A survey of flexible bronchoscopy practice at Kenyatta National Hospital, Kenya

INTRODUCTION

Since the discovery of flexible bronchoscopy (FB) by Ikeda in 1964, it has become an important diagnostic and therapeutic tool for the management of respiratory diseases. FB is an easy-to-perform and safer procedure compared to rigid bronchoscopy. The FB has replaced rigid bronchoscopy because it can be performed as an outpatient procedure and is safer. It can also be performed under conscious sedation and local anesthesia, thereby avoiding the attendant complications of general anesthesia and allowing it to be conducted as an outpatient procedure.

The practice of FB is not standardized even in the developed world despite the availability of guidelines concerned primarily with the safety of the procedure.^[1] It is for this reason that the British Thoracic Society in its 2013 guideline on the practice of diagnostic FB recommends periodic clinical audits of the practice for conformity with the standards.^[2]

Clinical audits and surveys on the practice of FB have shown considerable variation in physicians’ routines in patient preparation, drug therapy, sampling method, and occurrence of complications. This variability was thought to result from a combination of individual bronchoscopist experience and a lack of accepted standard practice guidelines. British Thoracic Society (BTS) 2001 guideline recommends several pre-procedure patient evaluations and preparation to enhance patient safety.

The Kenyatta National Hospital (KNH) Respiratory and Infectious Disease Unit has been performing bronchoscopy for over 20 years now, yet there is no published local data on the practice, indications, pattern of bronchoscopic findings, and safety. The goal of this study was to document bronchoscopy practice in KNH in terms of indications, patient evaluation, patient preparation, anesthetic technique, patient monitoring, sampling procedures, gross and histologic findings, and complications related to bronchoscopy.

MATERIALS AND METHODS

RESULTS

A total of 290 bronchoscopies were conducted between January 2016 and December 2020. Eight cases were under 18 years. Two hundred and eight-two cases were eligible for inclusion in our study. Of the 282 cases, 228 had complete documentation while 54 cases had incomplete documentation.

We analyzed the entire cohort of the 282 cases including patients with incomplete documentation; hence, the numerator was different for the different variables.

DISCUSSION

The 5-year retrospective audit demonstrated a gradually increasing trend of FB practice at KNH. The number of patients undergoing FB increased from 24 cases in 2016 to 104 and 74 cases in 2019 and 2020. The drop in cases in 2020 was attributed to the coronavirus disease outbreak and national movement restrictions. The increase in trends can be attributed to a better staffing of the unit with a dedicated bronchoscopy nurse and additional chest physicians. Similar trends in the uptake of FB services were observed in studies in Nigeria^[3] and Tanzania.^[4] We found that 51.4% of patients undergoing FB were male while 48.6% were female with an overall mean age of 51.3 years. Ndilanha et al. in Tanzania, Adewole et al. in Nigeria, Kontakiotis et al. in Greece, and Bennett et al. in the UK found a similar pattern. The mean age of patients in the above studies ranged between 54 years and 65 years.^[3-6] The finding that the majority of our patients (58.2%) were residents of counties outside Nairobi including as far as Meru county which is over 200km away from Nairobi could indicate the unavailability of bronchoscopy services in those counties and/or the concentration of expertise within Nairobi.

The most common indications for bronchoscopy were found to be the evaluation of lung mass (35.7%), pulmonary infiltrates (18.9%), suspected interstitial lung disease (13.9%), atelectasis (6.4%), and hemoptysis (3.9%). This is similar to findings of retrospective studies in Tanzania, Egypt, Nigeria, and Greece with the evaluation of lung mass being the most common indication.^[3-5,7]

Over 91% of the patients who underwent FB in the 5 years under study had their pre- and post-procedure BP, HR, respiratory rate (RR), and SPO₂ monitored and documented. The regular patient monitoring in our study was also attributable to the availability of a dedicated nurse for the bronchoscopy unit who ensures patient monitoring and documentation. This practice is in line with most guideline recommendations. Both BTS and the Joint Indian Chest Society (JICS) recommend that all patients should have their BP, HR, RR, and SPO₂ monitored, and recorded repeatedly before, during, and after the procedure. Furthermore, the JICS endorses that the RR and the depth of sedation should be documented during FB. Insufficient monitoring combined with too deep sedation and undetected respiratory depression can lead to substantial complications in bronchoscopic procedures.^[8]

The majority of the patients (98.2%), in our study, underwent FB under sedation with only 1.8% being done under local anesthesia without sedation. This was similar to findings in surveys among pulmonologists in the UK (96%) and Pakistan (98%).^[2,9] This sedation rate was however much higher when compared to 44% found by Kontakios et al. in an audit of private hospitals in Greece.^[5] The lower sedation rate in the Greece study was attributed to the concept of “Vocal Anesthesia” employed in that study. This concept involved pre-procedure patient preparation and education, and simple explanations of the maneuvers during the procedure. Smyth and Stead also note that extent of explanation and reassurance given to the patient is likely to affect the need for anesthesia.^[1] We found the technique of sedation to be variable, with the majority of the patients getting a combination of Ketamine and benzodiazepine (often midazolam) (74.3%), a combination of opioids and benzodiazepine (15.4%), while about 10% got local anesthesia with either monotherapy of ketamine only (0.7%), benzodiazepine only (8.9%), and opioid only (0.7%), this variability could be due to preference of an individual bronchoscopist, system factors as well as some patient factors. According to Credle et al., there is no consensus on whether to sedate or not and the choice of sedatives.^[10]

About 94.3% of the patients, in our study, underwent one or more bronchoscopic diagnostic procedures for microbiologic, cytological, and/or histologic diagnosis. Over 60% of these underwent combination diagnostic procedures. The most common combination is bronchial wash and bronchial biopsy. This is similar to the findings of studies done in Egypt and Saudi Arabia.^[11,12] The use of a combination of bronchoscopic sampling procedures is a BTS guideline recommendation. It recommends that biopsy (at least 5 biopsies), brushing, and washing be combined to increase the diagnostic yield of endobronchial tumors. In the evaluation of infectious etiologies, the guidelines recommend a combination bronchial wash and Broch alveolar lavage.^[2]

The most common macroscopic finding in our setting was visible tumors (38.5%), closely followed by normal bronchoscopic findings (29.5%), inflammatory changes (11.5%), and secretions (8.3%). This was similar to findings in Egypt and Nigeria^[3,13] but different from a study in Saudi Arabia by Alamoudi et al. that found 64% had inflammatory changes, 31% had narrowed segments, and 19% had visible tumors.^[12] The study in Saudi Arabia had an evaluation of pulmonary infiltrates (suspected infection) as the most common indication for flexible bronchoscopy, unlike our study where a third of the patient was due to suspected lung cancer.^[12]

Immediate procedure-related complications were rare and minor. We reported a complication rate of 3.2%. The global prevalence rate ranges between 0.5% and 5%. Pue and Pacht, in a 5-year retrospective review of bronchoscopy, found a complication rate of 0.5%, this low rate was attributed to the fact that most cases, in his study, did not undergo a biopsy.^[20] Dreisin et al., in a prospective study of 205 bronchoscopies, found a complication rate of 5%.^[14] The variability in the reported complications is influenced by the nature of the sampling procedures done, patient selection, and whether or not late complications were included in the study. Complication rates, in our study, may be an underestimation as it only focuses on reported immediate intraprocedure complications.

The most common bronchoscopic diagnosis was lung cancer (40.1%), interstitial lung disease (15.9%), and TB (6.2%). This is similar to the findings of studies done in Tanzania, Nigeria, Egypt, Saudi Arabia, Greece, and the UK which all found lung cancer to be the most common bronchoscopic diagnosis.^[3-6,11] The most common histologic subtype of lung cancer in our study was adenocarcinoma (45%), followed by squamous cell carcinoma (25.3%) and small cell lung cancer (7.7%). This was similar to a study in Tanzania that found adenocarcinoma as the most common histologic subtype.^[4] Our findings are in keeping with recent studies in the USA, Norway, and Japan that showed a trend toward a greater predominance of adenocarcinoma.^[15-17] However, studies in Egypt and Saudi Arabia found squamous cell carcinoma as a more common histologic subtype.^[7,12] These findings could be explained by the fact that the proportion of smokers, both current and ex-smokers is much higher in Saudi Arabia (>50%) and Egypt (61.5%)^[7,12] compared to our setting (31.8%).

The proportion of smokers (current and ex-smokers) among patients with malignancy was also lower at 36.3%, compared to 84% found by Alamoudi et al., who found squamous cell carcinoma as the most common histologic subtype.^[12] These findings suggest that a large proportion of patients with lung cancer in our setting were non-smokers. Multiple factors contribute to the development of lung cancer, including environmental toxin exposure, genetic predisposition, infectious comorbidities, and individual lifestyle.^[18] Long-term air pollution exposure increases the risk of lung cancer,^[19] in addition, increasing exposure to particulate matter (PM2.5) increases the risk of non-smoking individuals with pre-existing mutations.^[18] This could be a plausible explanation for our findings but this needs further work.

The overall diagnostic yield of bronchoscopy was 81.9%. This was similar to what was reported in studies in Egypt that found overall yields of 78.6%.^[7] The diagnostic yield for malignancy where there was a visible endobronchial tumor was 87.2%. This was similar to rates found in studies by Halima et al.^[7] and Mohamed et al.^[11] who reported diagnostic yield of 95% and 94.7%, respectively, for malignancy where there was an endobronchial tumor. Our study disagrees with a study by Adewole et al. in Nigeria that reported much lower rates of 55%.^[3] This variability in diagnostic yield can be explained by the possible difference in bronchoscopist experience, diagnostic sampling methods used, and the number of biopsies taken. The rate of 87.2%, in our study, meets the recommendation of the British thoracic guideline that recommends a diagnostic yield of more than 85% where there is a visible tumor.^[2] This is to minimize the need for repeat bronchoscopy for biopsy and ensure timely diagnosis of lung tumors.

The study was the first to review the practice in Kenya; however, it was a retrospective chart review and there were significant documentation gaps.

CONCLUSION

Our study showed that FB practice at KNH is safe and provides a good diagnostic yield; however, significant documentation gaps exist. There is a need to individualize the pre-procedure laboratory evaluation rather than as a routine workup.

The most common lung cancer histologic subtype was adenocarcinoma among mostly a non-smoker population.

Our practice is comparable to other countries in Africa and the rest of the world in terms of patient preparation, procedural practices, complication rates, and diagnostic yield.