Summary of Results

The goal of the European Survey of Foreign Bodies Injuries (ESFBI) was the collection of a large database of cases of foreign body injuries across Europe. To this purpose, the University Clinics of the capitals of the biggest extended European Union countries and in the major countries not belonging to the EU were enrolled in the project. Regarding the available data in Europe, most of the published work is based on single center case reports. In this reports, usually the objects are located in only one or two anatomic parts of the upper air and digestive ways (e.g. larynx, bronchus, trachea) [1-16]. Alternatively, the removal technique used for the extraction of the object [17-26] are described in more or less great detail. Beside a multicentric study [27] collecting 100 cases on 11 European hospitals, focused on the scintigraphy technique for the localization of foreign bodies, no wide systematic multicentric study exclusively devoted to the epidemiological aspect of foreign bodies accidents were conducted in Europe with the involvement of the othorinolaryngologists.

• Design of the study
   

  The Case Report Form

  The country selection
   

• Results

  Gender, Age and Location

  Foreign body typology

  Removal of the foreign body

  Complications

  Hospitalization

  The circumstances of the accident

  Arrival at hospital and discharge

  Risk of prolonged hospitalization

  The costs
   

• References

Design of the study

The coordinating center of the study was the Department of Statistics and Applied Mathematics of Torino (Italy) and the data quality control center the EUFOS board. The EUFOS board chose one Area Referent for each involved country. The contacted countries were: Portugal, Spain, France, United Kingdom, Germany, Switzerland, Austria, Italy, Belgium, the Netherlands, Denmark, Finland, Norway, Sweden, Croatia, Slovenia, Greece, Slovakia, Czech Republic, Romania, Bulgaria, Hungary, Turkey, Poland and Russia. In particular Russia met problems in participating, due to for the different type of classification they use. Collected data referred to accidents caused by the inhalation/ingestion/aspiration/insertion of foreign bodies in children aged 0-14. The recorded cases were the injuries referring to the International Classification of Disease ICD-9CM [28]with the following codes:

i. ICD931 foreign body in the ears;

ii. ICD932 foreign body in the nose;

iii. ICD933 foreign body in the pharynx and larynx;

iv. ICD934 foreign body in the trachea, bronchi and lungs;

v. ICD935 foreign body in the mouth, esophagus and stomach.

The Area Referents were asked to contact all the University Clinics in the capital (or in another big city in the country if the contact was not possible in the capital) and to collect all the data referring to the above-mentioned ICD codes for the period from January, the 1st, 2000 to December, the 31st, 2002, extended to the first semester in 2003 where data were available. Each Area Referent was asked to send a Center Information Form about the characteristics of each enrolled University Clinic. The requested information (data refers to the year 2002) about the hospitals was: the number of physicians working in the hospital; the number of beds available; the number of hospitalizations; the number of otorhinolaryngology (ORL) department; the number of people referring to the hospital; the number of people admitted to the first aid (FA) department; the number of otorhinolaryngologists working in the ORL department; the number of beds available in the ORL department; the number of hospitalizations in the ORL department; the number of physicians working in FA department; the number of beds available in the FA department; the number of hospitalizations in the FA department. To each participating clinic an Identification code (ID) was assigned and each reported case was progressive numbered by the compilers. So each case was univocally identified by the hospital ID and the patient ID, remaining however completely anonymous. The coordinating center (CC) received the data, made the quality case control in collaboration with the EUFOS board, the data entry and the statistical analysis.

... Top

The Case Report Form

The Case Report Form (CRF) was divided in four parts:

• the personal information of the child that were exclusively stored by the area referent;

• the information on the patient (birth date, gender, accident date, localization of the foreign body (ICD code), the presence of complications, the FB removal technique, the necessity of hospitalization, its lasting and the regime of hospitalization);

• the foreign body typology (the FB description, the shape, the size and the consistency, the association with other objects at the time of the injury and the details of the other object, the association with other objects at the moment of purchasing;

• the typology of assistance in the hospital (the means of transport to the hospital, the department first looked at the child and the department that discharged the child).

... Top

The feed back of the contacted countries

Out of the 25 contacted countries, 19 (76.0%) sent the data to the Coordinating Center. The countries that contributed to the survey are those in the table below. Most of the contacted countries which not gave a positive feed back, had old paper-based hospital discharge records. Thus, it was very difficult for them to review all the cases. Moreover, four countries (Bulgaria, Germany, Spain, and Austria supplied only the cases referring to the ICD934 (FB in the bronchi, trachea and lungs)) having in addition specific clinical databases only regarding data coming from endoscopies. The percentage of removal by endoscopy is thus resulting as an overestimate in terms of severity of cases. The need of anesthesia for this kind of technique explains also the high percentage of hospitalization observed in the ESFBI database.

United Kingdom and Denmark supplied the data about all the ICDs but not about ICD 934, because these cases were treated in other departments or hospitals. Turkey had cases only of ICD 934 and ICD 935 (FB in the mouth, esophagus and stomach) and Poland was lacking of the cases of objects extracted from the ears and nose. Belgium sent 5 cases about ICD 932 (FB in the nose). The heterogeneity of the distribution of collected ICD in the ESFBI Study indicates that European hospitals have different ways of organizing, managing and treating the emergency and that often this difference is evident also within the same country [29-41].

... Top

Gender, Age and Location

Before the statistical analysis the coordinating center carried out the data entry and the cleaning phase of the reported records. The cases not matching with the standards defined in the project (e.g. children older than 14 years or the cases referring to years different from 2000-2003) and having important errors (e.g. the children had the accident before his/her birth date) were excluded from the database. In this phase several checks were made, all aimed to verifying the consistency of the records. After the cleaning phase, the ESFBI dataset consisted in 2103 cases in these years. The distribution by gender of children involved in a foreign body injury was similar to the other published papers [1, 18, 19, 42-47]. The percentage of males in the ESFBI study was 56.3%, in accordance with the meta-analysis of Gregori [48], 59.7%, the work published by Reilly et al. [49], 53% (both in the Modern Study and in the Jackson Collection). This result confirmed that males had a higher risk to have an injury due to inhalation/aspiration of an object as compared to the females. The median age was 3 (2; 5, 1st and 3rd quartile), with 58.8% of the children below 3 years of age and 22.5% below the first year of life. The number of children younger than 3 years was smaller as compared to previous meta-analysis (76.5%). Also the mean age in the ESFBI study was higher than the mean age in the Modern Study (MS) (3.9 vs. 2) but it was quite similar to those of the Jackson Collection (JC) (3.9 vs. 3.2). As already stated, the major part of the published studies referred to only one location (esophagus, larynx, trachea and bronchi). The location of the objects in the ESFBI study was in 23.7% of the cases the ears (ICD931), 32.8% the nose (ICD932), 8.1% the pharynx and larynx (ICD933), 26.3% the trachea, bronchi and lungs (ICD934) and 8.9% the mouth, esophagus and stomach (ICD935).

... Top

Foreign body type

In the CRFs the field reserved to the type of foreign body and to its description was open, so each object needed to be re-classified for the purposes of the analysis. The main categories were the Inorganic and Organic objects, the first including balls, marbles and beads, toys, batteries, coins, buttons, pebbles, stones and grits, pins, screws, needles and nails, stationery, plastic, pearls, jewelers, foils and cellophane, FPCI and other inorganic; the second one comprised nuts, seeds, berries, peas, corns and beans, food, bones and fish bones, paper and other organics. The category called “balls, marbles and beans” included also wooden and metallic balls. The class “toys” comprise all the toys (bricks, toy cars, dolls, etc.) except those so called “FPCI” (Food Products Containing Inedible). The class “pebbles, stones and grit” comprised also sand and soil. The class “pins, screws, needles and nails” was formed by all the sharp objects like drawing pins, safety pins, and the class “stationery” collected pen caps, leads, erasers, part of pencils and pens, etc. The category “pearls” was formed by jeweler pearls and plastic pearls, and “jewelers” included earrings, chains, etc. The class “foils and cellophane” included also those rolled up to form a small ball. The FPCI class comprised toys, stickers and magnets sold packaged with food like chocolate, crisps, etc. The “other organics” are foreign bodies not falling in the other categories like caps, wax, glass, etc. The class “nuts, seeds, berries, peas, corns and beans” contained all the small round food and parts tied to the food that easily goes down to the wrong way, like dried fruits, peas, pop corns, fruit seeds and stones, etc. The category “food” was formed by all kinds of food products that were not in the former class: fruit, vegetables, meat, bread, etc. The “bones” usually were bone from chicken, and fish bones. “Paper” comprised also stickers, where the “other organics” class included insects, grass, etc. Moreover, all the foreign bodies that were not found and the missing fields were classified as ‘unknown’. The most frequent class of object was ‘nuts and seeds’ with 26.6% of the cases. Most of the previous studies showed that nuts and seeds were the most common foreign bodies causing injuries in children [1, 6, 17, 44, 50-69]. Balls, marbles and beads were representing the second most dangerous class with 13.8% of the cases. The other categories were bones and fish bones (7.3%), food (6.3%), toys (5.8%), pebbles and stones (5.4%) and coins (4.8) (an example of coins in our dataset were represented in Fig. I.1.5.1). 1.1% of the injuries due to the ingestion of batteries was not negligible as it could cause serious health problems, even death if located in the esophagus [70-81]. In the present study the complication due to ingestion of batteries were perforation, bleeding, fistula and secretion. Comparing the ESFBI study with the Reilly’s MS we found different results: the mean age of children in MS was 24 months and in the ESFBI is 46 months (C.I. 45-48). The most commons foreign bodies in the MS were coins (47%, C.I. 45.7%-48.3%), as compared to the ESFBI where they were about the 4.8% (C.I. 4.1%-5.5%). Food in MS was the second most frequent object, occurring in 28% (C.I 26.8%-29.2%) of the cases. In the ESFBI the percentage of food was 32.9% (C.I. 30.9%-34.9%). Moreover, in the MS database toys were 3% (C.I. 2.6%-3.4%) and in the ESFBI this percentage doubled at 5.8% (C.I. 4.8%-6.8%). If we considered the Jackson Collection, the mean age of children was 38 months and the percentage of coins, food and toys were respectively 10% (C.I. 8.4%-11.6%), 21% (C.I. 18.8%-23.2%) and 5% (C.I. 3.8%-6.2%). Looking over the Gregori’s meta-analysis we found that the percentage of children younger than 3 years was higher (76.5% C.I. 73.9-79.2) as compared to the ESFBI frequency (58.8%, C.I. 56.7%-60.9%). The percentage of toys was 4% (C.I. 2.1%-6.2%) being not significantly different from the percentage of toys in the ESFBI. Concluding, injuries involving organic objects were more frequent in the meta-analysis (73.2%, C.I. 64.5%-81.9%) than in the ESFBI (48.4%, C.I. 46.3%-50.5%). Considering the typology of FB by gender, quite surprising the percentage of nuts and seeds was higher in males as compared to females (29.3% vs. 23.4%, p-value=0.001) and the males’ risk of having an accident with organic objects was the 50.8% and females’ risk 45.3% (p-value 0.006). There are some differences in the distribution of objects by age classes: children younger than 3 years had a frequency of 37.9% of injuries caused by nuts, seeds, etc. as compared to older ones that had a frequency of 10.6% (p-value 0.000), on the contrary for older children was more probably to have an injury with balls (17.7% vs. 11.2%, p-value 0.000), pearls (3.3% vs. 1.6%, p-value 0.007), batteries (1.7% vs. 0.6%, p-value 0.009) and fish bones (13.5% vs. 3.0% p-value 0.000). Moreover, younger children had a smaller risk to have accidents with inorganics as compared to older children (OR 0.54, C.I. 0.45-0.65). The most commons foreign bodies in the ears were balls and pebbles with respectively 26.7% and 9.8%. The objects found with the highest frequency in the nose were balls (22.2%), nuts and seeds (14.8%) and toys (10.2%), and in the pharynx and larynx were fish bones and bones with the 80.0%. The foreign bodies causing more accidents in the trachea, bronchi and lungs were nuts and seeds (74.5%) and food (8.5%). At the end, coins (51.1%) and food (11.8%) were the most common objects found in mouth, esophagus and stomach. The most common shapes were spherical (46.1%) and three-dimensional (24.2%). For three-dimensional objects the volumes were estimated: the median volume was 113.04 mm3 (33.49, 268.0) and the foreign bodies with the biggest volume were FPCI with a median of 962.5 mm3 (800.0, 962.5) and on the contrary pins and screws, batteries, and nuts and seeds have the smallest volumes, respectively 27.0 mm3 (9.1, 48.0), 45.0 mm3 (30.0, 268.0) and 65.4 mm3 (36.0, 179.5). Ninety nine percent of the first dimension measured on the object was less or equal than 3.175 cm (the diameter of the cylinder test), and considering the same dimension in accidents of children under 3 years the percentage was 98.8 and in older children it was 99.7%. These results were quite similar to those of MS, where 99% of FBs were less than 1.25 in (about 3.125 cm) diameter. The location were the foreign body had the greatest median volume was the mouth, esophagus and stomach (1225.0 mm3 (574.2, 2634.3)) and the smallest were the ears with the median volume of 65.4 mm3 (30.9, 150.0). The consistency of the foreign bodies was rigid in 50.6% and semi-rigid in 27.7% and only in 12.7% it was conforming. Furthermore, comparing the consistency in the different objects locations, we noted that the rigid bodies were more frequent in the ears (30.3%), and in the nose (35.5%).

... Top

Removal of the foreign body

Several removal techniques could be applied for the foreign body extraction depending on the location, the type of the object and the age of the children. The endoscope technique was the most used (71.7%) in the ESFBI data set, followed by other removal methods (19.6%) usually carried out in ambulatory or with other optical techniques (e.g. microotoscopy). Only in 8.8% of the cases a surgical intervention was performed. Sixty-two percent of the surgeries were carried out on patients with the foreign bodies in the ears, 29.1% in the nose and 32.6% in the bronchi, trachea and esophagus. Females had a greater number of surgeries as compared to males (10.5 vs. 7.3, p-value: 0.009) and a smaller number of endoscopies (69.7 vs. 73.9, p-value: 0.025). Furthermore, the techniques were used in a significantly different way for patients younger than three years with respect to the older children: the percentage of endoscopies was respectively 81.4% and 58.6% (p-value: 0.000), that of surgery was 6.4% and 11.8% (p-value: 0.000) and that of other techniques 12.1% and 29.6%. Also in this case the literature does not present a wide range of papers comparing different kinds of extraction techniques.

... Top

Complications

Often injured children had serious consequences. Twelve percent of children had complications, equally distributed between males and females. The median age of children suffering complications was 4 (2, 6 1st and 3rd quartile), higher as compared to those that had solved the injury without any consequence (3; 2, 5 1st and 3rd quartile). The locations with the highest percentage of complications were ears (30.8%), and trachea, bronchi and lungs (13.7%). Most common consequences of foreign bodies in the trachea, bronchi and lungs were bronchitis (7 cases), bronchopneumonia (2), bronchospasm (7), pneumonia (12) and pneumothorax (4), these results being similar to those published in literature [10, 17, 46, 67, 87]. The most commons complications in the ears were otitis (7) and bleeding (6); in the nose were bleeding (8) and rinithis (5). Other important consequences of foreign body injuries were dyspnea (6), infection (7), inflamation (11), perforation (9) and dyphagia (5). Only one, a Finland male patient aged 3 died for the aspiration of a peace of apple in the trachea, bronchi or lungs. He suffered hypoacusis when he arrived at the hospital.

... Top

Hospitalization

Considering the distribution of hospitalization by gender, the males had a greater risk to be hospitalized than females (OR=1.31, C.I. 1.08-1.58). Moreover the risk for a younger children to be hospitalized was higher as compared to patients older than 3 years (OR =2.3, C.I. 1.9-2.9). The mean length of stay was 2.1 days, similar to the data presented in published literature [67, 88]. Nineteen percent of FB cases found in the trachea, bronchi and lungs, and 13.9% found in the pharynx and larynx had a length of stay greater than 3 days. In 87.6% of the hospitalized children the length of stay was less than 3 days.

... Top

The circumstances of the accidents

Several studies were carried out about the parental supervision on children [89-101] In the ESFBI study 48.9% of accidents happened under the supervision of a parent or an adult. Quite surprising males had the majority of injuries in presence of an adult (61%) as compared to females (55.6%). Moreover, children having an accident under the adult supervision had a median age of 2 (1, 5 1st and 3rd quartile), smaller as compared to that of children having an accident in absence of an adult (3; 2, 6 1st and 3rd quartile). Eighty-three percent of injuries in larynx and pharynx and the 68.0% of injuries in bronchi, trachea and lungs occurred under the adult supervision. The most common accidents taking place not in the presence of an adult were the FB in the ears (31.4%) and the FB in the nose (39.7%). The most common injury that happened under the parental supervision was the FB in the bronchi, trachea and lungs (34.9%). Another important aspect was the circumstance of occurrence of accident: in 34.3% of the cases the child was eating and in 58.6% the child was playing. Also in this case the median age was different for the two groups: the children that had an injury eating had a median age of 2 (1, 4 1st and 3rd quartile), smallest than those of children that were playing (3; 2, 5 1st and 3rd quartile). Females had a biggest number of accidents while playing (62.2%) than males (55.6%, p-value 0.002). Concluding, the most common FB location when the child was eating was the bronchi, trachea and lungs (62.8%), and the most common when the child was playing was the nose (45.3%).

... Top

Arrival at the hospital and discharge

The transports used to reach the hospital, the departments that first looked at the child and the department that discharged the child was another important question for understanding the severity of the injury and the organization of the hospital. The majority of children went to the hospital by private transport (62.6%) and 22.6% by public transport. The percentage of children which arrived at the hospital by ambulance was 13.6% and their median age was of 2 years (1, 3 1st and 3rd quartile) lower than those of children arrived by the other means of transport (3; 2, 6 1st and 3rd quartile). The departments that first looked after the children were in 49.9% otorhinolaryngology, in 28.2% pediatrics and in 19.4% the emergency department. The median age of the children in the pediatric department was 2 (1, 4 1st and 3rd quartile) smaller than in others. The otorhinolaryngology department looked mainly at the foreign bodies in the ears and nose (respectively 30.7% and 40.8%) like the emergency department (33.4% and 44.9%) while the pediatrics Departments those in bronchi, trachea and lungs (61.3%). The departments that discharged the children were otorhinolaryngology in 72.8% of the cases, pediatrics in 4.1%, the emergency department in 9.6% and other departments in the 13.5%.

... Top

Risk of prolonged hospitalization

Foreign body injuries caused most often hospitalization with high costs for the health system. Also delayed or overlooked diagnosis [1] causing the onset of complications usually reflects in a prolonged hospitalization. Published studies indicate an average length of stay (LOS) ranging from 1.97 [67] and 2.5 day [88] up to 3.7 days per admission. In the analysis, LOS was also categorized according to 3 days limit for prolonged hospitalization. Pitfalls in the management causing prolonged hospitalizations occurred because of incomplete radiographic evaluation, confusion due to prior or concurrent illness, improper patient transfer, unusual presentation, and inappropriate methods of FB removal [102]. Type of FB causing the injury are also of specific relevance in this context, nuts in particular being usually associated with more severe clinical presentation and difficulties in technical removal [103]. Association of FB characteristics and LOS was assessed using a multivariable logistic regression model and all effects were adjusted by age, gender and ICD9-CM code. Huber’s adjustment to the covariance matrix was used to take into account the clustering effect attributable to the country in estimating the effects in the logistic model.

... Top

The costs

The hospitalization and the extraction of foreign bodies caused high costs for the public health in all European countries and many clinics tried to use cost-effective techniques for curbing the costs [104-109]. The costs for the hospitalized children were estimated considering just the costs for the extraction of the foreign body and not those due to the. Therefore, the computed costs are most probably an underestimation of the true overall costs. Regardless of the country where the accident occurred, the estimation was based on the Italian DRG (Diagnosis Related Groups) system [110-114]. The mean cost estimated was thus about 1,017.37 € (S.E. 28.33). Analyzing the costs according to the gender, it emerged that the public health spent more money for the females’ foreign body injuries than for males (1,097.54 € vs. 974.12 €, p-value: 0.034). Moreover, children younger than 3 years had a mean costs equal to 1,081.48 €, higher as compared to the amount of 882.73 € spent for the older children (p-value: 0.001). The distribution of costs for the different ICDs was heterogeneous and the mean amounted to 514.86 € (S.E. 50.23) for ICD931, 475.68 € (S.E. 52.99) for ICD932, 800.74 € (S.E. 85.94) for ICD933, 1,308.02 € (S.E. 36.55) for ICD934 and at last 798.07 € (S.E. 63.63) for ICD935. The mean cost of ICD934 was significantly different from the mean costs of ICD931, ICD932, ICD933 and ICD935 (p-values <0.001). We found out significant results also comparing the mean costs of ICD935 with ICD931 and ICD932 and the mean costs of ICD933 with ICD931 and ICD932 (respectively p-value: 0.003 and 0.006 for ICD935, and p-value: 0.004 and 0.001 for ICD933).