Bronchoscopes with an external diameter of 4.9mm were used (FB15-V Pentax Europe GMBH). All patients had a venous catheter and were subjected to continuous monitorization of blood pressure, heart rate and oxyhemoglobin saturation determined by pulsioxymetry, during and until 1h after the end of the procedure. Arterial blood samples were collected and assessed in situ using a blood gas analyzer (Radiometer ABL 30) 5min before and 1h after the procedure. FBS was performed in the RICU and was supervised by one of the senior endoscopists (CA, AX). The endoscopic exploration was carried out under sedation with intravenous propofol (1–2mg/kg) or midazolam (0.1mg/kg)–the use of fentanyl and/or pancuronium proving necessary in some cases. The medication doses were varied on an individualized basis with the purpose of ensuring correct oxygenation and adaptation of the patient to ventilation. The bronchoscope was inserted through the nasopharynx, tracheostomy, orotracheal tube (8mm in diameter), or the laryngeal mask. In the latter three cases we used an adaptor to maintain ventilation with positive pressure. Patients requiring mechanical ventilation were maintained in volume control mode. The ventilator pressure limit was increased to ensure an adequate tidal volume during each respiratory cycle. FiO2 was increased to 100% from 20min before premedication and was maintained during the study period and in the immediate recovery period. Where present, the positive end-expiratory pressure (PEEP) was maintained at 5mmHg during the procedure. In patients with noninvasive ventilation (BiPAP vision, Respironics, Murrysville, PA, USA), bilevel positive airway pressure (BiPAP) was maintained.11

The endoscopic techniques performed included the following: diffuse bronchoaspiration (BAS), BAL (150ml of physiological saline solution divided into 3 aliquots of 50ml each, instilled in the lobe of the pulmonary infiltrate or in the middle lobe or lingula in the case of diffuse infiltration), bronchial biopsies and transbronchial biopsies. The samples obtained during the endoscopic study were individualized in each case. The techniques for microbiological sample processing and interpretation of the results have been described elsewhere.12

The indications of FBS were divided into diagnostic and therapeutic. The impact of FBS upon patient management was evaluated on the basis of the results obtained with the procedure. An endoscopic study was considered positive when it resulted in a definitive diagnosis. We documented both global mortality in the RICU and mortality among the patients subjected to FBS.

Complications of FBS were defined as those events occurring during the procedure and/or 24h after its conclusion. The adverse effects were documented prospectively. In this context we assessed fever, oxyhemoglobin desaturation with a need for increased oxygenation in the 4h after the procedure, the presence of bleeding requiring termination of the study, hypotension with or without the need for volume expansion or the use of vasoactive drugs, the development of hypertensive crises requiring the start of treatment and/or suspension of the procedure, and the evidence of cardiac arrhythmias during and 1h after conclusion of the technique. We also documented the time from the start to the end of the endoscopic exploration, and divided the explorations according to whether they had taken less than 5min, between 5 and 10min, and over 10min.

To the effects of the statistical analysis of the results obtained, quantitative variables were expressed as the mean±standard deviation, while qualitative variables were expressed as percentages and absolute frequencies. Proportions were compared using the chi-squared test. A p-value of <0.05 was regarded as statistically significant. The SPSS statistical package was used (Microsoft; Redmond, WA, USA).

During the inclusion period, a total of 297 patients were admitted to the RICU, and 107 FBS procedures were performed in 69 patients (23% of the total subjects admitted to the RICU)–of which 46 were males. The mean patient age was 60±14 years. A total of 47 patients underwent a single FBS procedure, 19 underwent two, one patient underwent three, another underwent 5, and two patients were subjected to 7 procedures each.

Tables 1 and 2 show the indications of admission to the RICU, and the comorbidities of the patients who required FBS.

Bronchoscopy was carried out for diagnostic purposes in 88 cases (82%), and with therapeutic intent in 19 cases (18%). The chest X-ray study proved normal in 25 of the 88 cases (28%) in which FBS was carried out for diagnostic purposes, while unilateral and bilateral infiltrates were observed in 24 (27%) and 39 cases (44%), respectively.

The clinical indications of diagnostic FBS are presented in Table 3. All therapeutic FBS procedures were carried out as treatment for secretion retention and atelectasis, with the exception of one patient in which a Fogarty catheter had to be placed in the context of a bronchopleural fistula.

Seventy-three (68%) of the 107 FBS procedures were carried out in patients with mechanical ventilation, and 34 (32%) in patients with spontaneous respiration and oxygen therapy. In 6 cases (6%) the procedure was performed in patients subjected to noninvasive ventilation.

FBS was performed by residents in Pneumology (always under supervision by a pneumologist) in 72 cases (67%), by endoscopists in 28 cases (26%), and by pneumologists in 7 cases (6%). In 53% of the FBS techniques the duration was under minutes, while in 31% it lasted between 5 and 10min, and in only 16% of the cases did the procedure take more than 10min.

In 68 cases (64%) different diagnostic bronchoscopic methods were used. In this context, BAL and BAS were the most common techniques (both performed in 34 cases, 32%). In only 7 cases (7%) were biopsy samples obtained.

In 45 of the 88 diagnostic FBS procedures (51%) the objective was to identify the causal agent in a patient with pneumonia. FBS proved able to yield a specific diagnosis in 19 of the 45 cases (42%). Specifically, 20 FBS procedures were performed in 18 immunocompetent patients (10 in patients with community-acquired pneumonia and 8 in patients with ventilation associated pneumonia), while 25 procedures were carried out in 15 patients with different types of immune suppression (Table 4). In the group of immunocompetent patients, a specific diagnosis was obtained in 6 of the 20 explorations carried out (30%). In contrast, the diagnostic performance of FBS in the immune depressed patients was significantly better – reaching a specific diagnosis in 12 of the 25 explorations performed (48%) (p<0.01) (Tables 4 and 5).

FBS did not trigger the need for mechanical ventilation in any patient under spontaneous respiration or subjected to noninvasive ventilation. There were no differences in PaO2/FiO2 before and after the procedure (Table 6). The changes in PaO2/FiO2 were not significant on dividing the patients into those subjected to mechanical ventilation and those with spontaneous respiration, or into those with unilateral or bilateral pulmonary infiltrates. In contrast, in those patients subjected to BAL, a significant decrease in PaO2/FiO2 was observed after the procedure (182±74 vs 163±79; p<0.005) in both the immune depressed patients (n=20) (203±79 vs 182±86; p<0.02) and in the immunocompetent individuals (n=14) (140±39 vs 125±42; p<0.02).

In 13 (12%) of the 107 FBS procedures a transient increase in FiO2 proved necessary (with respect to the FiO2 value before the endoscopic exploration) due to worsening of the oxyhemoglobin values during the procedure. In 10 cases (9%) we recorded reversible cardiorespiratory events (isolated supraventricular extrasystoles and/or variations in blood pressure), without the need for drug treatment. No other significant complications were observed in the hours close to the procedure.

The global mortality rate among the patients admitted to the RICU during the study period was 14% (42 of the 297 admitted patients). In the group of patients who did not require FBS the mortality rate was 11% vs 25% in the group of patients subjected to at least one FBS procedure (17 of the 68 included patients). In turn, the mortality rate was 15% (7 of 46 patients) in those requiring one FBS procedure, and 45% in the patients requiring more than one procedure (10 of 22 patients).

The mortality rate among the immune depressed patients subjected to diagnostic FBS was 60% (9 of 15 patients)–this being significantly higher than the percentage seen in the immunocompetent patients, where the mortality rate was 30% (6 of 20 patients) (p<0.05). In 7 of 9 immune depressed patients who died, the specific diagnosis was obtained by FBS.