Abstract
Introduction/Aim: Despite bronchitis being the most common finding at flexible bronchoscopy (FB) in many paediatric centres such as ours, no validated objective system exists. From previously recorded FB, we: (1) determined the correlation among the different macroscopic findings with airway neutrophilia (2) examined the inter‐rater repeatability of these findings and, (3) developed an experimental model of an objective FB‐derived bronchitis score (BScoreexp).
Methods: We reviewed 100 consecutive previous recordings (2016) from our database. We excluded FBs if: BAL data was unavailable, incomplete FB recording or FBs were on children who were immune‐compromised or had endotracheal tube, tracheostomy or foreign body. FB recordings were assessed (by 2 scorers independently,blinded to the clinical history) for 6 components: amount of secretions (scores 1‐6 from previous validated score), colour of secretions (0‐8 using BronkoTest), mucosal oedema (0‐3), ridging (0‐3), erythema (0‐3) and pallor (0‐3), based on pre‐determined criteria on a pictorial chart. The various models of BScoreexp were plotted against neutrophil% using a receiver operating characteristic (ROC) curve. Here we report our preliminary findings; on the first 65 children with valid FBs.
Results: Only secretion amount (rs=0.272, p=0.03) and colour (rs=0.342, p=0.005) significantly correlated with BAL %neutrophil but other macroscopic findings correlated with each other. For the 26 FBs examined for repeatability, kappa values for secretions (K=0.96, 95%CI 0.91‐1.0) and colour (K=0.84, 95%CI 0.73‐0.95) were excellent. Other K ranged from 0.38 to 0.67. Using BAL neutrophilia of 15% to define inflammation, the highest aROC (0.63, 95%CI 0.50‐0.76) was obtained by the giving three times weightage to secretion amount and colour and adding it to the other 4 components except pallor.
Conclusion: A repeatable FB‐defined bronchitis scoring system can be derived. However, a prospective study needs to be performed with larger numbers to further evaluate the different models to obtain aROC of >0.7.
Methods: We reviewed 100 consecutive previous recordings (2016) from our database. We excluded FBs if: BAL data was unavailable, incomplete FB recording or FBs were on children who were immune‐compromised or had endotracheal tube, tracheostomy or foreign body. FB recordings were assessed (by 2 scorers independently,blinded to the clinical history) for 6 components: amount of secretions (scores 1‐6 from previous validated score), colour of secretions (0‐8 using BronkoTest), mucosal oedema (0‐3), ridging (0‐3), erythema (0‐3) and pallor (0‐3), based on pre‐determined criteria on a pictorial chart. The various models of BScoreexp were plotted against neutrophil% using a receiver operating characteristic (ROC) curve. Here we report our preliminary findings; on the first 65 children with valid FBs.
Results: Only secretion amount (rs=0.272, p=0.03) and colour (rs=0.342, p=0.005) significantly correlated with BAL %neutrophil but other macroscopic findings correlated with each other. For the 26 FBs examined for repeatability, kappa values for secretions (K=0.96, 95%CI 0.91‐1.0) and colour (K=0.84, 95%CI 0.73‐0.95) were excellent. Other K ranged from 0.38 to 0.67. Using BAL neutrophilia of 15% to define inflammation, the highest aROC (0.63, 95%CI 0.50‐0.76) was obtained by the giving three times weightage to secretion amount and colour and adding it to the other 4 components except pallor.
Conclusion: A repeatable FB‐defined bronchitis scoring system can be derived. However, a prospective study needs to be performed with larger numbers to further evaluate the different models to obtain aROC of >0.7.
Original language | English |
---|---|
Article number | TO 015 |
Pages (from-to) | 28-28 |
Number of pages | 1 |
Journal | Respirology |
Volume | 23 |
Issue number | S1 |
Early online date | 14 Mar 2018 |
DOIs | |
Publication status | Published - Mar 2018 |