Article Text

Download PDFPDF

Protracted bacterial bronchitis: reinventing an old disease
  1. Vanessa Craven,
  2. Mark L Everard
  1. Department of Respiratory Medicine, Sheffield Children's Hospital, Sheffield, UK
  1. Correspondence to Professor Mark L Everard, Department of Respiratory Medicine, Sheffield Children's Hospital, Western Bank, Sheffield S10 2TH, UK; m.l.everard{at}


Chronic cough is common in the paediatric population, yet the true prevalence of this condition remains difficult to define. Protracted bacterial bronchitis (PBB) is a disease caused by the chronic infection of the conducting airways. In many children the condition appears to be secondary to impaired mucociliary clearance that creates a niche for bacteria to become established, probably in the form of biofilms. In others, immunodeficiencies, which may be subtle, appear to be a factor. PBB causes persistent coughing and disturbed sleep, and affects exercise tolerance, causing significant levels of morbidity. PBB has remained largely unrecognised and is often misdiagnosed as asthma.

  • Respiratory
  • General Paediatrics

Statistics from

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.


Unlike ‘wheeze’, persistent cough in childhood has received very little attention in the literature over the past four decades. Reviewing the relatively sparse literature relating to chronic cough in children suggests that its prevalence (coughing every day for many weeks or months) may be in the region of 8%, although some studies quote much higher figures.1 ,2 This represents a huge unrecognised level of morbidity,3 and for parents is a source of great frustration as clinicians frequently dismiss their concerns with statements such as ‘it's just another cold’ or ‘it's his asthma cough’. It is common for children to be seen more than 20 times because of their cough before being referred.4

As well as there being relatively few studies exploring the prevalence of persistent cough among children in the community, even fewer have attempted to determine its causes. Of those referred to one tertiary centre with a cough for at least 3 weeks, 54.5% were referred with a primary diagnosis of asthma. After investigations that included a bronchoscopy, the most common primary diagnosis (40%) was protracted bacterial bronchitis (PBB), with only 2% having this diagnosis at referral. The diagnosis of asthma was rarely substantiated,5 findings that are very similar to those observed in our clinic. It should be noted that the few studies in primary care generate a different pattern of causes, including a higher prevalence of asthma (because most asthmatics are not referred to secondary care) and unrecognised pertussis infection in school age children.

Despite PBB being an important cause of chronic morbidity, it is notable that this ‘disease’ did not feature in paediatric or even respiratory paediatric textbooks until very recently. Work highlighting the importance of this condition has led to its inclusion in national and international cough guidelines6 ,7 and the recent Scottish Intercollegiate Guidelines Network (SIGN) and the British Thoracic Society (BTS) asthma guidelines, as an alternative diagnosis in a wheezy child who has a wet cough.8 Much scepticism regarding the existence of PBB arises from the concept that a chronic bacterial infection would not be able to affect the lungs, an organ that has been traditionally thought of as sterile, unless the patient has a significant underlying disease such as cystic fibrosis, primary ciliary dyskinesia or an immunodeficiency. Recent advances in our understanding of the way bacteria behave in different settings are providing an opportunity to develop new conceptual models that permit us to understand the interaction of bacterial pathogens within the normal host bacterial flora in the upper and lower airways, thus challenging these beliefs.

What is PBB?

Historically, PBB9 has been labelled in a variety of ways, including ‘chronic bronchitis of childhood’,10 nomenclature that reflected the clinical phenotype. More recently terms have been used that describe the pathological process and the site of infection, namely persistent endobronchial infection and chronic suppurative lung disease.11 In essence the condition is, as these names suggest, a protracted or persistent infection of the conducting airways by pathogenic bacteria, most commonly by non-typable Haemophilus influenzae (NTHi), Streptococcus pneumoniae and less commonly Moraxella catarrhalis. In PBB it is often found that more than one organism is identified in bronchoalveolar lavage (BAL) samples, and viral pathogens including rhinovirus, adenovirus, respiratory syncytial virus (RSV) and parainfluenza virus are also commonly detected by PCR.12 It would appear that this condition, that was probably prevalent in the pre-antibiotic era and frequently led to bronchiectasis over time, is again becoming more common, having largely been forgotten for several decades.

Over the past few years there has been a relative explosion of publications relating to this disease. Clinically the diagnosis is based on typical features suggestive of a persistent infection of the conducting airways by pathogenic bacteria, most noticeably a persistent ‘wet’ cough in the absence of signs indicative of an alternative specific cause, that resolves with appropriate antibiotics. This is analogous to the diagnosis of asthma that is based on a suggestive history and an unequivocal response to therapy.

PBB appears to be a biofilm disease

An improved understanding of bacterial behaviour has provided us with a means to explain how some bacteria are able to persist so effectively within the conducting airways and in other related sites such as the middle ear and sinuses.13 Their ability to remain in potentially hostile environments has been linked to the production of biofilms. A biofilm is a matrix secreted by some bacteria that is thought to enhance attachment, facilitate access to nutrients and decrease antibiotic penetration, thus providing substantial protection against antibiotics and making the bacteria difficult to eradicate with standard length courses of antibiotics.14 In those with cystic fibrosis, the eradication of mucoid Pseudomonas aeruginosa is extraordinarily difficult once biofilms are established,15 providing the rationale for the early, aggressive use of antibiotics in order to eliminate it from the airways before it becomes part of a biofilm structure. Biofilm production by NTHi has been seen in patients with cystic fibrosis16 as well as in those with chronic otitis media17 and chronic obstructive pulmonary disease.18

It is important to recognise that there are two compartments within the lungs—the respiratory spaces (alveoli and respiratory bronchioles) and the conducting airways—each having very different structures and functions. The respiratory compartment has a huge cross-sectional area, with the thin wall of an alveolar epithelial cell and a capillary endothelial cell being all that separates pulmonary capillaries and the alveolar space. The conducting airways are much more complex structures, with multiple branches providing the structure for the large surface area, enabling gaseous exchange more peripherally while also protecting the fragile respiratory zone from pathogens and particulates.

Organisms behave quite differently in the two regions. S pneumoniae and H influenzae can cause acute life threatening pneumonias when they rapidly divide in the respiratory compartment of the lung. In the majority of cases, response to antibiotics is prompt and dramatic. S pneumoniae and NTHi can also cause an acute bronchitis in the conducting compartment; if not cleared, the bacteria appear to form biofilms in which they replicate slowly and are able to persist for prolonged periods, being difficult to eradicate. This is attributable to their slow replication, together with the structural and biological adaptations of organisms protected within biofilms.

What is the role of the host–pathogen interaction in PBB?

It is now recognised that we encounter NTHi, S pneumoniae and M catarrhalis frequently and regularly. Prospective studies show a high prevalence of these organisms ‘colonising’ the upper airway in infancy and childhood, with them being detected in some from as early as the neonatal period.19 ,20

It now appears that the ‘normal’ lower airways are not sterile, but also have their own resident microbiota consisting of large numbers of organisms, the majority of which are not able to be cultured using traditional microbiological techniques designed to isolate known pathogens.21 Establishing a presence, even transiently, involves potential pathogens interacting with the host's microbiota as well as its physical and immunological defences. The organisms use different strategies to deal with these various challenges. An example of the complexity of these interactions is studies in vitro suggesting that S pneumoniae will out compete NTHi using, for example, hydrogen peroxide to displace it.22 Conversely, in vivo, the neutrophilic inflammation induced by NTHi appears to confer an advantage over S pneumoniae. Both organisms appear to provide some protection against colonisation with Staphlococcus aureus.23 Furthermore, external factors such as antibiotic therapy and smoking can disrupt the normal flora of the upper and lower airways and may predispose to the emergence of potential pathogens, a process well described in ventilator acquired pneumonia.24

The importance of impaired mucociliary clearance

Stasis in any system enables pathogens to thrive, and the most common reason for these organisms to become established in the lower airway appears to be the opportunity provided by impaired mucociliary clearance. This might be due to specific diseases such as cystic fibrosis and primary ciliary dyskinesia, or airway problems such as tracheobronchomalacia.25 In asthma, particularly poorly controlled asthma, there may be poor mucociliary clearance and epithelial damage resulting from inflammation and mucus plugging. It appears that the loss of cilia and delayed recovery of normal function occurring with viral lower respiratory tract infections in early childhood may be a common predisposing factor permitting potential pathogens to exploit this opportunity.26

What are the clinical features of PBB?

Typically children with PBB are young, the majority of studies describing children less than 6 years old,5 ,27 although it can commence at any time including into adulthood. Without carefully exploring the reported symptoms, the initial history given by parents can appear to be typical of asthma, with night-time coughing, shortness of breath with exercise, ‘wheezing’ and exacerbations with upper respiratory tract infections being typical.

Affected children characteristically have a persistent ‘wet’ cough as opposed to the ‘dry’ nocturnal cough of asthma, although parents often find it difficult to make this distinction; the question ‘does he sound like a 60 a day smoker first thing in the morning?’ is often more helpful. The cough is typically persistent, with the question ‘when did he last not have a cough?’ being useful to distinguish a frequent intermittent cough from a persistent cough. Again the cough is typically worse when changing posture, just after lying down in bed and first thing in the morning, but can be present all night as occurs in asthma.

Parents often describe their child becoming short of breath and coughing with exercise. Children with PBB typically cough so much that they appear to be gasping for breath, while the shortness of breath observed in asthmatics is not directly related to coughing and hence a simple report such as ‘he gets short of breath and coughs with exercise’ needs to be explored in more detail. It is also common to report that a child has a ‘wheeze’, although in one study, on closer questioning, this almost always represented a ruttle (a coarse, non-musical noise generated in the larger airways as a result of secretions that can be felt on the chest).27

A viral infection will exacerbate both asthma and PBB. In PBB this is likely to be due to the release of planktonic forms of bacteria from biofilms, presumably in response to the associated inflammation which may help disseminate the organism as has been seen in cystic fibrosis.28 As with exacerbations of asthma, children with PBB who have a viral respiratory tract infection, and indeed those with symptoms due to other causes such as pertussis, will experience increased symptoms for a period and then improve as they ‘regress to the mean’. In these cases almost any intervention may be perceived to have been helpful, hence a very critical assessment of the response to treatment at the appropriate time is an essential part of the diagnostic process; one cannot simply provide 2 weeks of antibiotics and see the patient again several weeks later. In an area where response to treatment is a large component of the diagnostic process, this phenomenon can, and frequently does, lead to misdiagnoses. Unequivocal improvement following the introduction of a treatment such as an inhaled corticosteroid for a child with probable asthma or antibiotics for PBB is necessary to help confirm a presumptive diagnosis. One looks for a dramatic improvement that transforms the child's health; ‘he is a new child’, not simply ‘he is coughing less’.

Children with PBB generally do not look unwell, yet the morbidity resulting from disturbed sleep for both the child and their parents and the impact on their general well being is frequently significant if the right history is sought.29 When the condition is treated effectively, parents do report that the cough resolves for the first time, but more importantly for them, they generally say that they have ‘a new child’ because of the impact on their general health and activity.

Anecdotally PBB has become far more prevalent over the past 15 years, a period that has seen a progressive fall in antibiotic use.30 In primary care, treating suspected viral respiratory tract infections with antibiotics has been discouraged, but perhaps the caveat that a persistent cough may need treatment has been forgotten. Quite when antibiotics should be given is unclear, but a persistent wet cough (in contrast to the dry ‘post-viral’ or ‘non-specific’ cough frequently seen in primary care) that is not improving after 2–3 weeks might be a reasonable compromise. At this stage, a traditional 5–7-day course is probably effective; when later, such courses may have little impact. Parents often report that antibiotics have not helped, which is interpreted as making PBB unlikely, but on closer questioning it may be that the cough was improving, with symptoms worsening quickly when the antibiotics were stopped.

The role of defects in the immunological aspects of the host response remains unclear. In the majority of patients there are no identifiable major immunodeficiencies, but relatively minor or transient development defects may play a role in some.31

When and how to diagnose PBB?

An increasing number of clinical pathways and investigation guidelines for chronic cough in childhood exist.6 ,29 ,32 ,33 Chest radiographs are often normal or may have only minor abnormalities such as peribronchial wall thickening, findings consistent with other conditions such as viral lower respiratory tract infections, poorly controlled asthma and recurrent minor aspirations. Hyperinflation is uncommon and should again raise the question of asthma alone or the co-existence of PBB with asthma.29 Cough swabs can be useful but have a relatively low sensitivity.27 The definitive investigation, although invasive, is fibreoptic bronchoscopy with BAL, although care is needed when interpreting results. Timing of this is dependent on discussions with parents, and generally we reserve this for those who continue to relapse after three courses of antibiotics, but some parents prefer to have a definitive diagnosis at the outset.

Identifying a significant cause such as mild variant cystic fibrosis or an immunodeficiency is uncommon, thus we generally treat patients with two or three courses of antibiotics before proceeding to bronchoscopy and undertaking immunological investigations. In cases where there are concerning features or parental desire for a ‘definitive’ test, we may proceed directly to these, but this is influenced by the severity of symptoms.

Typically we find secretions and oedematous collapsible bronchi that collapse during suctioning while undertaking a BAL. Antibiotic usage within a month often results in a negative culture, even in a child with significant symptoms, hence the timing of bronchoscopy in relation to recent antibiotic courses is important although positive cultures can be obtained despite recent antibiotic use.34

How is PBB treated?

This is largely an evidence-free zone. A meta-analysis studying the treatment of ‘prolonged’ cough (greater than 10 days) in the absence of a diagnosis or an isolated infective cause, found only two small, randomised controlled trials that included a placebo or a control group. They both studied children under 7 years old and neither study was deemed of high quality. Their conclusion was that treatment with antibiotics was beneficial, with one clinical cure for every three children treated. The review supported the use of antibiotics in children with a prolonged wet cough, although called for further randomised controlled trials using a better study design and validated outcome measures.35 A more recent double blind study involving 50 children found that 2 weeks of co-amoxiclav, led to resolution of cough in 48% of children with a ‘wet’ cough lasting more than 3 weeks compared with 16% of those receiving placebo.36 None of the studies undertook longer follow-up.

The aim of treatment is to eradicate bacteria and to allow regeneration of the epithelium in the absence of infection. Two weeks of high dose antibiotics such as co-amoxiclav (40 mg/kg/day amoxicillin content) would normally lead to resolution of the cough and a dramatic improvement in the child's quality of life, however recurrence of symptoms is described, and repeated courses of antibiotics may be necessary to achieve clinical improvement. In a recent study, more than 70% of those treated with antibiotics for 2 weeks or longer relapsed.25 As the epithelium is likely to be damaged and impaired mucociliary clearance will persist beyond clearance of organisms, we empirically use prolonged courses of 6–8 weeks, but the optimal treatment remains to be determined. By definition the cough must resolve with a course of appropriate antibiotics; usually the effect of treatment is dramatic, with the child's family reporting a significant improvement in their symptoms. Shorter courses of antibiotics tend to result in a partial resolution of the cough or a cough that will relapse after a few days off treatment, differing from community acquired pneumonias for which 5–7 days’ treatment is usually sufficient. Donnelly found over half (51%) of their cohort were completely symptom-free after two courses of antibiotics, with only 13% requiring six or more courses or needing continuous prophylactic antibiotics for at least one winter.27

Unfortunately we do not know the best therapeutic approach, and there is a need to review the balance of the undoubted benefits to the individual in terms of quality, with potential problems resulting from development of antibiotic resistance, although anecdotally we observe a very low instance of apparent resistance. Inevitably prevention is always going to be better than cure, so early intervention with short courses of antibiotics would appear desirable, but is difficult to implement in the current climate without good diagnostic techniques. Whether vaccines against NTHi will ever prove to be effective is unclear. The use of pneumococcal conjugate vaccines has not reduced the incidence of this condition. In our cohort non-vaccine serotypes are predominantly isolated, although this does appear to be in part attributable to serotype replacement following the introduction of conjugate pneumococcal vaccines.37 Whether the less virulent non-vaccine serotypes are better suited to a biofilm existence is unclear.

What is the natural history of PBB?

The ‘vicious circle hypothesis’ is generally accepted as the most plausible explanation for the development of bronchiectasis in most patients.38 Implicit in this hypothesis is the development of impaired mucociliary clearance and chronic endobronchial infection, leading to inflammation and damage to the wall of the conducting airways that is eventually evident on high resolution computerised tomograph (HRCT) scans. Attempts to break this circle and permit full recovery of the epithelium include antibiotics to treat infection, physiotherapy to improve clearance, and in the case of cystic fibrosis, anti-inflammatory agents such as macrolides, DNAse and osmotic agents to help restore mucociliary clearance. The treatment of many viral upper respiratory tract infections with antibiotics possibly prevented the development of PBB and thus bronchiectasis in the years when cases appeared less, although this may also be secondary to an under-appreciation of the problem.

The progression to bronchiectasis may take many years or even decades. The progression from a ‘wet’ cough, to producing purulent sputum without radiological evidence of bronchiectasis, to established bronchiectasis was well described in the pre-antibiotic days when the term ‘pre-bronchiectasis’ was used.39 The rate of deterioration is thought to be determined by the extent and type of infection, the frequency of exacerbations, the nature of any underlying conditions, and the use, efficacy and adherence to medical therapy.

What is the danger of recognising this new/old diagnosis?

PBB is a real and significant cause of morbidity. Its prevalence in the community is unclear, but is likely to be growing with changes in antibiotic usage and possibly as a result of the H influenzae type B and pneumococcal conjugate vaccines.37 PBB accounts for nearly half of the referrals to our ‘general’ respiratory clinics at present, the majority of asthma being diagnosed and managed in primary care.

Despite this, it must not be forgotten that asthma is the commonest cause of significant respiratory symptoms in childhood and we must guard against returning to mistreating asthma with antibiotics. For those who are new to the area and who see the dramatic impact that antibiotics can have on the lives of children with PBB, there is a danger that we misdiagnose mild asthmatics as having PBB and begin to over-use antibiotics again to treat exacerbations of asthma. It is also important to recognise that asthma per se predisposes to PBB, particularly if poorly controlled, and hence in a patient with definite asthma who has a persistent wet cough, antibiotics may have an important role but they should not be used to treat exacerbations of asthma mislabelled as ‘chest infections’. Paediatric respiratory medicine continues to rely on accurate history taking and assessing the response to therapies. Asthma, recurrent wheezy bronchitis (wheezing due to a viral bronchitis without significant bronchospasm), frequent viral respiratory tract infections and PBB share many clinical features, and hence accurate assessment in the absence of diagnostic tests remains difficult. Many children and their families continue to experience unnecessary morbidity due to medical models of respiratory disease that do not embrace the complexities of the host's interactions with microorganisms, pollutants and allergens.


In a child in whom a persistent ‘wet’ cough is a prominent feature, the diagnosis of a persistent bacterial bronchitis should be considered, either alone or as a co-morbidity with asthma (antibiotics should not be used to treat exacerbations of asthma but may have a role in an asthmatic with a persistent wet cough). In recognising the existence of this condition, it is important that we do not go full circle and treat exacerbations of asthma with antibiotics; careful history taking and a robust approach to reviewing the response to therapy at the appropriate time are key components of the diagnostic process for asthma and alternative diagnoses. If the response is not as anticipated, the diagnosis must be revisited in order to try to minimise the morbidity associated with misdiagnosis, but if improvement is seen, the result is often a dramatic improvement in the child's and their parents’ quality of life.



  • Contributors The authors have contributed to this review as follows: VC: drafting of the article, changing it in light of review and referencing. ME: original concept, review of the first draft and final approval of the version to be published.

  • Competing interests None.

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