Updates in Hospital Medicine 2026

Nashville welcomed SHM Converge 2026, and with it one of the conference’s most anticipated annual traditions: the Updates in Hospital Medicine session. Kencee Graves, MD, FHM, clinical professor of medicine, academic hospitalist, and chief medical quality officer at the University of Utah School of Medicine in Salt Lake City, and Julia Caton, MD, EdM, associate professor of medicine and science education and academic hospitalist at Northwell Health and the Zucker School of Medicine at Hofstra Northwell in Hempstead, N.Y., delivered a rigorous, evidence-based review of the year’s most practice-changing inpatient literature. This year, the presenters embraced the local flavor, framing their session as a “Nashville Bachelorette” itinerary, a nod to the city’s reputation as the bachelorette capital of the U.S. The result was a high-yield review of the 2025 inpatient literature that, much like Nashville, rewarded those who showed up ready to listen.

The Case Against Reflexively Holding Metformin on Admission

Despite decades of accumulated evidence disputing any meaningful link between metformin and lactic acidosis, many hospitalists continue to hold the drug routinely on admission. A propensity-matched observational cohort study by Gallo et al., using Veterans Health Administration data from 114 hospitals and 163,858 hospitalizations, examined whether that practice causes downstream harm.¹ The study revealed massive facility-level variation, with continuation rates ranging from 6.3% to 76.6% across hospitals, driven by institutional culture rather than patient-level risk factors. After matching, patients whose metformin was continued (in those with estimated glomerular filtration rate [eGFR] greater than 30) had lower 90-day hypoglycemia (1.5% versus 1.8%; odds ratio, 0.83), lower 90-day mortality (6.4% versus 7.4%; odds ratio, 0.86), fewer readmissions, and an 11% reduction in new insulin prescriptions at discharge. As this is an observational study, it cannot establish causality, and inpatient safety outcomes were not captured. But taken together with the existing literature, the signal is hard to ignore: Continue metformin through hospitalization in patients with eGFR greater than 30 and no contraindications, and be deliberate when prescribing insulin at discharge.

Adding Aspirin to an Oral Anticoagulant in Stable Coronary Disease Increases Mortality and Bleeding

The AQUATIC trial examined a common clinical scenario.² In this prospective, double-blind, placebo-controlled, randomized, controlled trial across 51 centers in France, 872 patients with chronic coronary syndrome and prior stenting (more than six months ago) were randomized to add aspirin or placebo, while continuing their oral anticoagulation (either direct oral anticoagulants or vitamin K antagonist).

The trial was stopped early due to harm. Adding aspirin to anticoagulation led to higher rates of the primary composite outcome (cardiovascular events: 16.9% versus 12.1%), higher all-cause mortality (13.4% versus 8.4%; hazard ratio, 1.72; P=0.01), and three times the rate of major bleeding (10.2% versus 3.4%). The study was conducted in a predominantly male European population and excluded P2Y12 inhibitors, which limits broad generalizability. Still, for hospitalists reconciling medications in patients with stable coronary disease on anticoagulation at least six months post percutaneous coronary intervention, the evidence now clearly favors stopping aspirin.

Quick Hits: Calcium, Hyponatremia, and Steroid-Induced Leukocytosis

The presenters next delivered three rapid-fire practice pearls covering common inpatient laboratory and clinical scenarios.

First, stop correcting calcium for albumin; use total calcium or ionized calcium (iCa). A cross-sectional study of 22,658 patients with simultaneous total and iCa found that total calcium agreed with iCa 74% of the time, compared to only 63% for the correction formula. Crucially, the formula performed worst in hypoalbuminemia, the exact scenario for which it is most commonly used.³ Use total or iCa directly.

Second, urea is an effective and underused treatment for the syndrome of inappropriate antidiuretic hormone secretion (SIADH). A meta-analysis of 16 observational studies (518 patients) found that urea raised serum sodium by an average of 9.1 mEq/L, comparable to fluid restriction or vaptans, and was associated with less overcorrection than has been reported with vaptans.⁴ Consider it a safe, effective second-line option after fluid restriction.

Third, understanding the expected steroid-induced leukocytosis helps distinguish a drug effect from a new infectious process. A retrospective cohort of 28,425 non-surgical inpatients found that high-dose steroids raise the white blood cell count (WBC) by approximately 5 × 109/L, peaking at 48 hours. Lower-dose steroids, however, produce no clinically meaningful rise in WBC.⁵ Deviations from this pattern, such as earlier spikes, higher counts, or elevations on lower doses, should prompt evaluation for alternative causes.

Reduce the Apixaban Dose after Six Months in Patients with Cancer-Associated VTE

The API-CAT trial fills a genuine gap for hospitalists managing patients with active malignancy and venous thromboembolism (VTE). How long and at what dose should anticoagulation continue in patients with active cancer and VTE?⁶ This international, prospective, double-blind, non-inferiority, randomized, controlled trial randomized 1,766 patients with active cancer who had completed at least six months of anticoagulation without documented recurrent VTE to either reduced-dose apixaban (2.5 mg twice daily) or full-dose apixaban (5 mg twice daily) for 12 months. Recurrent VTE occurred in 2.1% of the reduced-dose group versus 2.8% in the full-dose group, meeting the non-inferiority threshold. Clinically relevant bleeding was lower with reduced dosing (12.1% versus 15.6%). Two-thirds of enrolled patients had metastatic disease, making these results highly applicable to the complex oncology patients commonly managed on hospitalist services. The main limitations are the absence of race and ethnicity data and the 12-month study window, leaving questions about longer durations unanswered. The practical takeaway is: after six months of full-dose therapy, reduce to apixaban 2.5 mg twice daily.

Check the CRP Before Deciding on Steroids in Pneumonia

The decision to add corticosteroids in community-acquired pneumonia has been complicated for years by inconsistent definitions of severity and conflicting society guidance. Smit and colleagues approached this with an individual-patient-data meta-analysis of eight randomized controlled trials, applying machine learning to identify which of 20 baseline variables predicted treatment response.⁷ C-reactive protein (CRP) was the lone predictor.

Among patients with CRP greater than 204 mg/L, adjuvant steroids reduced 30-day mortality from 13% to 6.1%, with a number needed to treat of 14. Below that threshold, there was no mortality benefit, and there was potential for harm. Pneumonia severity measured by Pneumonia Severity Index score did not predict steroid response, which runs counter to a widely held belief. The study is limited by heterogeneous steroid regimens, variability in CRP assays across participating institutions, and variable timing of CRP measurement, so the exact cutoff of 204 mg/L should not be treated as a hard line. The bottom line: for hospitalized pneumonia patients with markedly elevated CRP (roughly above 200 mg/L), consider adjuvant corticosteroids. For those with lower CRP, withhold them.

Consider Starting Carvedilol in Patients with Cirrhosis Admitted With New Ascites Without High-Risk Varices

The convention of reserving nonselective beta-blockers for secondary prevention of variceal hemorrhage may be due for revision. In the CARVE-AS trial, a prospective, single-center, open-label, randomized, controlled trial by Khajuria and colleagues, 104 patients with cirrhosis and new-onset uncomplicated ascites (Child-Pugh Class B, no high-risk varices) were randomized to carvedilol plus standard medical therapy or standard therapy alone.⁸ At one year, the composite rate of ascites-related complications was 38% in the carvedilol group versus 67% in the control group. Refractory ascites occurred in 17.3% versus 34.6%; acute kidney injury in 34.6% versus 63.4%. Mortality and portal pressure gradient improvement also favored carvedilol. The drug was well tolerated at a mean dose of 12.5 mg/day, achieving a mean heart rate of 64 bpm. The limitations are that this is a small (n=104), single-center, open-label trial, and its findings need replication before practice fully shifts. Even so, for hospitalists admitting patients with cirrhosis and a first episode of ascites without high-risk varices, this trial presents a compelling reason to consider starting carvedilol.

At Hospital Discharge, Oral and Injectable Naltrexone Are Equally Effective for AUD

Alcohol use disorder (AUD) affects a substantial proportion of hospitalized patients. A single-center randomized trial enrolled 248 patients with AUD (identified through clinician referral or addiction consult) and compared daily oral naltrexone (titrated to 50 to 100 mg) to extended-release injectable naltrexone (380 mg monthly intramuscular) initiated at hospital discharge.⁹ At three months, both groups showed meaningful reductions in heavy drinking days, with no statistically significant difference between formulations (P=0.14). Healthcare utilization and adverse events were comparable across arms. The study was unblinded, single-site, and enrolled a predominantly male English-speaking population, limiting generalizability. But the practical takeaway is that the choice between oral and injectable naltrexone can be guided by patient preference, adherence concerns, and formulary access, rather than a hierarchy of efficacy. The inpatient setting represents a valuable opportunity to initiate treatment.

Dr. Imoisili

Dr. Pawlowski

Dr. Choudry

Dr. Imoisili is a hospitalist and medical director of the hematology and oncology unit at MedStar Washington Hospital Center in Washington, D.C. Dr. Pawlowski is a hospitalist and unit medical director at MedStar Washington Hospital Center and an instructor of medicine at Georgetown University School of Medicine, both in Washington, D.C. Dr. Choudry is an academic hospitalist and associate division chief of hospital medicine at MedStar Washington Hospital Center and an assistant professor of medicine at Georgetown University School of Medicine, both in Washington, D.C.

References

1. Gallo RJ, et al. Inpatient metformin utilization and post-hospitalization clinical outcomes: an observational cohort study. J Gen Intern Med. 2025. doi:10.1007/s11606-025-09384-y.
2. Lemesle G, et al. Aspirin in patients with chronic coronary syndrome receiving oral anticoagulation. N Engl J Med. 2025;393(16):1578-1588. doi:10.1056/NEJMoa2507532.
3. Desgagnés N, et al. Use of albumin-adjusted calcium measurements in clinical practice. JAMA Netw Open. 2025;8(1):e2455251. doi:10.1001/jamanetworkopen.2024.55251. Erratum in: JAMA Netw Open. 20253;8(3):e256345. doi:10.1001/jamanetworkopen.2025.6345.
4. Chander S, et al. Urea to treat hyponatremia due to syndrome of inappropriate antidiuretic hormone secretion: a systematic review and meta-analysis. Am J Kidney Dis. 2025;85(3):303-319. doi:10.1053/j.ajkd.2024.07.011.
5. Sullivan E, et al. Elevation in white blood cell count after corticosteroid use in noninfected hospitalized patients. J Hosp Med. 2025;20(8):824-828. doi:10.1002/jhm.70008.
6. Mahé I, et al. Extended reduced-dose apixaban for cancer-associated venous thromboembolism. N Engl J Med. 2025;392(14):1363-1373. doi:10.1056/NEJMoa2416112.
7. Smit JM, et al. Predicting benefit from adjuvant therapy with corticosteroids in community-acquired pneumonia: a data-driven analysis of randomised trials. Lancet Respir Med. 2025;13(3):221-233. doi:10.1016/S2213-2600(24)00405-3.
8. Khajuria R, et al. Efficacy and safety of carvedilol in cirrhosis patients with new-onset uncomplicated ascites without high-risk esophageal varices (CARVE-AS trial). Am J Gastroenterol. 2025. Epub ahead of print. doi:10.14309/ajg.0000000000003650.
9. Magane KM, et al. Oral vs extended-release injectable naltrexone for hospitalized patients with alcohol use disorder: a randomized clinical trial. JAMA Intern Med. 2025;185(6):635-645. doi:10.1001/jamainternmed.2025.0522.