Rate of Adverse Outcomes During 6-Hour Observation for Asymptomatic Patients with Select Ingestions

Introduction

Of the 2,080,917 human exposures managed by America’s Poison Centers in 2021, 30.44% were managed at healthcare facilities, with most not requiring medical admission [1]. Unless delayed toxicity is anticipated based on the history of the ingestion, some poison centers recommend a 6-hour observation time for asymptomatic patients, even if significant toxicity is not expected. In unpublished data cited by Hollander et al., a 1994 survey of specialists in poison information (SPIs) revealed 79% recommend 6-h observation for suspected benign ingestions [2]. This recommendation was reportedly due to the possibility of co-ingestions. Where this generic 6-h observation time definitively originated is unclear, but it is likely at least in part derived from several studies [3–5].

Very little research has been done to validate the utility of the 6-h observation recommendation, and poison center-based studies are particularly lacking [2, 6]. There is likely a cohort of patients for which 6-h observation is unnecessary. This study investigated the rate of adverse outcomes during a 6-h observation period in patients reporting select ingestions managed by a single poison center.

Methods

An IRB exemption was granted for this study. A retrospective cohort study was performed of patients with a reported intentional or unintentional ingestion called to a single large regional poison center. Closed human (age 13–120 years) exposure cases either in or en route to a healthcare facility with any outcome were included in the study. This age range was chosen to allow reviewers to utilize a single set of normal vital signs when analyzing charts. A convenience sample of calls to a single poison center from January 1, 2021 to December 31, 2021 was collected. The size of the convenience sample was based on the resources available for data extraction. Data was abstracted from the Toxicall® electronic medical record.

Data was collected by the principal investigator (MK) and 3 additional reviewers (SR, EF, CJ) onto a standardized data abstraction form. All reviewers were trained by MK for uniform data extraction. The reviewers each reviewed the same 200 cases pre-reviewed by MK to ensure accuracy prior to starting formal chart review. Reviewers were encouraged to err on the side of including a case, and MK then went through all flagged cases to make final inclusion/exclusion decisions. Discrepancies were decided by MK and a senior investigator (JM). At least 10% of each reviewer’s charts were independently analyzed by JM to ensure data abstraction accuracy. Any discrepancies were flagged and discussed between MK and JM.

Cases were included if 6-h observation was recommended, and on intake the patient had normal vital signs, normal laboratory studies, normal EKG intervals, and a normal physical exam. Vital signs could be reported as “normal” or, if provided, were defined as heart rate less than or equal to 100 beats per minute (bpm) and greater than or equal to 60 bpm, systolic blood pressure greater than 100 mmHg, and respiratory rate greater than 8 respirations per minute (rpm) but less than or equal to 20 rpm. Laboratory studies had to include at minimum a basic metabolic panel, acetaminophen concentration, and salicylate concentration. Laboratory studies could be reported as “normal” or, if provided, sodium was 135–145 mmol/L, bicarbonate was 22–29 mmol/L, acetaminophen was not detected, and salicylate was not detected. The EKG intervals were considered normal if reported as “normal” or if QRS was less than 100 ms and QTc was less than 500 ms. A physical exam was considered normal if reported as “normal” or if the provider reported that the patient is asymptomatic. Vague subjective symptoms like “nausea” or “drowsy” were included, but more severe symptoms like vomiting were excluded. If a full physical exam was provided, it was considered normal if the reviewer felt it described an asymptomatic patient.

Patients were excluded from the study if the poison center requested repeat labs on intake or if the reported ingestion included any of the following xenobiotics: antihypertensive agents, digoxin, tricyclic antidepressants, toxic alcohol, caustics, extended-release products, venlafaxine, lithium, aspirin, acetaminophen, sulfonylureas, metformin, or bupropion. These xenobiotics were excluded because there is consensus that minimum observation time or the need for specific lab trends is indicated. Finally, follow-up had to be provided at or beyond 6 h from ED arrival to ensure no adverse outcomes were missed during the recommended observation period.

Once inclusion and exclusion criteria were met, age, sex, gastrointestinal (GI) decontamination, and suicidal intent were extracted. The xenobiotic was classified as an antidepressant, antipsychotic, analgesic, antiepileptic, antihistamine, antimicrobial, anxiolytic, recreational drug, miscellaneous, or polypharmacy. The following adverse events were recorded: seizure, mechanical ventilation, vasopressor requirement, death, need for observation longer than 6 h, and/or hospital admission related to ingestion (excluding psychiatric admission).

Results

Of 3308 charts reviewed, 107 charts met inclusion criteria/exclusion criteria for analysis (Fig. 1). At least 10% of each reviewers’ charts were assessed for accuracy as described in the methods. Average accuracy when compared to senior reviewer JM was 99.3% (range 99–100%).

Fig. 1.

Breakdown of included cases.

Patient characteristics are listed in Table 1. The most common age group was 18–30 (40.2%), and 73.8% were female. Most cases involved a single drug (71.0%). The most common class of drug involved in a single-drug overdose was analgesics (16.8%). Most of the cases were suicide attempts (88.8%). Most did not receive GI decontamination (82.2%). Of all 107 cases included, there were 4 cases (3.7%) meeting adverse outcome criteria. There were no seizures, deaths, or need for vasopressor or ventilatory support.

Table 1.

Case demographics and outcomes.

Age		N	Percentage
	13–18	32	29.9%
	18–30	43	40.2%
	31–40	8	7.5%
	41–50	7	6.5%
	51–60	11	10.3%
	61–70	4	3.7%
	71–80	2	1.9%
	> 80	0	0%
Sex		N	Percentage
	Male	25	23.4%
	Female	79	73.8%
	Transgender	3	2.8%
Class of drug		N	Percentage
	Antidepressant (single)	16 (12 SSRIs, 4 other)	15.0%
	Antipsychotic (single)	9	8.4%
	Analgesic (single)	18 (all NSAIDs)	16.8%
	Antiepileptic (single)	6 (1 pregabalin, 5 gabapentin)	5.6%
	Antihistamine (single)	9 (6 hydroxyzine and 3 diphenhydramine)	8.4%
	Antimicrobial (single)	0	0.0%
	Anxiolytic (benzodiazepines)	12	11.2%
	Recreational drug (single)	1 (1 “ecstasy”)	0.9%
	Misc (single)	5	4.7%
	Polypharmacy	31	29.0%
Suicidal intent		N	Percentage
	Yes	95	88.8%
	No	10	9.3%
	Unknown	2	1.9%
Activated charcoal documented as given		N	Percentage
	Yes	19	17.8%
	No	88	82.2%
Adverse outcome		N	Percentage
	Death	0	0.0%
	Mechanical ventilation	0	0.0%
	Vasopressor support	0	0.0%
	Seizures	0	0.0%
	> 6-h observation	2	1.9%
	Hospital admission	2	1.9%
`	No adverse outcome	103	96.3%

Reviewing the individual cases meeting adverse outcome criteria provides further insight. One case required a longer observation time (medically cleared at approximately 13.5 h after arrival) following a dose of benzodiazepine given for agitation in the department. Whether the agitation was secondary to a toxicologic cause is unclear. Another had vomiting greater than 6 h after ingestion of ibuprofen and was admitted for observation after this without further incident, but the authors feel this was unlikely due to the reported ingestion. The third case described a patient who was admitted for a clearly non-toxicologic cause for the patient’s presentation. Finally, the fourth case developed tachycardia during the 6-h observation period after an ingestion of fluvoxamine and immediate release amphetamine/dextroamphetamine. This ultimately required longer observation in the emergency department (ED), with the patient medically cleared approximately 16 h after arrival.

Discussion

In this very select cohort of asymptomatic patients presenting after an ingestion that does not include a xenobiotic listed in the exclusion criteria, low rates of adverse outcomes occurred. These results are not particularly surprising, but support what would likely be an intuitive hypothesis—that asymptomatic patients presenting after “less-concerning” ingestions often do well.

Other studies have had similar findings, albeit utilizing different methodologies. Hollander et al. investigated this question about appropriate observation time from an ED lens. They found that patients whom the emergency physician deemed safe to clear at 2 or 4 h did not go on to develop unexpected toxicity [2]. However, their ED study was completed more than 20 years ago, acknowledged several important biases in their limitations, and new trends of exposure and new xenobiotics limit the generalizability of that study to exposures managed in healthcare facilities today. More recently, Capua et al. also found that the disposition of asymptomatic pediatric ingestions could be determined before 6 h in a retrospective review of selected cases at a pediatric ED [6].

None of these studies to date, including this one, allow for broad alterations in observation protocol. It is impossible to account for all variables in patient presentation with the near limitless list of possible exposures or combination of exposures. Encouraged by these studies, however, future research could begin to look at more specific common exposure scenarios to determine if observation times could be safely shortened. As adverse outcomes are expected to be low in these select cohorts, these studies will require large numbers to avoid missing rare events. Our adverse event rate was low, but 4 cases out of 107 is not insignificant. More rigorous research would be useful to help minimize hospital resource utilization without affecting patient safety.

Limitations

First, the narratives used were obtained from poison center records, so the information contained is limited to the reports provided to the poison center. For example, confirmation of exposure was rarely available, and the poison center is unable to completely verify the accuracy of every report. Further, the sample size is small, and the convenience sampling method could be a source of introduced bias of which the degree cannot be determined. In addition, the retrospective nature of this study was a notable limiting factor. Reviewers had to make assessments on whether the patient was asymptomatic based on information in the chart. There is subjectivity on what constitutes “asymptomatic,” although it was attempted to minimize the subjectivity by having several levels of chart assessment, and the comparative accuracy of the reviewers for select charts was excellent. The resources and staffing available at the different healthcare facilities was not uniform and may have impacted the findings in this study. Finally, each poison center is different in its ability and frequency of follow-up, limiting generalizability between centers.

Conclusion

In this selected cohort of patients reported to a single regional poison center, there was a low rate of adverse outcomes during a 6-h observation period in a healthcare facility. Future research is needed to determine which exposures and clinical parameters may allow for 6-hour or shorter observation time. Minimizing observation time while not compromising patient safety is an important goal.

Sources of Funding

None.

Declarations

Conflicts of Interest

None.