Diagnostic yield and complications of flexible bronchoscopy performed at Charlotte Maxeke Johannesburg academic hospital: A 5 year restrospective study

INTRODUCTION

The lung is significantly exposed to airborne pollutants such as tobacco, biomass fuel, and various pathogens. This impacts the five common respiratory conditions that contribute to the global burden of disease: acute respiratory infections, chronic obstructive pulmonary disease (COPD), asthma, tuberculosis, and lung cancer.^[1] The use of flexible bronchoscopy (FB) in pulmonary medicine has provided access to the tracheobronchial tree both by direct visualization and through radiographic visualization, allowing therapeutic and diagnostic interventions that were not previously possible.^[2] Access is possible up to the fifth subsegmental bronchi, and when used appropriately, it is well tolerated by most patients with few complications.^[3]

The British Thoracic Society Guidelines (BTS) aimed to help those performing FB to better understand the procedure, its complications, and its role in the possible diagnosis and management of lung pathologies. FB is an integral part of pulmonary diagnostics and the constant evolution of new techniques allows safer and more effective ways to diagnose and treat patients. It is important, however, that they should be performed only in circumstances where the benefit outweighs the risk particularly when more invasive procedures are employed. The use of FB is limited by the paucity of skills and there is minimal literature in Africa, on the role, safety, and yield of FB. This study described the diagnostic yield, indications, and complications reported during flexible bronchoscopes at Charlotte Maxeke Johannesburg Academic Hospital (CMJAH) and determined the effectiveness of biopsy and bronchoalveolar washings in making specific diagnoses.

MATERIALS AND METHODS

This study was a retrospective cross-sectional review of all bronchoscopies performed on adult patients in the bronchoscopy suite at Charlotte Maxeke Academic Hospital during the period from 1 January 2015 and 31 December 2019. The population studied were all adults (above the age of 18 years) attending the respiratory clinic, referrals from respiratory, general wards, and peripheral hospitals with indications for bronchoscopy. Findings based on the clinical indications, medical background, chest X-ray features, computed tomography (CT) scan findings, and any complications were documented. Bronchoalveolar lavage (BAL) and endobronchial sampling were obtained based on the clinical situation, samples were sent off for histology and microbiology, and results were checked for the diagnosis obtained from the National Health Laboratory Services (NHLS) database. This data were tabulated using Excel spreadsheets and analyzed. Some records and results were not easy to trace, and so were not analyzed. There were 692 patients data entered and 647 cases analyzed due to 45 cases with inadequate data for analyses.

Age was categorized into four groups: 18–34, 35–44, 45–55, and 55+ years. Clinical data were comprised of comorbidities (such as human immunodeficiency virus (HIV) status, smoking, hypertension, malignancy, lung diseases), and in those that were HIV-positive, viral load and CD4 counts were included. CT scan findings, complications, and indications were also documented. Patients were asked about previous bronchoscopy (yes or no) and sputum Cepheid® GeneXpert (GXP) results (negative or positive) before bronchoscopy. Bacteriological analysis of specimens was conducted to identify organisms such as Klebsiella, Mycobacterium tuberculosis (MTB), Pseudomonas, and Candida. The diagnostic tests that were used were biopsy and BAL. Frequencies and percentages were reported for categorical data stratified by gender and HIV status. Not all patients had HIV tests done, so they were reported as HIV-positive or HIV-unknown. To compare categorical variables for statistical significance by gender and HIV status, Chi-square analysis or Fisher’s test was used. Descriptive statistics such as median, inter-quartile range (IQR), mean and standard deviation were presented for continuous measures. Student t-test and Kruskal–Wallis were used to compare continuous variables stratified by gender and HIV status. All statistical analysis was conducted in SAS Enterprise Guide 7.1 using the procedures FREQ, MEANS, NPAR1WAY, TTEST, and SGPLOT.

RESULTS

Indications for bronchoscopy

The majority of patients had suspected malignancy (63.99%, n = 414), followed by non-resolving pneumonia (14.53%, n = 94), chronic cough (7.88%, n = 51), and pulmonary infiltrates (6.03%, n = 39) [Figure 1].

Close

Figure 1:

Indications of bronchoscopy.

Export to PPT

DISCUSSION

FB is a crucial part of pulmonary medicine. Access to the procedure is unfortunately limited and is available only in tertiary/quaternary hospitals in Africa. Bronchoscopies need to be performed in a safe environment and are usually dependent on the expertise of the individual performing them. The BTS guidelines for diagnostic FB in adults and the local institutional protocol dictate that bronchoscopy should be performed under conscious sedation by a specialist pulmonologist or by a pulmonology trainee/physician.^[4]

During the five years of the study, we observed a male predominance of patients undergoing the procedure, but otherwise, the epidemiology was similar to other studies including the mean age years of 55.9.^[5,6] The overall diagnostic yield revealed a 36.6% biopsy and 33.9% BAL yield.

The most common indication for bronchoscopy was a suspicion of malignancy in 63.99% of cases, and 44.20% were smokers with an obvious mass on chest radiology , with masses or lung nodules reported in 235 CT scans. A finding that was similar to other studies in Africa.^[2,7] A survey reported in the United Kingdom also showed similar results, with malignancy being the most common indication for bronchoscopy.^[8]

The association between smoking and lung cancer has been well-established and documented since the 1930s in Germany and later in Great Britain and America in the 1960s.^[9] A systematic review and meta-analysis showed clear evidence of the association between the amount smoked, the duration of smoking, and all lung cancer and not only histological type.^[9] In this study, carcinoma was diagnosed in 203 cases, 120 on biopsy and 83 on BAL. Ninety-nine of the 203 were primary lung cancers, and of these, 33.33% were squamous, followed by adenocarcinomas at 30.30%. Of these, 35.38% were smokers and 29.41% were non-smokers, which was a finding similar to that of the meta-analysis in Great Britain. In contrast, a study conducted in Sao Paulo, Brazil, revealed that respiratory infections, such as MTB, were the most common indications for FB, with malignancy only third.^[10]

Non-resolving pneumonia is an important indication for FB as the causes include serious but potentially treatable conditions such as i.e., Staphylococcus and Klebsiella species, as well as some fungal and parasitic infections.^[11] Granulomatous diseases such as MTB infections and sarcoidosis, lung malignancies, vasculitides, and interstitial lung diseases are some of the treatable causes of non-resolving pneumonia that may be diagnosed using FB.^[11] Non-resolving pneumonia and chronic cough were also common indications for FB in 14.53% and 7.88% of cases, respectively. Our study, similar to a study performed in New Delhi, India, by Chaudhuri et al., also revealed that bacterial pneumonia, due to Klebsiella pneumoniae and Pseudomonas aeruginosa, were the most common organisms cultured causing non-resolving pneumonia, followed by bronchogenic carcinoma.^[11] In the current study in Johannesburg, South Africa, Pseudomonas spp. was the most common organism found in 70 cases, followed by MTB in 22 cases and K. pneumoniae in 19 cases. Twenty-three cultured Candida albicans, which was unlikely to be pathogenic in the lower respiratory tract.

Africa, and in particular, sub-Saharan Africa, has a very high burden of HIV infection along with this MTB.^[12] At CMJAH, not all patients who had bronchoscopy were tested for HIV, but of those that were tested, 20% were positive, with the remainder 79.90% being unknown. This was much lower in comparison to a study from the United States of America (USA) that recruited patients between 2012 and 2017, where 30.7% were found to be positive. Although the overall prevalence of HIV in the USA is lower than in SA, the low numbers reported might be due to the fact that many of our patients often decline testing due to fear or stigma.^[13] Those who were HIV-positive were primarily between the ages of 45 and 55, followed by 34–44, the majority of whom were on anti-retroviral therapy (ART) and had a suppressed viral load.

The evolution of TB diagnostics has led to the GXP test, which uses single cartridge-based nucleic acid amplification which can identify MTB and detect rifampicin resistance within 2–3 h.^[14] During the period of this audit, 21 cases were diagnosed with MTB by biopsy and BAL included. This is probably because the GXP assay performed on bronchial washings was more sensitive than induced sputum, as suggested by Lee et al.^[15]

There was a 14.32% complication rate during the study period (94 of 656 cases), the most common of which was any hemorrhage which occurred in 71 cases. Arrhythmias, bronchospasm, hypoxemia, pneumothorax, severe cough, and a single respiratory arrest, with survival, also occurred but were less common, and importantly, there was a zero mortality rate. A study in Egypt in 2013 reported that desaturation and bleeding were the most common complications, although this was on a sample size of only 100.^[16] Another Egyptian study, performed over seven years with 3980 cases, reported bronchospasm as one of the major complications.^[7] In the Nigerian Tertiary Centre, 163 diagnostic bronchoscopies were studied with an overall complication rate of 5%, with the majority of these being minor, such as spontaneously resolving epistaxis, vasovagal reaction, and hypoxemia.^[2]

CONCLUSION

The availability of FB has significantly impacted the diagnosis and treatment of pulmonary diseases. FB should be accessible to a wider range of public sector patients, especially those in whom there is a diagnostic dilemma. This study has shown that there is significant diagnostic utility in terms of lung malignancy and other pulmonary diseases such as PTB, which, if identified early, increase the likelihood of cure.