Predictors for extubation failure in COVID-19 patients using a machine learning approach.

INTRODUCTION: Determining the optimal timing for extubation can be challenging in the intensive care. In this study, we aim to identify predictors for extubation failure in critically ill patients with COVID-19. METHODS: We used highly granular data from 3464 adult critically ill COVID patients in the multicenter Dutch Data Warehouse, including demographics, clinical observations, medications, fluid balance, laboratory values, vital signs, and data from life support devices. All intubated patients with at least one extubation attempt were eligible for analysis. Transferred patients, patients admitted for less than 24 h, and patients still admitted at the time of data extraction were excluded. Potential predictors were selected by a team of intensive care physicians. The primary and secondary outcomes were extubation without reintubation or death within the next 7 days and within 48 h, respectively. We trained and validated multiple machine learning algorithms using fivefold nested cross-validation. Predictor importance was estimated using Shapley additive explanations, while cutoff values for the relative probability of failed extubation were estimated through partial dependence plots. RESULTS: A total of 883 patients were included in the model derivation. The reintubation rate was 13.4% within 48 h and 18.9% at day 7, with a mortality rate of 0.6% and 1.0% respectively. The grandient-boost model performed best (area under the curve of 0.70) and was used to calculate predictor importance. Ventilatory characteristics and settings were the most important predictors. More specifically, a controlled mode duration longer than 4 days, a last fraction of inspired oxygen higher than 35%, a mean tidal volume per kg ideal body weight above 8 ml/kg in the day before extubation, and a shorter duration in assisted mode (< 2 days) compared to their median values. Additionally, a higher C-reactive protein and leukocyte count, a lower thrombocyte count, a lower Glasgow coma scale and a lower body mass index compared to their medians were associated with extubation failure. CONCLUSION: The most important predictors for extubation failure in critically ill COVID-19 patients include ventilatory settings, inflammatory parameters, neurological status, and body mass index. These predictors should therefore be routinely captured in electronic health records.

Evaluation of ophthalmology residents' self-assessments and peer assessments in simulated surgery.

OBJECTIVE: To evaluate the accuracy of ophthalmology residents' self-assessment and peer assessment of surgical skills in a simulation setting. DESIGN: Simulation laboratory assessment. PARTICIPANTS: Ophthalmology residents novice to cataract surgery. METHODS: A modified International Council of Ophthalmology's Ophthalmology Surgical Competency Assessment Rubric: Phacoemulsification structured assessment tool for simulated cataract surgery was established by conventional Delphi method. Residents completed 10 independent simulated surgeries that were video-recorded. Two experts graded the videos using the assessment tool. Participants performed self-assessment of their own 10 videos, and peer assessment of 10 of their peers' videos. RESULTS: Nine cataract surgery experts provided feedback and modifications for the assessment tool. Agreement for the first round of the Delphi method ranged from 55.56% to 100%. Second round agreement was 80% or greater for all answers. The final assessment tool comprised (i) 4 procedural items scored from 0 (not performed) to 7 (competent), and (ii) a global rating scale (GRS) requiring yes/no answers to 4 performance-related questions. Eight residents participated in the study. There was excellent expert inter-rater reliability intraclass correlation ((ICC) = 0.844, 0.875, 0.809, 0.844) and fair to excellent inter-rater reliability between expert and peer scores (ICC = 0.702, 0.831, 0.521, 0.423), but systematic disagreement (ICC = -0.428, -0.038) or poor inter-rater reliability (ICC = 0.298, 0.362) between expert and self-scores. There was poor agreement for all GRS questions (κ statistic < 0.40) except 2 comparisons. CONCLUSIONS: In the simulation setting, experts were able to reliably assess trainees' performance using the assessment tool. Participants demonstrated inconsistency in assessing their own skills; however, they were adequate at assessing their peers' overall performance.

Temporal biomarker profiles and their association with ICU acquired delirium: a cohort study.

BACKGROUND: Neuroinflammation is thought to play an important role in the pathogenesis of ICU-acquired delirium, but the association between inflammatory and brain-specific proteins and ICU delirium is poor. We investigated whether or not serial determinations of markers may improve this association. METHODS: Critically ill patients with a high risk of ICU delirium and with an ICU length of stay of at least 6 days were included in the study. Blood was drawn on days 1, 2, 4 and 6 after ICU admission and analyzed for different markers of inflammation and several brain proteins. Differences in courses over time prior to and following the onset of delirium and absolute differences over time were analyzed in patients with and without delirium using repeated measurement analysis of variance. In addition, a cross-sectional analysis of levels of these markers before the first onset of delirium was performed. RESULTS: Fifty patients were included in this study. In the longitudinal analysis, there were no differences in the levels of any of the markers immediately prior to and following the onset of delirium, but overall, median levels of adiponectin (9019 (IQR 5776-15,442) vs. 6148 (IQR 4447-8742) ng/ml, p = 0.05) were significantly higher in patients with delirium compared to patients without delirium. In the cross-sectional analysis, median levels of the brain protein Tau (90 (IQR 46-224) vs. 31 (IQR 31-52) pg/ml, p = 0.009) and the ratio Tau/amyloid ß1-42 (1.42 ((IQR 0.9-2.57) vs. 0.68 (IQR 0.54-0.96), p = 0.003) were significantly higher in patients with hypoactive delirium compared to patients without. Levels of neopterin (111 (IQR 37-111) vs. 29 (IQR 16-64) mmol/l, p = 0.004) and IL-10 (28 (IQR 12-39) vs. 9 (IQR 4-12) pg/ml, p = 0.001) were significantly higher in patients with hypoactive delirium compared to patients with mixed-type delirium. CONCLUSIONS: While there are differences in markers (adiponectin and several brain proteins) between patients with and without delirium, the development of delirium is not preceded by a change in the biomarker profile of inflammatory markers or brain proteins. Patients with hypoactive delirium account for the observed differences in biomarkers. TRIAL REGISTRATION: ClinicalTrials.gov, NCT 01274819 . Registered on 12 January 2011.

A SECOND CASE OF BILATERAL RHEGMATOGENOUS RETINAL DETACHMENTS REPAIRED WITH SIMULTANEOUS BILATERAL PNEUMATIC RETINOPEXY.

BACKGROUND/PURPOSE: To present a case of a patient with simultaneous bilateral retinal detachments treated successfully with bilateral pneumatic retinopexy. STUDY DESIGN: Case report. RESULTS: This is a case of an otherwise healthy 49-year-old woman with no remarkable ocular history that presented with simultaneous phakic superior bilateral rhegmatogenous retinal detachments. Treatment on the day of presentation included laser retinopexy of the inferior lattice degeneration in the left eye and bilateral intravitreal injection of 0.4 cc of 100% C3F8 gas preceded by topical anesthesia. After 48 hours, both retinas were completely reattached, and bilateral laser retinopexy was performed to the superior tears. DISCUSSION: After a review of the literature, the authors could find only two reported cases of simultaneous bilateral retinal detachments treated successfully with pneumatic retinopexy. This is not only a cost-effective procedure but also allows treatment when there is no immediate operating room availability or a when a quick referral for surgery is not possible.

Embolization as treatment of choice for bleeding peptic ulcers in high-risk patients.

BACKGROUND/AIM: Peptic ulcers are the most common cause of acute upper gastrointestinal bleedings in the digestive tract. Most patients are poor surgical candidates. The aim was to describe the efficacy of embolization as the treatment of choice for bleeding peptic ulcers in high-risk patients when endoscopic treatment failed. METHODS: All patients who underwent a selective embolization of branches of the superior mesenteric artery and/or branches of the gastroduodenal artery for a bleeding peptic ulcer in the period January 2004 until December 2007 were included in this retrospective descriptive study. We examined the known risk factors for surgery and mortality in upper gastrointestinal bleeding due to peptic ulcers and describe the clinical course and outcome. Primary endpoints were the primary technical success and the clinical success rates. The secondary endpoint was the 30-day mortality. RESULTS: 16 patients were included. All patients had at least two risk factors for surgery and mortality. The clinical success rate was 81% (13/16). The first embolization failed in 3 patients, 1 was successful re-embolized and 2 were operated upon without re-embolization. The primary technical success rate, i.e. bleedings controlled by radiologic intervention, was 88% (14/16). 6 patients died due to non-embolization-related problems; 5 of them developed upper gastrointestinal bleeding during a stay in the hospital. CONCLUSION: Embolization was a successful, minimal invasive alternative for surgical intervention in high-risk patients with upper gastrointestinal bleeding after failure of endoscopic treatment.