Respiratory infections are the leading cause of morbidity and mortality in cystic fibrosis. Certain bacteria, such as Pseudomonas aeruginosa, are associated with a worse clinical outcome than others, but can be completely eradicated if identified and treated early. The diagnosis of lower respiratory tract infections can be challenging in the non-expectorating patient, in whom upper airway samples, such as cough swabs, are a surrogate for lower airway sampling. However, the results of these often do not fit with the clinical picture, presenting a management dilemma. Frequently, clinicians are faced with a negative culture result in a progressively symptomatic patient and vice versa. When judging the clinical significance of a positive upper airway culture result in an asymptomatic patient, it is important to consider the prognostic significance of the organism cultured. Given that the reported sensitivity of upper airway swabs (which includes throat swabs) is variable, ranging from 35.7% to 71% for Pseudomonas aeruginosa, 50% to 86% for Staphylococcus aureus and 11% to 92% for Haemophilus influenza, upper airway samples may fail to identify lower airway infections. Therefore, in symptomatic children, a repeatedly negative upper airway swab should not be considered as reassuring, and alternative sampling methods, such as induced sputum or bronchoalveolar lavage, should be considered. Here we use some examples of common scenarios to illustrate how best to use bacterial cultures to aid management decisions in cystic fibrosis.
- Cystic Fibrosis
Statistics from Altmetric.com
Recurrent bacterial infections are the leading cause of morbidity and mortality in cystic fibrosis (CF), accounting for over 80% of deaths.1 Early detection of infections is vital. Newborn screening allows most children to be diagnosed soon after birth providing a previously inaccessible window of opportunity for microbiological surveillance. The nutritional benefits of early diagnosis are clear and studies have also reported improvements in lung function, Pseudomonas aeruginosa acquisition rates and survival.2 Conversely chronic infection with P aeruginosa is associated with increased morbidity and mortality.3
It is important to detect lower airway infection accurately and promptly. CF Trust Guidelines4 recommends regular culture of expectorated sputum. In adults this is usually straightforward. Sputum is convenient, has a yield comparable with the ‘gold standard’ bronchoalveolar lavage (BAL),5 and has a high negative predictive value (NPV) and positive predictive value (PPV), up to 100%.6
More challenging is the non-expectorating patient, in particular when asymptomatic. Unfortunately, many patients, particularly young children, and also a proportion of adults, are unable to expectorate and cough swabs are performed. However, these have poor sensitivity and specificity and can lead to undertreatment or inappropriate antibiotic therapy.
This review focuses on the strengths and limitations of non-invasive airway sampling in CF, and when alternatives, such as sputum induction or BAL should be considered.
Physiological background: airway inflammation and infection
CF is due to a defect in the gene encoding the CF transmembrane conductance regulator protein, leading to either absence or lack of function of CF transmembrane conductance regulator, defective regulation of ion transport and airway surface dehydration. The current favoured hypothesis is that reduced airway surface liquid leads to reduced ciliary function, increased mucus viscosity and decreased mucociliary clearance.7 There is mucus hypersecretion, and once infection and inflammation are present, mucus viscosity is further increased by neutrophil-derived DNA and actin. Cough clearance is also impaired, resulting in the retention of secretions in the lower airways, where they provide a nidus for bacterial infection.
Early in life, bacteria such as Staphylococcus aureus predominate. With increasing age, P aeruginosa emerges as the most common and significant pathogen (figure 1).8 Persistence in the airway and a number of environmental triggers lead to mutations in P aeruginosa, overproduction of alginate and a mucoid phenotype. In addition, quorum sensing initiates biofilm formation, which enhances resistance to antibiotics.9 Although traditionally only a few organisms have been associated with CF airway infection, culture-independent techniques demonstrate greater diversity of microbes although the clinical significance of this is currently unclear.10 ,11
Lower airway cultures can be obtained by (A) spontaneous expectoration of sputum, (B) cough swab or plate, (C) induced sputum or (D) direct access to the lower airway (usually with bronchoscopy). Some clinics also measure serum antipseudomonal antibodies. Spontaneously expectorated sputum samples are relatively straightforward and will not be discussed further.
Cough swabs are performed by placing a swab into the posterior pharynx and asking the child to cough (or stimulating cough in a very young child); the aim is to avoid direct contact with the pharyngeal mucosa to prevent upper airway contamination. With older, compliant children, this is easy, although in a well child with a dry cough, yield may be low. Difficulties arise in younger children. The ideal timing of the test is debateable, with some studies showing increased sensitivity if oropharyngeal cultures are collected after physiotherapy12 or nebulised hypertonic saline.13 Throat swabs directly sampling from the upper airway are preferred in USA. Swabs should not be taken immediately after a child has been fed, due to the risk of gagging and potential contamination. Cough plates, taken by asking a child to cough directly onto a plate of culture medium, are sometimes used as an alternative, although repeated samples with specific agar plates are required. Their sensitivity is however debatable, with a recent study of 95 non-expectorating children demonstrating a pathogen isolation rate of only 8% with cough plates in comparison with 18.2% with cough swabs.14
Sputum induction is performed by ultrasonic nebulisation of hypertonic saline; even in healthy individuals, some ‘sputum’ can often be expectorated and the procedure is, in general, well tolerated although it may cause bronchoconstriction. However, it is time consuming, with a median duration of 49 min, and expensive, with an additional cost of US$150 in comparison with conventional techniques, thus limiting its use within the outpatient clinic.15 It is not performed routinely in most clinics, although some report high success rates even in small babies.16 Sputum induction, if successful, could obviate the need for BAL. Studies are underway to address this issue systematically.
BAL, often said to be the gold standard, has disadvantages; most importantly, it requires sedation or a general anaesthetic and cannot be performed frequently. There are regional differences in pathogens, so unless the entire bronchial tree is sampled, which is impractical, organisms may be missed.17 ,18
The temperature at which samples are stored can impact culture results. The optimum storage temperature would prevent overgrowth of organisms, but not lead to bacterial death. Ideally, samples would either be processed immediately or frozen. However, in clinical practice, particularly for samples collected in the community, storage at room temperature or refrigeration at 4°C is more practical. There is a lack of data comparing the effect of storage temperature on culture results in CF. Studies in bronchiectasis confirmed a negative impact of refrigeration of sputum compared with storage at room temperature for up to 24 h, with a 10-fold reduction in the number of pathogens present in 23% and 8% of samples, respectively.19 Conversely, May et al20 showed a 50% decrease in isolation of Haemophilus influenza in samples posted to the laboratory and arriving 24–36 h later. The CF Trust4 recommends as a compromise that samples be processed as soon as possible, and to use refrigeration where there is likely to be a delay of more than a few hours, and to interpret posted samples with caution.
The laboratory handling of samples can pose difficulties, and only experienced laboratories should be used. Mucolytics, such as dithiothreitol, can help in the culture of bacteria; mechanical homogenisation with sterile glass beads is preferred by other laboratories. The culture of ‘polymicrobial samples’ is difficult, because of bacterial overgrowth, particularly by mucoid P aeruginosa, and different pathogens will have different growth requirements, so it is important to use selective media4 ,21 especially for Burkholderia cepacia complex organisms. Culture duration is variable according to pathogen, with up to 5 days required for B cepacia complex.4
The use of P aeruginosa serology is controversial; the rationale is based on the premise that children will seroconvert during infection and revert once infection is eradicated. If true, serology would provide an attractive option but the data is variable, in part relating to differences in the nature of the sample (BAL vs sputum or oropharyngeal cultures), antigen assayed and cut-off values used (table 1).
Indications and limitations
In the chronic management of children with CF, how often should bacterial cultures be taken? Do regular surveillance cultures improve the rate of detection of bacterial infections?
Current, non-evidence based guidelines recommend culture of sputum or cough swab at every clinic visit (every 2–3 months) and, in addition, during exacerbations.21 ,22 There are no studies comparing outcomes of surveillance microbiology with culture only when patients are symptomatic. The rationale behind frequent surveillance is that significant bacterial infection is common in the absence of symptoms,23 with more than 50% of patients with P aeruginosa positive cultures reported to be asymptomatic,24 and that early attempts at eradication are likely to be more successful. Without regular surveillance, the detection of prognostically important pathogens may otherwise be missed.
In an asymptomatic child with CF, does a positive surveillance cough swab mandate treatment with antibiotics?
This is a common scenario, and the interpretation of the positive cough swab depends on the clinical context. The literature is (A) largely based on throat rather than cough swabs and (B) divided on issues of sensitivity and specificity. Cough swabs have usually been compared with sputum, whereas throat swabs have been compared with BAL. Importantly, the PPV of upper airway cultures can vary greatly for individual organisms, with reported ranges of 44% to 83.3% for P aeruginosa, 33% to 64% for S aureus and 50% to 81% for H influenza (see table 2), although this is predominantly for throat swabs. Hence, the accuracy of a positive upper airway culture to predict lower airway infections is questionable. There have been no randomised controlled trials of treating versus ‘ignoring’ positive cultures in this context.
The following details of the clinical context need to be considered: What is the pathogen that has been identified and what is its prognostic significance? Has this child grown this pathogen before? Is this a new isolate or does it reflect chronic infection? Is the laboratory able to give some idea of the number of organisms isolated? Some common clinical scenarios are discussed in more detail below.
A surveillance cough swab positive for P aeruginosa×10−3 cfu/mL in an infant with previously negative cultures
Ramsey et al25 have reported that quantitative oropharyngeal cultures are neither more sensitive nor specific than qualitative cultures for P aeruginosa, S aureus and H influenza. Given the adverse prognosis associated with P aeruginosa, treatment would be initiated by most specialists after even a single upper airway culture in a previously uninfected child, irrespective of quantity of organisms present. Good practice would be to attempt a second culture prior to initiating eradication treatment, although this may not always be practical. No trials have yet defined the optimal eradication regimen; success rates of around 80% have been achieved with combinations of oral ciprofloxacin/colomycin and with nebulised tobramycin. The UK is currently conducting a randomised controlled trial to compare oral or intravenous antipseudomonals (both arms also receive nebulised colomycin).26
Should attempts at eradication fail, guidelines recommend that regular inhaled antipseudomonal antibiotics (eg, colomycin) are continued for chronic P aeruginosa growths, with a 2 week course of oral ciprofloxacin reserved for early use with increased symptoms.21
Recurrent growth of Staphylococcus aureus on surveillance cough swabs despite flucloxacillin prophylaxis
Despite optimal management, some patients with CF become chronically infected with bacterial pathogens. The range of PPVs for methicillin-sensitive S aureus (MSSA) on oropharyngeal cultures is the lowest compared with P aeruginosa and H influenza, therefore this result is even less reliable than for other organisms (table 2). In this scenario, with a well child who is already receiving a prophylactic agent, attention should be focussed on determining adherence to prescribed medications, a particular problem in the teenage years. Although guidelines recommend treatment with oral antibiotics for initial growths of MSSA,21 for persistent infection, aggressive use of antibiotics may achieve little and carries with it the potential for side effects. The long term consequences of MSSA in CF are generally regarded as less severe than those of organisms, such as P aeruginosa,27 which would require continued treatment, as described above. Further courses of antistaphylococcal agents can be considered should symptoms develop.
To our knowledge, there have been no reports on the predictive values of cough swabs for the detection of methicillin-resistant S aureus (MRSA) specifically. The effect of MRSA on morbidity in CF is debateable. Two large observational studies have reported conflicting results. Dasenbrook et al28 observed an increased rate of decline in forced expiratory volume in 1 s of 0.5% predicted per year in 1732 patients with new persistent MRSA versus patients without MRSA. In contrast Sawicki et al29 reported a decline in lung function beginning prior to acquisition of MRSA in 593 new MRSA cases, suggesting that morbidity associated with MRSA lung infections may be a marker of disease severity rather than MRSA status itself. Similarly, Hubert et al27 suggested that MRSA may be more harmful only when there is co-infection with P aeruginosa, thus several confounders could contribute to the effect of MRSA on lung function decline. Given that a 1.27 times increased risk of death has been reported with MRSA infection, and no increased risk of death for patients in whom eradication is achieved within 1 year,30 eradication of MRSA seems beneficial and should be attempted. In the absence of randomised controlled trials to inform treatment strategies,31 guidelines recommend treatment of first isolates or return of MRSA following previous successful treatment with topical treatment and either combined oral rifampicin and fusidic acid or nebulised vancomycin or a combination of the three. Children with chronic infections may benefit from prolonged oral antibiotics until free from MRSA.21
A surveillance cough swab positive for B cepacia complex
Accurate identification of B cepacia complex is essential. Certain subtypes (eg, Burkholderia cenocepacia, Burkholderia dolosa) carry a poorer prognosis than others (eg, Burkholderia multivorans).32 There are also infection control implications; these organisms can be highly transmissible and patients harbouring them need to be strictly isolated.33 While awaiting full identification, isolation precautions are mandatory. The PPV of upper airway cultures increases with repeated testing.34 Further samples should be sent to the reference laboratory. Some will turn out to be another gram negative organism, often not present on repeated culture and of uncertain significance. If B cepacia complex is confirmed, guidelines recommend that eradication should be attempted,33 although the reported success of eradication regimens are moderate at best, with one group reporting successful treatment in only 4/14 (29%) of patients with B cepacia complex.35 A recent Cochrane review did not identify any randomised controlled trials to inform the choice of eradication strategies36 thus the choice of agents should be determined by sensitivities.33
A surveillance cough swab positive for Achromobacter xylosoxidans on two consecutive cough swabs taken 3 months apart
Pathogens not previously associated with CF are being isolated with increased frequency. It is difficult to know when to treat organisms such as Achromobacter xylosoxidans and Stenotrophomonas maltophilia. Current evidence available suggests that many of these rarer organisms may not be associated with clinical deterioration,37 thus guidelines21 do not recommend treatment unless patients are symptomatic. In this case, treatment choice should be guided by sensitivities. There is also uncertainty over non-aeruginosa pseudomonas species, which are treated aggressively in some centres and not at all in others. An exception to this is Pandoraea apista, which is sometimes misidentified as B cepacia complex,37 and has been associated with clinical deterioration.21 Guidelines recommend treatment where this is isolated.21
In a symptomatic child with CF, does a negative cough swab exclude a bacterial lower airway infection?
If a symptomatic child with CF (the patient) has a negative cough swab (the test), how confident can we be that this reflects the absence of lower airway infection (the outcome)? As discussed above, the literature is unclear. Although some studies have reported high NPVs of up to 99%,38 these are mostly for throat swabs. A study from our centre39 confirmed that a negative cough swab should not be reassuring and that further attempts should be made to obtain lower airway secretions, for example, by BAL or sputum induction. This is particularly true for certain organisms, such as non-tuberculous mycobacteria (NTM); in six patients known to have positive sputum samples for NTM, all cough swabs were negative,40 thus cough swabs do not detect NTM. Whether treatment is started depends on the clinical context. Questions that may be asked to guide decision making are outlined in box 1.
Box 1 Questions to consider with a symptomatic child with negative cough swabs
Has the cough swab been performed by an adequately trained and experienced person?
Was it processed in a laboratory experienced with cystic fibrosis samples?
Does this child have a history of previous lower airway infection which could be used to help guide treatment on the balance of probability?
Is the child so unwell that empirical treatment should be started in any case?
Does the clinical status of the child merit an invasive investigation, such as bronchoscopy?
Is there evidence of an intercurrent viral infection (for example, a positive NPA) that might predispose the child to secondary bacterial infections?
In a child with CF, who is undergoing P aeruginosa eradication on the basis of positive BAL cultures, does a negative cough swab taken at the end of treatment indicate that eradication has been successful?
This is very difficult; there is no convincing evidence. Studies in which infections have been diagnosed on BAL cultures have focussed on repeat BAL samples for confirmation of eradication.41 With early and aggressive treatment, successful eradication of P aeruginosa can be achieved, with one study reporting clearance of even mucoid P aeruginosa in 67/116 children for more than 1 year.42 However, given cough swabs have a low NPV, the cough swab result should not be relied on in isolation to confirm this. Particular caution should be taken with patients receiving inhaled antipseudomonal antibiotics for P aeruginosa eradication, as these may lead to false negative results. Supporting information should be sought including:
Were there any clinical features that could guide us, for example, symptoms improving or lung function returning to baseline? If so, and cough swabs were negative at the end of eradication, therapy is stopped with careful ongoing monitoring. Were there features to suggest chronic infection, such as the presence of a mucoid P aeruginosa? A paediatric CF specialist would be cautious and continue chronic suppressive treatment beyond 3 months. Should BAL be repeated to confirm eradication? Perhaps, but given the invasive nature of the intervention and the possibility of sampling errors, most clinicians (and parents) would be reluctant to do this; if the child was not doing well clinically, BAL or sputum induction should be considered.
In children who cannot expectorate sputum, is BAL superior to oropharyngeal cultures for monitoring changes in the respiratory flora?
A large randomised controlled trial attempted to answer this question. Surveillance was either by standard methods, such as oropharyngeal cultures, or with BAL; the latter was performed before 6 months of age, during exacerbations and following eradication for suspected P aeruginosa infection. The study found no difference in rates of P aeruginosa or any other end point between the two groups.41 Considering the risks associated with BAL, current evidence does not support the use of routine, regular BAL surveillance.
Non-culture based, molecular techniques are promising. PCR has been used for the early detection of P aeruginosa, although there have conflicting reports regarding its sensitivity and, where available, it should therefore be used as an adjunct to conventional culture.43 Sequencing of the 16S rRNA gene from the bacterial genome is perhaps the most exciting, as it has the potential to allow surveillance of entire bacterial communities, often comprising vast numbers of disparate bacteria. It remains to be determined which of these organisms are most important and whether in fact some of them may be clinically beneficial.
Non-invasive, breath-based detection systems
Bacteria produce an array of exoproducts, including volatile organic compounds. Such signals, if specific to an individual organism, could provide a valuable screening tool. Various markers have been proposed, such as 2-pentylfuran as a marker of infection with Aspergillus fumigatus and hydrogen cyanide as a marker of P aeruginosa.44 The technology and clinical utility of these tests are currently under investigation.
Clinical bottom line
Cough swabs are useful in surveying non-expectorating, asymptomatic patients, where more invasive testing would not be justified; their sensitivity and specificity is however suboptimal. A negative cough swab should not be taken as reassuring in an unwell child; induced sputum or bronchoalveolar lavage may be needed.
Eradication of Pseudomonas aeruginosa is successful in the majority of cases, provided the infection is detected promptly, so early identification is crucial.
Response to antibiotics may correlate poorly with in vitro sensitivities of cultured organisms.
Test your knowledge
A 6-year-old girl with cystic fibrosis presents with features suggestive of respiratory infection. Which of the following statements is true?
Expectorated sputum is likely to be easily obtained.
A negative cough swab rules out infection and treatment is not required.
If she is unable to expectorate, induced sputum samples may be a suitable alternative.
A cough swab shows a growth of Pseudomonas aeruginosa×10−3 for the first time. Which of the following statements is true?
If the child is asymptomatic, this is a likely contaminant and does not warrant treatment.
Only significant growths of Pseudomonas aeruginosa (of >×105) should be treated, irrespective of symptoms.
Any growth of Pseudomonas aeruginosa should be treated.
Which of the following is untrue of bronchoalveolar lavage?
A limitation is the ability to sample only small areas of the lower airway.
It is a useful sample for assessing inflammation.
It should not be performed repeatedly as a routine.
Answers to the quiz on page 186
Question 1: C.
Question 2: C.
Question 3: B.
Contributors BA, AB and JCD contributed equally to writing and editing the manuscript.
Competing interests None.
Provenance and peer review Commissioned; externally peer reviewed.
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.