Study rationale: Diagnosis of pulmonary tuberculosis (PTB) in children relies heavily on chest radiography as sputum samples are difficult to obtain and only yield positive results in 30-74% of children treated for PTB. However, radiological signs between lower respiratory tract infections (LRTI) and PTB overlap considerably and there is a wide inter-observer agreement in the detection of lymphadenopathy, considered the hallmark of PTB. Small pliable paediatric airways are easily compressed by enlarged lymph nodes. Unlike lymph nodes, however, the lucent airways contrast against the surrounding mediastinal structures on radiographs, thus airway compression may serve as a more objective criterion for diagnosing PTB. Many studies have reviewed the radiographic features of PTB in children but few included airway compression or used a control group and none have evaluated inter-observer agreement. Objective: To investigate frequency and inter-observer agreement of airway compression on chest radiographs in children with PTB compared to those with another LRTI. Methods: Chest radiographs of children admitted to Red Cross War Memorial Children’s Hospital with suspected PTB were read by two readers according to a standardised format and a 3rd when there was disagreement. Radiographs of children with definite PTB were compared to those with another LRTI. Frequency and location of airway compression were evaluated. Findings were correlated with human immunodeficiency virus (HIV) infection and age. Inter-observer agreement was assessed using kappa statistic. Results: Radiographs of 505 children (median age 25.9 months [IQR 14.3-62.2]) were reviewed; 97/505 (19%) children were HIV-infected. Airway compression occurred in 54/188 (28.7%) definite PTB cases versus 24/317 (7.6 %) of other LRTI cases (OR 4.9; 95%CI 2.9–8.3). The left main bronchus was most affected in 51/493 (10.3%). A higher frequency of airway compression occurred in infants at 22/101 (21.8%) compared to 56/404 (13.9%) in older children (OR 1.7; 95%CI 1.00–3.00). No association between airway compression and HIV infection was found. Inter-observer agreement ranged from none to fair (kappa of 0.0-0.4). Discussion: The overall frequency of airway compression in definite PTB is compatible with reports in the literature. Although airway compression used alone is not a specific sign, if seen on radiographs, there is a strong correlation with PTB compared to other LRTI with infants at higher risk due to their smaller airways. Contradictory to other studies, our study showed the left main bronchus to be affected twice more commonly than the bronchus intermedius in both age groups. This is thought to be due to different patient selection. Confirming reports in the literature, no significant association between airway compression and HIV status was found. A disappointing finding was the poor inter-observer agreement. Contributing aspects include the lack of standardised criteria in the definition of airway compression and suboptimal visualisation of the airways on standard chest radiographs due to patient, technical and post processing factors. Conclusion: There is a strong association between airway compression on chest radiographs and definite PTB, particularly in infants, irrespective of HIV status. However, its clinical use as an objective criterion in the diagnosis of PTB is limited by poor inter-observer agreement.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:uct/oai:localhost:11427/29690 |
| Date | 19 February 2019 |
| Creators | Richter-Joubert, Lisel |
| Publisher | University of Cape Town, Faculty of Health Sciences, Division of Radiology |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Master Thesis, Masters, MMed |
| Format | application/pdf |
Page generated in 0.0023 seconds