Annals of Internal Medicine最新文献

Background: Although glucagon-like peptide-1 receptor agonists (GLP-1RAs) have shown potential antidepressant effects, population studies yield inconsistent results.

Objective: To compare the risk for depression in older adults with type 2 diabetes (T2D) initiating treatment with GLP-1RAs versus sodium-glucose cotransporter-2 inhibitors (SGLT2is) or dipeptidyl peptidase-4 inhibitors (DPP4is).

Design: Target trial emulation study.

Setting: U.S. National Medicare administrative data from January 2014 to December 2020.

Patients: Adults aged 66 years or older with T2D initiating treatment with a GLP-1RA were matched 1:1 on propensity score with those initiating treatment with either an SGLT2i or a DPP4i.

Measurements: The primary end point was incident depression. Cox proportional hazards regression models were used to estimate the hazard ratio (HR) with 95% CI within matched groups.

Results: A total of 14 665 matched pairs of older adults were included in the cohort for GLP-1RAs versus SGLT2is; the rate difference of depression between GLP-1RA users and SGLT2i users was 3.48 (95% CI, -0.81 to 7.78) per 1000 person-years, with an HR of 1.07 (CI, 0.98 to 1.18). In the cohort for GLP-1RAs versus DPP4is (13 711 matched pairs), the rate difference was -5.78 (CI, -10.49 to -1.07) per 1000 person-years, with an HR of 0.90 (CI, 0.82 to 0.98).

Limitation: Unmeasured confounders (such as hemoglobin A_1c levels and body mass index), outcome misclassification, and limited generalizability to all GLP-1RA users (for example, younger populations or those without T2D receiving the drug for obesity treatment).

Conclusion: Among older adults with T2D, the incidence of depression was relatively low. Use of GLP-1RAs was associated with a modestly lower risk for depression compared with use of DPP4is, but not SGLT2is.

Primary funding source: National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health.

Background: Systematic reviews (SRs) are hindered by the initial rigorous article screen, which delays access to reliable information synthesis.

Objective: To develop generic prompt templates for large language model (LLM)-driven abstract and full-text screening that can be adapted to different reviews.

Design: Diagnostic test accuracy.

Setting: 48 425 citations were tested for abstract screening across 10 SRs. Full-text screening evaluated all 12 690 freely available articles from the original search. Prompt development used the GPT4-0125-preview model (OpenAI).

Participants: None.

Measurements: Large language models were prompted to include or exclude articles based on SR eligibility criteria. Model outputs were compared with original SR author decisions after full-text screening to evaluate performance (accuracy, sensitivity, and specificity).

Results: Optimized prompts using GPT4-0125-preview achieved a weighted sensitivity of 97.7% (range, 86.7% to 100%) and specificity of 85.2% (range, 68.3% to 95.9%) in abstract screening and weighted sensitivity of 96.5% (range, 89.7% to 100.0%) and specificity of 91.2% (range, 80.7% to 100%) in full-text screening across 10 SRs. In contrast, zero-shot prompts had poor sensitivity (49.0% abstract, 49.1% full-text). Across LLMs, Claude-3.5 (Anthropic) and GPT4 variants had similar performance, whereas Gemini Pro (Google) and GPT3.5 (OpenAI) models underperformed. Direct screening costs for 10 000 citations differed substantially: Where single human abstract screening was estimated to require more than 83 hours and $1666.67 USD, our LLM-based approach completed screening in under 1 day for $157.02 USD.

Limitations: Further prompt optimizations may exist. Retrospective study. Convenience sample of SRs. Full-text screening evaluations were limited to free PubMed Central full-text articles.

Conclusion: A generic prompt for abstract and full-text screening achieving high sensitivity and specificity that can be adapted to other SRs and LLMs was developed. Our prompting innovations may have value to SR investigators and researchers conducting similar criteria-based tasks across the medical sciences.

Primary funding source: None.

Background: The federal government spends billions of dollars per year on payments to Medicare Advantage (MA) plans based, in part, on beneficiaries' risk scores. Despite this, little is known about the true burden of chronic medical conditions among MA beneficiaries compared with those in fee-for-service (FFS) Medicare.

Objective: To determine whether the prevalence of chronic medical conditions is higher among MA compared with FFS beneficiaries.

Design: Cross-sectional.

Setting: Population based.

Participants: Adults aged 65 years or older enrolled in MA or FFS Medicare.

Measurements: Using direct physical examination and laboratory data from the National Health and Nutrition Examination Survey (2015 to 2018), we compared the age- and sex-standardized prevalence of obesity, hypertension, hyperlipidemia, diabetes, and chronic kidney disease between MA and FFS beneficiaries.

Results: The unweighted study population included 2446 respondents corresponding to a weighted total of 45 426 711 adults (34.4% MA, 65.6% FFS Medicare). The prevalence of obesity (41.1% vs. 40.6%; standardized difference [SDiff], 0.48 percentage points [pp] [95% CI, -5.2 to 6.2 pp]), hypertension (70.9% vs. 71.0%; SDiff, -0.05 pp [CI, -5.8 to 5.7 pp]), hyperlipidemia (79.4% vs. 82.3%; SDiff, -2.86 pp [CI, -7.0 to 1.3 pp]), and chronic kidney disease (19.2% vs. 22.8%; SDiff, -3.48 pp [CI, -9.2 to 2.3 pp]) was not higher among MA beneficiaries compared with FFS beneficiaries. However, the prevalence of diabetes was higher in MA (33.3% vs. 26.3%; SDiff, 7.00 pp [CI, 3.3 to 10.7 pp]).

Limitation: Differences in the severity of specific medical conditions between groups could not be assessed.

Conclusion: In this nationally representative study from 2015 to 2018, the prevalence of obesity, hypertension, hyperlipidemia, and chronic kidney disease was not higher among MA compared with FFS beneficiaries; however, the prevalence of diabetes was higher among MA beneficiaries.

Primary funding source: National Heart, Lung, and Blood Institute (NHLBI) and American Heart Association (AHA).

Background: The role of transitioning from short dual antiplatelet therapy (DAPT) to potent P2Y12 inhibitor monotherapy in patients with acute coronary syndrome (ACS) undergoing drug-eluting stent (DES) implantation remains inconclusive.

Purpose: To compare the effects of de-escalating DAPT to ticagrelor monotherapy versus standard DAPT from randomized clinical trials in patients with ACS.

Data sources: PubMed, EMBASE, Scopus, and ClinicalTrials.gov from inception to 12 December 2024.

Study selection: Randomized clinical trials comparing de-escalating DAPT to ticagrelor monotherapy versus ticagrelor-based standard DAPT for 12 months, specifically in patients with ACS undergoing DES implantation.

Data extraction: The coprimary end points were an ischemic end point (composite of death, nonprocedural [spontaneous] myocardial infarction, or stroke) and a bleeding end point (Bleeding Academic Research Consortium types 3 or 5 bleeding).

Data synthesis: Individual patient data were obtained from 3 trials (TICO [Ticagrelor Monotherapy After 3 Months in the Patients Treated With New Generation Sirolimus-Eluting Stent for Acute Coronary Syndrome], T-PASS [Ticagrelor Monotherapy in Patients Treated With New-Generation Drug-Eluting Stents for Acute Coronary Syndrome], and ULTIMATE-DAPT [Ticagrelor alone versus ticagrelor plus aspirin from month 1 to month 12 after percutaneous coronary intervention in patients with acute coronary syndromes]), including 9130 randomized patients with ACS; 3132 had ST-segment elevation myocardial infarction (STEMI), 3023 had non-STEMI (NSTEMI), and 2975 had unstable angina. The rate of the primary ischemic end point was not different between the ticagrelor monotherapy and standard DAPT groups (1.7% vs. 2.1%; hazard ratio [HR], 0.85 [95% CI, 0.63 to 1.16]). The rate of the primary bleeding end point was lower in the ticagrelor monotherapy group (0.8% vs. 2.5%; HR, 0.30 [CI, 0.21 to 0.45]). These findings were consistent in patients with STEMI, NSTEMI, and unstable angina.

Limitation: Other de-escalation strategies for modulating antiplatelet therapy were not included.

Conclusion: In patients with ACS undergoing DES implantation, de-escalating DAPT to ticagrelor monotherapy was associated with a lower risk for major bleeding compared with standard DAPT, without an increase in ischemic events, regardless of the type of ACS.

Primary funding source: None. (PROSPERO: CRD42024565855).

When randomized trials are not available to answer a causal question about the comparative effectiveness or safety of interventions, causal inferences are drawn using observational data. A helpful 2-step framework for causal inference from observational data is 1) specifying the protocol of the hypothetical randomized pragmatic trial that would answer the causal question of interest (the target trial), and 2) using the observational data to attempt to emulate that trial. The target trial framework can improve the quality of observational analyses by preventing some common biases. In this article, we discuss the utility and scope of applications of the framework. We clarify that target trial emulation resolves problems related to incorrect design but not those related to data limitations. We also describe some settings in which adopting this approach is advantageous to generate effect estimates that can close the gaps that randomized trials have not filled. In these settings, the target trial framework helps reduce the ambiguity of causal questions.