The Clinical Toxicology Recommendations Collaborative: purpose, organization, and methodology.

INTRODUCTION: Scientific societies aim to provide a collective voice and unified stance on important issues. The Clinical Toxicology Recommendations Collaborative was formed in 2016 to develop evidence- and consensus-based recommendations for the management of patients exposed to common and/or serious poisonings for which the management is unclear or controversial. ORGANIZATION: The Clinical Toxicology Recommendations Collaborative is led jointly by the American Academy of Clinical Toxicology, the Asia Pacific Association of Medical Toxicology, and the European Association of Poison Centres and Clinical Toxicologists. The Governance Committee is chaired by a Past-President of one of these Societies and comprised of the six Presidents and Immediate Past-Presidents of the three Societies. A Steering Committee oversees the process of each project workgroup. METHODOLOGY: The overall process is guided by standards set forth by the Institute of Medicine for developing trustworthy guidelines and the Appraisal of Guidelines for Research and Evaluation Instrument. Systematic reviews are produced using the framework set in the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) methodology. Workgroup members jointly review the evidence and prepare statements on which they vote anonymously using a 9-point Likert scale. A two-round modified Delphi method is used to reach a consensus on clinical recommendations using the RAND/UCLA Appropriateness Method. Final recommendations are approved by unanimous consent of the workgroup and are expressed as both levels of evidence and strength of recommendations. LIMITATIONS: The major limitations of the Clinical Toxicology Recommendations Collaborative process centre around the amount and quality of evidence, the assessment of that evidence, and the voting of the panel. CONCLUSIONS: By using a transparent evidence- and consensus-based approach to produce systematic reviews and clinical recommendations, the Clinical Toxicology Recommendations Collaborative aims to create an international framework for clinical toxicology education and decision-making and foster positive change for the benefit of poisoned patients.

The role of QRS complex prolongation in predicting severe toxicity in single-xenobiotic overdose.

OBJECTIVE: The QRS complex duration is commonly used to prognosticate severity, predict outcomes, and indicate treatment in overdose. However, literature to support this practice is mixed in tricyclic antidepressant overdoses and absent in non-tricyclic antidepressant overdoses. Our objective was to assess the validity of QRS complex duration as a prognostic marker in overdose. METHODS: This was a secondary analysis of cases reported to the Toxicology Investigators Consortium between January 1, 2010, and December 31, 2022. Cases were assessed to determine the six xenobiotics most associated with QRS complex prolongation. All cases involving these six xenobiotics, regardless of QRS complex duration, constituted the study cohort. Inclusion criteria were cases of patients older than 12 years old with single-xenobiotic exposures. Clinical outcomes evaluated were seizure, ventricular dysrhythmia, metabolic acidosis, and death. RESULTS: Of 94,939 total cases, diphenhydramine, amitriptyline, bupropion, quetiapine, nortriptyline, and cocaine were most associated with QRS complex prolongation. Inclusion criteria were met by 4,655 cases of exposure to these xenobiotics. QRS complex prolongation was associated with increased odds ratio of seizure in all included xenobiotics, of ventricular dysrhythmia in all included xenobiotics except nortriptyline, and of metabolic acidosis or death in all included xenobiotics except nortriptyline and quetiapine. A normal QRS complex duration had a negative predictive value of greater than or equal to 93.0 percent of developing metabolic acidosis and 98.0 percent of developing a ventricular dysrhythmia or death from the xenobiotics studied. DISCUSSION: This study demonstrates that patients with QRS complex prolongation from all six xenobiotics studied had an increased prevalence and odds of developing severe outcomes. Furthermore, patients who did not develop QRS complex prolongation were unlikely to develop a ventricular dysrhythmia, metabolic acidosis, or death. These findings were noted in six xenobiotics that mechanistically can cause QRS complex prolongation through sodium channel or gap junction inhibition. CONCLUSION: Identification of patients at risk for severe outcomes after overdose can be aided by measuring the QRS complex duration. If prospectively validated, these outcomes have implications on risk stratification, disposition level of care, and appropriateness of treatments.

Patients with Altered Mental Status May Benefit from Medical Toxicology Consultation.

BACKGROUND: Presentation of intoxicated patients to hospitals is frequent, varied, and increasing. Medical toxicology expertise could lead to important changes in diagnosis and treatment, especially in patients presenting with altered mental status. OBJECTIVES: To describe and analyze clinical scenarios during a 1-year period after the establishment of a medical toxicology consultation service (MTCS). METHODS: Cases of 10 patients with altered mental status at presentation were evaluated. Medical toxicology consultation suggested major and significant changes in diagnosis and management. RESULTS: Of 973 toxicology consultations performed during the study period, bedside consultation was provided for 413 (42%) patients. Of these 413, 88 (21%) presented with some level of altered mental status. We described 10 patients in whom medical toxicology consultation brought about major and significant changes in diagnosis and management. CONCLUSIONS: Benefits may be derived from medical toxicology consultations, especially in patients with altered mental status. Medical toxicology specialists are well positioned to provide high value and expedited patient care.

Outcome prediction of methadone poisoning in the United States: implications of machine learning in the National Poison Data System (NPDS).

Methadone is an opioid receptor agonist with a high potential for abuse. The current study aimed to compare different machine learning models to predict the outcomes following methadone poisoning. This six-year retrospective longitudinal study utilizes National Poison Data System (NPDS) data. The severity of outcomes was derived from the NPDS Coding Manual. Our database was divided into training (70%) and test (30%) sets. We used a light gradient boosting machine (LGBM), extreme gradient boosting (XGBoost), random forest (RF), and logistic regression (LR) to predict the outcomes of methadone poisoning. A total of 3847 patients with methadone exposures were included. Our results demonstrated that machine learning models conferred high accuracy and reliability in determining the outcomes of methadone poisoning cases. The performance evaluation showed all models had high accuracy, precision, specificity, recall, and F1-score values. All models could reach high specificity (more than 96%) and high precision (80% or more) for predicting major outcomes. The models could also achieve a high sensitivity to predict minor outcomes. Finally, the accuracy of all models was about 75%. However, XGBoost and LGBM models achieved the best performance among all models. This study showcased the accuracy and reliability of machine learning models in the outcome prediction of methadone poisoning.

2023 American Heart Association focused update on the management of patients with cardiac arrest or life-threatening toxicity due to poisoning: an update to the American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care

In this focused update, the American Heart Association provides updated guidance for resuscitation of patients with cardiac arrest, respiratory arrest, and refractory shock due to poisoning. Based on structured evidence reviews, guidelines are provided for the treatment of critical poisoning from benzodiazepines, ß-adrenergic receptor antagonists (also known as ß-blockers), L-type calcium channel antagonists (commonly called calcium channel blockers), cocaine, cyanide, digoxin and related cardiac glycosides, local anesthetics, methemoglobinemia, opioids, organophosphates and carbamates, sodium channel antagonists (also called sodium channel blockers), and sympathomimetics. Recommendations are also provided for the use of venoarterial extracorporeal membrane oxygenation. These guidelines discuss the role of atropine, benzodiazepines, calcium, digoxin-specific immune antibody fragments, electrical pacing, flumazenil, glucagon, hemodialysis, hydroxocobalamin, hyperbaric oxygen, insulin, intravenous lipid emulsion, lidocaine, methylene blue, naloxone, pralidoxime, sodium bicarbonate, sodium nitrite, sodium thiosulfate, vasodilators, and vasopressors for the management of specific critical poisonings.

2023 American Heart Association Focused Update on the Management of Patients With Cardiac Arrest or Life-Threatening Toxicity Due to Poisoning: An Update to the American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care.

In this focused update, the American Heart Association provides updated guidance for resuscitation of patients with cardiac arrest, respiratory arrest, and refractory shock due to poisoning. Based on structured evidence reviews, guidelines are provided for the treatment of critical poisoning from benzodiazepines, ß-adrenergic receptor antagonists (also known as ß-blockers), L-type calcium channel antagonists (commonly called calcium channel blockers), cocaine, cyanide, digoxin and related cardiac glycosides, local anesthetics, methemoglobinemia, opioids, organophosphates and carbamates, sodium channel antagonists (also called sodium channel blockers), and sympathomimetics. Recommendations are also provided for the use of venoarterial extracorporeal membrane oxygenation. These guidelines discuss the role of atropine, benzodiazepines, calcium, digoxin-specific immune antibody fragments, electrical pacing, flumazenil, glucagon, hemodialysis, hydroxocobalamin, hyperbaric oxygen, insulin, intravenous lipid emulsion, lidocaine, methylene blue, naloxone, pralidoxime, sodium bicarbonate, sodium nitrite, sodium thiosulfate, vasodilators, and vasopressors for the management of specific critical poisonings.

Antivenom supply and demand: An analysis of antivenom availability and utilization in South Africa.

Introduction: Snakebites are a neglected tropical disease. In many areas, envenoming incidence and antivenom administration rates are unknown. This study compared antivenom (AV) availability to rates of envenoming and recommendations to treat (RTT) in South Africa. Methods: This retrospective study identified, extracted, and reviewed all cases of envenoming (snake bites and spits) reported to the Poisons Information Helpline of the Western Cape of South Africa (PIHWC) from June 1, 2015 to May 31, 2020 by public hospitals in the Western Cape. A standardized interview was administered to the pharmacies of the 40 hospitals in winter and summer to determine how many vials of monovalent and polyvalent AV they had on hand at the time of the call and their expiration dates. Descriptive analysis was used to compare rates of envenoming and recommendations to treat to antivenom stock in winter and summer and by hospital type and location. Results: Public hospitals reported 300 envenomings, 122 from snakes. The PIHWC recommended antivenom administration in 26% of cases (N = 32). All hospital pharmacies queried answered our questions. Our study demonstrates urban district hospitals have higher ratios of AV vials compared to mean annual rates of envenoming and RTT than rural district hospitals at both the winter and summer timepoints. Conclusion: This study evaluates antivenom supply and demand in a province of South Africa. The findings suggest South African urban hospitals have a relative excess of antivenom, and thus more capacity to meet demand, than their rural counterparts. It supports consideration of a redistribution of antivenom supply chains to match seasonal and local rates of envenoming. It indicates a need for higher quality, prospective data characterizing envenoming incidence and treatment.

A Case Report of Venezuelan Suntiger Tarantula (Psalmopoeus Irminia) Envenomation and Review of Tarantula Exposures.

BACKGROUND: Tarantula envenomations are encountered infrequently but may increase with increased exotic animal ownership. This case report presents the first documented toxicity from a Venezuelan suntiger tarantula (VST), Psalmopoeus irminia, and provides a general framework for approaching patients with tarantula exposures. CASE REPORT: A 35-year-old man presented to an emergency department 4 h after experiencing a bite from his pet VST. He developed erythema, pain, and edema to the bite site on the left thenar eminence that extended proximally. Within 4 h, he developed abdominal pain, nausea, vomiting, throat itching, and tightness. The patient had a blood pressure of 131/105 mm Hg, heart rate of 102 beats/min, 36.6°C, respiratory rate of 20 breaths/min, and SpO2 of 94%. Laboratory evaluations were within normal limits (other than chronically elevated but improved transaminases). The patient received 0.5 mg epinephrine intramuscularly, 50 mg diphenhydramine IV, 20 mg famotidine IV, 0.4 mg ondansetron IV, and 1 L of normal saline for a suspected anaphylactic reaction. Shortly after epinephrine administration, his gastrointestinal and upper airway symptoms resolved. All symptoms resolved within 1 week. WHY SHOULD AN EMERGENCY PHYSICIAN BE AWARE OF THIS?: Little is known about VST toxicity. Therefore, providers should rely on a general framework for approaching patients with tarantula exposures. Morbidity from tarantula exposures is mediated by mechanical injury, venom effects, and hypersensitivity reactions. Typical clinical findings include local pain, pruritis, edema, erythema, and burning. Muscle cramping, ophthalmia nodosa, and hypersensitivity reactions may occur. Treatment is primarily supportive and includes decontamination, cool compresses, analgesia, treatment of anaphylaxis, and ophthalmology evaluation if ocular exposure.

Toxic Tetrahydrocannabinol (THC) Dose in Pediatric Cannabis Edible Ingestions.

OBJECTIVE: The study characterizes cannabis toxicity in relation to tetrahydrocannabinol (THC) dose in pediatric edible cannabis ingestions. METHODS: This is a retrospective review of children aged <6 years presenting with edible cannabis ingestions of known THC dose within a pediatric hospital network (January 1, 2015-October 25, 2022). Cannabis toxicity was characterized as severe if patients exhibited severe cardiovascular (bradycardia, tachycardia/hypotension requiring vasopressors or intravenous fluids, other dysrhythmias), respiratory (respiratory failure, apnea, requiring oxygen supplementation), or neurologic (seizure, myoclonus, unresponsiveness, responsiveness to painful stimulation only, requiring intubation or sedation) effects. Cannabis toxicity was characterized as prolonged if patients required >6 hours to reach baseline. The relationship between THC dose and severe and prolonged toxicity was explored using multivariable logistic regression and receiver operator characteristic curve analyses. RESULTS: Eighty patients met inclusion. The median age was 2.9 years. The median THC ingestion was 2.1 mg/kg. Severe and prolonged toxicity was present in 46% and 74%, respectively. THC dose was a significant predictor of severe (adjusted odds ratio 2.9, 95% confidence interval: 1.8-4.7) and prolonged toxicity (adjusted odds ratio 3.2, 95% confidence interval: 1.6-6.5), whereas age and sex were not. Area under the curve was 92.9% for severe and 87.3% for prolonged toxicity. THC ingestions of ≥1.7 mg/kg can predict severe (sensitivity 97.3%) and prolonged toxicity (sensitivity 75.4%). CONCLUSIONS: The THC dose of edible cannabis correlates to the degree of toxicity in children <6 years old. The threshold of 1.7 mg/kg of THC may guide medical management and preventive regulations.

Severe outcomes following pediatric cannabis intoxication: a prospective cohort study of an international toxicology surveillance registry.

INTRODUCTION: An increasing number of jurisdictions have legalized recreational cannabis for adult use. The subsequent availability and marketing of recreational cannabis has led to a parallel increase in rates and severity of pediatric cannabis intoxications. We explored predictors of severe outcomes in pediatric patients who presented to the emergency department with cannabis intoxication. METHODS: In this prospective cohort study, we collected data on all pediatric patients (<18 years) who presented with cannabis intoxication from August 2017 through June 2020 to participating sites in the Toxicology Investigators Consortium. In cases that involved polysubstance exposure, patients were included if cannabis was a significant contributing agent. The primary outcome was a composite severe outcome endpoint, defined as an intensive care unit admission or in-hospital death. Covariates included relevant sociodemographic and exposure characteristics. RESULTS: One hundred and thirty-eight pediatric patients (54% males, median age 14.0 years, interquartile range 3.7-16.0) presented to a participating emergency department with cannabis intoxication. Fifty-two patients (38%) were admitted to an intensive care unit, including one patient who died. In the multivariable logistic regression analysis, polysubstance ingestion (adjusted odds ratio = 16.3; 95% confidence interval: 4.6-58.3; P < 0.001)) and cannabis edibles ingestion (adjusted odds ratio = 5.5; 95% confidence interval: 1.9-15.9; P = 0.001) were strong independent predictors of severe outcome. In an age-stratified regression analysis, in children older than >10 years, only polysubstance abuse remained an independent predictor for the severe outcome (adjusted odds ratio 37.1; 95% confidence interval: 6.2-221.2; P < 0.001). As all children 10 years and younger ingested edibles, a dedicated multivariable analysis could not be performed (unadjusted odds ratio 3.3; 95% confidence interval: 1.6-6.7). CONCLUSIONS: Severe outcomes occurred for different reasons and were largely associated with the patient's age. Young children, all of whom were exposed to edibles, were at higher risk of severe outcomes. Teenagers with severe outcomes were frequently involved in polysubstance exposure, while psychosocial factors may have played a role.

Deep learning neural network derivation and testing to distinguish acute poisonings.

INTRODUCTION: Acute poisoning is a significant global health burden, and the causative agent is often unclear. The primary aim of this pilot study was to develop a deep learning algorithm that predicts the most probable agent a poisoned patient was exposed to from a pre-specified list of drugs. RESEARCH DESIGN & METHODS: Data were queried from the National Poison Data System (NPDS) from 2014 through 2018 for eight single-agent poisonings (acetaminophen, diphenhydramine, aspirin, calcium channel blockers, sulfonylureas, benzodiazepines, bupropion, and lithium). Two Deep Neural Networks (PyTorch and Keras) designed for multi-class classification tasks were applied. RESULTS: There were 201,031 single-agent poisonings included in the analysis. For distinguishing among selected poisonings, PyTorch model had specificity of 97%, accuracy of 83%, precision of 83%, recall of 83%, and a F1-score of 82%. Keras had specificity of 98%, accuracy of 83%, precision of 84%, recall of 83%, and a F1-score of 83%. The best performance was achieved in the diagnosis of single-agent poisoning in diagnosing poisoning by lithium, sulfonylureas, diphenhydramine, calcium channel blockers, then acetaminophen, in PyTorch (F1-score = 99%, 94%, 85%, 83%, and 82%, respectively) and Keras (F1-score = 99%, 94%, 86%, 82%, and 82%, respectively). CONCLUSION: Deep neural networks can potentially help in distinguishing the causative agent of acute poisoning. This study used a small list of drugs, with polysubstance ingestions excluded.Reproducible source code and results can be obtained at https://github.com/ashiskb/npds-workspace.git.

Using a decision tree algorithm to distinguish between repeated supra-therapeutic and acute acetaminophen exposures.

BACKGROUND: This study aimed to compare clinical and laboratory characteristics of supra-therapeutic (RSTI) and acute acetaminophen exposures using a predictive decision tree (DT) algorithm. METHODS: We conducted a retrospective cohort study using the National Poison Data System (NPDS). All patients with RSTI acetaminophen exposure (n = 4,522) between January 2012 and December 2017 were included. Additionally, 4,522 randomly selected acute acetaminophen ingestion cases were included. After that, the DT machine learning algorithm was applied to differentiate acute acetaminophen exposure from supratherapeutic exposures. RESULTS: The DT model had accuracy, precision, recall, and F1-scores of 0.75, respectively. Age was the most relevant variable in predicting the type of acetaminophen exposure, whether RSTI or acute. Serum aminotransferase concentrations, abdominal pain, drowsiness/lethargy, and nausea/vomiting were the other most important factors distinguishing between RST and acute acetaminophen exposure. CONCLUSION: DT models can potentially aid in distinguishing between acute and RSTI of acetaminophen. Further validation is needed to assess the clinical utility of this model.

Comparison of decision tree with common machine learning models for prediction of biguanide and sulfonylurea poisoning in the United States: an analysis of the National Poison Data System.

BACKGROUND: Biguanides and sulfonylurea are two classes of anti-diabetic medications that have commonly been prescribed all around the world. Diagnosis of biguanide and sulfonylurea exposures is based on history taking and physical examination; thus, physicians might misdiagnose these two different clinical settings. We aimed to conduct a study to develop a model based on decision tree analysis to help physicians better diagnose these poisoning cases. METHODS: The National Poison Data System was used for this six-year retrospective cohort study.The decision tree model, common machine learning models multi layers perceptron, stochastic gradient descent (SGD), Adaboosting classiefier, linear support vector machine and ensembling methods including bagging, voting and stacking methods were used. The confusion matrix, precision, recall, specificity, f1-score, and accuracy were reported to evaluate the model's performance. RESULTS: Of 6183 participants, 3336 patients (54.0%) were identified as biguanides exposures, and the remaining were those with sulfonylureas exposures. The decision tree model showed that the most important clinical findings defining biguanide and sulfonylurea exposures were hypoglycemia, abdominal pain, acidosis, diaphoresis, tremor, vomiting, diarrhea, age, and reasons for exposure. The specificity, precision, recall, f1-score, and accuracy of all models were greater than 86%, 89%, 88%, and 88%, respectively. The lowest values belong to SGD model. The decision tree model has a sensitivity (recall) of 93.3%, specificity of 92.8%, precision of 93.4%, f1_score of 93.3%, and accuracy of 93.3%. CONCLUSION: Our results indicated that machine learning methods including decision tree and ensembling methods provide a precise prediction model to diagnose biguanides and sulfonylureas exposure.

Comparison of Two-Bag Versus Three-Bag N-Acetylcysteine Regimens for Pediatric Acetaminophen Toxicity.

BACKGROUND: Acetaminophen overdose is a leading cause of liver failure, and a leading cause of pediatric poisoning requiring hospital admission. The antidote, N-acetylcysteine (NAC), is traditionally administered as a three-bag intravenous infusion. Despite its efficacy, NAC is associated with high incidence of nonallergic anaphylactoid reactions (NAARs). Adult evidence demonstrates that alternative dosing regimens decrease NAARs and medication errors (MEs). OBJECTIVES: To compare NAARs and MEs associated with two- versus three-bag NAC for acetaminophen overdose in a pediatric population. METHODS: This is a retrospective observational cohort study comparing pediatric patients who received three- versus two-bag NAC for acetaminophen toxicity. The primary outcome was incidence of NAARs. Secondary outcomes were rates of MEs and relevant hospital outcomes (length of stay [LOS], intensive care unit (ICU) admission, liver transplant, death). RESULTS: Two hundred forty-three patients met inclusion criteria (median age of 15 years): 150 (62%) three-bag NAC and 93 (38%) two-bag NAC. There was no difference in overall NAARs (p = 0.54). Fewer cutaneous NAARs were observed in the two-bag group, three-bag: 15 (10%), two-bag: 2 (2%), p = 0.02. MEs were significantly decreased with the two-bag regimen, three-bag: 59 (39%), two-bag: 21 (23%), p = 0.01. No statistical differences were observed in LOS, ICU admissions, transplant, or death. CONCLUSION AND RELEVANCE: A significant decrease in cutaneous NAARs and MEs was observed in pediatric patients by combining the first two bags of the traditional three-bag NAC regimen. In pediatric populations, a two-bag NAC regimen for acetaminophen overdose may improve medication tolerance and safety.

Decision tree algorithm can determine the outcome of repeated supratherapeutic ingestion (RSTI) exposure to acetaminophen: review of 4500 national poison data system cases.

This study is aimed at establishing the outcome of RSTI exposure to acetaminophen based on a decision tree algorithm for the first time. This study used the National Poison Data System (NPDS) to conduct a six-year retrospective cohort analysis, which included 4522 individuals. The patients had a mean age of 26.75 ± 16.3 years (1-89). 3160 patients (70%) were females. Most patients had intentional exposure to acetaminophen. Almost all the patients had acetaminophen exposure via ingestion. In addition, 400 (8.8%) experienced major outcomes, 1500 (33.2%) experienced moderate outcomes, and 2622 (58%) of the patients experienced mild ones. The decision tree model performed well in the training and test groups. In the test group, the accuracy was 0.813, precision of 0.827, recall being 0.798, specificity 0.898, and an F1 score 0.80. In the training group, accuracy was 0.831, recall was 0.825, precision was 0.837, specificity was 0.90, and F1 score was 0.829. Our results showed that serum liver enzymes being present at elevated levels (Alanine aminotransferase (ALT), Aspartate aminotransferase (AST) greater than 1000 U/L followed by ALT, AST between 100 and 1000 U/L), prothrombin time (PT) prolongation, bilirubin increase, renal failure, confusion, age, hypotension, other coagulopathy (such as partial thromboplastin time (PTT) prolongation), acidosis, and electrolyte abnormality were the effective factors in determining the outcomes in these patients. The decision tree algorithm is a dependable method for establishing the prognosis of patients who have been exposed to RSTI acetaminophen and can be used throughout the patients' hospitalization period.

Classification of acute poisoning exposures with machine learning models derived from the National Poison Data System.

The primary aim of this pilot study was to develop a machine learning algorithm to predict and distinguish eight poisoning agents based on clinical symptoms. Data were used from the National Poison Data System from 2014 to 2018, for patients 0-89 years old with single-agent exposure to eight drugs or drug classes (acetaminophen, aspirin, benzodiazepines, bupropion, calcium channel blockers, diphenhydramine, lithium and sulfonylureas). Four classifier prediction models were applied to the data: logistic regression, LightGBM, XGBoost, and CatBoost. There were 201 031 cases used to develop and test the algorithms. Among the four models, accuracy ranged 77%-80%, with precision and F1 scores of 76%-80% and recall of 77%-78%. Overall specificity was 92% for all models. Accuracy was highest for identifying sulfonylureas, acetaminophen, benzodiazepines and diphenhydramine poisoning. F1 scores were highest for correctly classifying sulfonylureas, acetaminophen and benzodiazepine poisonings. Recall was highest for sulfonylureas, acetaminophen, and benzodiazepines, and lowest for bupropion. Specificity was >99% for models of sulfonylureas, calcium channel blockers, lithium and aspirin. For single-agent poisoning cases among the eight possible exposures, machine learning models based on clinical signs and symptoms moderately predicted the causal agent. CatBoost and LightGBM classifier models had the highest performance of those tested.

Adolescent US Poison Center Exposure Calls During the COVID-19 Pandemic.

PURPOSE: The objective of this study was to evaluate trends and characteristics in adolescent poison center (PC) exposure calls before and during the COVID-19 pandemic. METHODS: A retrospective review of PC calls for adolescents aged 13-17 years from January 1, 2018 through June 30, 2021. RESULTS: During the pandemic, US PCs had a higher proportion of adolescent exposure calls managed in a healthcare facility (71.9% vs. 67.4%) and hospital admissions (27.2% vs. 25.7%) than prior to the pandemic. There was a higher proportion with suicide intent (55.8% vs. 48.8%), moderate/major clinical effects (22.8% vs. 20.1%), and deaths (0.07% vs. 0.05%). Monthly calls significantly increased from 30 calls/month to 204 calls/month (p < .001). The slope of hospital admissions significantly increased (0.19% per month, p < .001) during the pandemic. DISCUSSION: During the COVID-19 pandemic, US PCs observed an increase in adolescent suicidal intent exposure calls with more severe outcomes, hospitalizations, and deaths.

Utility of support vector machine and decision tree to identify the prognosis of metformin poisoning in the United States: analysis of National Poisoning Data System.

BACKGROUND: With diabetes incidence growing globally and metformin still being the first-line for its treatment, metformin's toxicity and overdose have been increasing. Hence, its mortality rate is increasing. For the first time, we aimed to study the efficacy of machine learning algorithms in predicting the outcome of metformin poisoning using two well-known classification methods, including support vector machine (SVM) and decision tree (DT). METHODS: This study is a retrospective cohort study of National Poison Data System (NPDS) data, the largest data repository of poisoning cases in the United States. The SVM and DT algorithms were developed using training and test datasets. We also used precision-recall and ROC curves and Area Under the Curve value (AUC) for model evaluation. RESULTS: Our model showed that acidosis, hypoglycemia, electrolyte abnormality, hypotension, elevated anion gap, elevated creatinine, tachycardia, and renal failure are the most important determinants in terms of outcome prediction of metformin poisoning. The average negative predictive value for the decision tree and SVM models was 92.30 and 93.30. The AUC of the ROC curve of the decision tree for major, minor, and moderate outcomes was 0.92, 0.92, and 0.89, respectively. While this figure of SVM model for major, minor, and moderate outcomes was 0.98, 0.90, and 0.82, respectively. CONCLUSIONS: In order to predict the prognosis of metformin poisoning, machine learning algorithms might help clinicians in the management and follow-up of metformin poisoning cases.

Distinguishing characteristics of exposure to biguanide and sulfonylurea anti-diabetic medications in the United States.

OBJECTIVES: Biguanides and sulfonylureas are anti-hyperglycemic drugs commonly used in the United States. Poisoning with these drugs may lead to serious consequences. The diagnosis of biguanide and sulfonylurea poisoning is based on history, clinical manifestations, and laboratory studies. METHODS: This study is a six-year retrospective cohort analysis based on the National Poison Data System. Clinical effects of 6183 biguanide and sulfonylurea exposures were identified using binary logistic regression. RESULTS: The mean age of patients with biguanide and sulfonylurea exposure was 39.27 ± 28.91 and 28.91 ± 30.41 years, respectively. Sulfonylurea exposure is most commonly seen via unintentional exposure, while biguanide exposure frequently occurs as a result of intentional ingestion. Minor and moderate outcomes commonly developed following biguanide and sulfonylurea exposure, respectively. Sulfonylurea exposure was less likely to develop clinical effects abdominal pain, metabolic acidosis, diarrhea, nausea, vomiting, and elevated creatinine than patients ingesting biguanides. However, sulfonylurea exposure was more likely to cause dizziness or vertigo, tremor, drowsiness or lethargy, agitation, diaphoresis, and hypoglycemia. CONCLUSIONS: Our study is the first to use a wide range of national data to describe the clinical characteristics that differentiate the toxicologic exposure to biguanides and sulfonylureas. Sulfonylurea exposure is commonly seen via unintentional exposure, while metformin exposure is frequently seen via intentional exposure. Sulfonylurea toxicity is more likely to cause agitation, dizziness or vertigo, tremor, diaphoresis, and hypoglycemia, while metformin exposure induces abdominal pain, acidosis, diarrhea, nausea, vomiting, and elevated creatinine.