Machine learning for the real-time assessment of left ventricular ejection fraction in critically ill patients: a bedside evaluation by novices and experts in echocardiography

BACKGROUND

Machine learning algorithms have recently been developed to enable the automatic and real-time echocardiographic assessment of left ventricular ejection fraction (LVEF) and have not been evaluated in critically ill patients.

METHODS

Real-time LVEF was prospectively measured in 95 ICU patients with a machine learning algorithm installed on a cart-based ultrasound system. Real-time measurements taken by novices (LVEF_Nov) and by experts (LVEF_Exp) were compared with LVEF reference measurements (LVEF_Ref) taken manually by echo experts.

RESULTS

LVEF_Ref ranged from 26 to 80% (mean 54 ± 12%), and the reproducibility of measurements was 9 ± 6%. Thirty patients (32%) had a LVEF_Ref < 50% (left ventricular systolic dysfunction). Real-time LVEF_Exp and LVEF_Nov measurements ranged from 31 to 68% (mean 54 ± 10%) and from 28 to 70% (mean 54 ± 9%), respectively. The reproducibility of measurements was comparable for LVEF_Exp (5 ± 4%) and for LVEF_Nov (6 ± 5%) and significantly better than for reference measurements (p < 0.001). We observed a strong relationship between LVEF_Ref and both real-time LVEF_Exp (r = 0.86, p < 0.001) and LVEF_Nov (r = 0.81, p < 0.001). The average difference (bias) between real time and reference measurements was 0 ± 6% for LVEF_Exp and 0 ± 7% for LVEF_Nov. The sensitivity to detect systolic dysfunction was 70% for real-time LVEF_Exp and 73% for LVEF_Nov. The specificity to detect systolic dysfunction was 98% both for LVEF_Exp and LVEF_Nov.

CONCLUSION

Machine learning-enabled real-time measurements of LVEF were strongly correlated with manual measurements obtained by experts. The accuracy of real-time LVEF measurements was excellent, and the precision was fair. The reproducibility of LVEF measurements was better with the machine learning system. The specificity to detect left ventricular dysfunction was excellent both for experts and for novices, whereas the sensitivity could be improved.

TRIAL REGISTRATION

NCT05336448. Retrospectively registered on April 19, 2022.

Use of Venous Excess UltraSound (VExUS) score in hyponatraemia management in critically ill patient

Hyponatraemia is the most prevalent electrolyte disorder in the neurocritical care setting and is associated with a significant morbimortality. Cerebral salt wasting and inappropriate antidiuretic hormone secretion syndrome have been classically described as the two most frequent entities responsible for hyponatraemia in neurocritical care patients. An accurate aetiological diagnosis of hypotonic hyponatraemia requires a proper volume status assessment. Nevertheless, determination of volume status based on physical examination, laboratory findings and imaging modalities have several limitations and can lead to improperly diagnosis and hyponatraemia mismanagement. Point-of-care ultrasound (POCUS), specifically Venous Excess UltraSound (VExUS) score, is a fast and valuable tool to evaluate venous congestion at the bedside and identify hypervolaemia, helping the physicians in therapeutic decision making in a patient with hyponatraemia. We report a case where the use of POCUS, and more specifically VExUS, can be helpful in volume status assessment, complementing the complex management of multifactorial hyponatraemia in a neurocritical patient.

Impact of Phosphatemia Variability in Neurological Outcomes in Patients With Spontaneous Subarachnoid Hemorrhage

Introduction: Electrolyte disturbances, such as dysnatremia, hypokalemia, and hypomagnesemia, are frequently observed during acute spontaneous subarachnoid hemorrhage (sSAH). However, there are limited data concerning hypophosphatemia.

Objective: To analyze the frequency of phosphate (Pi) disturbances in sSAH patients and assess their influence on neurological outcomes compared with that in patients without sSAH.

Methods: We conducted a retrospective study of patients with sSAH admitted to a neurocritical care unit in two years. We also included nonneurocritical patients admitted to a general intensive care unit (ICU). Serum Pi levels and daily Pi repletion data were collected during the first 10 days after admission. The primary endpoint was neurologic outcome using the Glasgow Outcome Scale at six months (GOS-6M) and the Glasgow Coma Scale at ICU discharge (GCS-ICUd). The effect of phosphatemia variability on mortality and ICU length of stay (ICU-LOS) was also analyzed.

Results: Patients with sSAH had lower mean Pi level and median Pi dose repletion than that of nonneurocritical patients (3.1 ± 0.4 vs. 3.9 ± 1.3, p < 0.001). In the sSAH group, patients with hypophosphatemia had lower GCS-ICUd (12 ± 3.3 vs. 14 ± 2.4). Also, GOS-6M was lower in patients with hypophosphatemia but was not statistically significant (p = 0.09). By contrast, a higher mean Pi level in nonneurocritical patients was significantly associated with higher ICU mortality (4.8 ± 1.6 mg/dL vs. 3.6 ± 1.0 mg/dL, p = 0.003) and higher ICU-LOS (r = 0.231, p = 0.028). In the sSAH group, we found the opposite. In a multivariate analysis of the sSAH group, the increase in the Pi level was associated with higher GCS-ICUd (unstandardized coefficient in multiple linear regression [B] 1.79; 95% CI 0.43-3.15). The opposite was found in nonneurocritical patients. A Pi concentration higher than 2.5 mg/dL was associated with a better GCS-ICUd. We also found that creatinine, urea, chloride, need for Pi substitution, therapy intensity level, and pH were independent predictors of the mean Pi level during ICU stay in the sSAH group.

Conclusions: Patients with sSAH had lower mean Pi levels and required significantly higher daily Pi replacement compared with those of nonneurocritical patients. Since hypophosphatemia may be associated with poor neurological outcomes, patients with sSAH need cautious phosphate repletion.

Vascular Ehlers-Danlos syndrome presenting in the ICU as aneurysmal subarachnoid haemorrhage

Vascular Ehlers-Danlos syndrome is caused by mutations of COL3A1 gene coding for type III collagen. The main clinical features involve a propensity to arterial tears leading to several life-threatening conditions and intensive care unit admission. We, herein, report the case of a 34-year-old woman presenting with an aneurysmal subarachnoid haemorrhage. Endovascular coil treatment was attempted; however, the procedure was complicated by dissection of the left iliac artery and abdominal aorta. Hospital management was marked by a series of vascular and haemorrhagic complications. These events, together with some distinctive physical features and medical history, raised the suspicion of vascular type of Ehlers-Danlos syndrome. Neurological evolution was not favourable, and the patient evolved to brain death. Genetic testing was available postmortem and identified a mutation in the COL3A1 gene. This case illustrates the importance of medical history and clinical suspicion for diagnosis, which often goes unnoticed until major complications occur.

Alveolar proteinosis due to toxic inhalation at workplace

We report a clinical case of a 39-year old male, without any known previous medical condition but with occupational exposure to paints and dust cement, who presented an autoimmune pulmonary alveolar proteinosis (PAP) triggered by exposure to toxic inhalation at his workplace. PAP is a rare lung disease characterized by intra-alveolar abnormal accumulation of surfactant. The presence of a crazy-paving pattern in high-resolution computed tomography scan brings the suspicion of PAP although histopathology results of bronchoalveolar lavage are always required for its final diagnosis. The autoimmune form of PAP due to toxic inhalation, such as the one here described, is rare and it is usually difficult to establish a causal relationship.