What's new in hospital medicine

The following represent additions to UpToDate from the past six months that were considered by the editors and authors to be of particular interest. The most recent What's New entries are at the top of each subsection.

GENERAL HOSPITAL MEDICINE

Initiation of medications for alcohol use disorder during hospitalizations (May 2024)

Medications for the management of alcohol use disorder are effective but underutilized. Hospitalization may be an opportune time to initiate them. In an observational study of over 6700 individuals hospitalized at least once for alcohol-related disorders, initiation of medications for alcohol use disorder at discharge was associated with a 42 percent lower rate of all-cause death, emergency department visits, and readmission at 30 days and a 51 percent lower rate of alcohol-related emergency visits or hospitalization [1]. These findings support efforts to initiate medications for alcohol use disorder at hospital discharge. (See "Alcohol use disorder: Pharmacologic management", section on 'Patients hospitalized for alcohol-related disorder'.)

HOSPITAL CARDIOVASCULAR MEDICINE

Liberal transfusion strategy for acute myocardial infarction (December 2023)

Restrictive transfusion (transfusing at a lower hemoglobin, typically <7 or 8 g/dL) is appropriate for most patients based on evidence from randomized trials, but trial data for patients with acute myocardial infarction (MI) have been slower to accumulate. In the MINT trial, which randomly assigned 3504 patients with acute MI and anemia to a restrictive or liberal (transfusing for hemoglobin <10 g/dL) strategy, there was a trend toward better outcomes with the liberal strategy without an increased risk of adverse events [2]. We now suggest a liberal strategy for acute MI. A slightly lower hemoglobin may be reasonable for stable, asymptomatic patients, and patients with hemodynamic instability may require a higher hemoglobin. (See "Indications and hemoglobin thresholds for RBC transfusion in adults", section on 'Acute MI'.)

HOSPITAL ENDOCRINOLOGY AND DIABETES

Epidemiology of myxedema coma in the United States (May 2024)

Myxedema coma is a rare presentation of hypothyroidism. A 2024 analysis of a national inpatient database in the United States provides new information on the epidemiology and prognosis of myxedema coma [3]. The estimated incidence in the United States is 2.6 cases per million persons per year. The estimated in-hospital mortality was 6.8 percent, compared with 0.7 percent for patients hospitalized for hypothyroidism without myxedema coma. This mortality rate is substantially lower than prior mortality estimates, which ranged from 30 to 50 percent. The reason for the lower mortality rate in this analysis is uncertain but may be related to inclusion of patients with less severe myxedema coma or to recent improvements in intensive care unit management. (See "Myxedema coma", section on 'Prognosis'.)

HOSPITAL GASTROENTEROLOGY

Antibiotics are required for nonoperative management of uncomplicated appendicitis (April 2024)

In uncomplicated appendicitis, antibiotic therapy alone is a widely accepted alternative to appendectomy. Whether antibiotic therapy can be omitted in some patients is unknown. In a trial that randomly assigned 100 patients with mild appendicitis (defined as white blood cell count <13,000/microliter, C-reactive protein <60 mg/dL) to piperacillin-tazobactam or observation for either disease regression or the need for surgical exploration, antibiotic therapy reduced the need for appendectomy both during the initial hospitalization (28 versus 53 percent) and at three-year follow-up (50 versus 63 percent) [4]. Thus, we continue to suggest routine antibiotic therapy for all patients managed nonoperatively for acute appendicitis. (See "Management of acute appendicitis in adults", section on 'Protocols'.)

Pancreatic stenting for preventing pancreatitis after ERCP (February 2024)

Patients with difficult biliary access are at increased risk for pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP), which may be prevented by prophylactic pancreatic stenting. In a trial including 1950 patients at high risk for post-ERCP pancreatitis, individuals assigned to pancreatic stent placement plus indomethacin had lower rates of post-ERCP pancreatitis compared with those assigned to indomethacin alone (11 versus 15 percent) [5]. Rates of serious adverse events were not significantly different between groups. These data support our approach of prophylactic pancreatic stenting in high-risk patients undergoing ERCP. (See "Management of difficult biliary access during ERCP in adults", section on 'Efficacy'.)

HOSPITAL INFECTIOUS DISEASES

Pyogenic liver abscess and colorectal cancer (February 2024)

Studies from Asia suggest that pyogenic liver abscess is associated with increased incidence of colorectal cancer, but studies outside Asia are scarce. In a 10-year retrospective study from 120 hospitals in the United States, the incidence of colorectal cancer among over 8000 patients with liver abscess was almost fourfold higher during the first six months after diagnosis of liver abscess compared with 23,000 matched controls without liver abscess [6]. The correlation was not observed among patients whose liver abscess was due to cholangitis or cholecystitis, and the type of organism causing the abscess did not correlate with the incidence of cancer. These findings support prompt screening for colorectal cancer in patients with pyogenic liver abscess, particularly in patients without an underlying hepatobiliary cause. (See "Pyogenic liver abscess", section on 'Association with colorectal cancer'.)

Emerging microbiologic colonization in mechanically ventilated patients (January 2024)

Mechanically ventilated patients act as reservoirs for hospital-acquired pathogens, including Staphylococcus, Pseudomonas, and Aspergillus species. However, a recent surveillance study of 51 acute care and long-term health care facilities reported the emergence of two additional species in mechanically ventilated patients, Acinetobacter baumannii (31 percent of patients, and one-half were carbapenem-resistant) and Candida auris (7 percent, and one-third were newly identified) [7]. Clinicians should be aware of emerging microbiologic species in their local facility so that appropriate surveillance can be conducted and antimicrobial therapy initiated, if indicated. (See "Clinical and physiologic complications of mechanical ventilation: Overview", section on 'Aspiration and ventilator-associated pneumonia and microbial colonization'.)

Nasal decolonization in intensive care units (November 2023)

To reduce hospital-acquired infections, many hospitals provide nasal decolonization with either mupirocin or an iodophor to all patients in intensive care units (ICUs). In a cluster-randomized trial in over 130 hospitals that used universal nasal mupirocin and daily chlorhexidine bathing for ICU patients, switching to nasal iodophor was associated with a higher rate of Staphylococcus aureus growth on clinical cultures than continuing with mupirocin [8]. There was no difference in the rate of bloodstream infection from any pathogen. For hospitals that elect to use nasal decolonization in the ICU, we suggest mupirocin rather than iodophors. This practice may be particularly beneficial in ICUs with high rates of S. aureus infections, including methicillin-resistant strains. (See "Nosocomial infections in the intensive care unit: Epidemiology and prevention", section on 'Patient bathing plus decolonization'.)

HOSPITAL NEPHROLOGY

Diagnosis of arginine vasopressin deficiency (central diabetes insipidus) versus primary polydipsia (February 2024)

It can be difficult to distinguish arginine vasopressin deficiency (AVP-D) from primary polydipsia in patients who present with polyuria and polydipsia; if reliable plasma copeptin measurements are available, these two disorders can be differentiated by measuring plasma copeptin after hypertonic saline infusion (to induce hypernatremia) or after intravenous arginine infusion. In a trial that compared both approaches, the diagnostic accuracy was higher when combining copeptin measurement with hypertonic saline infusion as compared with arginine infusion (96 versus 74 percent) [9]. Despite the superior accuracy of hypertonic saline infusion, the response to arginine infusion can be used as the initial test because it is easier to perform, generally preferred by patients, and highly accurate if the postarginine copeptin level remains below 3 pmol/L (which is specific for AVP-D) or rises above 5.2 pmol/L (which is highly specific for primary polydipsia). If the postarginine copeptin level is intermediate, then hypertonic saline infusion can be performed. (See "Evaluation of patients with polyuria", section on 'If water restriction is nondiagnostic'.)

HOSPITAL NEUROLOGY

Reversal strategy for intracerebral hemorrhage associated with direct factor Xa inhibitors (May 2024)

The optimal reversal strategy for direct factor Xa inhibitors (apixaban, edoxaban, rivaroxaban) in acute intracerebral hemorrhage (ICH) is uncertain. In the ANNEXA-I trial, which randomly assigned 530 patients with factor Xa inhibitor-associated ICH to andexanet alfa or standard care (typically including a prothrombin complex concentrate [PCC]), patients assigned to andexanet had higher rates of hemostasis than those assigned to standard therapy (67 versus 53 percent) [10]. However, thrombotic events, including ischemic stroke and myocardial infarction, were more common with andexanet (10.3 versus 5.6 percent). Mortality and functional outcomes at 30 days were similar. Based on these results, we individualize selection of andexanet alfa or PCC for direct factor Xa inhibitor reversal in acute ICH and other life-threatening bleeding; previously, we favored andexanet in most cases. Andexanet may restore hemostasis more effectively than PCC but is associated with higher thrombotic risk. (See "Reversal of anticoagulation in intracranial hemorrhage", section on 'Reversal agent options'.)

Minimally invasive surgery for acute hemispheric intracerebral hemorrhage (May 2024)

Surgical evacuation is used for select patients with large hemispheric intracerebral hemorrhage (ICH) to reduce risk of mortality, but the effect on functional outcomes is less certain. In a clinical trial of 300 patients with acute ICH, functional outcome at 180 days was better in patients assigned to minimally invasive surgery than those assigned to guideline-based medical therapy [11]. The 30-day death rate was also lower in the surgery group (9 versus 18 percent). These results support the role of minimally invasive surgery to improve functional outcomes and reduce mortality risk for properly selected patients with acute ICH. (See "Spontaneous intracerebral hemorrhage: Acute treatment and prognosis", section on 'Supratentorial hemorrhage'.)

Cerebral amyloid angiopathy as a risk for isolated subdural hematoma (February 2024)

Cerebral amyloid angiopathy (CAA) commonly presents with acute intracerebral hemorrhage that may extend into the subarachnoid or subdural spaces in some instances, but the risk of isolated spontaneous subdural hematoma (SDH) from CAA is uncertain. In a retrospective study of data from two large population-based cohorts, CAA was associated with an elevated risk of SDH after adjustment for patient demographics, cardiovascular risks, and antithrombotic medication use [12]. Leptomeningeal amyloid deposition may predispose such patients to spontaneous SDH. These results expand our understanding of the varied hemorrhagic presentations associated with CAA. (See "Cerebral amyloid angiopathy", section on 'Imaging features'.)

Updated guideline on postoperative delirium in adults (February 2024)

The European Society of Anaesthesiology and Intensive Care Medicine has published an updated guideline on postoperative delirium (POD) [13]. Recommendations include preoperatively screening older adults for risk factors for POD and multicomponent nonpharmacological interventions for all patients with risk factors. In addition, review of recent evidence showed that perioperative use of dexmedetomidine was associated with a lower incidence of POD, particularly when administered postoperatively in the intensive care unit. We agree with the recommendations and often use dexmedetomidine in the perioperative period to reduce the incidence of POD in high-risk patients. (See "Perioperative neurocognitive disorders in adults: Risk factors and mitigation strategies", section on 'Intravenous agents associated with lower risk'.)

Time window to start dual antiplatelet therapy for high-risk TIA or minor ischemic stroke (January 2024)

There is evidence from several randomized trials that early initiation of short-term dual antiplatelet therapy (DAPT) for select patients with high-risk transient ischemic attack (TIA) or minor ischemic stroke reduces the risk of recurrent ischemic stroke. The evidence comes from trials that started DAPT within 12 to 24 hours of symptom onset. Results from the recent INSPIRES trial suggest that DAPT is still beneficial when started up to 72 hours after symptom onset [14]. Although the time window is extended by the results from INSPIRES, we start DAPT as soon as possible for patients with high-risk TIA or minor ischemic stroke. (See "Early antithrombotic treatment of acute ischemic stroke and transient ischemic attack", section on 'High-risk TIA and minor ischemic stroke'.)

HOSPITAL PULMONOLOGY AND CRITICAL CARE MEDICINE

Incidence of transfusion-related acute lung injury (April 2024)

Transfusion-related acute lung injury (TRALI) is a potentially fatal complication of transfusion characterized by rapid-onset noncardiogenic pulmonary edema. The incidence is challenging to determine due to differing case definitions and reliance on passive reporting (requiring the clinician to notify the transfusion medicine service). A new meta-analysis that included approximately 176 million transfused blood components provides estimates from active surveillance studies [15]. For red blood cells, TRALI occurred with 0.17 of 10,000 units; for platelets, 0.31 of 10,000 units; and for plasma, 3.19 of 10,000 units (the incidence for plasma was much lower when two outlier studies were removed). TRALI remains rare and has been significantly reduced by mitigation measures such as excluding plasma from multiparous female donors; nevertheless, these numbers suggest it is more common than estimated by passive surveillance. (See "Transfusion-related acute lung injury (TRALI)", section on 'Epidemiology'.)

Ideal oxygen targets in COVID-19 (April 2024)

In patients with acute respiratory failure due to coronavirus-2019 (COVID-19), ideal oxygenation targets are unclear. A recent study of spontaneously breathing or mechanically ventilated hospitalized adults with acute respiratory failure due to COVID-19 reported that targeting an arterial oxygen tension (PaO₂) ≥60 mmHg was associated with more days alive without ventilatory support compared with a target ≥90 mmHg [16]. However, there was no overall mortality benefit. Although the study was limited by lack of blinding and early cessation for slow enrollment, it supports our recommendation of targeting a peripheral oxygen saturation between 90 and 96 percent or PaO₂ ≥60 mmHg, when feasible. (See "COVID-19: Respiratory care of the nonintubated hypoxemic adult (supplemental oxygen, noninvasive ventilation, and intubation)", section on 'Oxygenation targets'.)

Extracorporeal cardiopulmonary resuscitation (December 2023)

Extracorporeal cardiopulmonary resuscitation (ECPR) is being increasingly used, but data are limited and the benefits are uncertain. In a recent meta-analysis of 11 studies (10,000 patients) who underwent CPR, compared with standard CPR, ECPR was associated with decreased in-hospital mortality and increased long-term favorable neurologic outcome and survival at one year [17]. The benefit of ECPR was confined to patients with in-hospital cardiac arrest. These data support the growing practice of ECPR in select patients likely to benefit. (See "Extracorporeal life support in adults: Management of venoarterial extracorporeal membrane oxygenation (V-A ECMO)", section on 'Sudden cardiac arrest (extracorporeal cardiopulmonary resuscitation)'.)

Heart rate control in septic shock (December 2023)

Beta blockade has the potential to limit harm from the adrenergic overdrive that occurs in septic shock. However, data to support heart rate control in patients with septic shock are limited. In a recent, unblinded randomized trial of 126 patients with septic shock-related tachycardia (heart rate ≥95/min) who were receiving norepinephrine, the beta blocker landiolol did not reduce organ failure as measured by the sequential organ failure assessment score [18]. Furthermore, landiolol was associated with increased 28-day mortality compared with standard care (37 versus 25 percent). We continue to avoid the routine use of beta blockers in patients with septic shock. (See "Investigational and ineffective pharmacologic therapies for sepsis", section on 'Heart rate control'.)

QUALITY AND SAFETY

Use of a formal handoff tool during intraoperative transfer of anesthetic care (March 2024)

Intraoperative handoffs of care from one anesthesia provider to another have been associated with adverse outcomes, possibly due to inadequate communication of clinical information. A retrospective study noted that an intraoperative handoff (defined as transfer of care lasting more than 35 minutes) occurred in 40 percent of approximately 120,000 noncardiac surgical cases performed from 2016 to 2021 [19]. Adverse outcomes (defined as a composite of postoperative mortality or major morbidity) occurred more often when a handoff occurred (7.2 versus 6.2 percent with no handoff). After implementation of a structured electronic handoff tool in 2019, a correlation over time was noted between increased handoff tool usage and decreased likelihood of the composite adverse outcome. We use a formal handoff protocol during all phases of perioperative care. (See "Handoffs of surgical patients", section on 'Use of a formal handoff procedure'.)

Tailored clinical alerts in the intensive care unit (March 2024)

Clinical decision support systems are commonly used in hospital settings to assist in medication ordering, but numerous low-yield alerts can lead to alarm fatigue and lack of benefit. In a randomized trial conducted in nine intensive care units (ICUs) in the Netherlands among 9887 patients, ICU-tailored drug-drug interaction alerts reduced administrations of high-risk drug combinations relative to control (26.2 versus 35.6 per 1000 drug administrations, respectively) [20]. Tailoring alert systems to be relevant to the clinical setting may result in meaningful improvements in care. (See "Prevention of adverse drug events in hospitals", section on 'Computerized physician order entry'.)

Diagnostic errors in hospitalized patients (February 2024)

Diagnostic errors are important causes of preventable morbidity and mortality in hospitalized patients. In a retrospective cohort study conducted in 29 hospitals of 2428 adults who were transferred to an intensive care unit (ICU), 23.0 percent were judged to have experienced a diagnostic error [21]. In approximately 80 percent of these patients, errors were thought to have contributed to harm or death. Diagnostic errors in hospitalized patients can have serious consequences and are targets for safety improvements. (See "Diagnostic errors", section on 'Adult medicine'.)