Article Text

Download PDFPDF

Bronchoalveolar lavage in children with cystic fibrosis: how many lobes should be sampled?
  1. F J Gilchrist1,
  2. S Salamat1,
  3. S Clayton1,
  4. J Peach1,
  5. J Alexander2,
  6. W Lenney1,3
  1. 1Academic Department of Child Health, University Hospital of North Staffordshire, Stoke-on-Trent, UK
  2. 2Paediatric Intensive Care Unit, University Hospital of North Staffordshire, Stoke-on-Trent, UK
  3. 3Institute of Science and Technology in Medicine, Keele University, Stoke-on-Trent, UK
  1. Correspondence to Dr Francis J Gilchrist, Academic Department of Child Health, University Hospital of North Staffordshire, Newcastle Road, Stoke-on-Trent ST4 6QG, UK; fjgilchrist{at}


Background The European Respiratory Society guidance on bronchoalveolar lavage (BAL) in children was published in 2000. It recommended taking one BAL specimen from the most affected lobe or from the right middle lobe in diffuse disease. In 2007, the European Respiratory Society modified the recommendations for children with cystic fibrosis (CF), suggesting two BAL specimens (right middle lobe and the lingula or the most affected lobe).

Objective To determine if BAL samples from one or two lobes give the full picture of lower airway infection in children with CF.

Design, setting and patients A retrospective review of all paediatric patients with CF who underwent flexible bronchoscopy between May 2007 and May 2009 was undertaken.

Main outcome measures As BAL specimens from all six lobes were collected, the BAL results were reviewed to identify if positive cultures would have been missed if only one lobe (right middle or most affected) or two lobes (right middle plus the lingula or most affected) had been sampled.

Results The results of 39 bronchoscopic procedures in 31 children were reviewed. The BAL samples were taken from 6 lobes in all 39 procedures. Had only one lobe been used, 26 positive cultures (14 organisms) would have been missed in 11 patients. Had two lobes been used, 12 positive cultures (8 organisms) would have been missed in 7 patients.

Conclusion A single-lobe BAL is insufficient in assessing patients with CF for lower airway infection. Even when BAL specimens are taken from two lobes, a number of infections may be missed.

View Full Text

Statistics from

Cystic fibrosis (CF) is characterised by recurrent respiratory infections and chronic inflammation, which contribute to a progressive deterioration in lung function. It has been shown that this process begins in infancy, often in the absence of symptoms.1,,3 Infection with Pseudomonas aeruginosa (PA) is of particular importance because if it is not diagnosed and treated promptly, chronic infection can develop. This is associated with clinical deterioration and reduced life expectancy.4

Diagnosing lower airway infection in infants and young children, who are unable to expectorate, is difficult. Traditionally, this has been done using oropharyngeal (cough) swabs, but these have been shown to be unreliable when compared with bronchoalveolar lavage (BAL) or sputum samples.5 6 Induced sputum is safe and effective, but it is our experience that children often find it unpleasant, and there is limited experience of its use in patients with CF.7 BAL is the main method to gain reliable information on lower respiratory tract infections in young children with CF.

What is already known on this topic

A single-lobe bronchoalveolar lavage is insufficient in assessing patients with cystic fibrosis for lower airway infection.

What this study adds

Even when bronchoalveolar lavage samples are taken from two lobes, a number of infections may be missed.

Flexible bronchoscopy in children is a safe procedure with a low incidence of complications. BAL performed during the flexible bronchoscopy procedure is also safe, but it does increase the incidence of postbronchoscopy fever.8 A European Respiratory Society (ERS) task force published guidance on BAL in children in 2000.9 It recommended taking a single BAL from the most affected lobe (identified radiologically and/or endoscopically) or if there was diffuse disease from the right middle lobe (RML). It recommended that the BAL should consist of three aliquots of normal saline, the first to be sent for microbiological surveillance and the second and third for inflammatory studies.

A study in 2001 compared BAL results from the RML and the lingula in children with CF. It showed that although the inflammatory indices were similar in both lobes, there were significant differences in the bacterial culture results, suggesting that bacterial distribution is inhomogeneous.10 An ERS-sponsored workshop held in Leuven, Belgium in 2007 modified the BAL recommendations for children with CF—it advised taking a three-aliquot BAL from the RML and a single-aliquot BAL from the lingula or the most affected lobe.11

In line with recently published data, we have been performing bronchoscopies more frequently in our paediatric patients with CF.11 12 This is usually to obtain a definitive microbiological diagnosis and guide management. For patients with no CF, we follow the ERS recommendations for BAL. In our patients with CF, we have traditionally taken single-aliquot BALs from all six lobes because therapeutic lavages were often necessary and we do not routinely perform inflammatory studies. We were planning to modify our BAL practice to that recommended by the aforementioned 2007 ERS workshop. Before doing this, we thought we should review our bronchoscopic findings to see if we would have missed any positive cultures if we had only sampled a single lobe (most affected lobe or RML in diffuse disease) or two lobes (RML plus the most affected lobe or lingula in diffuse disease).


For our paediatric patients with CF, BALs are performed under general anaesthesia. A bronchoscope (BF3C30/BF3C40 3.6 mm; Olympus America Inc, Center Valley, PA, USA) is introduced to the lower airway through a laryngeal mask, and the suction port is not used until the tip of the bronchoscope is below the level of the carina. BALs are obtained by wedging the tip of the scope into a lobar bronchus and gently instilling 0.5–1 ml/kg (maximum 20 ml) of room temperature 0.9% saline, under direct vision. The saline is then immediately aspirated into a sterile suction trap. In patients with diffuse disease, the lobes are sampled in a set order: right upper, right middle, right lower, left upper, lingula and left lower. In patients with an identified most affected lobe, this is sampled first and then the above order is followed.

BAL samples are sent to the microbiology laboratory for semiquantitative bacterial cultures. Visible growth is categorised into the following groups: no growth, scanty growth, moderate growth and heavy growth. After each procedure, the bronchoscope is manually cleaned with detergent/enzymatic solution at the bedside and then reprocessed using an automatic endoscope reprocessor. The bronchoscope is reprocessed before use. Weekly surveillance swabs are taken from the scopes, which have always been negative.

We undertook a retrospective case note review of all paediatric patients with CF who had undergone flexible bronchoscopy from May 2007 to May 2009. We recorded the BAL culture results taken during bronchoscopy and the surveillance culture results over the preceding 12 months.


We identified 39 bronchoscopies performed on 31 patients. The median age at bronchoscopy was 8.5 years (interquartile range 5.3–11.5 years), and the median time from CF diagnosis to the first bronchoscopy was 7 years (IQR 4.5–8.7 years). At the time of bronchoscopy, 85% of patients had symptoms (46% productive cough, 33% non-productive cough, and 6% nocturnal cough). In all 39 bronchoscopies, BAL samples were taken from 6 lobes. Positive BAL cultures were obtained from 31/39 bronchoscopies. A total of 56 organisms were identified; of these, only 15 had been identified by surveillance cultures (cough swabs or sputum samples) in the same patient, over the preceding 4 months. PA was cultured in 9 patients (2 for the first time), and Staphylococcus aureus was isolated in 17 patients (5 for the first time).

Thirty-three of the procedures were performed in patients with diffuse disease and six in patients with an identifiable most affected lobe. This lobe was the left upper on three occasions and the right upper, right lower and lingula once. It was identified radiologically on four occasions (all by computerised tomography scan) and twice at bronchoscopy. Given this, when the BAL results from 6 lobes were compared with those from a single lobe (as per the 2000 ERS guidance), the RML was used 33 times; the left upper, 3 times; and the right upper, right lower and lingual, once. When the BAL results from the six lobes were compared with the results from two lobes (as per the 2007 ERS guidance), the RML and lingula were used 34 times; the RML and the left upper, 3 times; the RML and the right upper, once; and the RML and the right lower, once.

If we had relied solely on the BAL culture results from a single lobe, we would have missed 26 positive cultures (14 organisms), in 11 patients. All these patients had diffuse disease. The organisms were four Haemophilus influenzae, three Stenotrophomonas maltophilia, two PA, two Sphingomonas paucimobilis and one each of Moraxella catarrhalis, Aspergillus fumigatus and S aureus. If we had relied on BAL cultures from 2 lobes, we would have missed 12 positive cultures (8 organisms), in 7 patients. These organisms were three S maltophilia, two H influenzae, two S paucimobilis and one A fumigatus (see table 1). In the six patients with an identified most affected lobe, no organisms would have been missed if we only had BAL results from one or two lobes. Five patients had documented postprocedure fever >40°C (12.8%), and no clinical significant episodes were recorded during the procedures.

Table 1

BAL results from the 11 patients in whom positive cultures would have been missed (shaded results) if we had relied on results from a single lobe (RML in these 11 cases)


The severity of bronchiectasis varies between lobes when examined radiologically13 and at postmortem.14 Bacterial distribution is also heterogeneously distributed throughout the lungs, with PA abundant in areas of active inflammation and absent in areas without inflammation.15 As mentioned previously, Gutierrez et al10 demonstrated significant differences in the BAL bacterial culture results between the RML and lingula. They concluded that single-lobe BAL may not be representative of other regions of the lung and that sampling from multiple lung segments or lobes should be recommended. This contributed to the guidance from the ERS in 2007 regarding BAL in children with CF, which recommends taking BAL samples from two lobes (a triple-aliquot BAL from the RML and a single-aliquot BAL from the lingula or most affected lobe).

We believe ours is the first reported study in children with CF to have BAL culture results from multiple lobes. We have compared our results from six lobes with those from one or two lobes that would have been sampled if the ERS guidance had been followed. Our data support previous findings that a single-lobe BAL is insufficient in assessing patients with CF for lower airway infection. If we had relied only on a BAL sample from the RML or the most affected lobe, we would have missed 26 positive cultures from 11 patients including 2 patients with PA and 1 with S aureus. Our results also indicate that when BAL samples are taken from two lobes, not all lower airway infections are identified. If we had relied on culture results from the RML plus the lingula or most affected lobe, we would still have missed positive cultures in seven patients, including three with H influenza. Although some of the organisms that would have been missed are not regarded as respiratory pathogens, the sampling of all six lobes did alter the choice of antibiotic regimens on a number of occasions.

As we take a single-aliquot BAL from each lobe, the samples are likely to be mainly bronchial in origin and not the more distal samples that subsequent aliquots would yield. We do not believe that this affects the culture results because culture samples taken in accordance with the ERS guidance on BAL in children with CF are also proximal bronchial washings. This is because it is only the first of the three aliquots that are sent for culture in the triple aliquot BAL (the second and third are pooled and sent for cytological and inflammatory studies) and the single-aliquot sample that is suggested uses a method similar to our own. In larger children, because of the comparative sizes of the bronchoscope and the airways, when taking BAL samples, the scope was wedged into a segmental bronchus rather than a lobar bronchus. In these patients, we would only have been sampling part of the lobe rather than the full lobe. Because of this, we may not have obtained the full microbiological picture.

During a bronchoscopic procedure where the scope is passed from one lobe to the next, there is a risk of cross contamination of organisms between lobes. There is also a theoretical risk that when taking multiple lavage samples, a later sample may be contaminated by residual fluid in the suction channel from an earlier one. As we perform the BALs in a set order, we were able to analyse the results and they do not suggest that this occurred. Despite this, as it is a retrospective study, it is not possible to definitively exclude cross contamination. Taking BAL from multiple lobes is likely to prolong the anaesthetic time and increases the total volume of lavage fluid. Despite this, there were no clinically significant events recorded during the procedures, and the incidence of postprocedure fever was similar to other published data.8 16

Our data confirm that the bacterial distribution differs between the two lungs and within lobes of the same lung. Hence, BAL samples from one or two lobes are insufficient to give a full picture of lower airway infection in CF. This needs to be considered when interpreting studies, which have reported the presence of inflammation but no organisms in BAL samples from one or two lobes. For individual patients, missing lower airway infections using this gold-standard method of diagnosis is clinically important. Infections may go untreated for many weeks leading to chronic inflammation and permanent structural changes. For organisms such as PA and S aureus, this may lead to colonisation of the airways with its associated morbidity and mortality. Failure to identify organisms that are present in the lower airways may lead the clinician to undertake unnecessary investigation and commence inappropriate treatments.

In summary, the data from this small study suggest that limiting BAL samples to one or two lobes may not give the full picture of lower airway infection in young children with CF. Although it does not provide strong enough evidence to suggest a change to the current ERS guidance, further work is required to clarify from which lobes BAL samples should be taken, thereby ensuring that infections are not missed and patients' lung health is optimised.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.