Prediction of intracranial pressure crises after severe traumatic brain injury using machine learning algorithms.

OBJECTIVE: Avoiding intracranial hypertension after traumatic brain injury (TBI) is a foundation of neurocritical care, to minimize secondary brain injury related to elevated intracranial pressure (ICP). However, this approach at best is reactive to episodes of intracranial hypertension, allowing for periods of elevated ICP before therapies can be initiated. Accurate prediction of ICP crises before they occur would permit clinicians to implement preventive strategies, minimize total time with ICP above threshold, and potentially avoid secondary injury. The objective of this study was to develop an algorithm capable of predicting the onset of ICP crises with sufficient lead time to enable application of preventative therapies. METHODS: Thirty-six patients admitted to a level I trauma center with severe TBI (Glasgow Coma Scale score < 8) between April 2015 and January 2019 who underwent continuous intraparenchymal ICP monitor placement were retrospectively identified. Continuous ICP data were extracted from each monitoring period (range 4-96 hours of monitoring). An ICP crisis was treated as a binary outcome, defined as ICP > 22 mm Hg for at least 75% of the data within a 5-minute interval. ICP data preceding each ICP crisis were grouped into four total data sets of 1- and 2-hour epochs, each with 10- to 20-minute lead-time intervals before an ICP crisis. Crisis and noncrisis events were identified from continuous time-series data and randomly split into 70% for training and 30% for testing, from a subset of 30 patients. Machine learning algorithms were trained to predict ICP crises, including light gradient boosting, extreme gradient boosting, and random forest. Accuracy and area under the receiver operating characteristic curve (AUC) were measured to compare performance. The most predictive algorithm was optimized using feature selection and hyperparameter tuning to avoid overfitting, and then tested on a validation subset of 5 patients. Precision, recall, F1 score, and accuracy were measured. RESULTS: The random forest model demonstrated the highest accuracy (range 0.82-0.88) and AUC (range 0.86-0.88) across all four data sets. Further validation testing revealed high precision (0.76), relatively low recall (0.46), and overall strong predictive performance (F1 score 0.57, accuracy 0.86) for ICP crises. Decision curve analysis showed that the model provided net benefit at probability thresholds above 0.1 and below 0.9. CONCLUSIONS: The presented model can provide accurate and timely forecasts of ICP crises in patients with severe TBI 10-20 minutes prior to their occurrence. If validated and implemented in clinical workflows, this algorithm can enable earlier intervention for ICP crises, more effective treatment of intracranial hypertension, and potentially improved outcomes following severe TBI.

A case report of type 1 diabetes mellitus and factitious hypoglycemia in a patient with Munchausen Syndrome / Hipoglucemia facticia: el caso de un paciente con diabetes mellitus tipo 1 y síndrome de Munchausen

SUMMARY Munchausen syndrome is rarely considered as a first diagnosis, especially in a type 1 diabetic patient presenting with hyperinsulinemic hypoglycemia. The diagnosis should be considered when episodes of hypoglycemia are persistent, and tests suggest a possible exogenous source of insulin. We report a case of a 26-year-old man with multiple hypoglycemic episodes and a long known diagnosis of diabetes type 1 who was referred to our institution after multiple in and out patient consultations in other institutions. He arrived with persistent hypoglycemia, even after withdrawal of insulin therapy on medical record, but persistent self-administration and misuse, without health care professional knowledge, of insulin therapy. He was diagnosed with factitious hypoglycemia after psychiatric evaluation. The patient improved with psychotherapy and family support as well as strict vigilance of insulin administration.

RESUMEN El síndrome de Munchausen rara vez es considerado como primer diagnóstico, especialmente en pacientes diabéticos tipo 1 con cuadro de hipoglicemia hiperinsulinémica. Debe pensarse en este diagnóstico cuando los episodios de hipoglicemia sean persistentes y los exámenes paraclínicos sugieran una fuente exógena de insulina. El siguiente es un reporte de caso de un paciente masculino de 26 años con múltiples episodios de hipoglicemia y diagnóstico conocido de diabetes mellitus tipo 1, quien fue referido a nuestro hospital universitario después de haber consultado en varias ocasiones y haber sido hospitalizado y dado de alta en otras instituciones. Ingresa por múltiples episodios de hipoglicemias, y que incluso al retirar las insulinas por orden médica, persistían los síntomas. Se encontró auto-administración de uso de insulinas sin el conocimiento de los profesionales de la salud, llegando al diagnóstico de hipoglicemia facticia después de valoración por psiquiatría. El paciente presentó mejoría con psicoterapia y apoyo familiar, además de vigilancia estricta de la administración de insulinas.

Reduction of glycemic variability with Degludec insulin in patients with unstable diabetes.

INTRODUCTION: Degludec (IDeg) is an ultralong-acting insulin, with stable pharmacodynamic profile which leads to lower fluctuations in glucose levels. The effect of IDeg has not been specifically assessed in patients with unstable diabetes, defined as increased glycemic variability (GV). METHODS: A prospective before-after pilot study was conducted, including patients managed at Hospital Universitario San Ignacio in Bogotá, Colombia. The impact of the switch from a Glargine or Detemir insulin to a basal insulin regimen with IDeg for 12 weeks on GV measured by continuous glucose monitoring, on A1c levels, and on the incidence of episodes of global and nocturnal hypoglycemia was assessed in a group of patients with (coefficient of variation >34%) or without increased basal GV using a Generalised Estimating Equation (GEE) analysis. RESULTS: 60 patients with basal bolus therapy and history of hypoglycemia were included. 18 patients had High GV (HGV). In this group a significant reduction of 11.1% of CV (95% CI: 6.3, 15.9, p = 0.01) was found. GEE analysis confirmed a higher impact over time on patients with HGV (p < 0.001). The percentage of patients with at least 1 episode of hypoglycemia decreased from 66.6% to 22.2% (p = 0.02) and from 37.14% to 5.71% (p < 0.01) for global and nocturnal hypoglycemia, respectively. Changes were not significant in patients with low GV. A reduction of A1c was observed in both groups (p < 0.001). CONCLUSIONS: The results suggest that treatment with IDeg reduces GV, A1c levels and the incidence of global and nocturnal hypoglycemia events in patients with HGV, but not in patients with low GV.