Survey of flexible fibreoptic bronchoscopy in the United Kingdom

It is now >30 yrs since Ikeda 1 introduced the technique of flexible fibreoptic bronchoscopy, and this procedure has been established throughout the UK for >20 yrs. Despite this, bronchoscopic practice in UK hospitals is not standardized. National guidelines on flexible bronchoscopy concentrate predominantly on safety aspects of the procedure, rather than practical aspects of the technique 2. In view of this, the authors undertook a national survey of flexible bronchoscopy which focused primarily on physicians' preferences for patient preparation, monitoring during the procedure, drug therapy and tissue sampling methods. Information concerning physicians' experiences of complications and of performing therapeutic procedures with the flexible bronchoscope was also sought. The aims of the study were to assess the extent of variation in physicians' common practice, to compare activity with evidence-based data, and to assess the degree of adherence to accepted standards of practice.

Method

Consultant physicians in adult respiratory medicine in the UK were identified by the national directory of the British Thoracic Society. A structured questionnaire was sent by post to the hospital addresses of all doctors listed, although it was anticipated that not all physicians contacted performed bronchoscopy. The questionnaire dealt with several areas of practice, including: 1) patient preparation and investigations prior to bronchoscopy, duration of fasting prior to bronchoscopy and patient consent; 2) drug administration for bronchoscopy, including methods of sedation, topical anaesthesia, and prophylactic antibiotic therapy; 3) patient monitoring during the procedure; 4) endoscopist clothing/protection; 5) sampling methods via the bronchoscope; 6) experience of therapeutic procedures performed via the bronchoscope; 7) experience of complications/death.

Responders were asked to provide detailed comment on all areas covered. To encourage a high response rate, replies were anonymous and nontraceable. Neither second questionnaires nor reminders were sent. Data were obtained only from physicians' practicing in state-funded hospitals; thus there was no financial incentive to perform bronchoscopy. Comparisons were made using Fisher's exact test, where appropriate.

Results

A total of 547 questionnaires were sent to consultant respiratory physicians. Three-hundred and twenty-eight were returned, of which 27 were incompletely answered. A further 15 physicians replied, but did not perform bronchoscopy; thus an overall response rate of 62.7% was attained. Over the 12‐month period prior to answering the questionnaire, approximately 60,100 bronchoscopic examinations had been performed or supervised by the 328 responders (median 175 procedures·physician⁻¹·yr⁻¹ (range: 5–1000)) (table 1⇓).

Discussion

These findings confirm that the practice of flexible bronchoscopy varies greatly. This is likely to result from a combination of individual familiarity with learned techniques, and the presence of conflicting data regarding many aspects of bronchoscopic procedure. However, when areas of management for which standards of care are already established were examined, considerable variation was noted also.

Patient consent was obtained by a junior doctor in almost one-third of responses. It is likely that a proportion of these doctors did not discuss intricate details/possible complications of bronchoscopy, due to lack of familiarity with the technique. A recent review of litigation cases arising from gastrointestinal endoscopy in the UK revealed more than one-third of claims related to poor patient explanation when obtaining consent 3. The authors believe that consent to bronchoscopy should only be obtained by personnel familiar with the procedure and its possible complications.

The authors were surprised by the number of physicians who fasted patients for <4 h before bronchoscopy. When administering sedation and topical anaesthetic agents, it is probably prudent to fast patients for at least 4 h. Prior assessment of blood-clotting function has been shown to be poorly predictive of haemorrhage in patients without risk factors for bleeding, who undergo bronchoscopy without TBB. The current finding that 18% of UK physicians routinely check blood clotting before bronchoscopy in such patients was much lower than the 44% of physicians in a recent North American survey who did likewise 4. Assessment of spirometric performance is unlikely to influence bronchoscopy, provided continuous assessment of oxygen saturation is made. Although almost all physicians now monitor oxygen saturation routinely during flexible bronchoscopy, the proportion requesting prior pulmonary function testing is unchanged from a previous British survey in 1983 5.

There has been debate concerning the need for sedation in flexible bronchoscopy. Sedation has been implicated as the cause of up to 50% of major complications resulting from bronchoscopy 6. Hatton et al. 7, in a double-blind trial, showed no improvement in patient comfort when comparing midazolam 70 µg·kg⁻¹ versus placebo. However, other studies have not confirmed these findings. Putinati et al. 8 showed significant improvement in patient tolerance when intravenous diazepam was given. This study also demonstrated a significant reduction in the number of abandoned procedures due to patient intolerance when sedation was given. The need for sedation is likely to vary among patients, and may be affected by the extent of explanation and reassurance given by medical staff. The finding that 7.3% of physicians performed bronchoscopy without sedation is similar to the 4% of physicians who did likewise in a recent American study 4. The fact that most physicians administered sedation suggests that it is generally felt that this practice is relatively safe. The combined use of opioid with benzodiazopine, however, can be questioned as this has not been shown to improve patient tolerance compared to either agent alone and may increase the risk of cardiorespiratory depression 9, 10.

Most endoscopists used lidocaine for topical anaesthesia. Only a minority favoured intranasal lidocaine spray, which produces greater patient discomfort than lidocaine gel 11. Lidocaine was administered via nebulizer by a small number of physicians. Despite high patient acceptability for this method 12, it may induce bronchospasm in susceptible patients 13. A large number of physicians reported use of total lidocaine doses in excess of previously recommended amounts (5 mg·kg⁻¹). Despite this, there were no reported cases of complications that could be directly attributed to lidocaine toxicity. Earlier studies have shown that doses of lidocaine well in excess of 5 mg·kg⁻¹ for bronchoscopy are tolerated by patients and rarely produce toxic blood levels 11, 14. However, in a report of 48,000 bronchoscopies by Suratt et al. 15, there were six documented cases of convulsion, which were felt to be due to topical anaesthetic toxicity, although drug doses were not stated.

Use of antimuscarinics (29%) prior to bronchoscopy was less common than in a previous British study (87%) 5, but contrasts with a recent North American survey, which showed that more than two-thirds of physicians still prescribed these agents 4. There is no convincing evidence that antimuscarinic drugs provide any benefit to bronchoscopy.

Administration of antibiotic prophylaxis against endocarditis was common. Although endocarditis after bronchoscopy has been described 16, it would appear rare. Bronchoscopy is, however, a recognized cause of bacteraemia, and in view of this, antibiotic prophylaxis is probably justified.

Patient monitoring throughout the procedure was generally in accordance with established recommendations 2. However, elecrocardiogram (ECG) monitoring was never undertaken by more than one-half of responders, which contrasts sharply with North American practice, where 85% of physicians routinely monitor the ECG during bronchoscopy 4. Continuous ECG monitoring is recommended, as myocardial ischaemia during the procedure is common, particularly in the elderly 17.

Hypoxaemia during bronchoscopy results from a combination of sedation, topical anaesthesia, and mechanical obstruction from the endoscope. Supplemental oxygen can avoid hypoxaemia in most cases [l8], and its routine use seems logical.

Low levels of clinical staffing during bronchoscopy (single assistant in 5% responses) may prove inadequate in the event of an emergency, and the authors feel that at least two trained assistants should be present to aid specimen handling and ensure adequate patient monitoring throughout the procedure. Reported wearing of gloves during bronchoscopy was high, but routine use of other protective clothing as previously recommended 2, was low.

Differences in the sampling methods performed were noted, particularly when the bronchoscopic appearances were normal or showed VOT. When sampling from VOTs any additional yield from cytological samples will depend partly on the number of endobronchial biopsies taken. As a cost-effective means of diagnosis, some authors recommend obtaining cytological samples when occluding tumour is seen, but analysing them only if the biopsy sample is nondiagnostic. Use of cytological samples in the diagnosis of malignancy when bronchoscopic appearances are normal or show only airway oedema/compression, provide a greater input to the overall diagnostic yield. Cytology alone provided the only means of diagnosis via the bronchoscope in 34.2% and 34.5% of bronchoscopically nonvisible tumours in two series 19, 20, with washings and brushing giving complementary yields. Furthermore, the yield from wash and brush cytology specimens appears even greater when fluoroscopic guidance is used 21, 22.

Only a small number of physicians performed bronchial biopsies (blindly or with fluoroscopic guidance) in patients with a pulmonary mass lesion and normal bronchoscopic appearances. This is perhaps surprising since diagnostic rates of ⩽69% for malignant lesions have been reported with biopsy under radiographical screening, in such patients 23, 24.

Controversy exists over the need for fluoroscopic guidance when performing TBB. The safety of TBB without fluoroscopy has been documented 25. However, Simpson et al. 5 found a significant reduction in the reported incidence of all pneumothorax when TBB was performed with fluoroscopy, compared to TBB without fluoroscopy. The current study showed a significant reduction only in the frequency of pneumothorax requiring drain insertion, following TBB when fluoroscopy was used.

Reported use of TBNA was low. This may have been due partly to the cost of disposable needles, but is also likely to be attributable to previous experience of a low yield, as well as a lack of familiarity and training with the technique. Low rates of utilization of TBNA have been reported previously in North America 4, 26. This is despite a number of series highlighting the sensitivity and safety of TBNA via the flexible bronchoscope in diagnosing and staging central and peripheral lung tumours, with and without radiographical control. Furthermore, Haponik et al. 27 reported a significantly increased yield from TBNA after endoscopists underwent a period of training in the technique, combined with improved preparation of cytology samples.

A small number of large airway stent insertions and other specialist procedures were reported. This is likely to be due to the low number of physicians regularly performing these therapeutic interventions, some of whom may not have replied. Nonetheless, the finding that 15% of physicians had recent experience of airway stenting via the flexible bronchoscope is higher than the corresponding figure from a recent survey of North American chest physicians where only 4.6% were found to stent patients 4. Although the current questionnaire did not ask the type of stent inserted by physicians, in the UK uncovered stents appear to be used in greater numbers than covered stents, a practice which is unusual in Europe (Personal communication, Boston Scientific). Complication rates reported from airway stents were similar to those reported previously 28.

The mortality rate reported from bronchoscopy of 0.045% was similar to that of a previous British survey (0.04%) 5, but somewhat higher than other multicentre studies (0.01–0.02%) 6, 15.

Problems inherent with surveys of this nature include dependance on physicians' reported answers, which may have relied on memory alone. Furthermore, in view of the anonymous nature of the questionnaire, replies could not be verified and some physicians may also have been reluctant to provide full information. The completed response rate of 60% compares favourably with previous bronchoscopy survey returns of 31% 15 and 53% 5. Furthermore, as it is likely that many nonresponders did not perform bronchoscopy, the authors feel that this work represents a reasonable assessment of current bronchoscopic practice in the UK.

Despite the diversity apparent in this study, bronchoscopy appears in the main, to be a safe procedure. Further prospective studies are required to establish optimal anaesthetic and sedation regimes, and to evaluate the most efficient and cost-effective combinations of sampling procedures for bronchoscopic investigation of patients with suspected lung tumours.