systems, this opens up future perspectives for the application of sex-specific diagnostic patterns for prediabetes screening in urine.
BibTeX
@inproceedings{li2022diagnostic,
  author = {Zaifang Li and Yanhui Zhang and Miriam Hoene and Louise Fritsche and Sijia Zheng and Andreas L. Birkenfeld and Andreas Fritsche and Andreas Peter and Xinyu Liu and Xinjie Zhao and Lina Zhou and Ping Luo and Cora Weigert and Xiaohui Lin and Guowang Xu and Rainer Lehmann},
  title = {{Diagnostic Performance of Sex-Specific Modified Metabolite Patterns in Urine for Screening of Prediabetes}},
  booktitle = {Frontiers in Endocrinology},
  year = {2022},
  doi = {10.3389/fendo.2022.935016},
}
Figure