Daily Endocrinology Research Analysis
Analyzed 46 papers and selected 3 impactful papers.
Summary
Across endocrinology, real-world target trial emulation from England strengthens the case for SGLT2 inhibitors as the preferred second-line therapy after metformin, improving glycemia, weight, blood pressure, and key cardiorenal outcomes. A randomized crossover trial in adults with type 1 diabetes shows adjunctive dapagliflozin increases CGM time-in-range versus acarbose without excess hypoglycemia, while an ED clinical decision support alert markedly shortens time to hydrocortisone in adrenal insufficiency.
Research Themes
- Second-line pharmacotherapy optimization in type 2 diabetes
- Adjunctive SGLT2 inhibitor therapy in type 1 diabetes with CGM endpoints
- Health systems innovation: CDS to prevent adrenal crisis in the ED
Selected Articles
1. PERsonalised Medicine for Intensification of Treatment (PERMIT) in type 2 diabetes mellitus: a target trial emulation from routine data.
Using target trial emulation and instrumental variables in 75,739 adults with type 2 diabetes initiating second-line therapy, SGLT2 inhibitors achieved greater reductions in HbA1c (−2.5 mmol/mol vs SU; −3.2 mmol/mol vs DPP-4i at 1 year), BMI, and systolic blood pressure than comparators. They also reduced heart failure hospitalization (vs DPP-4i) and ≥40% eGFR decline (vs SU), with microsimulation predicting fewer long-term complications.
Impact: This large-scale, methodologically rigorous real-world emulation provides decision-grade comparative effectiveness evidence supporting SGLT2 inhibitors as the preferred second-line therapy after metformin.
Clinical Implications: For most patients requiring intensification beyond metformin, SGLT2 inhibitors should be prioritized given superior glycemic, weight, blood pressure, and cardiorenal benefits. These data can inform formularies and guideline updates.
Key Findings
- At 1 year, SGLT2i vs SU: HbA1c −2.5 mmol/mol (95% CI −3.7 to −1.3); SGLT2i vs DPP-4i: −3.2 mmol/mol (95% CI −4.6 to −1.8).
- SGLT2i reduced BMI and systolic blood pressure more than SU or DPP-4i at 1 and 2 years.
- Lower hazards with SGLT2i for heart failure hospitalization (vs DPP-4i) and ≥40% eGFR decline (vs SU).
- Microsimulation projected fewer cases of end-stage kidney disease, heart failure, and eye disease with SGLT2i.
Methodological Strengths
- Target trial emulation with instrumental variable analyses to mitigate confounding and channeling bias.
- Large national linked datasets with hard outcomes and complementary microsimulation of long-term complications.
Limitations
- Instrumental variable assumptions were only partially testable; residual confounding cannot be excluded.
- Observational emulation limits causal inference; some outcomes may be underpowered over available follow-up.
Future Directions: Head-to-head pragmatic RCTs and emulations including newer agents (e.g., GLP-1/GIP co-agonists) and cost-effectiveness analyses across risk strata are warranted.
BACKGROUND: Evidence is required on the relative effectiveness of sulphonylureas, dipeptidyl peptidase-4 inhibitors or sodium-glucose cotransporter 2 inhibitors added to metformin for people with type 2 diabetes mellitus. OBJECTIVES: To assess disparities in the initiation of second-line antidiabetic treatments prescribed among people with type 2 diabetes mellitus in England according to ethnicity and social deprivation. To compare the effectiveness of sulphonylureas, dipeptidyl peptidase-4 inhibitors and sodium-glucose cotransporter 2 inhibitors added to metformin for people with type 2 diabetes mellitus who require second-line treatment in routine clinical practice. To examine heterogeneity in the comparative short-term (12 months) effectiveness of sulphonylureas versus dipeptidyl peptidase-4 inhibitors combined with metformin on levels of glycated haemoglobin across the entire target population and subpopulations of decision-making relevance. To assess the comparative effectiveness of sulphonylureas, dipeptidyl peptidase-4 inhibitors or sodium-glucose cotransporter 2 inhibitors added to metformin according to individual risk-factor profiles of multiple long-term conditions. To calibrate the RAPIDS microsimulation model to UK data and then use the resultant RAPIDS-UK model to predict probabilities of long-term complications for people with type 2 diabetes mellitus in England after second-line treatment with sulphonylureas, dipeptidyl peptidase-4 inhibitors or sodium-glucose cotransporter 2 inhibitors added to metformin. METHODS: We included adults with type 2 diabetes mellitus who initiated second-line antidiabetic treatment with sulphonylureas, dipeptidyl peptidase-4 inhibitors or sodium-glucose cotransporter 2 inhibitors added to metformin monotherapy. We used data from the Clinical Practice Research Datalink linked to Hospital Episode Statistics and the Office of National Statistics. We applied target trial emulation and instrumental variable analyses to reduce the risks of biases, including confounding. The primary outcome was change in mean glycated haemoglobin (mmol/mol) at 1-year follow-up. Secondary outcomes: change in mean body mass index, systolic blood pressure, estimated glomerular filtration rate and time to major adverse kidney event, major adverse cardiovascular event, heart failure hospitalisation, eye disease, amputation and all-cause mortality. We assessed treatment effect heterogeneity according to multiple long-term conditions. We used a microsimulation model to report the impact on long-term complications. RESULTS: After the instrumental variable analysis, the mean 95% confidence interval differences in glycated haemoglobin change between baseline and 1 year were: -2.5 mmol/mol (-3.7 to -1.3) for sodium-glucose cotransporter 2 inhibitors versus sulphonylureas, and -3.2 mmol/mol (-4.6 to -1.8) for sodium-glucose cotransporter 2 inhibitors versus dipeptidyl peptidase-4 inhibitors. Sodium-glucose cotransporter 2 inhibitors were more effective in reducing body mass index and systolic blood pressure compared to either sulphonylureas or dipeptidyl peptidase-4 inhibitors. Sodium-glucose cotransporter 2 inhibitors were also more effective at reducing mean glycated haemoglobin at 2-year follow-up, at reducing body mass index and systolic blood pressure at 1- and 2-year follow-ups and at reducing the hazards of heart failure hospitalisation (vs. dipeptidyl peptidase-4 inhibitors) and ≥ 40% decline in estimated glomerular filtration rate (vs. sulphonylureas). We did not find evidence of treatment effect heterogeneity by baseline cardiovascular disease status or multiple long-term condition profiles. The microsimulation model found that sodium-glucose cotransporter 2 inhibitors led to lower predicted incidence of end-stage kidney disease, heart failure and eye disease. Public and patient involvement translation workshop participants provided valuable insights on how best to share our findings. LIMITATIONS: We could only partially evaluate the main instrumental variable assumptions. CONCLUSIONS: We found that sodium-glucose cotransporter 2 inhibitors were better than dipeptidyl peptidase-4 inhibitors and sulphonylureas at improving important risk factors and at reducing the risk of complications among a general population of people with type 2 diabetes mellitus. FUTURE WORK: Newer antidiabetic treatments should be evaluated. FUNDING: This synopsis presents independent research funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme as award number NIHR128490. In the United Kingdom, around 4 million people have been diagnosed with type 2 diabetes, which causes high levels of sugar (glucose) in the blood. Most people start treatment with a drug called metformin, but some will later need to take additional drugs when metformin is not effective enough. We do not know drugs are best to use with metformin. We looked at the health care information routinely collected about people with type 2 diabetes when they use the National Health Service. We used the information collected between 2015 and 2021 to investigate which treatments worked best. We included 75,739 people in our study, all of whom received one of the three drug types below in addition to metformin: sulphonylureas (e.g. Diamicron) dipeptidyl peptidase-4 inhibitors (e.g. Januvia) sodium–glucose cotransporter 2 inhibitors (e.g. Forxiga). There was a lot of variation across different groups of general practices in the treatment they prescribed. We found that sodium–glucose cotransporter 2 inhibitors were more effective than sulphonylureas or dipeptidyl peptidase-4 inhibitors in reducing blood sugar levels, body mass index and blood pressure and preventing hospitalisation due to heart failure. For other outcomes (like preventing or delaying heart attack, stroke or death), there was no clear difference between drug types. Our results show that using sodium–glucose cotransporter 2 inhibitors alongside metformin may help patients manage some of the common symptoms and side effects of their diabetes. However, some of these effects are quite small. There will be people whose personal circumstances mean that the drug they are taking works especially well for them. Where we did not see a significant difference between drug types, this may be because we did not have data for enough people over a long enough period.
2. Dapagliflozin as an Adjunct to Insulin Improves Time-in-Range and Glycemic Variability Versus Acarbose in Adults With Type 1 Diabetes: A Randomized Crossover Trial.
In a randomized open-label crossover trial with blinded CGM, adjunctive dapagliflozin increased time-in-range (68.2% vs 56.0% vs acarbose; P<0.001), reduced glycemic variability, and lowered insulin dose without increasing hypoglycemia or causing DKA over 4-week periods.
Impact: The study provides controlled CGM evidence that SGLT2 inhibition can meaningfully improve short-term glycemic profiles and insulin requirements in adults with type 1 diabetes.
Clinical Implications: Adjunctive SGLT2 inhibitors may enhance time-in-range and reduce insulin dose in T1D; if pursued, careful DKA risk mitigation and patient selection are essential pending longer-term safety/efficacy trials.
Key Findings
- Dapagliflozin increased TIR from 55.4% to 68.2% (P<0.001); vs acarbose, TIR 68.2% vs 56.0% (P<0.001).
- Reduced glycemic variability (SDBG, LAGE; P<0.01) and mean glucose with dapagliflozin.
- Lower total daily insulin dose during dapagliflozin vs acarbose (31.94 U vs 35.63 U; P=0.001), and reduced from baseline only with dapagliflozin.
- No increase in hypoglycemia; no diabetic ketoacidosis events occurred.
Methodological Strengths
- Randomized crossover design with within-subject comparisons and blinded CGM endpoints.
- Direct head-to-head comparison versus an active comparator (acarbose) with standardized dosing periods.
Limitations
- Short treatment periods (4 weeks each) and open-label design may limit generalizability.
- Sample size modest (n=44); longer-term safety including DKA risk not assessed.
Future Directions: Longer-term, adequately powered RCTs evaluating DKA mitigation strategies, renal outcomes, and patient-reported outcomes are needed to define the role of SGLT2 inhibitors in T1D.
OBJECTIVE: To compare dapagliflozin versus acarbose as adjuncts to insulin on continuous glucose monitoring (CGM)-derived time in range (TIR), glycemic variability, and safety in adults with type 1 diabetes (T1DM). METHODS: In this prospective, randomized, open-label, crossover trial, adults with T1DM (aged 18-60 years, HbA1c ≤9.0%) received dapagliflozin (10 mg once daily) and acarbose (50 mg three times daily), each for 4 weeks with a 2-week washout. Blinded CGM was worn during the final 14 days of each period. The primary endpoint was TIR (3.9-10.0 mmol/L). RESULTS: Forty-four participants completed both periods. Dapagliflozin significantly increased TIR from baseline (55.4% to 68.2%; P<0.001) and reduced mean glucose and glycemic variability (SDBG, LAGE; P<0.01). Acarbose reduced postprandial glucose excursions (P=0.001). Compared directly with acarbose, dapagliflozin achieved a higher TIR (68.2% vs. 56.0%; P<0.001) and resulted in a lower total daily insulin dose during treatment (31.94 U vs. 35.63 U; P=0.001). Notably, only dapagliflozin was associated with a significant reduction in insulin requirements from baseline (31.94 U vs 38.45 U, P=0.009). Hypoglycemia frequency did not differ between treatments (P=0.467), and no diabetic ketoacidosis events occurred. CONCLUSIONS: In this short-term controlled study of adults with T1DM, adjunctive dapagliflozin was associated with improved CGM-derived glycemic control and reduced insulin dose versus acarbose, without DKA events. Longer-term studies are warranted.
3. Impact of Alert Systems on Glucocorticoid Administration for Adrenal Insufficiency in the Emergency Department.
In a pre–post quasi-experiment at a tertiary ED, a CDS alert for known adrenal insufficiency reduced median time to intravenous hydrocortisone from 202 to 77 minutes and increased 1-hour administration from 0% to 42%, with fewer new-onset hypotension events.
Impact: Timely stress-dose steroids are lifesaving in adrenal crisis; this study demonstrates a practical, scalable informatics intervention that substantially improves emergency care timelines.
Clinical Implications: EDs should consider CDS alerts that flag known adrenal insufficiency to prompt immediate hydrocortisone, standardized order sets, and staff education to reduce delays and prevent deterioration.
Key Findings
- Median time from ED arrival to IV hydrocortisone fell from 202 to 77 minutes after CDS implementation (p=0.01).
- Hydrocortisone within 1 hour increased from 0% pre-CDS to 42% post-CDS (p=0.002).
- New-onset hypotension decreased from 21% to 4% among treated encounters.
- Overall treatment rates with IV hydrocortisone were similar (89% vs 92%; p=0.81), indicating improvement driven by timeliness.
Methodological Strengths
- Objective time-stamp outcomes and clinically meaningful process metrics in a high-risk population.
- Real-world implementation evaluation with sufficient pre/post encounters to detect change.
Limitations
- Single-center quasi-experimental design susceptible to temporal confounding and workflow changes.
- Relatively small number of encounters requiring IV hydrocortisone limits precision for clinical outcomes.
Future Directions: Multicenter stepped-wedge or cluster RCTs should test CDS bundles (alerts, order sets, education) on morbidity, ICU transfer, and cost-effectiveness in adrenal insufficiency.
CONTEXT: Inappropriate glucocorticoid management in adrenal insufficiency (AI) can lead to adrenal crises, particularly in an emergency setting. Clinical decision support (CDS) can be helpful, but its impact on AI management in emergency departments (EDs) remains limited. OBJECTIVE: To evaluate the efficacy of CDS on glucocorticoid administration time in known AI patients presenting to the ED. DESIGN: Quasi-experimental (pre- and post- intervention) design. SETTING: Tertiary academic hospital. PATIENTS: A total of 211 patients with AI; 145 had no ED visits, 66 patients had 187 ED encounters. MAIN OUTCOME MEASURES: Time from ED arrival to intravenous hydrocortisone administration pre- versus post-CDS implementation. RESULTS: The incidence of ED visits was 42 events per 100 patient-years, and the incidence of ED visits presenting with hypotension was 7 events per 100 patient-years. Among patients with ED visits, the number of visits was 73 before and 114 after CDS implementation. Events requiring intravenous hydrocortisone were 19 before and 24 after implementation, with treatment given in 89% vs. 92%, respectively (p=0.81). The median time from ED arrival to hydrocortisone administration decreased from 202 (IQR: 99-306) minutes to 77 (IQR: 45-153) minutes (p=0.01). The proportion of patients receiving hydrocortisone within 1 hour increased from 0% to 42% (p=0.002). New-onset hypotension occurred in 4 cases (21%) before implementation and in 1 case (4%) after implementation. CONCLUSIONS: Following CDS implementation, the time to hydrocortisone administration was significantly shorter among patients with known AI. These findings highlight the value of CDS in improving AI recognition and management in high-risk settings.