Objectives: Sepsis is a leading global cause of mortality, and predicting its outcomes is vital for improving patient care. This study explored the capabilities of ChatGPT, a state-of-the-art natural language processing model, in predicting in-hospital mortality for sepsis patients. Methods: This study utilized data from the Korean Sepsis Alliance (KSA) database, collected between 2019 and 2021, focusing on adult intensive care unit (ICU) patients and aiming to determine whether ChatGPT could predict all-cause mortality after ICU admission at 7 and 30 days. Structured prompts enabled ChatGPT to engage in in-context learning, with the number of patient examples varying from zero to six. The predictive capabilities of ChatGPT-3.5-turbo and ChatGPT-4 were then compared against a gradient boosting model (GBM) using various performance metrics. Results: From the KSA database, 4,786 patients formed the 7-day mortality prediction dataset, of whom 718 died, and 4,025 patients formed the 30-day dataset, with 1,368 deaths. Age and clinical markers (e.g., Sequential Organ Failure Assessment score and lactic acid levels) showed significant differences between survivors and non-survivors in both datasets. For 7-day mortality predictions, the area under the receiver operating characteristic curve (AUROC) was 0.70–0.83 for GPT-4, 0.51–0.70 for GPT-3.5, and 0.79 for GBM. The AUROC for 30-day mortality was 0.51–0.59 for GPT-4, 0.47–0.57 for GPT-3.5, and 0.76 for GBM. Zero-shot predictions using GPT-4 for mortality from ICU admission to day 30 showed AUROCs from the mid-0.60s to 0.75 for GPT-4 and mainly from 0.47 to 0.63 for GPT-3.5. Conclusions GPT-4 demonstrated potential in predicting short-term in-hospital mortality, although its performance varied across different evaluation metrics.

Background: Worldwide, sepsis is the leading cause of death in hospitals. If mortality rates in patients with sepsis can be predicted early, medical resources can be allocated efficiently. We constructed machine learning (ML) models to predict the mortality of patients with sepsis in a hospital emergency department. Methods: This study prospectively collected nationwide data from an ongoing multicenter cohort of patients with sepsis identified in the emergency department. Patients were enrolled from 19 hospitals between September 2019 and December 2020. For acquired data from 3,657 survivors and 1,455 deaths, six ML models (logistic regression, support vector machine, random forest, extreme gradient boosting [XGBoost], light gradient boosting machine, and categorical boosting [CatBoost]) were constructed using fivefold cross-validation to predict mortality. Through these models, 44 clinical variables measured on the day of admission were compared with six sequential organ failure assessment (SOFA) components (PaO 2 /FIO 2 [PF], platelets (PLT), bilirubin, cardiovascular, Glasgow Coma Scale score, and creatinine).The confidence interval (CI) was obtained by performing 10,000 repeated measurements via random sampling of the test dataset. All results were explained and interpreted using Shapley’s additive explanations (SHAP). Results: Of the 5,112 participants, CatBoost exhibited the highest area under the curve (AUC) of 0.800 (95% CI, 0.756–0.840) using clinical variables. Using the SOFA components for the same patient, XGBoost exhibited the highest AUC of 0.678 (95% CI, 0.626–0.730). As interpreted by SHAP, albumin, lactate, blood urea nitrogen, and international normalization ratio were determined to significantly affect the results. Additionally, PF and PLTs in the SOFA component significantly influenced the prediction results. Conclusion Newly established ML-based models achieved good prediction of mortality in patients with sepsis. Using several clinical variables acquired at the baseline can provide more accurate results for early predictions than using SOFA components. Additionally, the impact of each variable was identified.

Background: Prolonged length of hospital stay (LOS) is associated with an increased risk of hospital-acquired conditions and worse outcomes. We conducted a nationwide, multicenter, retrospective cohort study to determine whether prolonged hospitalization before developing sepsis has a negative impact on its prognosis. Methods: We analyzed data from 19 tertiary referral or university-affiliated hospitals between September 2019 and December 2020. Adult patients with confirmed sepsis during hospitalization were included. In-hospital mortality was the primary outcome. The patients were divided into two groups according to their LOS before the diagnosis of sepsis: early- (< 5 days) and late-onset groups (≥ 5 days). Conditional multivariable logistic regression for propensity score matched-pair analysis was employed to assess the association between lateonset sepsis and the primary outcome. Results: A total of 1,395 patients were included (median age, 68.0 years; women, 36.3%).The early- and late-onset sepsis groups comprised 668 (47.9%) and 727 (52.1%) patients.Propensity score-matched analysis showed an increased risk of in-hospital mortality in the late-onset group (adjusted odds ratio [aOR], 3.00; 95% confidence interval [CI], 1.69–5.34).The same trend was observed in the entire study population (aOR, 1.85; 95% CI, 1.37–2.50).When patients were divided into LOS quartile groups, an increasing trend of mortality risk was observed in the higher quartiles (Pfor trend < 0.001). Conclusion Extended LOS before developing sepsis is associated with higher in-hospital mortality. More careful management is required when sepsis occurs in patients hospitalized for ≥ 5 days.

Background: Administration of adequate antibiotics is crucial for better outcomes in sepsis. Because no uniform tool can accurately assess the risk of multidrug-resistant (MDR) pathogens, a local antibiogram is necessary. We aimed to describe the antibiogram of MDR bacteria based on locations of sepsis onset in South Korea. Methods: We performed a prospective observational study of adult patients diagnosed with sepsis according to Sepsis-3 from 19 institutions (13 tertiary referral and 6 universityaffiliated general hospitals) in South Korea. Patients were divided into four groups based on the respective location of sepsis onset: community, nursing home, long-term-care hospital, and hospital. Along with the antibiogram, risk factors of MDR bacteria and drug-bug match of empirical antibiotics were analyzed. Results: MDR bacteria were detected in 1,596 (22.7%) of 7,024 patients with gram-negative predominance. MDR gram-negative bacteria were more commonly detected in long-termcare hospital- (30.4%) and nursing home-acquired (26.3%) sepsis, whereas MDR grampositive bacteria were more prevalent in hospital-acquired (10.9%) sepsis. Such findings were consistent regardless of the location and tier of hospitals throughout South Korea. Patients with long-term-care hospital-acquired sepsis had the highest risk of MDR pathogen, which was even higher than those with hospital-acquired sepsis (adjusted odds ratio, 1.42; 95% confidence interval, 1.15–1.75) after adjustment of risk factors. The drug-bug match was lowest in patients with long-term-care hospital-acquired sepsis (66.8%). Conclusion Gram-negative MDR bacteria were more common in nursing home- and long-term-care hospital-acquired sepsis, whereas gram-positive MDR bacteria were more common in hospital-acquired settings in South Korea. Patients with long-term-care hospitalacquired sepsis had the highest the risk of MDR bacteria but lowest drug-bug match of initial antibiotics. We suggest that initial antibiotics be carefully selected according to the onset location in each patient.

Background: This study aimed to evaluate whether the effect of tachycardia varies according to the degree of tissue perfusion in septic shock. Methods: Patients with septic shock admitted to the intensive care units were categorized into the tachycardia (heart rate > 100 beats/min) and non-tachycardia (≤ 100 beats/min) groups. The association of tachycardia with hospital mortality was evaluated in each subgroup with low and high lactate levels, which were identified through a subpopulation treatment effect pattern plot analysis. Results: In overall patients, hospital mortality did not differ between the two groups (44.6% vs. 41.8%, P = 0.441), however, tachycardia was associated with reduced hospital mortality rates in patients with a lactate level ≥ 5.3 mmol/L (48.7% vs. 60.3%, P = 0.030; adjusted odds ratio [OR], 0.59, 95% confidence interval [CI], 0.35–0.99, P = 0.045), not in patients with a lactate level < 5.3 mmol/L (36.5% vs. 29.7%, P = 0.156; adjusted OR, 1.39, 95% CI, 0.82–2.35, P = 0.227). Conclusion In septic shock patients, the effect of tachycardia on hospital mortality differed by serum lactate level. Tachycardia was associated with better survival in patients with significantly elevated lactate levels.

Background/Aims: Patients in the intensive care unit (ICU) are at high risk for developing pressure injuries, which can cause severe complications and even increase the mortality risk. Therefore, prevention of pressure injuries is most important. In this study, we investigated the risk factors of pressure injury development in patients admitted to the ICU. Methods: We retrospectively analyzed patients ages > 18 years admitted to the medical ICU in a tertiary hospital between January and December 2019. We collected patient baseline characteristics, medications received, mechanical ventilation or hemodialysis use, laboratory findings, and date of pressure injury onset and characteristics. Results: We analyzed 666 patients who did not have pressure injuries at ICU admission. Pressure injuries developed in 102 patients (15%). The risk of pressure injury development increased as the administration days for neuromuscular blocking agents (NMBAs; odds ratio [OR], 1.138; p = 0.019) and opioids (OR, 1.084; p = 0.028) increased, and if the patient had problem with friction and shear (OR, 2.203; p = 0.011). Conclusions The prolonged use of NMBAs and opioids can increase the risk of pressure injury development. Because these medications are associated with immobilization, using both should be minimized and patient early mobilization should be promoted. Among the Braden subscales, “friction and shear” was associated with the development of pressure injuries in ICU patients.

BACKGROUND: Hepatitis B virus (HBV) infection leads to hepatic and extrahepatic manifestations including chronic kidney disease (CKD). However, the association between HBV and CKD is not clear. This study investigated the association between chronic HBV infection and CKD in a nationwide multicenter study. METHODS: A total of 265,086 subjects who underwent health-check examinations in 33 hospitals from January 2015 to December 2015 were enrolled. HBV surface antigen (HBsAg) positive cases (n = 10,048), and age- and gender-matched HBsAg negative controls (n = 40,192) were identified. CKD was defined as a glomerular filtration rate (GFR) < 60 mL/min/1.73 m² or proteinuria as at least grade 2+ of urine protein. RESULTS: HBsAg positive cases showed a significantly higher prevalence of GFR < 60 mL/min/1.73 m² (3.3%), and proteinuria (18.9%) than that of the controls (2.6%, P < 0.001, and 14.1%, P < 0.001, respectively). In the multivariate analysis, HBsAg positivity was an independent factor associated with GFR < 60 mL/min/1.73 m² along with age, blood levels of albumin, bilirubin, anemia, and hemoglobin A1c (HbA1c). Likewise, HBsAg positivity was an independent factor for proteinuria along with age, male, blood levels of bilirubin, protein, albumin, and HbA1c. A subgroup analysis showed that HBsAg positive men but not women had a significantly increased risk for GFR < 60 mL/min/1.73 m². CONCLUSION: Chronic HBV infection was significantly associated with a GFR < 60 mL/min/1.73 m² and proteinuria (≥ 2+). Therefore, clinical concern about CKD in chronic HBV infected patients, especially in male, is warranted.
Anemia ; Antigens, Surface ; Bilirubin ; Case-Control Studies* ; Female ; Glomerular Filtration Rate ; Hepatitis B Surface Antigens ; Hepatitis B virus ; Hepatitis B, Chronic* ; Hepatitis, Chronic* ; Humans ; Male ; Multivariate Analysis ; Prevalence ; Proteinuria ; Renal Insufficiency, Chronic*

PURPOSE: To evaluate the efficacy and safety of once-daily ciclesonide in comparison to both levocetirizine alone, and a ciclesonide/levocetirizine combination in patients with seasonal allergic rhinitis (SAR) and perennial allergic rhinitis (PAR). METHODS: Subjects exhibiting moderate to severe allergic rhinitis for longer than 1 year were randomized in an open-label, 3-arm, parallel group, multicenter study. Subjects received 200 microg ciclesonide, 5 mg levocetirizine, or a combination of both. Changes from baseline until the end-of-study visit (2 weeks following) were evaluated by reflective total nasal symptom scores (rTNSSs), reflective total ocular symptom scores (rTOSSs), physician-assessed overall nasal signs and symptoms severity (PANS), and rhinoconjunctivitis quality-of-life questionnaires (RQLQ). RESULTS: Significant improvements in rTNSS, PANS, and RQLQ in the ciclesonide monotherapy group were observed in comparison to the levocetirizine alone group. Three individual symptoms of rTNSS, including runny nose, nasal itching, and congestion, were improved in the ciclesonide-treated group. rTOSS scores for ciclesonide monotherapy improved from baseline, but no superiority over levocetirizine was shown. The absolute score and changes in rTNSS and PANS were positively correlated. Ciclesonide spray was more effective than levocetirizine in reducing nasal symptoms in both SAR and PAR patients. Ciclesonide and levocetrizine were well tolerated alone and in combination. CONCLUSIONS: Our results provide support for an AR and its Impact on Asthma (ARIA) recommendation stipulating that ciclesonide is superior to levocetirizine for the treatment of AR, with tolerable safety. Addition of levocetirizine to ciclesonide did not give further clinical benefit over monotherapy.
Asthma ; Estrogens, Conjugated (USP) ; Humans ; Nose ; Pruritus ; Rhinitis* ; Rhinitis, Allergic, Seasonal ; Surveys and Questionnaires

BACKGROUND: There are conflicting findings regarding the association between hepatitis B (HB) virus (HBV) infection and atherosclerosis. CASE REPORT: A 34-year-old man was admitted for transient dysarthria and facial palsy. Ten years previously he had been diagnosed with HBV infection and treated with lamivudine (100 mg/day). Reactivation of HBV was detected 6 months before this recent admission. Serologic tests revealed that he was positive for HB early antigen, HB surface antigen, and anti-HB core. Brain magnetic resonance images were normal, but magnetic resonance angiograms revealed severe stenosis of the right middle cerebral artery, both external carotid arteries, and the basilar artery. CONCLUSIONS: This case report reveals that a young patient with reactivated HBV developed multiple arterial stenoses even though he had no risk factors for this disease.
Adult ; Antigens, Surface ; Atherosclerosis ; Brain ; Carotid Artery, External ; Constriction, Pathologic ; Dysarthria ; Facial Paralysis ; Hepatitis ; Hepatitis B ; Hepatitis B virus ; Humans ; Lamivudine ; Magnetic Resonance Spectroscopy ; Middle Cerebral Artery ; Risk Factors ; Serologic Tests ; Viruses

BACKGROUND: Discontinuation of antiepileptic drug (AED) is valuable in patients of epilepsy who have been seizure free for a considerable time. However, there are few established methods to successfully stop AED. The purpose of study was to find the prognostic factors and the method of desirable AED withdrawal. METHODS: Seventy-eight patients who were seizure free for more than 18 months were enrolled in this retrospective study. The patients were followed for more than 2 years after starting withdrawal, or until seizure recurrence within 2 years. To clarify the prognostic factors affecting the seizure recurrence related to the withdrawal, statistical analyses were done about AED types, EEG before withdrawal, sex, age, onset age of seizure, febrile convulsion, family history, brain CT/MRI, trauma history, loss of consciousness history, seizure nature, and duration of seizure free. Furthermore, survival analysis between groups over the duration of withdrawal period and AED decrement rate was done to elucidate the most successful methods of AED withdrawal. RESULTS: Fifty-seven patients (73%) experienced seizure recurrence after starting withdrawal within 2 years. Over 12 months of withdrawal duration (p=0.037) and under 20 percent per month of the decrement rate of carbamazepine (p=0.019) were related to reduce the seizure recurrence. These were also confirmed via survival curve analysis. CONCLUSIONS: This study is a new attempt to apply reduction rate in withdrawal. AED withdrawal duration and decrement rate are important factor to prognostic factors of desirable AED withdrawal and minimize the recurrence after withdrawal.
Age of Onset ; Anticonvulsants ; Brain ; Carbamazepine ; Electroencephalography ; Epilepsy ; Humans ; Recurrence ; Retrospective Studies ; Seizures ; Seizures, Febrile ; Unconsciousness