Prediction of naloxone dose in opioids toxicity based on machine learning techniques (artificial intelligence)

BACKGROUND

Treatment management for opioid poisoning is critical and, at the same time, requires specialized knowledge and skills. This study was designed to develop and evaluate machine learning algorithms for predicting the maintenance dose and duration of hospital stay in opioid poisoning, in order to facilitate appropriate clinical decision-making.

METHOD AND RESULTS

This study used artificial intelligence technology to predict the maintenance dose and duration of administration by selecting clinical and paraclinical features that were selected by Pearson correlation (filter method) (Stage 1) and then the (wrapper method) Recursive Feature Elimination Cross-Validated (RFECV) (Stage2). The duration of administration was divided into two categories: A (which includes a duration of less than or equal to 24 h of infusion) and B (more than 24 h of naloxone infusion). XGBoost algorithm model with an accuracy rate of 91.04%, a prediction rate of 91.34%, and a sensitivity rate of 91.04% and area under the Curve (AUC) 0.97 was best model for classification patients. Also, the best maintenance dose of naloxone was obtained with XGBoost algorithm with R2 = 0.678. Based on the selected algorithm, the most important features for classifying patients for the duration of treatment were bicarbonate, respiration rate, physical sign, The partial pressure of carbon dioxide (PCO2), diastolic blood pressure, pulse rate, naloxone bolus dose, Blood Creatinine(Cr), Body temperature (T). The most important characteristics for determining the maintenance dose of naloxone were physical signs, bolus dose of 4.5 mg/kg, Glasgow Coma Scale (GCS), Creatine Phosphokinase (CPK) and intensive care unit (ICU) add.

CONCLUSION

A predictive model can significantly enhance the decision-making and clinical care provided by emergency physicians in hospitals and medical settings. XGBoost was found to be the superior model.

A case of lacosamide and mirtazapine self-poisoning

Lacosamide is a relatively new antiepileptic drug that exerts its anticonvulsant effect by selectively inactivating sodium channels. Since its launch, it has been used widely for the treatment of intractable epilepsy, but there are scant data on the toxic or lethal blood concentrations. Here, we report a case of drug poisoning following simultaneous high-dose self-administration of lacosamide and mirtazapine. We developed and validated an approach that uses liquid chromatography coupled with electrospray ionization-tandem mass spectrometry to determine the concentrations of lacosamide and mirtazapine in cadaveric blood, urine and liver. Calibration curves showed good linearity (r2 > 0.995), and our method enabled repeatable and accurate quantification, with intra- and inter-assay coefficients of variation not exceeding 10.9 % and 12.8 %, respectively, for each target drug. We used the method to measure the drug concentrations in the blood of a dead victim and found a lacosamide concentration of 91.9 μg/mL and a mirtazapine concentration of 12.0 μg/mL. The blood mirtazapine concentration was in the lethal range, and that of lacosamide was about 10 times the therapeutic range. The synergistically central nervous system depressive and cardiotoxic effects of these drugs may have contributed to the cause of death. We concluded that the cause of death in this case was lacosamide and mirtazapine poisoning.

Geographically specific associations between county-level socioeconomic and household distress and mortality from drug poisoning, suicide, alcohol, and homicide among working-age adults in the United States

Background

Mortality rates from drug poisoning, suicide, alcohol, and homicide vary significantly across the United States. This study explores localized relationships (i.e., geographically specific associations) between county-level economic and household distress and mortality rates from these causes among working-age adults (25-64).

Methods

Mortality data were from the National Vital Statistics System for 2014-2019. County-level socioeconomic distress (poverty, employment, income, education, disability, insurance) and household distress (single-parent, no vehicle, crowded housing, renter occupied) were from the 2009-2013 American Community Survey. We conducted Ordinary Least Squares (OLS) regression to estimate average associations and Geographically Weighted Regression (GWR) to estimate localized spatial associations between county-level distress and working-age mortality.

Results

In terms of national average associations, OLS results indicate that a one standard deviation increase in socioeconomic distress was associated with an average of 6.1 additional drug poisoning deaths, 3.0 suicides, 2.1 alcohol-induced deaths, and 2.0 homicides per 100,000 population. A one standard deviation increase in household distress was associated with an average of 1.4 additional drug poisonings, 4.7 alcohol-induced deaths, and 1.1 homicides per 100,000 population. However, the GWR results showed that these associations vary substantially across the U.S., with socioeconomic and household distress associated with significantly higher mortality rates in some parts of the U.S than others, significantly lower rates in other parts of the U.S., and no significant associations in others. There were also some areas where distress overlapped to influence multiple causes of death, in a type of compounded disadvantage.

Conclusions

Socioeconomic and household distress are significant and substantial predictors of higher rates of drug poisoning mortality, suicide, alcohol-induced deaths, and homicide in specific regions of the U.S. However, these associations are not universal. Understanding the place-level factors that contribute to them can inform geographically tailored strategies to reduce rates from these preventable causes of death in different places.

Preference for hotline versus mobile application/countdown-based mobile overdose response services: a qualitative study

BACKGROUND

In response to the exacerbated rates of morbidity and mortality associated with the overlapping overdose and COVID-19 epidemics, novel strategies have been developed, implemented, operationalized and scaled to reduce the harms resulting from this crisis. Since the emergence of mobile overdose response services (MORS), two strategies have aimed to help reduce the mortality associated with acute overdose including staffed hotline-based services and unstaffed timer-based services. In this article, we aim to gather the perspectives of various key interest groups on these technologies to determine which might best support service users.

METHODS

Forty-seven participants from various interested groups including people who use substances who have and have not used MORS, healthcare workers, family members, harm reduction employees and MORS operators participated in semi-structured interviews. Transcripts were coded and analyzed using a thematic analysis approach.

RESULTS

Four major themes emerged regarding participant perspectives on the differences between services, namely differences in connection, perceived safety, privacy and accessibility, alongside features that are recommended for MORS in the future.

CONCLUSIONS

Overall, participants noted that individuals who use substances vary in their desire for connection during a substance use session offered by hotline and timer-based service modalities. Participants perceived hotline-based approaches to be more reliable and thus potentially safer than their timer-based counterparts but noted that access to technology is a limitation of both approaches.

A cost benefit analysis of a virtual overdose monitoring service/mobile overdose response service: the national overdose response service

BACKGROUND

The overdose crisis continues across Canada which calls for novel harm reduction strategies. Previous research indicates that a majority of eHealth solutions are cost-effective however current literature on the cost-benefit of eHealth for harm reduction is sparse. The National Overdose Response Service (NORS) is a Canada-wide telephone-based harm reduction service. Service users can call the phone number and connect to a peer who can virtually monitor the substance use session and dispatch appropriate interventions in the case of overdose.

OBJECTIVES OF THE RESEARCH/PROJECT

We aim to assess the cost-benefit of NORS by comparing the estimated cost-savings from prevented overdose mortality to the operating costs of the program, alongside healthcare costs associated with its operation.

METHODS

Data around systems costs and operational costs were gathered for our calculations. Our primary outcome was cost-benefit ratios, derived from estimates and models of mortality rates in current literature and value of life lost. We presented our main results across a range of values for costs and the probability of death following an unwitnessed overdose. These values were utilized to calculate cost-benefit ratios and value per dollar spent on service provision by NORS over the length of the program's operation (December 2020-2022).

RESULTS

Over the total funded lifespan of the program, and using a Monte Carlo estimate, the benefit-to-cost ratio of the NORS program was 8.59 (1.53-15.28) per dollar spent, depending on estimated mortality rates following unwitnessed overdose and program operation costs. Further, we conservatively estimate that early community-based naloxone intervention results in healthcare system savings of $4470.82 per overdose response.

CONCLUSIONS

We found the NORS program to have a positive benefit-to-cost ratio when the probability of death following an unwitnessed overdose was greater than 5%. NORS and potentially other virtual overdose monitoring services have the potential to be cost-effective solutions for managing the drug poisoning crisis.

Changing dynamics of drug overdoses in the United Kingdom: An attempt to replicate the Jalal et al. findings of steady exponential growth

BACKGROUND

Jalal et al. discovered that between 1979 and 2020 total rates and counts of fatal drug overdoses in the United States exhibited exponential growth at a very steady rate even though deaths from individual drugs did not. That is a startling result because it means that the different drugs are in effect "taking turns", with one growing faster just as another drug's death rate growth ebbs. That raises the question of whether this steadiness in the all-drug death rates is in some sense just a coincidence peculiar to the United States or whether it might reflect some more general phenomenon and so manifest in other countries.

METHODS

We fit the same model used by Jalal et al. to data on drug-related death rates for the countries of the United Kingdom.

RESULTS

The main finding is largely a failure to replicate the United States result. Simple graphical display of the trends and a number of statistical measures show that the growth in the United Kingdom was not only slower than in the United States, it was also less steady, with the exception of Northern Ireland.

CONCLUSIONS

Steady exponential growth in the all-drugs mortality rate may be a phenomenon specific to certain contexts. It remains an open question whether the explanation of steady exponential growth in the United States and Northern Ireland relates to demand and supply mechanisms, to social and political conditions, or to coincidence.

Drugs involved in Kentucky drug poisoning deaths and relation with antecedent controlled substance prescription dispensing

BACKGROUND

The shift from prescription to illicit drugs involved in drug poisoning deaths raises questions about the current utility of prescription drug monitoring program (PDMP) data to inform drug poisoning (overdose) prevention efforts. In this study, we describe relations between specific drugs involved in Kentucky drug poisoning deaths and antecedent controlled substance (CS) dispensing.

METHODS

The study used linked death certificates and PDMP data for 2,248 Kentucky resident drug poisoning deaths in 2021. Death certificate literal text analysis identified drugs mentioned with involvement (DMI) in drug poisoning deaths. We characterized the concordance between each DMI and the CS dispensing history for this drug at varying timepoints since 2008.

RESULTS

Overall, 25.5% of all decedents had dispensed CS in the month before fatal drug poisoning. Over 80% of decedents were dispensed opioid(s) since 2008; the percentage was similar regardless of opioid involvement in the poisoning death. One-third of decedents had dispensed buprenorphine for treatment of opioid use disorder since 2008, but only 6.1% had dispensed buprenorphine in the month preceding death. Fentanyl/fentanyl analogs were DMI in 1,568 (69.8%) deaths, yet only 3% had received a fentanyl prescription since 2008. The highest concordance in the month preceding death was observed for clonazepam (43.6%).

CONCLUSION

Overall, concordance between CS dispensing history and the drugs involved in poisoning deaths was low, suggesting a need to reevaluate the complex relationships between prescription medication exposure and overdose death and to expand harm reduction interventions both within and outside the healthcare system to reduce drug poisoning mortality.

[Rapid determination of acetaminophen in plasma by LC-MS/MS]

Objective: To establish a method for the rapid determination of acetaminophen (APAP) in human plasma by LC-MS/MS. Methods: The plasma samples were extracted by methanol and acetonitrile (1: 1) and purified directly. C(18) column was used for sample separation. The mobile phase were methanol (5 mmol/L ammonium acetate) and water (5 mmol/L ammonium acetate). Samples were analyzed by LC MS/MS with the electrospray ionization multi reaction monitoring (MRM) mode. Results: The calibration curves of APAP was linear in the concentration range of 0~10 mg/L, the correlation coefficient (r) was greater than 0.999 0. The relative standard deviation within and between batches was less than 10%. The recovery rate were 96.81%~101.7%. The detection limit of the method was 0.1 μg/L and the lower limit of quantification was 0.3 μg/L. Conclusion: This method has strong specificity, high sensitivity and reliable determination results. It is suitable for the rapid analysis of clinical plasma samples.

The Role of Angiotensin II in Poisoning-Induced Shock-a Review

Background

Shock in drug poisoning is a life-threatening condition and current management involves fluid resuscitation and vasopressor therapy. Management is limited by the toxicity of high-dose vasopressors such as catecholamines. Clinical trials have shown the efficacy of angiotensin II as an adjunct vasopressor in septic shock. The aim of this review is to assess the use of angiotensin II in patients with shock secondary to drug overdose.

Methods

Medline (from 1946), Embase (from 1947) and PubMed (from 1946) databases were searched until July 2021 via OVID. Included studies were those with shock due to drug poisoning and received angiotensin II as part of their treatment regimen. Of the 481 articles identified, 13 studies (case reports and scientific abstracts) were included in the final analysis with a total of 14 patients. Extracted data included demographics, overdose drug and dosage, angiotensin II dosage, time of angiotensin II administration, haemodynamic changes, length of hospital stay, mortality, complications, cardiac function and other treatment agents used.

Results

Thirteen studies were included consisting of 6 case reports, 6 scientific abstracts and 1 case series. Overdose drugs included antihypertensives (n = 8), psychotropics (n = 4), isopropanol (n = 1) and tamsulosin (n = 1). Out of a total of 14 patients, 3 patients died. Ten patients had their haemodynamic changes reported. In terms of MAP or SBP changes, three patients (30%) had an immediate response to angiotensin II, four patients (40%) had responses within 30 min, one patient (10%) within two hours and two patients (20%) did not have their time reported. Two patients were shown to have direct chronotropic effects within 30 min of angiotensin II administration. The median hospital stay for patients was 5 days (IQR = 4). The time from overdose until angiotensin II administration ranged from 5 to 56 h. Other vasopressors used included phenylephrine, noradrenaline, adrenaline, vasopressin, dobutamine, dopamine, methylene blue and ephedrine. A median of 3 vasopressors were used before initiation of angiotensin II. Twelve patients received angiotensin II as their final treatment.

Conclusions

Angiotensin II may be useful as an adjunct vasopressor in treating shock secondary to drug poisoning. However, the current literature consisted of only very low-quality studies. To truly assess the utility of angiotensin II use in drug-induced poisoned patients, further well-designed prospective studies are required.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13181-022-00885-4.

Sex-based differences in timely emergency department evaluations for patients with drug poisoning

OBJECTIVES

Unintentional poisoning was the leading cause of injury-related death in the United States in 2017. Prescribed and illicit drugs are the most common cause of poisoning, and timely management in the emergency department (ED) is important. Our aim was to identify any disparities in wait times associated with sex for drug poisoning-related ED visits.

STUDY DESIGN

We examined ED visits using data from the 2009-2017 National Hospital Ambulatory Medical Care Survey (NHAMCS).

METHODS

Drug poisoning-related visits were identified using the International Classification of Diseases, Ninth or Tenth Revision, Clinical Modification codes. Delayed assessment was defined as wait times exceeding the recommended triage time. Weighted logistic regression was used.

RESULTS

The average age was 36 years (standard error = 1.1), 54% female, 87% White and 29% had delayed assessment. Most common drugs were psychotropics (45%) and opioids (32%). Adjusting for race, payment source, urgency, multiple drug types and NSAIDs, females who had poisoning by substances other than opioids had 2.1 times higher likelihood of having a delayed assessment compared with males (odds ratio [95% confidence interval]: 2.1 [1.03-4.2]), although there was no difference between sexes among visits with opioid poisoning (P = 0.27). Neither race (P = 0.23) nor payment source (P = 0.22) were associated with delayed assessment, and the sex association was consistent across these groups.

CONCLUSIONS

Females with non-opioid drug poisoning were more likely to have delayed assessment than men. None of the other demographic factors demonstrated a correlation. Identifying more populations vulnerable to delays in the ED can help guide the development of interventions and policies to expedite care and attenuate existing disparities.

[Accountability of anti-cholinergic drugs in increase of psychotic episodes: A case of tropatepine overdose]

INTRODUCTION

Corrective treatments for some neurological side effects due to antipsychotic treatments can induce anticholinergic adverse effects. However, the risk of triggering or increasing psychotic symptoms induced by these drugs is unknown with only one case-report concerning the tropatepine. In addition, recommendations for the use of these drugs remain imprecise regarding the management of this type of adverse effect.

CASE REPORT

We report the case of a psychotic episode in an 18-year-old patient potentially acutised after an auto-intoxication with tropatepine.

CONCLUSION

At high doses, anticholinergic treatments, including tropatepine, might increase psychotic episodes. In addition, the available epidemiological data reveal an inappropriate and excessive prescription of these drugs. Their use and risks should be better known and need an update of the available recommendations.

Racial/ethnic differences in opioid-involved overdose deaths across metropolitan and non-metropolitan areas in the United States, 1999-2017

BACKGROUND

From 1999 to 2017, more than 400,000 Americans died from a drug overdose death involving an opioid. Early surveillance studies have observed large variations in opioid-involved overdose deaths among different geographic regions and racial/ethnic groups. The purpose of this study was to characterize trends in racial/ethnic opioid-involved overdose deaths across metropolitan and non-metropolitan areas in the United States from 1999 to 2017.

METHODS

The analysis used National Vital Statistics System data from 1999 to 2017 that were accessed through the CDC WONDER online database. Drug overdose deaths involving any opioid were identified using the International Classification of Diseases, Tenth Revision, codes and were represented as age-adjusted rates per 100,000 population. Joinpoint regression was used to examine trends in opioid-involved overdose deaths among racial/ethnic groups (non-Hispanic white, non-Hispanic black, Hispanic, non-Hispanic other) by metropolitan and non-metropolitan status (large metropolitan areas, medium-small metropolitan areas, and non-metropolitan areas).

RESULTS

The annual age-adjusted death rates for drug overdose deaths that involved any opioid significantly increased for all racial/ethnic groups in metropolitan and non-metropolitan areas from 1999 to 2017. The largest average annual increases in rates occurred among non-Hispanic whites in non-metropolitan areas (13.6% increase per year) and medium-small metropolitan areas (12.3% increase per year), followed by non-Hispanic blacks in medium-small metropolitan areas (11.3% increase per year).

CONCLUSIONS

The variations in opioid-involved overdose deaths among different racial/ethnic groups across geographic regions support the existence of multiple sub-epidemics in the current opioid overdose crisis and provide directions for targeted intervention efforts.

Methodological Complexities in Quantifying Rates of Fatal Opioid-Related Overdose

PURPOSE OF REVIEW

Effective responses to the US opioid overdose epidemic rely on accurate and timely drug overdose mortality data, which are generated from medicolegal death investigations (MDI) and certifications of overdose deaths. We identify nuances of MDI and certification of overdose deaths that can influence drug overdose mortality surveillance, as well as recent research, recommendations, and epidemiological tools for improved identification and quantification of specific drug involvement in overdose mortality.

RECENT FINDINGS

Death certificates are the foundation of drug overdose mortality surveillance. Accordingly, counts and rates of specific drug involvement in overdose deaths are only as accurate as the drug listed on death certificates. Variation in systematic approaches or jurisdictional office policy in drug overdose death certification can lead to bias in mortality rate calculations. Recent research has examined statistical adjustments to improve underreported opioid involvement in overdose deaths. New cause-of-death natural language text analysis tools improve quantification of specific opioid overdose mortality rates. Enhanced opioid overdose surveillance, which combines death certificate data with other MDI-generated data, has the potential to improve understanding of factors and circumstances of opioid overdose mortality.

SUMMARY

The opioid overdose crisis has brought into focus some of the limitations of US MDI systems for drug overdose surveillance and has given rise to a sense of urgency regarding the pressing need for improvements in our MDI data for public health action and research. Epidemiologists can stimulate positive changes in MDI data quality by demonstrating the critical role of data in guiding public health and safety decisions and addressing the challenges of accurate and timely overdose mortality measures with stakeholders. Education, training, and resources specific to drug overdose surveillance and analysis will be essential as the nation's overdose crisis continues to evolve.

Drug poisoning and associated factors in western Saudi Arabia: A five-year retrospective chart review (2011-2016)

Objective:

Drug poisoning is a globally common cause of emergency-room admissions. This study explores drug-poisoning prevalence patterns, associated risk factors (gender, age and exposure circumstances), and outcomes in western Saudi Arabia.

Methods:

Retrospective analysis of Clinical drug poisoning cases (2011-2016). The data were retrieved from the Saudi Ministry of Health’s record and Patients’ medical charts were analyzed.

Results:

The Ministry of Health received 1,474 reports of drug poisoning during 2011–2016. More than half involved females (n=885, 60%) or young children (0-4 years old) (n=764, 51.8%) and occurred accidentally (n=786, 53.3%); almost all had an oral route of poisoning (n=1,466, 99.5%). The cases most frequently involved analgesic and non-steroidal anti-inflammatory drugs (n=373, 25.2%); antiepileptic, antipsychotic, psychoactive, and anxiolytic drugs (n=229, 16.3%); antihistamine, asthma, flu, and cough drugs (n=157, 12.0%); and antibiotic, anti-fungal; and antiprotozoal drugs (n=74, 5.0%). Antidotes were administered in only 2.2% of cases, and no deaths were reported.

Conclusion:

The drug poisoning cases involved females and young children (younger than 5 years old) and the most cases were accidental, and the most commonly used drugs were analgesics (Panadol), followed by antipsychotics, antihistamines, and antiepileptics (Tegretol).

Epidemiology of overdose episodes from the period prior to hospitalization for drug poisoning until discharge in Japan: An exploratory descriptive study using a nationwide claims database

Background

Little is known about the nationwide epidemiology of the annual rate, causative substance, and clinical course of overdose-related admission. We aimed to describe the epidemiology of overdose episodes from the period prior to hospitalization for drug poisoning until discharge to home.

Methods

We assessed all cases of admission due to overdose (21,663 episodes) in Japan from October 2012 through September 2013 using the National Database of Health Insurance Claims and Specific Health Checkups of Japan.

Results

The annual rate of overdose admission was 17.0 per 100,000 population. Women exhibited two peaks in admission rates at 19–34 years (40.9 per 100,000) and ≥75 years (27.8 per 100,000). Men exhibited one peak in the admission rate at ≥75 years (23.7 per 100,000). Within 90 days prior to overdose, ≥60% and ≥9% of patients aged 19–49 years received a prescription for benzodiazepines and barbiturates, respectively. In addition, 59% of patients aged ≥75 years received a prescription for benzodiazepines prior to overdose, 47% had a history of congestive heart failure, and 24% had a diagnosis of poisoning by cardiovascular drugs. The proportion of patients with recent psychiatric treatments decreased with age (65.1% in those aged 35–49 years and 13.9% in those aged ≥75 years).

Conclusions

The findings emphasize the need for overdose prevention programs that focus on psychiatric patients aged 19–49 years who are prescribed benzodiazepines or barbiturates and on non-psychiatric patients aged ≥75 years who are prescribed benzodiazepines or digitalis.

•

Benzodiazepines were prescribed to ≥59% of overdose patients aged ≥19 years.

•

Cardiovascular drugs were frequently ingested among overdose patients aged ≥75 years.

•

The proportions of patients with recent psychiatric treatments decreased with age.

•

Overdose prevention should be optimized according to the age-related differences.