Bisso, Indalecio Carboni

Background: The new Global definition of ARDS recently introduced a subgroup known as non-intubated ARDS. This study aimed to assess the risk of progression from noninvasive oxygen support to intubation and ARDS severity based on the S pO 2 /F IO 2 among non-intubated subjects with ARDS. Methods: This retrospective study included subjects with COVID-19 admitted to 7 hospitals (5 in the United States and 2 in Argentina) from January 2020–January 2023. Subjects meeting the new non-intubated ARDS definition (high-flow nasal cannula [HFNC] with an S pO 2 /F IO 2 ≤315 [with S pO 2 ≤97%] or a P aO 2 /F IO 2 ≤300 mm Hg while receiving ≥30 L/min O 2 via HFNC) were included. The study evaluated the proportion of subjects who progressed to intubation, severity levels using the S pO 2 /F IO 2 cutoff proposed in the new ARDS definition, and mortality. Results: Nine hundred sixty-five non-intubated subjects with ARDS were included, of whom 27% ( n = 262) progressed to meet the Berlin criteria within a median of 3 d (interquartile range 2–6). The overall mortality was 23% (95% CI 20–26) ( n = 225), and among subjects who progressed to the Berlin criteria, it was 37% (95% CI 31–43) ( n = 98). Additionally, the worst S pO 2 /F IO 2 within 1 d of ARDS diagnosis was correlated with mortality, with mortality rates of 26% (95% CI 23–30) ( n = 177) for subjects with S pO 2 /F IO 2 ≤148, 17% (95% CI 12–23) ( n = 38) for those with S pO 2 /F IO 2 between 149–234, and 16% (95% CI 8–28) ( n = 10) for subjects maintaining an S pO 2 /F IO 2 higher than 235 ( P < .001). Conclusions: The non-intubated ARDS criteria encompassed a broader spectrum of subjects with lower in-hospital mortality compared to the Berlin criteria. The S pO 2 /F IO 2 and ARDS severity cutoff proposed in the new Global ARDS definition were valuable predictors of in-hospital mortality in these subjects.

IMPORTANCE: The scientific community debates Generative Pre-trained Transformer (GPT)-3.5’s article quality, authorship merit, originality, and ethical use in scientific writing. OBJECTIVES: Assess GPT-3.5’s ability to craft the background section of critical care clinical research questions compared to medical researchers with H-indices of 22 and 13. DESIGN: Observational cross-sectional study. SETTING: Researchers from 20 countries from six continents evaluated the backgrounds. PARTICIPANTS: Researchers with a Scopus index greater than 1 were included. MAIN OUTCOMES AND MEASURES: In this study, we generated a background section of a critical care clinical research question on “acute kidney injury in sepsis” using three different methods: researcher with H-index greater than 20, researcher with H-index greater than 10, and GPT-3.5. The three background sections were presented in a blinded survey to researchers with an H-index range between 1 and 96. First, the researchers evaluated the main components of the background using a 5-point Likert scale. Second, they were asked to identify which background was written by humans only or with large language model-generated tools. RESULTS: A total of 80 researchers completed the survey. The median H-index was 3 (interquartile range, 1–7.25) and most (36%) researchers were from the Critical Care specialty. When compared with researchers with an H-index of 22 and 13, GPT-3.5 was marked high on the Likert scale ranking on main background components (median 4.5 vs. 3.82 vs. 3.6 vs. 4.5, respectively; p < 0.001). The sensitivity and specificity to detect researchers writing versus GPT-3.5 writing were poor, 22.4% and 57.6%, respectively. CONCLUSIONS AND RELEVANCE: GPT-3.5 could create background research content indistinguishable from the writing of a medical researcher. It was marked higher compared with medical researchers with an H-index of 22 and 13 in writing the background section of a critical care clinical research question.

Abstract Background The criteria for the selection of COVID‐19 patients that could benefit most from ECMO organ support are yet to be defined. In this study, we evaluated the predictive performance of ECMO mortality predictive models in patients with COVID‐19. We also performed a cost–benefit analysis depending on the mortality predicted probability. We conducted a retrospective cohort study in COVID‐19 patients who received ECMO at two tertiary care hospitals between March 2020 to July 2021. Materials and Methods We evaluated the discrimination (C‐statistic), calibration (Cox calibration), and accuracy of the prediction of death due to severe ARDS in V‐V ECMO score (PRESERVE), the Respiratory Extracorporeal Membrane Oxygenation Survival Score (RESP) score, and the PREdiction of Survival on ECMO Therapy‐Score (PRESET) score. In addition, we compared the RESP score with Plateau pressure instead of Peak pressure. Results We included a total of 36 patients, 29 (80%) of them male and with a median (IQR) APACHE of 10 (8–15). The PRESET score had the highest discrimination (AUROCs 0.81 [95%CI 0.67–0.94]) and calibration (calibration‐in‐the‐large 0.5 [95%CI −1.4 to 0.3]; calibration slope 2.2 [95%CI 0.7/3.7]). The RESP score with Plateau pressure had higher discrimination than the conventional RESP score. The cost per QALY in the USA, adjusted to life expectancy, was higher than USD 100 000 in patients older than 45 years with a PRESET > 10. Conclusion The PRESET score had the highest predictive performance and could help in the selection of patients that benefit most from this resource‐demanding and highly invasive organ support.