INTRODUCTION
Bronchoscopy has various diagnostic and therapeutic indications. Diagnostic indications include the evaluation of symptoms and signs such as haemoptysis, stridor and hoarseness, evaluation of pulmonary infections, interstitial lung disease and suspected pulmonary malignancies.(1,2) Therapeutic indications include relief of airway obstruction, treatment of extrinsic tumours, and closure of tracheoesophageal and bronchopleural fistulae.(1,2)
There is a paucity of data in South Africa assessing the indications and diagnostic yield of bronchoscopies, with one retrospective review from Kwa-Zulu Natal which showed that the commonest indications for bronchoscopy were investigation of a lung mass (35.8%), non-resolving lower respiratory tract infection (15%) and suspected tuberculosis (TB) (15%).(3) Due to the paucity of data in the South African context, as well as the increased burden of HIV-associated infections (e.g. tuberculosis and non-tuberculous mycobacterial disease), data from international cohorts may not be relevant to the local setting. Additional local data is therefore needed.
METHODOLOGY
Standardised bronchoscopy reports for a 8 year period from 2011–2018 archived within the Division of Pulmonology at Chris Hani Baragwanath Academic Hospital (CHBAH) were accessed. Data points included demographic details (age, gender), associated clinical conditions (HIV status and if positive, CD4 count), bronchoscopy details (indication, finding, specimens obtained, complications) and final diagnosis. Associated histological, cytological and microbiological results were accessed from the National Health Laboratory Service. All data were logged into a Microsoft Excel spreadsheet and statistical analysis was performed using Stata 15®. Ethical approval for the study was granted by the Human Research Ethics Committee (Medical) of the University of the Witwatersrand (M190758).
Categorical variables were described using frequencies and proportions. Normally distributed variables were described with mean and standard deviation and non-normal distribution was described using the median and interquartile ranges. Univariate analyses were used to determine associations between bronchoscopy diagnoses (malignancy, inflammation, infections and sarcoidosis) and age, gender and HIV status. Multivariate logistic regression models were adjusted for age, gender and HIV status. Statistical significance was determined as p<0.05.
RESULTS
A total of 830 patient records were accessed over the 8 year period.
Demographic Outcomes (Table 1)
| Characteristics | Mean(SD)/Median(IQR) | N (%) |
|---|---|---|
| Demographics | ||
| Age (years) | 56.2 (13.3) | |
| Sex | ||
| Male | 546 (65.8) | |
| Female | 282 (34) | |
| Unknown | 2 (0.2) | |
| Clinical Features | ||
| In-patient | 484 (58.3) | |
| Outpatient | 298 (35.9) | |
| Unknown | 48 (5.8) | |
| HIV Status | ||
| Negative | 430 (52.0) | |
| Positive | 183 (22.0) | |
| Unknown | 216 (26.0) | |
| CD4 Count (cells/mm3) | 233 (85-434) | |
| Biopsy | ||
| Endobronchial biopsy | 416 (50.1) | |
| No biopsy | 220 (26.5) | |
| Transbronchial biopsy | 113 (13.6) | |
| Both | 60 (7.2) | |
| Unknown | 21 (2.5) |
Male patients comprised 65.8% of the total. The mean age of patients was 56.2 (+/−13.3) years. A total of 613 patients (74%) had documented HIV serology results. Of these, 183 (30.0%) were seropositive, with a median CD4 cell count of 233 cells/mm3 (IQR: 85-434 cells/mm3). Flexible bronchoscopy was performed in 484 (58.4%) inpatients and 298 (35.9%) outpatients. In the remainder, the location was not recorded. About half of the study population (416 patients; 50.1%) had an endobronchial biopsy; 220 (26.5%) had bronchial washings only, 113 (13.7%) had transbronchial biopsies alone, and 60 (7.2%) had both transbronchial and endobronchial biopsies performed.
Clinical suspicion of an endobronchial lesion was the most common indication for bronchoscopy (52% of bronchoscopies) (Table 2). Bronchoscopy was performed for the evaluation of pulmonary infiltrates in 12%, cavitary lung lesions in 3% and haemoptysis in 1%.
| Indications for flexible bronchoscopy at CHBAH | Number (%) |
|---|---|
| Suspected endobronchial lesion | 433 (52.2) |
| Pulmonary infiltrates | 100 (12.1) |
| Hilar/Mediastinal mass/Adenopathy | 71 (8.6) |
| Ancillary investigation for pleural effusion | 50 (6.0) |
| Non resolving pneumonia | 49 (5.9) |
| Cavitating lung lesions | 29 (3.5) |
| Investigation of lung pathology in extra-pulmonary malignancy | 29 (3.5) |
| Haemoptysis | 11 (1.3) |
| Persistent unexplained cough | 11 (1.3) |
| Exclusion of tracheoesophageal fistula | 10 (1.2) |
| Previous non-diagnostic bronchoscopy | 7 ( 0.8) |
| Suspected occupational lung disease | 5 ( 0.6) |
| Missing data | 25 (3.0) |
The sub-categories for suspected endobronchial lesions were: mass lesions on radiological imaging (55%), atelectasis (29%), hemithorax white-out (6.5%), bronchial cut-off (6%) and localised bronchiectasis (3.7%). (Table 3)
| Indication | N % |
| Mass lesion on X-ray | 239 (55.1) |
| Atelectasis | 124 (28.6) |
| White out hemi-thorax | 28 (6.5) |
| Bronchial cut off on X-ray | 26 (6.0) |
| Localised bronchiectasis | 16 (3.7) |
Table 4 highlights the findings at bronchoscopy. In 84% of patients with suspected endobronchial lesions, this suspicion was confirmed on bronchoscopy.
Final Diagnoses (Table 5)
| Diagnosis | N | (%) |
|---|---|---|
| Lung malignancy | 320 | (38.5) |
| Primary lung malignancy | 240 | (28.9) |
| Possible or probable malignancy | 63 | (7.6) |
| Secondary lung malignancy | 17 | (2.0) |
| Endobronchial lesion excluded | 146 | (17.6) |
| Normal lung biopsy/no pathological changes | 92 | (11.1) |
| Non-diagnostic | 73 | (8.8) |
| Non- diagnostic - Inadequate biopsy | 65 | (7.8) |
| Non-diagnostic - biopsy not done | 8 | (1.1) |
| Chronic nonspecific inflammation | 64 | (7.7) |
| Infections | 49 | (5.9) |
| TB | 36 | (4.3) |
| Others | 9 | (1.1) |
| NTM | 2 | (0.2) |
| Aspergillosis | 1 | (0.1) |
| Salmonella | 1 | (0.1) |
| Missing data | 24 | (2.9) |
| Sarcoidosis | 15 | (1.8) |
| Transoesophageal fistulae excluded | 9 | (1.1) |
| Squamous dysplasia | 7 | (0.84) |
| ILD other than sarcoidosis | 6 | (0.7) |
| Endobronchial KS excluded | 5 | (0.6) |
| Foreign body inhalation | 4 | (0.48) |
| Transoesophageal fistulae observed | 3 | (0.36) |
| Combination | 2 | (0.24) |
| CA+TB | 2 | (0.24) |
| Hydatid disease | 1 | (0.12) |
| Organising pneumonia | 1 | (0.12) |
Lung malignancy was the most common final diagnosis in our study. Primary lung malignancy was diagnosed in 75% of such cases, with non-pulmonary malignancy comprising an additional 5%. In those with primary lung malignancy, squamous cell carcinoma was the commonest subtype (42.5%), followed by adenocarcinoma in 31.3%. Interestingly, 2 patients had combinations of two different histological subtypes (small and large cell carcinoma (n = 1) and squamous cell and neuroendocrine carcinoma (n = 1).
A variety of infections were diagnosed. Tuberculosis (TB) was the most common n=36; (4.3%). Cases of non-tuberculous mycobacteria (NTM), Aspergillus, Salmonella and hydatid lung disease together comprised 6 cases (0.5%). Other infectious organisms were identified in p patients.
Sarcoidosis was identified in 15 patients (1.8%).
Bronchoscopy proved non-diagnostic in 8.8% of procedures, including 7.8% where the biopsy was inadequate. In 8 patients (1%), a biopsy was not performed. In 24 patients (2.9%), the final sdiagnosis was not determined due to incomplete records
Complications associated with bronchoscopy
Flexible bronchoscopy is an inherently safe procedure when appropriate precautions are taken. Overall, 97.8 % of all procedures were completed without any complications. The most frequent complication was pneumothorax (0.7%), this despite the absence of fluoroscopic guidance which was not available during the study period. (Table 6)
Factors associated with specific bronchoscopic diagnosis
There was a 5% increased odds of a diagnosis of malignancy with each additional year of age. [OR 95% CI, 1.05 (1.04-1.06)]. Women were less likely to be diagnosed with malignancy compared to men, and HIV positive patients were less likely to have a diagnosis of malignancy compared to HIV negative patients. The association between age and malignancy [OR 95% CI, 1.05 (1.03-1.06)] and between age and gender [OR 95% CI, 0.67 (0.51-0.96)] remained significant in multivariable analyses (p<0.001 and 0.026, respectively).
There were no associations found between age and gender with chronic non-specific inflammation.There were no significant association between gender and HIV status with infections. There was no association between HIV status and a diagnosis of sarcoidosis.
DISCUSSION
The study analysed all bronchoscopies performed at a large urban public hospital over a 8 year period. The mean age of patients undergoing the procedure in our centre was 56.2 ±13.3 years. This finding is in line with a similar study in Brazil where the mean age of patients involved in the study was 57.7±21.9 years.(4) Males comprised a significant proportion of our study population (66%), which is similar to other studies.(3–5)
Twenty-nine percent of our study population, who had their HIV status documented were HIV seropositive with a median CD4 count of 233 cells/mm3 (IQR 85-434 cells/mm3). In a study by Galego et al. assessing bronchoscopies performed in an HIV positive population, 82% of the study population had a stage 3 Centre for Diseases Control (CDC) (CD4 200-499) cell count classification, with 67% of participants having a CD4 count <50 cells/mm.(6) Our study showed that HIV positive patients were less likely to be diagnosed with a malignancy, compared to HIV negative patients.
In our series, a suspected endobronchial lesion was the indication for bronchoscopy in 52% of patients. Cavitary lung lesions and non-resolving pneumonia accounted for 9% of the indications for bronchoscopy. This finding differed from a Brazilian study by Costa et al, where infections and secretions (42.4%) were the predominant indication for bronchoscopy and bronchoscopy was performed for a suspected neoplasm in only 11% of patients.(4)
Endobronchial lesions were noted in 44% of our study population. This finding is in keeping with a study by Kumar et al in India, where an endobronchial lesion was noted in almost 60% of the study population.(5)
In our series, lung malignancy was the commonest diagnosis, comprising 39%. A study in India showed that lung malignancies constituted the most predominant diagnosis (64%).(5) Primary lung malignancy is classified into small cell lung cancer and non-small cell lung cancer, with major histological subtypes listed as adenocarcinoma, squamous cell carcinoma, large cell lung carcinoma and small cell lung carcinoma.(7) Of these subtypes, adenocarcinoma is the predominant subtype of non-small cell lung cancer, accounting for between 25-30% of cases. Worldwide, recent reports indicate that squamous cell carcinoma incidence is on a downward trend, with adenocarcinoma on an upward trend;(8) however, this finding was not supported by our series. Our study showed that squamous cell carcinoma was the commonest histological subtype of primary lung malignancy (42%); followed by adenocarcinoma (31%) and neuroendocrine tumours (11%). However, our study is limited to only those who underwent flexible bronchoscopy. The contributions made by mixed histological subtypes such as large cell carcinoma and small cell carcinoma were negligible (0.42%), confirming the rarity of such cancers. The predominance of squamous cell carcinoma in our study was supported by similar findings by Kumar et al in 2017. In their series, the incidence of squamous cell carcinoma was 62% with adenocarcinoma in 27%.(9) Another Indian study revealed squamous cell carcinoma to be the most common histological subtype (34.5%) with adenocarcinoma and small cell carcinoma in 9% and 7%, respectively.(10) These findings may be due to the fact that adenocarcinoma is more commonly located peripherally, therefore making them generally less accessible to bronchoscopic diagnosis.
In our study, chronic non-specific inflammation, infections and organising pneumonia were diagnosed in 14% of the study population. Unsurprisingly, given the local prevalence of TB, TB remained the most common infective diagnosis. Bronchoscopy is not routinely indicated in the workup of lung infections, and is reserved for cases with uncertain diagnoses at our centre. This finding is similar to that obtained in an Indian study in which pulmonary tuberculosis was diagnosed in 15 % of the study population on bronchoscopy.(5)
Certain interstitial lung diseases (ILD) may be evaluated by bronchoscopy. In our series less than 3% of patients who underwent bronchoscopy were diagnosed with ILD. Fibreoptic bronchoscopy is certainly useful in certain interstitial lung diseases such as sarcoidosis; however surgical biopsy is indicated for a definitive diagnosis in most patients. In the current study, women were 7 times more likely to be diagnosed with sarcoidosis than men. This association persisted even after adjusting for age and HIV status.
Overall, bronchoscopy proved to be a relatively safe procedure with an overall complication rate of 2.16%, which is comparable with other studies, with complication rates during bronchoscopy of between 0.08% and 2%.(4,11,12)
In this study cohort, 3 fatalities were reported following the procedure. One was a case of sudden unexpected cardiac arrest, one patient had unexplained respiratory arrest post-bronchoscopy and the third was a case of sudden demise after transfer back to the medical ward after bronchoscopy.
Due to the retrospective nature of this study, there were limitations due to missing or incomplete data. However, this represented less than 2% of the overall study population. Selection bias may have been introduced, owing to the setting of the study in a tertiary hospital, and so overall findings may not reflect the pathology seen in other centres.
CONCLUSION
To date, this is the largest study in South Africa assessing patients that have had a bronchoscopy procedure. An endobronchial lesion was the indication for bronchoscopy in half of the patients and was visualised on bronchoscopy in 44% of the study population. In our cohort, squamous cell carcinoma was the commonest primary lung malignancy diagnosed, followed by adenocarcinoma. Flexible bronchoscopy was found to be a safe and useful tool in the diagnosis of respiratory diseases. It may also be worthwhile to study if the addition of an endobronchial ultrasound (EBUS) would improve the overall utility of bronchoscopy.
Author contributions – A.O.A, M.V and M.L.W conceived the study, A.O.A implemented the study. M.LW and M.V assisted with data analysis and reviewed multiple drafts of the manuscript. All authors approved the final manuscript.
Funding – None
Conflict of interest – None
