Intoxication (or poisoning) that necessitates hospitalization remains an important source of morbidity in children.
To determine changes, during an 11-year period (1987-1997), in the incidence of hospitalization due to intoxication among children in Washington State and circumstances of ingestion, agents used, hospital length of stay, charges, and mortality.
A computerized database of all hospital discharges (Comprehensive Hospital Abstract Reporting System [CHARS] database) in Washington was used. Cases included all children younger than 19 years with a primary or secondary diagnosis for an intoxication or with an external cause of injury code (E code) for an intoxication from 1987 to 1997.
There were 7322 hospitalizations (45 per 100 000 children per year); the annual rate significantly decreased during the study period. Most patients (75%) were teenagers. Sixty-five percent were female. Pharmaceutical agents were used in 80% of cases. Analgesics were the most commonly used (34%), followed by antidepressants (12%) and psychotropic drugs (8%). Nonpharmaceutical agents were more prevalent in children younger than 12 years than in teenagers. Self-inflicted intoxication was the most frequent cause identified by E codes (47%). Median length of stay was 1 day, and median hospital charges were $2096. Mortality was low (0.2%) and did not change significantly over time.
Acute intoxication continues to be an important cause of hospitalization in children. The type of agent involved did not change significantly over time. Teenage girls continue as the highest risk group for suicide attempt from ingestions. Self-inflicted intoxications were associated with higher costs, length of stay, and readmissions. Although preventive measures anddevelopment of poison centers have contributed to decrease mortality from acute intoxication in children in the last 50 years, efforts need to be targeted toward suicide prevention, especially among teenage girls.
ACUTE INTOXICATION (or poisoning) in children can lead to serious complications, hospitalization, and even death. Pediatric mortality from intoxications has declined considerably in the last 30 years because of multiple preventive measures. In 1950, there were 834 deaths from intoxications among US children younger than 5 years; this number decreased to less than 50 per year in 1997.1 Nevertheless, acute intoxication remains an important cause of illness in children. In 1997, 66 participating US poison centers reported 2 192 088 human exposure cases.2
Children younger than 20 years were involved in two thirds of the cases.
The kind of agents used, the cause of intoxication, and the population at risk may change over time; recognition of such changes might enhance preventive measures and treatments to reduce morbidity and mortality related to childhood intoxications. Few studies have evaluated childhood intoxication that results in hospitalization. Our study describes the population of children hospitalized for intoxication, from 1987 through 1997, in Washington State. We determined the incidence of acute intoxications leading to hospitalization in children and evaluated changes in intoxicating agents during the study period. We also evaluated the circumstances of the intoxication to determine if changes occurred during the study time. Finally, fatalities, hospital length of stay (LOS), and hospital charges associated with acute intoxication were calculated.
After institutional review board approval by the University of Washington and the State of Washington Health and Human Services Department, we obtained computerized information for persons hospitalized in Washington from 1987 through 1997. Data from 100 hospitals in Washington were included. All hospitalized patients (younger than 19 years) with a discharge diagnosis of intoxication were identified in the Comprehensive Hospital Abstract Reporting System (CHARS) database. Subjects were included if they had a primary or secondary diagnosis code for intoxication by drugs, medicinal, and biological substances (International Classification of Diseases, Ninth Revision
[ICD-9] codes 960-979); toxic effects of substances chiefly nonmedical (codes 980-989); or nondependent abuse of drugs (code 305).3 However, cases with only nondependent use of tobacco (code 305.1) were excluded because we did not consider it to be a primary reason for hospitalization. The following external causes of injury codes (ICD-9 E codes) were also identified to determine the cause of the intoxication: unintentional poisoning by drugs, medicinal substances, and biological substances (E850-E858); unintentional poisoning by other solid and liquid substances, gases, and vapors (E860-E869); suicide and self-inflicted poisoning by solid or liquid substances or by gases (E950-E952); assault by poisoning (E962); and poisoning undetermined whether unintentionally or purposely inflicted (E980-E982).3 Children with E codes for intoxication were included as cases. All newborns and patients transferred from other hospitals were excluded to avoid counting the same patient twice. We also excluded all intoxication ICD-9
codes that were associated with an adverse reaction E code (E930-E949) to avoid cases where the intoxication was iatrogenic or occurred during a hospitalization.
Demographic data (age, sex), hospital data (total charges, LOS, mortality), and agents used were analyzed. Patients records were linked, and all subsequent hospital admissions for intoxication during the study period were also evaluated.
The incidence of hospitalization for intoxication was determined using census data in Washington (average of annual populations for a specified period). Normally distributed continuous data were compared using the t test and 1-way analysis of variance. The Tukey B test was used to adjust for multiple pairwise comparisons among the years of study. Nonparametric data are reported as medians and 25th to 75th quartiles. Nonparametric data were compared with the Kruskal-Wallis test and the Mann-Whitney test. Categorical data were examined using the χ2 test. SPSS 9.0 for Windows (SPSS Inc, Chicago, Ill) was used for all statistical calculations. Statistical significance was defined as P<.05.
There were 7322 hospitalizations for intoxication from 1987 to 1997 in Washington State. The average incidence of hospitalization for intoxication in children was 45 per 100 000 children per year, and intoxication accounted for 0.06% of all pediatric hospitalizations during the 11-year study period. Children aged 12 to 18 years were the largest patient group (75%), followed by children aged 0 to 5 years (20%) and children aged 6 to 11 years (5%). Adolescents constituted most of the sample, and global results mainly reflect this category of patients. Ingestions with more than 1 agent were reported in 10% of cases. The median LOS was 1 day (range, 1-3 days); the median hospital charges were $2096 (range, $1246-$3519). Only 15 children (0.2%) died.
Pharmaceutical agents were identified in 80% of the intoxications (Table 1). Analgesic agents were used in a third of cases (34%), and acetaminophen was the most common medication in this category (18%). Antidepressant agents constituted the next most common category of medication leading to hospital admission (12%), followed by psychotropic agents (8%) and anti-allergy or antiemetic drugs (6%). Nonpharmaceutical agents were identified in 22% of patients with an intoxication. The most common agents were alcohol (6%), followed by "street" drugs (4%) and fumes (3%).
Features of the intoxication are presented by age categories in Table 2. Girls were significantly more likely to have a hospital admission due to an intoxication in the teenager group (2.5 times more frequent); boys were more frequently involved in all younger groups. Comparing the 3 different age categories, analgesic ingestions were significantly more common in teenagers (42%). Nonpharmaceutical agents were more frequently involved in children younger than 12 years (36%) than in teenagers (17%). Nonpharmaceutical agents included mainly bites or venom, fumes, cleaning agents, and solvents or hydrocarbons. Multiple agents and self-inflicted intoxications were significantly more common in teenagers compared with younger children. Unintentional intoxications were least common among teenagers (13%) compared with younger children (58% in the 6- to 11-year-old group and 75% in the 0- to 5-year-old group). Number of readmissions, LOS, and hospital charges were also significantly higher in teenagers.
Features of the intoxications and the relationship to years of study are presented in Table 3. The number of admissions significantly decreased during the study period (786 in 1987 to 592 in 1997). The mortality rate decreased from 0.3% to 0.1% during the study period; however, this decline did not achieve statistical significance. Median hospital LOS did not change; however, hospital charges significantly increased. The causes of intoxication did not change between 1990 and 1993 compared with 1994 to 1997 (data before 1990 are not precise enough to conclude the cause for this period). The most common agents involved in children hospitalized for intoxication remained fairly constant during the 11-year period. However, intoxication with analgesics, antidepressants, and psychotropic drugs was significantly more frequent in the last period compared with the first 2 periods. Acetaminophen ingestions became significantly more frequent over time (increased from 15% to 22%; P<.001), whereas salicylic acid use declined significantly (14% to 8%; P<.001).
Causes of the intoxications are presented in Table 4. E codes were not included in the statewide data set for 1987 and 1988. When these years were included, the cause of the intoxication was undetermined in 25% of cases. Patients without an identified cause were excluded from the table. Self-inflicted intoxication or suicide attempt was the most frequent cause of intoxication (48%) and was more common in teenagers (99%) and girls (77%). Self-inflicted intoxications were associated with significantly longer hospital LOS and higher charges compared with unintentional intoxications and assaults. Analgesics were the most frequent agents involved in all categories, but were significantly more common in self-inflicted intoxications compared with unintentional intoxications. Likewise, antidepressants were significantly more common in self-inflicted intoxications compared with unintentional intoxications.
There were 15 fatalities. Deaths were more frequent in girls (73%) and were more frequently caused by self-poisoning (47%). The agents involved in the deaths were fumes (n = 4), barbiturates (n = 2), tricyclic antidepressants (n = 2), unspecified agents (n = 2), and alcohol, plant, anticonvulsant, cardiac medication, and acetaminophen (n = 1 each). The highest mortality rate was for children aged 6 to 11 years (1.3%); the higher mortality in this group was due to cases of fatal fume intoxications. Multiple ingestions were not reported in any child who died. Hospital charges were significantly greater in children who died ($9078) compared with children who did not ($2093); however, LOS was similar in children who died (2 days) and those who survived (1 day).
A total of 244 patients had multiple admissions for intoxication during the study period (520 admissions). Twenty-four had more than 2 admissions (79 admissions). Six patients were hospitalized 4 to 5 times. Readmissions were significantly more frequent among teenagers (93%) and girls (73%). In patients with multiple admissions, self-inflicted intoxications were significantly more common (66%) compared with unintentional poisoning (11%). Antidepressant medications were the most common agents ingested by children with more than 2 hospital admissions. Hospital LOS and hospital charges also significantly increased with subsequent readmission; however, mortality did not increase.
We found that the number of children in Washington State admitted to a hospital for acute intoxication declined during the 11 years studied. The medications and toxic agents involved in pediatric intoxications that required hospitalization did not substantively change during the study period. Likewise, the percentage of children with intoxication leading to hospital admission due to self-inflicted intoxications did not change during the study period.
Our results differ from those of Rodriguez and Sattin,4
who found no significant difference between annual rates of hospitalizations from poisonings between 1979 and 1983 in children (0-9 years old) in the United States. We cannot fully explain these differences; however, they may be due to regional differences in health care, our study population, or the longer period of our study. Washington has a larger proportion of managed care (average of 16% during the study period) than many other states, which may lead to differences in access to primary and emergency care. Differences in referral patterns and treatment advice provided by poison centers may exist. Differences may in part be due to ethnic or socioeconomic differences in the populations. Washington has a relatively homogeneous ethnic population, with 91% white Americans in 1990. In the study by Rodriguez and Sattin,4
children of other ethnic backgrounds had a hospitalization rate for poisoning that was 2.4 times that of white children.
In our study, the mortality rate was low (0.2%) and declined slightly over time. This could be due to better recognition of intoxications or improved prehospital or hospital care; however, we cannot determine which factors accounted for the decline. It could also be because deaths occurring in the emergency department or at home were not included. The fatal ingestions in our study were similar to fatal intoxications reported in the Toxic Exposure Surveillance Systems of the American Association of Poison Control Centers (TESS) database.2
In our study, analgesics constituted the most common cause of acute intoxication leading to hospital admission in children. Our findings agree with prior reports.2,5- 9
Ferguson et al10 reported that the most common agents involved in intoxication leading to hospitalizations were analgesics and that the number of salicylic acid intoxications was declining, whereas the number of intoxications from acetaminophen had increased. We found the same decline in salicylic acid intoxication, probably because salicylic acids have largely been replaced by acetaminophen as the first-line antipyretic and analgesic therapy, to prevent Reye syndrome. We also found that analgesic ingestions were more common in older children compared with younger children. This is probably because these medications are present in almost every home and teenagers tend to act impulsively. Trinkoff and Baker11
pointed out that the availability of an agent was important in both unintentional intoxications of preschool children and intentional intoxications among adolescents.
Similar to other reports,4,5,12
we found that children aged 12 to 18 years and 0 to 5 years were the most likely to be admitted to the hospital with acute intoxication. As expected, toddlers and school-aged children were usually involved in unintentional intoxication, whereas teenagers generally had intentional intoxications. Most patients were adolescents, thus influencing the agents used and the cause of intoxications for the whole group.
Self-inflicted intoxication was the leading identified cause of intoxication-related hospitalizations in Washington for children younger than 19 years. Poisoning is a well-known method for suicide attempt in adolescent girls.9
From 1990 to 1997, the incidence of hospitalization for self-inflicted intoxications did not significantly change. Hospitalization secondary to self-inflicted intoxication did not increase in younger age groups either (<12 years old). The number of self-inflicted intoxications from 1986 to 1989 was lower than in the later periods. Many undetermined E codes were used in this period; therefore, we cannot estimate the actual number of self-inflicted intoxications or unintentional intoxications. Nevertheless, from 1986 to 1989, the ratio of self-inflicted intoxication or unintentional intoxication is approximately the same as in the other 2 periods (1.7:1 vs 1.5:1). No prior population-based study was available to compare these results. A hospital-based study by McEvedy13 in London, England, showed a 108% increase of suicide attempts by poisoning during a 4-year period and an increase in suicide attempts in younger age groups over time. Unfortunately, our results cannot be compared with those of this study because the populations are different. Furthermore, the cause of the intoxications in our study was identified by E codes. Assignment of the E codes relies on the medical record and was limited by missing data that we could not verify.
Ingestion of multiple agents was more frequent in teenagers and when the cause of intoxication was intentional. The use of multiple agents did not increase during the study. These findings agree with the data from the TESS report, where multiple ingestions account for 11% of cases.2
Multiple agent intoxications were more common in children who were admitted more than twice with intoxications but were not associated with increased mortality. Readmission was associated with use of potentially more toxic medications but was not associated with higher mortality.
By using a hospital-based data set, we evaluated intoxication leading to hospitalizations only. Thus, we report on a subset of intoxications, the ones necessitating prolonged medical supervision and treatment. We cannot determine how many intoxication-related deaths occurred outside the hospital or in the emergency department. A recent study14
on poisoning-related visits to the emergency department in the United States reported that for every 147 poison exposures, there are 59 emergency department visits, 13 hospitalizations, and 1 death.
Our study has several limitations. First, the discharge codes recorded in the discharge data were not verified. However, audits of the ICD-9 codes have reported that they are reliably reported in statewide hospital discharge data.15 Second, we could not confirm the E codes on which the cause of ingestion was reported. Finally, we did not have information about children from Washington who received inpatient care for acute intoxication in other states.
In conclusion, acute intoxications continue to be an important cause of childhood hospitalization in Washington and other countries.16- 19
Pharmaceutical agents are more frequently involved in teenager intoxications, whereas nonpharmaceutical agents are more prevalent in younger children. Importantly, the type of agent involved did not change significantly during the study period, and acetaminophen remains the most commonly ingested medication. Our data also show that although intoxications are still a major health problem, the hospitalization rate declined from 1987 to 1997 in Washington. This may be secondary to preventive measures, such as the development of effective poison centers and improvement in treatment and hospital care. The US Childhood Intoxication Prevention Packaging Act of 1972 increased the safety of medications and home products, and significant declines in emergency department visits for intoxication (5.7 per 1000 in 1973 to 3.4 per 1000 in 1978) and in the mortality rate (from 2.1 per 100 000 to 0.5 per 100 000) were observed.20
In our study, the mortality rate is low but did not change significantly over time. Prevention efforts should be continued to further decrease intoxication-related hospitalizations and deaths. Prevention should include parental education about transferring medications or household products from their original container and use of child-resistant containers.21 Use of a locked cabinet even when children are older and education of grandparents is also important. Because female teenagers continue as the highest risk group for a suicide attempt by ingestion of pharmacologic agents, prevention efforts should be targeted to this population. The incidence of self-inflicted intoxications did not decline in the last 8 years of our study. Self-inflicted intoxications were associated with the highest costs (hospital charges), LOS, and number of readmissions among children with poisonings that required hospitalization in Washington. Suicide prevention in teenagers is essential but difficult to realize. Implication of teachers, parents, and physicians should be emphasized.
Accepted for publication March 29, 2001.
Presented at the Third World Congress on Pediatric Intensive Care, June 28, 2000, Montreal, Quebec.
Acute intoxication (or poisoning) in children can lead to serious complications, hospitalization, and even death. Pediatric mortality from intoxications has declined considerably in the last 30 years because of multiple preventive measures. Nevertheless, acute intoxication remains an important cause of illness in children. Our study describes the population of children hospitalized for intoxication, from 1987 through 1997, in Washington State. The incidence of hospitalizations for intoxication was 45 per 100 000 children per year; the annual rate decreased during the 11-year period. Mortality was low (0.2%) and did not change significantly over time. The type of agent involved did not change significantly over time; acetaminophen remained the most common ingestion. Teenage girls continue as the highest risk group for suicide attempt from ingestions. Self-inflicted intoxications were associated with higher costs, length of stay, and number of readmissions. Although preventive measures and development of poison centers have contributed to decrease mortality from acute intoxication in children, efforts need to be targeted toward suicide prevention, especially among teenage girls.
Corresponding author and reprints: France Gauvin, MD, Hôpital Ste-Justine, Department of Pediatrics, 3175 Côte Ste-Catherine, Montreal, Quebec, Canada H3T 1C5.
France Gauvin, Benoît Bailey, Susan L. Bratton. Hospitalizations for Pediatric Intoxication in Washington State, 1987-1997. Arch Pediatr Adolesc Med. 2001;155(10):1105–1110. doi:10.1001/archpedi.155.10.1105