Metabolic Markers Improved During Telehealth Care With or Without PMOS

Key Takeaways

• Among women with a body mass index (BMI) of at least 25 receiving physician- and dietitian-supported telehealth care, weight and glucose-related measures, liver enzymes, and total cholesterol improved significantly from baseline; laboratory results come only from patients who had abnormal values at baseline.
• After adjustment, polyendocrine metabolic ovarian syndrome (PMOS) status was not associated with differential improvement in any outcome.
• GLP-1 receptor agonist use was associated with greater weight loss, and improvements also occurred without medication, though some insulin-related and HbA1c changes in the no-medication groups were not significantly different from zero.
• The analysis included only 919 of 3,117 women who started care, with no control group, so it cannot show that the care model caused the improvements.

Among women with elevated BMI, does PMOS distinguish how glucose regulation, liver enzymes, lipids, and weight change over time? Those trajectories matter when metabolic health is assessed across several measures rather than weight alone. Women with and without PMOS provided a basis for comparing those changes during care.

Investigators retrospectively analyzed a multidisciplinary telehealth cohort at Allara Health, a commercial, direct-to-consumer platform that combines visits with a medical provider and a registered dietitian. Of 3,117 women under 50 with BMI ≥25 who began care in 2024, the analysis included 919 who had at least six months of follow-up data for at least one laboratory value (557 with PMOS and 362 without); mean follow-up was 9.6 months. The 2,197 excluded women were, according to the authors, less likely to have PMOS and had a less severe metabolic profile. Mean age was 34.6 years and mean BMI 34.6. PMOS status came from diagnosis codes that providers were trained to assign using the Rotterdam criteria. Electronic health records supplied baseline and last-measured weight (self-reported), hemoglobin A1c (HbA1c), homeostatic model assessment of insulin resistance (HOMA-IR), fasting insulin, fasting glucose, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total cholesterol. Weight was analyzed in the full eligible cohort (870 in the adjusted model), while each laboratory outcome was analyzed only among women with an abnormal baseline value for that marker, for example HOMA-IR above 4 (401 women), HbA1c above 5.7% (202), and fasting glucose above 100 mg/dL (165). Linear regression adjusted each outcome for its baseline value, age, BMI, race/ethnicity, follow-up time, and medication group.

All measured outcomes improved significantly from baseline within the cohort and within both PMOS groups after Bonferroni correction. Among women with abnormal baseline values, HOMA-IR declined by 14.0%, fasting insulin by 15.6%, fasting glucose by 13.0%, HbA1c by 6.2%, ALT by 21.2%, AST by 25.5%, and total cholesterol by 8.0%; weight fell 4.9% across the cohort. Because laboratory analyses were limited to participants whose values started above the study thresholds, some of the change may reflect regression toward the mean. These changes describe the overall cohort rather than differences between women with and without PMOS.

After adjustment, PMOS status was not associated with differential improvement in any outcome, and unadjusted percent changes also did not differ significantly between groups. A statistical interaction between PMOS status and medication group for HbA1c (p = 0.011) did not remain significant after correction for eight tests (p = 0.090), and the authors regard it as exploratory.

Women with PMOS began care with a more adverse profile: elevated HOMA-IR in 50.6% versus 32.9%, elevated fasting insulin in 29.3% versus 15.2%, elevated ALT in 36.3% versus 18.8%, and elevated AST in 18.1% versus 10.8%, along with a higher mean BMI (35.4 versus 33.4). They were younger (33.5 versus 36.3 years) and more often prescribed metformin alone or with a GLP-1 receptor agonist (83.5% versus 65.8%), while GLP-1 receptor agonist alone was less common (8.8% versus 15.2%).

Compared with no GLP-1 receptor agonist or metformin, recorded GLP-1 receptor agonist use was associated with greater weight loss: −10.1 lb (95% confidence interval [CI] −14.0 to −6.2) when used alone and −7.1 lb (95% CI −10.4 to −3.9) with metformin. Metformin alone was not associated with weight change. All three medication groups were associated with greater HbA1c reduction, but after adjustment medication group was not significantly associated with changes in HOMA-IR, fasting insulin, fasting glucose, liver enzymes, or cholesterol. Improvements were also seen among women taking neither drug, particularly in weight, fasting glucose, liver enzymes, and cholesterol, but changes in HOMA-IR, HbA1c, and fasting insulin among women with PMOS taking no medication, and in fasting insulin among women without PMOS taking none, were not significantly different from zero. The authors noted that women with PMOS taking no medication saw HbA1c rise while comparable women without PMOS saw it fall, an exploratory finding.

These retrospective, nonrandomized before-and-after observations cannot establish that telehealth care caused the improvements. There was no control group, and the authors could not separate the effect of the multidisciplinary model from medication, medical provider care alone, or dietitian care alone. The study drew on one commercial platform serving a mostly privately insured population, weight was self-reported, prescriptions were not confirmed as taken, and PMOS and medication classification may be imprecise. Only women who stayed in care and had follow-up laboratory tests are represented, and about 9 months of follow-up says nothing about durability. The absence of a detected adjusted difference by PMOS status does not establish statistical equivalence between groups, although the authors describe the improvements as equivalent.

The authors interpreted the metabolic improvements among women with PMOS as comparable to those among women without PMOS, despite a more adverse metabolic profile at baseline in the PMOS group.

Clinician Questions

Is polyendocrine metabolic ovarian syndrome the same term as polycystic ovary syndrome?

Investigators identify PMOS as formerly polycystic ovary syndrome (PCOS) and use both names, sometimes interchangeably; this terminology alone does not establish a change in diagnostic criteria.

Does GLP-1–associated weight loss establish a medication effect in women receiving PMOS telehealth care?

No. Because medication use was recorded rather than randomly assigned, the observed association cannot distinguish a drug effect from other differences in care or patient characteristics. The authors noted that prescribing reflected clinical indication, severity, and patient preferences.

Recommended Reading

• For more on lifestyle support alongside GLP-1 therapy: Enhancing the Metabolic Benefits of GLP-1s with Diet and Exercise