Childhood acute poisoning in the Italian North-West area: a six-year retrospective study

This is a retrospective study on ascertained acute poisonings on consecutive patients aged from 0 to 14 who presented at the Regina Margherita Hospital from January 1^st, 2012 to December 31^st, 2017. This is an inner-city teaching tertiary hospital in Turin, Piedmont, that serves local residents and out-of-state referrals with an estimated catchment area of about 500,000 potential users, being therefore one of the largest children’s hospitals in Italy.

This study was approved by the Ethics Review Board of Hospital (Comitato Etico Interaziendale A.O.U. Città della Salute e della Scienza di Torino, number 00156/2019) in compliance with the Italian regulations in force for data collection and in accord with the principles of declaration of Helsinki.

The unidentified information and related data on individual patients were obtained from the database of Hospital discharge cards reporting. Being the study retrospective, data extractors were not aware of the hypothesis investigated. All data are available in Figshare Public repository. (Accession number https://​doi.​org/​10.​6084/​m9.​figshare.​11382054.​v1). Patients were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Substances were classified in two categories: Pharmaceuticals and Nonpharmaceuticals. Pharmaceuticals consist of analgesics (including anti-inflammatories, opioid analgesics and antispasmodics), psychotropics (including barbiturates, benzodiazepines, neuroleptics, antidepressant), respiratory drugs, antimicrobials (antibacterials, antivirals and antifungals), cardiovascular drugs (including diuretics), antihistamines, hormones and gastrointestinal drugs, vitamins, supplements and herbal remedies. The Nonpharmaceutical agents include alcohol, carbon monoxide (CO)/fumes, household cleaning products, plants/mushrooms, cosmetics, recreational drugs, pet food and viper bite. Adverse drug reactions, food intoxications and chronic poisoning were excluded from the study.

Four age groups were defined according to psychophysical development: i) under 1 year of age; ii) 1 to 4 years old; iii) 5 to 9 years old; iiii) 10 to 14 years old.

At arrival at the CE-RMH, patients were tagged according to the seriousness of their clinical situation and identified with a colored code according to the Italian guidelines for the correct execution of triage in emergency rooms: i) white, no urgency, no alteration of vital functions, no critical symptomatology or at risk of aggravation; ii) green, minor emergency, at the time of evaluation vital functions are intact and vital parameters in the norm; iii) yellow, serious emergency, patients with impending threat of failure of the vital functions (consciousness, breath, circulation); iiii) red, life threatening.

Statistical analysis of the data was performed using the Statistical Package for the Social Sciences (SPSS), version 24.0 software (SPSS, Inc., Chicago, IL, USA). Values of quantitative variables were expressed as mean ± standard deviation or median and interquartile range (IQR), while values of categorical variables were presented as frequencies and percentages. The χ2 test was used to evaluate any potential association between categorical variables and the Mann-Whitney U-test or the Kruskal-Wallis test were used to assess differences of quantitative variables between two or more groups of children, as appropriate. When there were significant differences according to the Kruskal-Wallis test, pairwise multiple comparisons were carried out using the Dunn test. Multiple logistic regression models were developed to identify predictors of hospitalization (yes vs. no) and severity of clinical status (minor -white or green code-, vs. serious - yellow and red codes). Purposeful selection of statistically (p values ≤0.2 in the univariate analyses) and clinically relevant was conducted. Data were presented as odds ratio (OR) and 95% confidence intervals (CI). Two-tailed p values less than 0.05 were considered statistically significant.

From January 1^st, 2012 to December 31^st, 2017, a total of 1030 (567 males and 463 females) consecutive acute intoxicated patients below the age of 14 years (mean 3.4 ± 3.1, median 2.2 and IQR 2.3) presented to the Children’s Emergency at the Regina Margherita Hospital (CE-RMH). These represent the 0.38% of the total accesses at the CE-RMH in the same period (n = 271,654 patients). Their distribution according to the countries of origin and age are detailed in Fig. 1a and b, respectively. Children were grouped in the above-mentioned age-categories as reported in Fig. 1c. Under the age of 9, males and females equally accessed the CE-RMH; on the contrary, over the age of 10, the majority of cases were girls (57.1%, p = 0.05 males vs. females).

An average of 85.8 ± 9.95 patients per month accessed the CE-RMH due to acute intoxication. Accesses were less frequent between 1 a.m. and 9 a.m. (only 5.6% of the total) and very frequent between 7 p.m. and 11 p.m. (37.7% of the total).

Children were tagged according to the severity of their clinical situations and identified with a colored code. Their distribution by age, type of agent and output is detailed in Table 1.

Children requiring a medical intervention (n = 982, 95.3%) had a median age of 2.2 years (IQR 2.4) and patients requiring surgery (n = 48, 4.7%) a median age of 2.1 years (IQR 1.9).

The number of exposures to the subclasses of Non- and Pharmaceutical agents were stratified according to the age groups or sex and detailed in Table 2. Six hundred and eight patients (59%) were exposed to Nonpharmaceutical compounds, 63% of which were males. Among household cleansers, 96 cases were due to exposure to bleach or bleach-containing products, 55 to surface cleaners, 37 to dish washing liquids and 18 to drain cleaners. The more frequent intoxicating product among cosmetics was nail polish remover (16 out of 54 cases). Tobacco was responsible for 25 out of 30 exposure to recreational drugs.

Fifty percent of alcohol and 37% of plant and mushroom intoxications occurred in summer, 63.3% of drug and tobacco intoxication in spring / summer.

Among analgesics, 48 children were exposed to acetaminophen, 21 to a FANS and 2 to opioid analgesics. Benzodiazepines were the most frequent agent among psychotropics (24 out of 77), 17 children were intoxicated by antiepileptics and 15 by antidepressants. Levothyroxine was responsible for 13 out of 26 exposures to hormonal drugs. Trend by day of the week, hours or season were not evidenced for Pharmaceutical intoxications.

In any age group, intoxications with Nonpharmaceuticals were more frequent than with Pharmaceuticals. In general, over the age of 10, intoxication in girls was more frequent than in males although at the limit of significance.

The number of poisoning substances and the percentage of hospitalizations are detailed in Table 3.

When comparing children exposed to one single substance to the rest of the population considered, it appears that they were younger (median age 2.1 years, IQR 2.3) than children poisoned by two or more agents (median age 3.5 years, IQR 3.9, p = 0.005). Twenty-five patients (2.4%, 12 males and 13 females) had already experienced previous episodes of intoxication; their median age (3.9 years, IQR 11.6) was significantly higher than that of children intoxicated for the first time (2.2 years, IQR 2.3, p < 0.01) and ten of them were at least 10 years old with the higher probability to access as a red code (p < 0.001).

The mode of exposure to Nonpharmaceuticals and Pharmaceuticals is reported in Table 4, while the different classes are detailed in Fig. 2.

Detergents were the main cause of intoxication (Fig. 2a). Among the Pharmaceuticals (Fig. 2b), analgesics and psychotropics were the main cause of both accidental poisonings and therapeutic errors. Nevertheless, there was no significant correlation between any class of Pharmaceutical and the mode of exposure.

Twelve out of 16 suicide attempts were performed with Pharmaceutical agents by children older than 10, and 75% of them used psychotropics drugs.

Therapeutic errors were more frequent in children younger than 1 year (26/83 children younger than 1 year, 31.3%), while the 1–4 years old were mostly intoxicated accidentally (687/751 children aged 1–4, 91.5%). Notably, the only cases of intentional intoxication were found within the > 10 years old group. Children who attempted suicide were older (median 13.2 years, IQR 0.63) than those who had accidental exposure (median 2.1 years, IQR 1.9, p < 0.001), therapeutic errors (median 2.3 years, IQR 3.9, p < 0.001) and recreational purposes (median 12.6 years, IQR 10.7, p < 0.001); moreover, children who had recreational purposes were older than those who underwent accidental poisonings (p < 0.001) and therapeutic errors (p < 0.001). The ingestion resulted the most frequent route in all age groups (p < 0.001).

Although data were available only for 330 out of the 882 accidentally exposed patients, in 206 intoxications (62.4%) the product packaging was original, while in 51 situations (15.4%) it was not. Moreover, 41 children (12.4%) subtracted the agent from the appropriate place, while 32 (9.7%) unintentionally assumed the toxic agent considering it as a game. The main place of exposure resulted own home (n = 190, 18.4%), for 29 children (2.8%) it happened in public places, for 10 (1%) in a youth center (mainly with detergents and cosmetics) and for 5 (0.5%) at school (mostly with plants and ink). The route of exposure was oral in the majority of cases (n = 915, 88.8%), whereas inhalation and cutaneous contacts occurred in 49 (4.8%) and 20 (1.9%) children, respectively.

Interventions of caretakers, if any, and dispositions are detailed in Table 5.

In 33 out of the 608 cases of exposure to Nonpharmaceuticals, the intervention of caretakers is unknown, while the substance category in which the action was more frequent was detergents (101/161, 62.7%); no action was taken in the cases of carbon monoxide poisoning or in viper bite.

Fifty-one children, 17 of which poisoned by plants or mushrooms and 14 poisoned by recreational drugs, required activated charcoal therapy. Sixteen patients underwent gastric lavage, among which 4 exposed to detergents and 5 to recreational drugs. In one case of intoxication with a cleanser (sodium hypochlorite) it was necessary to administer an antidote (sodium thiosulfate).

The parents’ action was about two and a half times more frequent in the case of Nonpharmaceutical than in Pharmaceutical poisoning (p < 0.001). Among the situations requesting CAV advice, 69 events were due to analgesics, 58 to psychotropics and 44 to cardiovascular drugs. One hundred twenty-six children required activated carbon therapy, most frequently in case of contact with analgesics (n = 27) and cardiovascular drugs (n = 25). Moreover, 35 children, of whom 12 exposed to painkillers, underwent gastric lavage. In one intoxication by a benzodiazepine, an antidote (Flumazenil) was administered. In general, activated carbon therapy was used much more often in case of exposure to Pharmaceuticals than to Nonpharmaceuticals (29.8% vs 8.34%, p = 0.001). The same for gastric lavage (3.5 times more).

Among exposure to Nonpharmaceuticals, detergents and household products represented the main cause of hospitalization (31 cases), while the Pharmaceuticals causing the highest number of hospitalizations were analgesics and psychotropics (13 and 10, respectively).

In general, discharge was more frequent than hospitalization (90.4% vs 9.6%) even if in patients over 10 years the percentage of hospitalization was significantly higher (41.2%) than in any other age group (p < 0.001). Among the 96 hospitalized patients, 78 intoxications occurred through the oral route, 11 by inhalation, 4 by dermal contact and 3 through parenteral/transdermal route (p < 0.05, oral vs. any other route). The largest number of hospitalizations occurred in the category of accidental exposure (57.6%) and 60% were due to Nonpharmaceuticals. In none of the 1030 cases of our series there was death.

The results of multivariable regression models for severity of clinical status, hospitalization and time spent at the ER are presented in Table 6. The odds of having a severe clinical status (yellow or red code) was higher in children requiring an advice from the CAV (OR 4.89, p < 0.001), and in those exposed to CO/fumes compared to Pharmaceutical poisoned children (OR 10.18, p = 0.029), while it was lower in those exposed to cosmetics (OR 0.40, p = 0.05). The hospitalization was three times more likely to occur in children from extra-European countries (p < 0.001) and in those intoxicated due to therapeutic error (OR 2.74, p = 0.01) compared to accidentally exposed children. The odds increased to 3.22 (p = 0.009) for cases of intentional poisoning.