US Hospital Stays in Children and Adolescents With Acetaminophen Poisoning Free

Pediatric suicide attempts by medication self-poisoning have risen substantially in recent years.^1,2  Acetaminophen is one of the most common over-the-counter medications sold in the United States, and its easy accessibility may be influential in children and adolescents who attempt suicide by self-poisoning. Data characterizing acetaminophen-related hospitalizations and outcomes are needed to help address rising rates of suicide and develop novel interventions to prevent the tragedy of a successful suicide attempt.

Whether intentional or not, acetaminophen poisonings occur in all age groups and have drawn heavily on the resources of poison control centers, emergency departments, and inpatient units.^3,4  The literature to date largely describes the care of adults with acetaminophen poisoning^5,6  or combines data for children and adults^3,7–10  such that pediatric specific conclusions are difficult to ascertain. Less detailed information about outcomes is available about hospitalizations of children and adolescents,^3,8,11  particularly in the context of self-poisoning.^4,7  Previous studies that included hospitalizations of children and adolescents were comprised of predominately adult patients.^3,8  In these studies, rates of admission, gender, intentionality, hepatotoxicity, and mortality were not separately reported from the adult population. One pediatric study of adverse drug-related hospitalizations and emergency department visits included those with acetaminophen poisoning but excluded those related to self-poisoning.⁴  In addition, the rate and outcomes of acute care hospitalizations attributed to all acetaminophen poisonings in children have not been evaluated in the last 10 years.

The objectives of this study were to use a nationally representative inpatient data set to describe the incidence, patient characteristics, and outcomes of pediatric hospitalizations related to acetaminophen poisoning in the United States and evaluate the contribution of intentionality. These findings may inform development of medication poisoning, suicide prevention strategies, and targeted interventions.

This retrospective analysis used hospital discharge data from the 2016 Kids’ Inpatient Database (KID), which is part of the Healthcare Cost and Utilization Project (HCUP) sponsored by the Agency for Healthcare Research and Quality.¹²  The KID is a public-use administrative data set comprised of: (1) a 10% random sample of discharge records for uncomplicated birth hospitalizations, and (2) an 80% random sample of discharge records for all other pediatric inpatient and observation stays in each nonfederal, short-term general and specialty hospital in participating states. The KID does not include stays in psychiatric facilities. For 2016, the KID contains data on >3 million (unweighted) discharges of children aged <21 years in 4200 U.S. hospitals in 46 states. It includes discharge-level information, including patient demographics, payer, hospital-assigned diagnostic and procedure codes, length of stay (LOS), total hospital charges, and discharge disposition. The KID sample can be used to generate national estimates for hospital stays of children and adolescents in the United States. Exact results are suppressed when ≤10 to maintain anonymity in accordance with the HCUP data use agreement.

We used previously validated International Classification of Diseases, 10th Revision, (ICD-10) discharge diagnostic codes to identify U.S. hospitalizations associated with acetaminophen¹³  and self-harm medication poisonings.¹⁴  ICD-10 codes for intentional self-harm poisoning previously demonstrated a positive predictive value of up to 91% in adolescents.¹⁵  A complete list of included ICD-10 codes is available in the Supplemental Table 3. There were no exclusions from the identified population.

Demographic and clinical variables included: age, sex, payer, household income (quartile by zip code), discharge codes indicating mental health diagnosis, intentionality, mechanical ventilation, hepatotoxicity, liver transplantation (during the same stay), LOS (unweighted), unadjusted hospital charges, and discharge disposition, which included in-hospital mortality.

Analysis was restricted to discharges of patients aged 0 to 19 years. We used survey methods recommended by the HCUP to generate weighted national population estimates and to describe characteristics and outcomes. Variances were estimated within domains and strata using models that included all data in the KID sample and accounted for sampling weights and clustering at the hospital level. Specifically, we calculated annual hospitalization rates using the estimated number of acetaminophen ingestion-related discharges (and their confidence limits) as numerators and U.S. Census estimates (assuming no variance) for children aged 0 to 19 years as denominators.¹⁶  We further restricted our secondary analysis to patients aged 10 to 19 years and assessed characteristics associated with intentionality using unadjusted odds ratios. All estimates are based on nonmissing values in the KID sample, and associated confidence intervals (CIs) were generated under the (conservative) assumption that missing values were not missing completely at random.¹⁷ 

Analyses were conducted in SAS (SAS Institute, Cary, NC). Our institutional review board waived formal review of this study protocol.

Of the 31 632 hospital discharges in 2016 associated with a principal or secondary diagnosis of self-harm medication poisoning in children aged 0 to 19 years, acetaminophen was the most commonly implicated agent and accounted for 27.0% (95% CI, 26.2–27.8) of discharges. Of 37 404 discharges assigned to a diagnosis-related group indicating poisoning and toxic effects of drugs, the single most common principal diagnosis code (T391 × 2A) was acetaminophen-related and was listed for 15.7% (95% CI, 15.2%–16.2%) of discharges.

There were 9935 discharges (95% CI, 9252–10 619) from acute care hospitals for acetaminophen poisoning in children aged 0 to 19 years during 2016 (Table 1). This corresponded to a population rate of 12.1 hospitalizations per 100 000 children in the United States (95% CI, 11.3–12.9). The rate reached a peak of 35.5 discharges per 100 000 (95% CI, 33.2–37.7) in adolescents aged 15 to 19 years. In all, 85.9% of discharges for acetaminophen poisoning had a code indicating self-harm. Almost two-thirds (60.9%) were attributed to acetaminophen ingestion alone, as opposed to in combination with another medication. Similarly, 84.8% of all discharges for acetaminophen poisoning had a concurrent mental health diagnosis (eg, depression, anxiety) upon discharge. The majority of hospitalizations for both intentional and unintentional acetaminophen poisoning occurred in females, and the sex-distribution patients was strongly related to age (Fig 1). Median LOS was 2 days (interquartile range [IQR], 1–4 days), but almost half of discharges were associated with transfer to another type of facility, such as a psychiatric unit or hospital. The median LOS was 2 days (IQR, 1–4 days) for stays associated with discharge to home care and 2 days (IQR, 1–3 days) for those associated with transfer to another type of health care facility. Liver transplantation occurred in ≤10 discharges. Similarly, in-hospital death occurred in ≤10 discharges. Median unadjusted hospital charges for acute care were $14 379 (IQR, $9162–$23 114), totaling $204.7 million (95% CI, $187.4 –$221.9) in aggregate.

Acetaminophen poisoning hospitalizations were then restricted to ages 10 to 19 years old and stratified by intentionality (Table 2). There were significant univariate associations between intentional self-poisoning and being 10 to 14 years old, female, privately insured, in the top income quartile, and having a comorbid mental health diagnosis. Adolescents with intentional poisoning were also more likely to be transferred to another type of facility (including psychiatric facilities).

Our findings demonstrated that acetaminophen poisoning was the most common cause of U.S. hospitalizations associated with medication poisoning in children and adolescents in 2016. These hospitalizations were common, occurring an average of 27 times per day or >1 every hour. The majority occurred in adolescent females and were related to self-harm with acetaminophen alone or in combination with other medications. Although these hospital stays were typically brief, associated with charges comparable to typical pediatric stays,¹⁸  and rarely linked to severe outcomes (transplantation, death), almost half were transferred to other types of facilities (eg, psychiatric units) that would be expected to add (uncaptured) charges to total charges for each episode of care.

U.S. hospital discharges of children with acetaminophen poisoning have remained high and constant since the late 1990s.³  Discharges in 2016 were more often in females (79.8%) and attributed to intentional ingestion (85.0%), which are similar to findings reported in studies of previous time periods and/or countries.^{3,7,8,19–21}  We found very few adolescent admissions for unintentional acetaminophen self-poisoning. This is in contrast to the sizable proportion of adults hospitalized for “therapeutic misadventures.”^19,22  Discharges without an associated intentionality code had a significantly higher likelihood (10.5%) of having a diagnosis code for hepatotoxicity; the clinical significance of this, if any, is unknown. Discharges with codes indicating intentional poisoning were more likely to occur in the younger adolescent age group (10–14 years) than discharges coded for nonintentional ingestion. Our results show that U.S. hospitalizations related to acetaminophen poisoning peaked at age 16 years in females but continued to rise throughout adolescence in males, particularly in those with intentional self-harm. This may suggest that different interventions are needed based on gender and age.

Acetaminophen poisoning remains prevalent despite extensive prevention efforts, including altering packaging, adding package insert warnings, and regulating acetaminophen-containing prescription products.²³  Although child-resistant packaging has been effective in reducing accidental medication poisoning in young children,²⁴  effective strategies to reduce intentional acetaminophen poisoning in adolescents have not been established. As adolescent suicide attempts by self-poisoning continue to rise substaintially,^1,2  development of prevention strategies should be a top priority. Moreover, intentional ingestions may reveal modifiable risk factors for future self-harm attempts because the majority of these encounters are also coded for a mental health diagnosis. Admissions for self-poisoning may provide an opportunity to identify and treat underlying mental health conditions,²⁰  and hopefully intervene before a successful suicide attempt.

Findings of this observational study should be taken in the context of its limitations. KID discharge-level data limit evaluation of patient risk factors associated with acetaminophen poisoning and preexisting mental health diagnoses and preclude describing the full extent of care across health care settings. Although we speculate that those transferred to another acute care facility were transferred to a tertiary care children's hospital, it is not possible to establish person-level linkage of related hospital discharge records and we cannot differentiate the type of acute care hospital in KID. Thus, it is possible that some patients who were transferred to another acute care hospital were captured more than once in our analysis. KID data elements also do not distinguish between transfers to skilled nursing facilities and transfers to other types of nonacute-care facilities such as psychiatric facilities. Finally, identification of acetaminophen poisoning hospitalizations depends on discharge diagnosis coding accuracy. Although we applied previously validated codes for acetaminophen ingestion^13,14  and self-harm ingestion¹⁵  with a positive predictive value of up to 91.3% in adolescents (95% CI, 86.6–94.8), it is surprising that 54% of hospitalizations for unintentional ingestion had a mental health diagnosis at discharge. The limitations of discharge-level observational data precluded us from investigating these relationships further in this study. Recent changes in discharge diagnosis coding from ICD-9 to ICD-10 prevented us from assessing trends in acetaminophen poisoning over time.

Acetaminophen poisoning remained a frequent reason for pediatric hospitalization in the United States in 2016, particularly in adolescents. Future studies that evaluate the trend of acetaminophen poisonings, as well as their outcomes, are important next steps. Additional prevention strategies, identification of at-risk individuals, and targeted interventions are needed to reduce the burden of this common problem.