Daily Endocrinology Research Analysis
Analyzed 98 papers and selected 3 impactful papers.
Summary
Three studies stood out today in endocrinology: a high-quality network meta-analysis of insulin regimens in type 2 diabetes clarified trade-offs between glycemic benefit and safety; a massive Nordic cohort established increased cardiovascular disease risk in women with PCOS independent of obesity; and a multi-site EHR analysis showed that COVID-19–diabetes associations are highly sensitive to phenotype misclassification, urging caution in observational inference.
Research Themes
- Diabetes therapy optimization and safety
- Cardiometabolic risk in endocrine disorders
- EHR phenotyping bias in endocrinology research
Selected Articles
1. Effects of insulin regimens for type 2 diabetes mellitus: a systematic review and network meta-analysis.
Across 58 RCTs (n=19,122), basal–bolus, biphasic, and prandial regimens achieved modest additional HbA1c lowering versus basal insulin but at the cost of greater weight gain and suggested higher hypoglycemia risk. Evidence supports individualized regimen selection balancing efficacy, safety, and patient priorities.
Impact: Provides high-level comparative evidence to guide insulin regimen choice in T2DM, clarifying efficacy–safety trade-offs that directly influence clinical decision-making.
Clinical Implications: Prefer basal insulin as foundational therapy when appropriate; escalate to more complex regimens only when incremental HbA1c benefit outweighs weight gain and hypoglycemia risks, and align decisions with patient preferences and quality-of-life considerations.
Key Findings
- 58 RCTs (n=19,122) compared basal, basal–bolus, biphasic, and prandial regimens for ≥12 weeks.
- Complex regimens yielded modest HbA1c improvements over basal insulin.
- Complex regimens were associated with greater weight gain and a suggested higher hypoglycemia risk.
- Risk of bias assessed with RoB-2; PROSPERO-registered methodology.
Methodological Strengths
- Network meta-analysis integrating 58 RCTs with frequentist random-effects modeling
- Rigorous risk-of-bias assessment (Cochrane RoB-2) and protocol registration (PROSPERO)
Limitations
- Heterogeneity in trial designs, insulin titration algorithms, and background therapies
- Some outcomes (e.g., hypoglycemia severity definitions) may not be fully harmonized across RCTs
Future Directions: Prospectively compare regimen strategies incorporating modern CGM-guided titration, hypoglycemia-aware metrics, and patient-reported outcomes to refine individualized insulin therapy.
AIMS/HYPOTHESIS: Insulin therapy is essential for managing hyperglycaemia in type 2 diabetes mellitus when oral or non-insulin injectable agents are no longer effective. However, the comparative effectiveness and safety of different insulin regimens remain uncertain. We aimed to compare the effects of basal, basal-bolus, biphasic and prandial insulin regimens on glycaemic management, weight, severe hypoglycaemia, insulin dose and quality of life in adults with type 2 diabetes. METHODS: We conducted a systematic review and network meta-analysis of RCTs. PubMed, EMBASE, the Cochrane Library and ClinicalTrials.gov were searched through to September 2025. The inclusion criteria for studies were RCTs enrolling adults with type 2 diabetes that compared at least two of the specified insulin regimens over ≥12 weeks. Pairs of reviewers independently screened studies, extracted data and assessed risk of bias using the Cochrane RoB-2 tool. Network meta-analyses were performed using a frequentist random-effects model, with basal insulin as the reference. RESULTS: Fifty-eight RCTs involving 19,122 participants were included. Compared with basal insulin, HbA CONCLUSIONS/INTERPRETATION: Complex insulin regimens provide modest glycaemic benefits over basal insulin but are associated with greater weight gain and a suggested higher risk of hypoglycaemia. These results are based on evidence of moderate confidence. These trade-offs support the need for individualised regimen selection, informed by clinical context, patient preferences and treatment goals. TRIAL REGISTRATION: PROSPERO registration no. CRD42020181473. FUNDING: This research was supported by the Committee for the Development of Higher Education Personnel (CAPES) and the Hospital de Clínicas de Porto Alegre - FIPE HCPA.
2. Increased prospective cardiovascular disease risk in 127,517 Nordic women with polycystic ovary syndrome. A national cohort study.
In a national register-based cohort of 127,517 women with PCOS and 587,810 controls, PCOS was associated with a 32% higher adjusted risk of CVD, robust across Denmark, Finland, and Sweden. Elevated risk persisted in women with BMI <25 kg/m2 and without type 2 diabetes.
Impact: Defines PCOS as an independent cardiometabolic risk state at a population level, strengthening rationale for early cardiovascular risk assessment and prevention in PCOS.
Clinical Implications: Incorporate systematic CVD risk assessment (lipids, blood pressure, lifestyle, thrombosis history) into PCOS care pathways regardless of BMI; consider earlier preventive interventions and long-term monitoring.
Key Findings
- PCOS was associated with increased CVD risk across three Nordic countries (adjusted HR ~1.32 overall).
- Risk elevation remained significant after adjusting for obesity and education.
- CVD risk persisted in lean women (BMI <25 kg/m2) without type 2 diabetes (adjusted HR 1.40).
- Median follow-up 8–10 years with large, register-based cohorts.
Methodological Strengths
- Massive population-based registers with age-matched controls across three countries
- Consistent Cox modeling with adjustment for BMI and education; meta-analysis across cohorts
Limitations
- Administrative coding may introduce misclassification of PCOS and CVD events
- Residual confounding (e.g., lifestyle, family history) cannot be fully excluded
Future Directions: Intervention trials testing targeted cardiometabolic prevention in PCOS, including lean phenotypes, and mechanistic studies to dissect pathways independent of adiposity.
BACKGROUND: Cardiovascular disease (CVD) risk factors are prevalent in women with PCOS, but prospective data regarding CVD in population-based cohorts are limited. AIM: To investigate prospective risk of CVD in Nordic women with PCOS. DESIGN: National register-based study in women with PCOS and age-matched controls originating from Denmark (PCOS Denmark, N = 27,298, controls, N = 135,019), Finland (PCOS Finland, N = 20,765, controls, N = 59,122), and Sweden (PCOS Sweden, N = 79,454, controls, N = 393,669). The main study outcome was CVD. CVD was defined according to ICD-10 diagnostic codes for major adverse cardiac events, pulmonary embolism and/or deep venous thrombosis. Cox regression analyses estimated HR with 95% CI and adjusted analyses included BMI and education. RESULTS: The median age at cohort entry was 28 years (Denmark) and 29 years (Finland and Sweden) and the median follow-up time was 8.0 to 10.0 years. The unadjusted hazard ratio (HR, 95% CI) for CVD in women with PCOS was 1.30 (1.20; 1.41) in Denmark, 1.45 (1.31; 1.60) in Finland, and 1.52 (1.44; 1.60) in Sweden. Models remained significant after adjusting for obesity and level of education. In a combined meta-analysis including all countries (PCOS, N = 127,517, controls, N = 587,810, the adjusted HR for CVD in women with PCOS was 1.32 (1.25; 1.39). In women with BMI < 25 kg/m2 and no type 2 diabetes, the adjusted HR for CVD risk was 1.40 (1.26; 1.55). CONCLUSION: The risk of CVD was increased in women with PCOS across the three Nordic countries, also among women with BMI < 25 kg/m2.
3. Multi-site analysis of COVID-19 and new-onset diabetes reveals need for improved sensitivity of EHR-based COVID-19 phenotypes-a DiCAYA network analysis.
Across multiple U.S. sites, EHR-documented COVID-19 was associated with higher incident diabetes risk, but COVID-19 phenotype sensitivity was low and varied widely. Bias analyses showed that differential misclassification could bias results away from the null, challenging causal interpretation.
Impact: Sets methodological guardrails for EHR-based endocrine epidemiology by quantifying how phenotype misclassification can distort associations between infections and diabetes.
Clinical Implications: Interpret COVID-19–diabetes observational findings with caution; where possible, improve case definitions using multi-source validation (labs, diagnoses, testing records) and perform bias analyses to assess robustness.
Key Findings
- EHR-documented COVID-19 had low and heterogeneous prevalence across sites (children 4.4–7.7%, young adults 6.2–22.7%).
- Documented COVID-19 associated with higher incident diabetes risk, but results varied across sites.
- Bias analyses indicated results were highly sensitive to misclassification assumptions and could be biased away from the null.
Methodological Strengths
- Multi-site design with predefined bias analysis scenarios
- Cox models and pooled meta-analytic synthesis across sites
Limitations
- Retrospective EHR data with potential under-ascertainment of COVID-19 exposure
- Heterogeneity in coding and testing practices across sites
Future Directions: Standardize high-sensitivity EHR phenotypes for infections; link to external testing databases and deploy probabilistic bias analyses to improve causal inference in endocrine outcomes.
OBJECTIVE: We discuss implications of potential ascertainment biases for studies examining diabetes risk following SARS-CoV-2 infection using electronic health records (EHRs). We quantitatively explore sensitivity of results to misclassification of COVID-19 status using data from the U.S.-based Diabetes in Children, Adolescents and Young Adults (DiCAYA) Network on children (≤17 years) and young adults (18-44 years). MATERIALS AND METHODS: In our retrospective case study from the DiCAYA Network, SARS-CoV-2 was identified using labs and diagnoses from 6/1/2020-12/31/2021. Patients were followed through 12/31/2022 for new diabetes diagnoses. Sites examined incident diabetes by COVID-19 status using Cox proportional hazards models. Results were pooled in meta-analyses. A bias analysis examined potential impact of COVID-19 misclassification scenarios on results, guided by hypotheses that sensitivity would be < 50% and would be higher among those who developed diabetes. RESULTS: Prevalence of documented COVID-19 was low overall and variable across sites (children: 4.4%-7.7%, young adults: 6.2%-22.7%). Individuals with documented COVID-19 were at higher risk of incident diabetes compared to those with no documented infection, but results were heterogeneous across sites. Findings were highly sensitive to COVID-19 misclassification assumptions. Observed results could be biased away from the null under several differential misclassification scenarios. DISCUSSION: Although EHR-based documentation of COVID-19 was associated with incident diabetes, COVID-19 phenotypes likely had low sensitivity, with considerable variation across sites. Misclassification assumptions strongly impacted interpretation of results. CONCLUSION: Given the potential for low phenotype sensitivity and misclassification, caution is warranted when interpreting analyses of COVID-19 and incident diabetes using clinical or administrative databases.