Articles You Might Have Missed

Article #1: Capua M, Amlicke M, Esposito E, Et al: Time of observation in xenobiotic ingestions in children: is 6 hours too long? Pediatr Emerg Care 2023; 39(1):e24-e29

Background: Pediatric ingestions are common presentations to the emergency department (ED) and calls to poison control centers. At present, most asymptomatic patients that present to the ED after ingestion are observed for six hours. Adult studies suggest that this observation period may be unnecessary, and patients who present asymptomatic infrequently develop symptoms. Thus, a shorter period may be sufficient.

Research Questions: Do observations of pediatric patients with xenobiotic ingestions, who present to an ED asymptomatic, change management or outcomes? Do initial vital signs (VS) impact these patients’ outcome and management?

Methods: This was a retrospective chart review of patients < 18 years presenting to a single tertiary care pediatric ED for ingestion from May 2014 to December 2018. Patients were identified using ICD 9/10 codes for toxic ingestions and included if they presented with normal VS and no documented toxidrome. Transfers from outside hospitals, symptomatic patients, patients with abnormal blood glucose (not 70–150 mg/dL) and the following ingestions were excluded: acetaminophen, caustics, hydrocarbons, salicylates, sulfonylureas, and extended-release products. In the secondary analysis, a set of patients with abnormal VS as the only outlying variable was used to compare to the original group. Interventions were defined as laboratory tests, EKGs, medications, or imaging studies.

Results: From a total of 2,817 charts, 430 asymptomatic subjects were identified. Of these, 109 subjects (mean 4.7 years, median 2.7 years) with 103 known and six unknown xenobiotic ingestions were included. Another 321 subjects with abnormal triage VS only underwent secondary analysis. The average time from ingestion to triage was 2.3 hours. The most common ingestions were non-steroidal anti-inflammatory drugs (NSAIDs, 13.6%) and cardiac medications (9.7%). Only two (1.8%) cases were suicide attempts, and eight (7.3%) were multiple xenobiotic ingestions. Twenty-four (22%) subjects underwent an intervention within the first hour; 12 subjects had an intervention, and one was admitted, between one and two hours. In the third hour of observation, 12 more subjects received and intervention; during the fourth hour, eight subjects had an intervention, and three were admitted. After four hours of observation, seven more subjects had an intervention, and one was admitted; no further interventions were performed after six hours. All five (4.6%) admitted subjects were discharged within 24 hours; one returned within 72 hours, with emesis and loose stools related to a laxative ingestion.

In the abnormal VS subset of 321 subjects (mean 6.6 years, median 3.1 years): 22 (6.8%) were admitted, including nine subjects after six hours. The most common abnormal VS was elevated blood pressure. The most common ingestions were NSAIDs (13.2%) and cardiac medications (7.3%). There were 36 (11.2%) intentional (suicidal) ingestions and 26 (8%) that involved multiple xenobiotics. After six hours, 16 (5%) subjects received interventions, and nine (2.8%) were admitted. No subjects returned to the ED for a complaint related to the initial exposure; three returned, as instructed, for repeat tests or consultation. The odds ratio for admission was not significantly different between the normal and abnormal VS groups (odds ratio [OR] 1.5; 95% confidence interval [CI] 0.57–4.15) despite abnormal VS being more likely following suicidal ingestions (OR 6.8; 95% CI 1.6–28.6).

Conclusion: In this study, which excluded many common ingestions, most asymptomatic pediatric patients with xenobiotic ingestions required interventions or admission within the first four hours of ED presentation. Prospective studies are needed to better identify cases that can be safely managed with shorter observation periods.

Critique: This was a retrospective chart review with the potential for missing or incorrect data and cases, and the reported ingestions were not confirmed. This single-center study might lack generalizability; re-admissions to other facilities were not ruled out. This small, heterogeneous study population might be underpowered to detect rare but severe outcomes. The need for interventions or admission were not analyzed; the effects of intentional (suicidal) ingestions on interventions/admission were not characterized.

Implication for Toxicologists: Although this study suggests that low-risk asymptomatic pediatric exploratory ingestions may not require 6-hour observation, more work is required. Toxicologists and poison centers may help shorten observation periods in some cases, but the required data to do this safely is presently unknown.

Article #2: Marco CA, Studebaker H, Repas SJ, et al.: Clinician assessment of blood alcohol levels among emergency department patients. Am J Emerg Med 2023; 63:110–112.

Background: Alcohol intoxication is a frequent problem in the ED, resulting in millions of ED visits per year. Blood alcohol levels (BALs) are frequently measured but do not correlate well with clinical effects.

Research Questions: The primary question was the accuracy of clinicians estimates of patients’ BALs and actual (laboratory derived) BALs in an ED? Secondary analyses included correlations between BAL accuracies with patient demographics (age, gender, ethnicity, insurance type or mode of arrival) and clinicians’ roles. For example: was BAL accuracy affected by clinician’s role?

Methods: This was a prospective survey study at a single United States Level 1 Trauma Center’s ED. Study participants included attending and resident physicians, physician assistants (PAs), registered nurses (RNs), and medical students (MSs) caring for patients with a measured BAL. Patients in the ED with a pending BAL were identified; a survey was administered by study investigators before BAL results reporting. The primary outcome measure (accuracy of clinicians’ estimated BAL) was calculated as the difference between actual and estimated BALs. Pearson correlations were measured for patient and clinician variables affect on accuracy.

Results: The study included 243 clinician data points from 237 clinicians (35.8% RNs, 25.5% resident physicians, 26.8% attending physicians, 9.7% MSs, and 1.7% PAs). The range of clinicians’ experience was measured between > 5 years (70%) and > 20 years (9.2%). Most patients were male (66.3%), white (69.4%), and brought in via ambulance (75.3%). The average patient age was 45.2 years (standard deviation [SD] 15 years). Most clinicians overestimated the BAL by an average of 17.4 (95% CI 4.7–30.1). No significant difference was found in the accuracy stratified by clinician role or any patient characteristic except mode of arrival. Clinicians tended to underestimate the BAL for walk-ins (-14.9; 95% CI -32.8–3.1) and overestimate for ambulance arrivals (28.3; 95% CI 12.7–44). Interestingly, 107 patients had a BAL of zero; clinicians’ BAL estimates ranged from 0–350 (17% of these were non-zero estimates). A “Tox screen” (method/assay not reported) was performed on 90 patients with the most common results showing THC (52.2%), nothing detected (37.8%), and cocaine (14.4%).

Conclusion: Clinicians’ estimates of BAL were often inaccurate, regardless of clinical role and experience. All clinicians tended to overestimate the BAL, even in patients without a measurable BAL.

Critique: This survey study was collected as an unblinded convenience sample at a single center, which introduced selection biases on subjects (clinicians) and their patients. The reasons for an ordered BAL were not provided. This omission, as well as the finding of non-detectable BAL in 44% of patients may have affected accuracy. The actual ethanol assay (serum or whole blood) and units (e.g., mg/dL) were not reported. The presence of trauma, abnormal glucose concentrations, or signs/symptoms suggesting concurrent drug intoxication were not reported. The limited data concerning the “Tox screen” is also problematic. The total time and timing (at arrival or sometime later) of the clinicians’ patient evaluations are unknown; and the interval between the blood draw and clinician survey was not standardized or reported. Finally, ethanol in isolation may have added difficulty for clinical estimation because the degree of clinical intoxication may be related to the patient’s tolerance for ethanol. In acute intoxication, the Mellanby Effect describes the paradox where an individual appears to be more clinically intoxicated in the absorption phase than the elimination phase for the same BAL.

Implication for Toxicologists: This study confirms that BAL cannot be accurately estimated by medical professions. Toxicologists and poison center staff must remind bedside providers to consider other etiologies for suspected alcohol intoxication.

Article #3: Marcotte TD, Umlauf A, Grelotti DJ, et al.: Driving performance and cannabis users' perception of safety: a randomized clinical trial. JAMA Psychiatry 2022; 79(3):201–209.

Background: As the legal landscape across the US lessens restrictions on cannabis, consumption has increased and will likely continue. Delta-9-tetrahydrocannabinol (THC; active cannabis component) negatively impacts cognitive functioning and in turn, driving performance. The data on the effect of cannabis on driving are inconsistent; questions remain regarding the degree and length of impairment after THC use.

Research Questions: What is the effect of ad-lib cannabis smoking on driving performance as a function of time? Secondary questions included: does THC concentration impact drive?; how does the subject’s self-perception of impairment correlate with driving performance?; and how do blood THC concentrations correlate with driving performance when stratified by degree of THC use?

Methods: This was a double-blind, placebo-controlled parallel randomized control trial conducted in San Diego, CA from February 2017 to June 2019. Participants were recruited from the community and eligible if they were 21 to 55 years of age, used cannabis more than four times in the past month, held an active driver’s license, had driven 1000 miles in the past year, and were willing to not use cannabis for two days before the study. Exclusion criteria included any history of traumatic brain injury, significant medical comorbidities, pregnancy, concurrent substance use, past year diagnosis of a substance use disorder, active suicidal ideations, history of schizophrenia or bipolar with mania, or evidence of THC or alcohol use on the day of study (e.g., THC concentration > 5 ng/mL or positive urine drug screen; positive alcohol breathalyzer).

Participants were stratified by frequency of cannabis use and then randomized into three groups of THC-cigarette concentration (13.4%, 5.9%, or 0.02% = placebo). Participants were given a one-hour simulator training session and a 25-minute baseline simulated drive (that included common city and rural traffic challenges) before the experiment day. Participants were told to “smoke the [assigned THC concentration] cigarette the way you do at home to get high.” They were given ten minutes and had to take at least four puffs. During this, whole blood was collected to measure THC concentrations 15 minutes after smoking. Before each driving session, participants were asked a series of questions about their perceived intoxication and ability to drive safely. A Composite Driving Score (CDS) comprised key driving simulator variables; higher scores indicated worse performance. Simulations were performed at 0.5, 1.5, 3.5, and 4.5 hours after smoking.

Results: A total of 191 (out of an initial 199) subjects were included in the final study. The amount of material smoked during sessions did not differ between the THC and placebo groups. The venous THC concentrations measured 15-minutes after use differed significantly between the three groups with the 5.9 THC% group having the highest (mean [SD]: placebo, 1.3 [1.9] ng/mL; 5.9 THC%, 50.6 [40.8] ng/mL; 13.4 THC%, 32.7 [29.3] ng/mL). Blood THC concentrations were statistically different between the “low” regular use group as compared to the “middle” and “highest” groups (based on quartile stratification).

The two THC groups were combined for many comparisons with a placebo. Compared to the placebo group, both THC groups had a significant decline in CDS performance at 0.5 and 1 hour, the borderline difference at 3.5 hours, and no differences at 4.5 hours. THC content or use intensity did not affect CDS scores over the prior 6 months. While the THC group self-reported more perceived driving impairment at all times; 45.7% would have driven at 0.5 hours and 68.6% at 1.5 hours. There were no correlations between blood THC concentrations (r = 0.025) and CDS scores.

Conclusion: Smoking THC impairs simulated driving for at least 3.5 hours. This objective impairment lasts longer than the perceived impairment and is not predicted by prior smoking frequency or blood THC concentrations.

Critique: The study was well designed but did have several limitations, including a narrow population of cannabis users (smoking only and excluding non-THC users) and could not control for the subjects’ desired ‘highness’ during ad-lib smoking. The study was not able to address the effects of repeated smoking (“dose stacking”) on CDS scores. It would be interesting to see the effects on a naive or infrequent group of users.

Implication for Toxicologists: Toxicologists and poison center staff must be aware of THC-induced driving impairment for at least 3.5 hours and that THC concentration or blood levels do not predict these effects. Subjective perception of THC effects is also a poor predictor of complex behaviors.

Article #4: Pistore A, Penney S, Bryce R, et al.: A retrospective evaluation of phenobarbital versus benzodiazepines for the treatment of alcohol withdrawal in a regional Canadian emergency department. Alcohol 2022; 102:59–65.

Background: Alcohol withdrawal syndrome (AWS) is a common presentation in the ED with a high morbidity if left untreated. Many protocols include the Clinical Institute Withdrawal Assessment for Alcohol-Revised (CIWA-Ar) scale and utilize benzodiazepine monotherapy for treatment despite growing evidence of benzodiazepine resistance in some patients. Few studies have compared the use of benzodiazepines versus barbiturates for treating AWS.

Research Question: Among adult patients experiencing AWS in the ED, does phenobarbital versus diazepam monotherapy impact ED length of stay (LOS) or rate of admission?

Methods: This study was a retrospective, observational quality improvement project of ED patients ≥ 18 years treated for AWS between June 2019-January 2021 at a single, small hospital in Saskatchewan, Canada. Subjects were included if they were treated with either phenobarbital or diazepam monotherapy algorithms for AWS, as chosen at the discretion of ED physicians. Before the study period, clinical management algorithms with asymptomatic and maintenance periods were developed locally for monotherapy with diazepam or phenobarbital (with the intent to increase use of phenobarbital). Phenobarbital treatment included a loading dose of 10 mg/kg intravenous (IV), 120–240 mg 60 minutes later, then doses every 30 minutes as needed followed by maintenance phase (100–200 mg intramuscularly or orally [PO] every hour, as needed). The diazepam algorithm received as needed diazepam for CIWA-Ar score > 7; 10 mg PO or IV every 15 minutes until stabilized, then as needed, hourly doses.

Subject demographics, psychosocial factors, and clinical course (including protocol adherence, additional therapies, ED LOS, disposition, and return to the ED) were recorded. Descriptive statistics, log-rank testing, and univariate/multivariate logistic regressions were conducted to detect meaningful associations between pathways and outcomes.

Results: Thirteen physicians treated 83 subjects over 184 ED visits for AWS during the study period. Subjects’ average age was 39.8 years (SD 9.9 years), 64.7% were male, and half (54.2%) presented for AWS treatment once (range 1–12 ED presentations). Fifty-six (30.4%) were treated with phenobarbital, and 128 (69.6%) with diazepam. A total of 9.4% of phenobarbital versus 17.1% of diazepam visits resulted in hospitalization (p = 0.20). Those treated with phenobarbital presented with greater severity (CIWA-Ar > 15); 76.8% compared to 66.4% (p = 0.04). Median ED LOS was 4.4 hours for both groups (phenobarbital interquartile range [IQR] 3.8–5.4; diazepam IQR 3.7–5.1; p = 0.21). At time of ED departure, the mean CIWA-Ar scores for phenobarbital and diazepam treated subjects were, respectively: discharged 5.0 versus 5.3; admitted 2.0 versus 3.9. Regression analysis did not find a significant difference in ED LOS between treatment groups. The phenobarbital algorithm resulted in 9.4% (5/53) admissions compared to 17.1% (21/123) for those treated with diazepam. When adjusted for confounders, the phenobarbital treatment pathway was associated with a 71.3% (95% CI 9.8%-90.9%; p = 0.03) decreased likelihood of admission. There were 10 treatment pathway deviations (e.g., use of non-algorithmic GABA-agonists).

Conclusion: Monotherapy of AWS with phenobarbital performed similarly to diazepam and may result in a lower rate of hospitalization.

Critique: This was a single-center non-randomized and non-blinded observational study with a small population which limits its generalizability and may have been underpowered to identify small differences. The implementation of the new phenobarbital pathway lacked formal education or training for providers, which could have led to underutilization. The study enrolled subjects with repeated ED treatments by different physicians that appeared to have been biased by prior treatment. Subjects treated with phenobarbital were sicker (e.g., had higher CIWA-Ar scores) at presentation. The study was conducted during the COVID-19 pandemic, where admission and discharge criteria were variable depending on each hospital. Finally, the study did not address the frequency of adverse effects from the two therapies or outcomes beyond the ED.

Implication for Toxicologists: Phenobarbital is useful as treatment for AWS, both as an adjunct and monotherapy. Toxicologists should assist with educating other providers in its use for AWS.

Sources of Funding

No sources of funding were utilized.

Declarations

Conflicts of Interest

None.