A new clinical algorithm scoring for management of suspected foreign body aspiration in children

Based on the inclusion criteria, a total of 300 children were included in this study. Of the 300 patients, 163 were boys (54.3%) and 137 were girls (45.7%), giving a male-to-female ratio of 1.2:1. A total of 139 children (46.3%) were Qatari nationals. The mean age was 2.1 ± 1.7 years (range 7 months to 14 years) with a peak incidence at 1 to 2 years of age. Seventy patients (23.3%) were less than 12 months of age.

These 300 patients cumulatively underwent a total of 410 bronchoscopies, which included both flexible and rigid bronchoscopies (88 patients underwent 2 bronchoscopies, 17 had 3, 4 patients had 4 and one patient underwent 5 bronchoscopies). Only 190 (66.7%) had a definite history of FBA. FB was found in the airway in 91 of the 300 children (30.3%). In all of these children, the FB was successfully extracted. The demographic characteristics, clinical, laboratory and bronchoscopy findings of the study population are shown in Table 1.

Positive history of noisy breathing/stridor/dysphonia and recurrent/persistent wheezing were notably more common in children with FBA compared to children without FBA. There was no significant difference observed in the occurrence of other clinical signs and symptoms between the two groups (P > 0.05) as shown in Table 2. In children with bronchoscopically proven FBA, the two most common presenting symptoms were; cough in 79.1% and witnessed choking in 71.4%. However, both witnessed choking and cough had a low specificity i.e. 36 and 31.1% respectively (Tables 3 and 4). Therefore, it can be concluded that the presence of either of these cannot be relied upon for an accurate diagnosis of FBA.

Decreased air entry (58.9% vs 27%; P = 0.0001) and wheezing (32.2% vs 18.5%; P = 0.011) were significantly more common in children with FBA compared to children without FBA. Children without FBA were also more likely to have a completely normal physical exam when compared to children with FBA (56% vs 23.3%; P = 0.0001). No statistically significant differences were observed in the frequencies of other related physical exam findings between children with FBA and those without FBA (P > 0.05); refer to Table 2.

Of the 91 children with bronchoscopically proven FBA, 13 (17.1%) had normal chest radiographs, 31 (40.8%) and 10 (13.2%) had unilateral and bilateral hyperinflation, respectively. Ten patients had either right/left or bilateral lung infiltrates, six patients (7.9%) had collapse-consolidation, and 4 (5.3%) had radio-opaque FB and 2 had other findings. In 15 cases the CXR was either not done or not reported. Of note, children in the FBA group were significantly less likely to have a normal CXR compared to children without FBA (P < 0.0001). There were no significant differences observed in the occurrence of other radiological findings between the two groups (P > 0.05) as shown in Table 2, apart from the finding of a radio-opaque FB, which had a specificity and positive predictive value of 100% (Tables 3 and 4).

As stated previously, 91 (30.3%) of the 300 patients had a bronchoscopically confirmed diagnosis of FBA. In all of these cases, the FB was successfully removed; in 67 cases (74.4%) with rigid bronchoscopy and in 23 patients (25.6%) using a flexible bronchoscope. One patient required a thoracotomy. The flexible bronchoscopy was conducted by an experienced pediatric pulmonologist in most cases via the nasal route under moderate sedation in an intensive care setting. In 6 patients, it was performed in the operating theatre through an endotracheal tube under general anesthesia with a surgeon on standby. On the other hand, rigid bronchoscopies were performed by pediatric surgeons in the operating theatre under general anesthesia. It is of note that all but 6 cases that had their FB extracted via rigid bronchoscopy underwent a diagnostic flexible bronchoscopy by a pediatric pulmonologist before the rigid bronchoscopy was undertaken for FB extraction. Post bronchoscopy complications were reported in only 2.6% cases, with no procedure related mortalities (Table 1).

Of the 300 cases, 82 patients underwent multiple bronchoscopies. Twelve of these were patients who did not have a FB visualized during their initial procedure, but had a subsequent bronchoscopy performed a day or two later to exclude any missed FB as they remained symptomatic. The remaining 70 patients had a FB, but required multiple bronchoscopies due to various reasons. 45 patients had a flexible bronchoscopy performed for diagnostic purposes, and once the FB was located, a rigid bronchoscopy was done for FB extraction. In 8 patients, there was an unsuccessful attempt to extract the FB via flexible bronchoscopy, necessitating a second rigid bronchoscopy. In another 8 patients, the initial bronchoscopy (either rigid or flexible) extracted the FB incompletely; needing a second bronchoscopy for extracting the remnants. Nine patients were complicated cases who were admitted to the intensive care unit from the outset, they underwent multiple bronchoscopies including follow up bronchoscopies performed to rule out retained FB. One of these 9 patients eventually required a thoracotomy.

The most common site of FB lodging was the right main stem bronchus (47.3%), followed by the left bronchus (32, 35.2%), bilateral bronchi (5, 5.5%), carina (5, 5.5%) and the trachea (6, 6.6%). The rest were recovered from segmental smaller airways. Organic FB accounted for 62.6% of the FB removed, with peanuts (n = 15) being the most common type, followed by seeds (n = 13) and beans (n = 4), pistachios, and other food materials. Eight (8.8%) of the FB were inorganic; most of them were either metal or plastic objects, out of which 4, as stated earlier, were radio-opaque. In the remaining 28.6% cases, the patients’ records did not specify the type of FB. Table 1 outlines the bronchoscopy findings.

Historical, clinical, physical exam and radiological findings associated with FBA are shown in Table 2. The derivation of the weighted risk score for FBA is shown in Table 3. The five predictors selected for the risk score were history of witnessed choking (adjusted OR 2.1, 95% CI 1.1–4.3; P = 0.041), noisy breathing/stridor/ dysphonia (adjusted OR = 2.7, 95% CI 1.2–6.2; P = 0.015), new onset wheezing/ recurrent/persistent wheeze (adjusted OR 4.6, 95% CI 1.8–11.8; P = 0.002), abnormal chest x-ray findings (adjusted OR 4.0, 95% CI 1.9–8.5; P < 0.001), and unilateral reduced air entry (adjusted OR 2.9, 95% CI 1.5–5.5; P = 0.001) were significant predictors associated with an increased risk for FBA (P < 0.05). The clinical feature of witnessed choking on history was used as the reference regression coefficient and assigned a weighted value of 1. The discriminative ability of the model was found to be good with an area under the ROC curve value of 0.76 (95% CI 0.70, 0.82). The sensitivity and specificity values at a cutoff score point of ≥4 were 49 and 85%, and at a cutoff score point of ≥5 were 22 and 98% (positive likelihood ratio 6.6), respectively (Fig. 1). The predicted cutoff score derived using ROC analysis was found to co-relate well with known clinically significant predictors of FBA (Fig. 2).

Based on the aforementioned results and statistically significant predictors associated with FBA derived from multivariate logistic regression, we designed a clinical algorithm with a scoring system. In our proposed algorithm, (Fig. 2), each of the significant FBA predictors namely witnessed choking, noisy breathing/stridor/dysphonia, new onset/recurrent/persistent wheeze, unilateral reduced air entry and abnormal CXR are assigned a weighted risk score of 1, 1, 2, 1 and 2 respectively. Patients with a score of 5 or more had nearly 72% bronchoscopy proven FBA. For a child with a score range of 4–5, our study demonstrated that 59% had bronchoscopy proven FBA. In patients with a score of 2–3, the bronchoscopy proven FBA was ~29.8%, while with a score of ≤1, the bronchoscopy proven FBA dropped to no more than 9.6%.

Patients were also grouped based on the cumulative presence of 4 of the above 5 mentioned FBA predictors (witnessed choking crisis, Noisy breathing/stridor/dysphonia, New onset wheezing/recurrent/persistent wheeze, and/or decreased air entry). Figure 3a shows the cumulative proportion of children with proven FBA according to these 4 risk factors. Only 7.7% of the children without any of these risk factors had a proven FBA. The likelihood increased significantly with an increasing number of risk factors. When all 4 risk factors were present, the likelihood of FBA was 100%. A total of 242 of these 300 patients were also analyzed and grouped based on whether they had normal or abnormal physical and/or radiological findings on presentation. The remainder of the patients did not have a documented exam and/or chest radiological findings on record review. The results were as follows: (Fig. 3b) in children with both abnormal physical and radiological findings, 47.2% had a proven FBA. If only one was abnormal (i.e. either physical exam or chest x-ray), the likelihood of FBA reduced to 32–33.3%. In children who presented with a completely normal physical exam and chest radiograph, only 7.4% had a FB removed by bronchoscopy (Fig. 3b).