Skip to main content
SMFM 2024 Pregnancy Meeting Main banner
Final Program


Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Presentation Icons
Additional registration fee required
Additional registration fee required
Faculty have requested this content not be shared on social media
Faculty have requested this content not be shared on social media
CME Credit Offered
CME Credit Offered
Poster Icons
Foundation Awardee
Foundation Awardee
No Social Media
No Social Media
Back
Q&A

Computers

Poster Session 4

(1141) T2DM identified as risk factor for IHCP using a machine learning approach with pregnancy records

Wednesday, February 14, 2024
1:00 PM – 2:30 PM  
Location: Prince George's Exhibit Hall CDE, Lower Atrium Level

    Primary & Presenting Author(s)

  • Maxim D. Seferovic, PhD

    Assistant Professor
    Baylor College of Medicine and Texas Children’s Hospital
    Houston, TX, United States

    • Coauthor(s)

  • Kjersti M. Aagaard, MD, PhD

    Professor and Vice Chair of Research Department of Obstetrics and Gynecology, Division of MFM
    Texas Children's Hospital, Baylor College of Medicine
    Houston, TX, United States

Objective:

Intrahepatic cholestasis of pregnancy (IHCP) is a poorly understood disease known to have hormonal and metabolic links. In a discovery approach, we sought to naively assess a large database of pregnancies using machine learning and natural language processing (NLP) for novel insights into pathology and risk.

Study Design: Peribank, a research database of 46,641 pregnancies containing 12,017 data entry fields, was natural language processed using spaCy pretrained models, with Boolean and scalar data retained. Medications were corrected against MEDI-Ensemble using Levenshtein distance fuzzy matching. Post processing filtering left ~450,000 unique features that were assessed for predictive feature selection with scikit-learn. Logit regression was used for select features.  

Results: There were 1374 IHCP pregnancies discovered in the database. Feature selection revealed top scoring predictors of IHCP (positive or negative) in decreasing score as T2DM as indication for antenatal testing, T2DM, metformin use at conception, white maternal race, white paternal race (both inclusive of Hispanic ethnicity), no medications at conception, tetanus vaccination, high alkaline phosphatase, high glucose, flu vaccination, and GA at admittance. Diabetes and metformin use were then assessed independently for risk. T2DM patients had increased risk for IHCP (OR 1.44, 95% CI 1.05-1.98, p=0.023) as did GDM (1.54, CI 1.31-1.81, p=3.5x10-7), while there was no association with Type I diabetes or metformin use independently. The associations were confirmed by Χ2 for T2DM (p=0.0029), GDM (p=1.3 x10-7) and metformin use (p=0.056) (Fig. A). A sub analyses of only diabetic patients of all types (n=5110) found metformin treated patients more likely to develop IHCP than others (14.0% vs. 10.8%) though it did not reach significance (p=0.17, Χ2) (Fig. B).

Conclusion:

The novel association of risk for IHCP with T2DM adds to the known association of GDM. Given the absence of association with T1DM, and the possible influence of metformin, the results support the theory that IHCP pathology is linked to hepatic insulin resistance.