The central pathway for cough is a brainstem reflex linked to control of breathing (the central respiratory pattern generator) 3, which undergoes a maturation process such that the reference values for normal respiratory rate in children are different to those in adults 4 and reaches adult values in adolescence. In early life, cough is related to primitive reflexes (laryngeal chemoreflex), that undergo maturation resulting in significant differences in swallowing between young children and adults 5. Plasticity (modulation) of the cough reflex has been shown 36, although it is unknown if the young have greater plasticity (propensity to modulate or change). Like other organs directly relevant to cough (eg the systemic and mucosal immune system) 78 or not directly related to cough (eg the renal system), one can speculate that the cough reflex has maturational differences as well. Indeed children differ from adults in some immunological response to lipopolysaccharides 9. Also, children, especially their neurological system, are more sensitive than adults to certain environmental exposures 10. For example, in children, the utility of CT scans has to be balanced with the reported increased lifetime cancer mortality risk, which is age and dose dependent. Although the risk is relatively negligible, children have 10 times increased risk compared to middle aged adults 10. Lastly, the distinct differences in respiratory physiology and neuro-physiology between young children and adults include maturational differences in airway, respiratory muscle and chest wall structure, sleep characteristics, respiratory reflexes and respiratory control 111213.

Cough can be cortically modulated 14. In adults, chronic cough is associated with anxiety as an independent factor 15; such data are unavailable in children. Adults seeking medical attention are primarily self-driven but in children, parental and professional expectations influence consulting rates and prescription of medications 161718. Reporting of childhood respiratory symptoms is biased and parental perception of childhood cough plays an important role 1920. In asthma, parental psychosocial factors (in particular anxiety) were strongest predictors for emergency attendances for children whereas in adults, asthma severity factors were the risk factors 21. In cough, use of cough medications and presentation to doctors were less likely in children with higher educated mothers 22. Hutton and colleagues' described "parents who wanted medicine at the initial visit reported more improvement at follow-up, regardless of whether the child received drug, placebo, or no treatment" 23. Rietveld and colleagues showed that children were more likely to cough under certain psychological settings 2425.

'Normal' children occasionally cough as described by two studies that objectively measured cough frequency 2627. Normal children without a preceding upper respiratory infection in the last 4 weeks have up to 34 cough epochs per 24 hours 26. In another study, 0–141 cough epochs/24 hours (median 10) were recorded in 'controls' (these children were considered well by parents and attending school and were age, gender and season matched 27). Medicalisation of an otherwise common symptom can foster exaggerated anxiety about perceived disease and lead to unnecessary medical products and service 28. Cough in this situation is termed 'expected cough'. Such data are unavailable in adults.

However, concerns of parents presenting to general practitioners for their children's cough can be extreme (fear of child dying, chest damage) 2930. Other parental concerns were disturbed sleep and relief of discomfort 29. However the burden of illness on children and their family has not been well described. In contrast adult data have shown that chronic cough causes a significant burden of illness (physical and psychosocial) that is often not appreciated by physicians 20 as reflected in adult cough-QOL scores 3132.

The utility of definitions depends on the intention of use. In adults, chronic cough is defined as cough lasting >8 weeks 33. In children the definition of chronic cough varies from 3-weeks duration 34 to 12-weeks 3536. There are no studies that have clearly defined when cough should be defined chronic or persistent. As studies have shown that cough related to ARIs resolves within 1 to 3 weeks in most children 1737 it would be logical to define chronic cough as daily cough lasting >4 weeks.

Paediatric cough can be classified in several ways, based on aetiology 38, timeframe 35 and characteristic (moist vs dry). For practical reasons, guidelines based on cough duration, combined with cough quality have been developed 35. An evidence based guideline specific for paediatrics will be published as part of the American College of Chest Physicians' Guidelines on the Management of Cough in Adults and Children 39. The previous guidelines which stated that "the approach to managing cough in children is similar to the approach in adults" 34 was arguably inaccurate.

Unlike cough in adults, paediatric cough has also been classified into specific and non-specific cough (with an overlap) for practical reasons (figure 1). Indeed, the most common paradigm encountered in clinical paediatrics when cough is a presenting feature is the differentiation between specific and non-specific cough. Specific cough refers to cough in the presence of pointers (table 1) that suggest the presence of an underlying aetiology. A thorough history and examination to elucidate these points are necessary when assessing children with cough and in the majority of situations, specific cough aetiologies can be defined. While some of these symptoms and signs are common in adults (such as haemoptysis), others are not (such as failure to thrive). Unlike in adults, where cough characteristics has been shown to be of little diagnostic value 40, paediatricians often recognise certain cough qualities such as staccato cough (table 2). A chronic moist cough is always abnormal and represents excessive airway secretions 41. However in a small group of children natural resolution may occur 42 and a specific paediatric diagnostic category may not be found 43. A chronic dry cough however may represent a dry phase of an otherwise usually moist cough or airway secretions too little to influence the cough quality 41. Chronic dry cough in the absence of specific pointers (table 1) in the history and examination is termed 'non-specific cough' or 'isolated cough', ie cough is the sole symptom. In non-specific cough, the aetiology is ill defined and we suspect that the majority are related to post viral cough and/or increased cough receptor sensitivity 4445. However in the majority of children, it is most likely related to a non serious aetiology 38 or may spontaneously resolve as evidenced in the placebo arms of RCTs 464748 and cohort studies 495051. Thus if one assumes that the natural resolution of non-specific cough occurs in 50% of children, 85 children per study arm is required in a randomised controlled trial to detect a 50% difference between active and placebo groups, for a study powered at 90% at the 5% significance level.

Children with specific cough usually require a variety of investigations which include chest CT, bronchoscopy, barium meal, video fluoroscopy, nuclear scans, sweat test, etc. The role of these tests for evaluation of lung disease is beyond the scope of this article as it would encompass the entire spectrum of paediatric respiratory illness. The more common problem of non-specific cough is further briefly discussed. In general investigations are rarely needed in non-specific cough.

Cough severity indices, broadly divided into subjective and objective outcomes, measure different aspects of cough. In children, measures of CRS have a weak relationship with cough frequency. Subjective cough scores have a stronger and consistent relationship with cough frequency 107. The choice of indices depends on the reason for performing the measurement 107.

Answers to questions on isolated cough are largely poorly reproducible 108 and nocturnal cough in children is unreliably reported 5253. The kappa value relating the chance-corrected agreement to questions on isolated cough is poor (0.02–0.57) 19108109 in contrast to isolated wheeze (0.7–1.0) 108. Biased reporting of cough has been shown; parents who smoke under-report cough in their children 19. Diary cards for cough have been validated against an objective method and children aged >6 years are better than their parents at quantifying their cough severity 54. Cough-specific QOL questionnaires exist for adults but not for children. There is a clear need for a paediatric cough specific QOL scores, as adult QOL scores cannot be applied to children. Cough specific objective tests include ambulatory and non-ambulatory objective cough meters, CRS and cough peak flows (reviewed elsewhere) 110. Adult type instruments require modification for use in children 111.

Although some diseases are common to both adults and children, the pattern of many respiratory illnesses in children is clearly different to adults; eg viruses associated with the common cold in adults can cause serious respiratory illnesses such as bronchiolitis and croup in previously well young children 112. Both of these respiratory syndromes are non existent in adults. Conversely, common causes of cough and respiratory diseases in adults such as chronic bronchitis 113 and chronic obstructive pulmonary disease are not recognised diagnostic entities in paediatric respiratory literature and main textbooks 114115. The following highlights some of the differences between children and adults.

Providing parents with information on the expected time length of resolution of acute respiratory infections may reduce anxiety in parents and the need for medication use and additional consultation 120. Appreciation of specific concerns and anxieties, and an understanding of why they present are thus important when consulting children with non-specific cough. Educational input is best done with consultation about the child's specific condition 213. A RCT 214 examining the effect of a pamphlet and a videotape promoting the judicious use of antibiotics, found that their simple educational effort was successful in modifying parental attitudes about the judicious use of antibiotics.

In contrast to adults where OTC medications, in particular codeine and its derivatives have been shown to be useful, systemic reviews for children have concluded that cough OTCs have little, if any, benefit in the symptomatic control of cough in children 215216. Moreover OTCs have significant morbidity and mortality 217218 and are common unintentional ingestions in children aged <5 years 219. Use of diphenhydramine is also non beneficial for symptomatic treatment of cough related to pertussis 220. The use of steam inhalation, vitamin C, zinc, and echinacea for upper RTI has been summarised 221 with little benefit, if any, for symptomatic relief of cough for adults and children.

The efficacy of anti-histamines in relieving cough in children is minimal, if at all 222. Thus, unlike adults, the use of anti-histamine therapy for chronic cough in children is mostly unjustified. Whether this difference between children and adults is related to atopic states is unknown. A RCT on ketotifen did not show any clinical benefit in the treatment of 113 infants and children with chronic cough and/or wheeze 223. A systematic review of anti-histamine and nasal decongestion combinations, and anti-histamines in OTC medications has shown that these pharmaceuticals were no more likely than placebo in reducing acute cough in children 215. The use of these medications that contain H₁ receptor antagonist has to be balanced with adverse events 217224225 which includes reported death from toxicity in young children 217218. The latest published RCT (n = 100) also showed that diphenhydramine (a first generation H₁-antagonist) and dextromethorphan were no different to placebo in reducing nocturnal cough or sleep disturbance in both the children and parent(s) 225. Like other RCTs there was a significant improvement in both placebo and active arms for the cough outcomes measured 225.

Old cohort studies describing that asthma therapy for that era (oral orciprenaline, salbutamol syrup 226227 theophylline 97227 and metaproterenol with theophylline) 88 was useful in abolishing cough included children with clinically recognisable asthma. For example, 8 of 11 children in Konig's study had cough with co-existant chest pain or dyspnea on exertion) 88 10 of 32 children in another study had abnormal examination findings 89.

In ambulatory children with acute cough (1–10 days) with no history of asthma and a normal chest examination, oral albuterol was not effective in reducing cough frequency or duration 48. In a meta-analysis, Smucy et al likewise concluded that "there is no evidence to support using beta2-agonists in children with acute cough and no evidence of airflow obstruction" 228. There is only one study on use of inhaled salbutamol in chronic cough (median of 8-weeks) which also showed no benefit 47. There is no evidence for the use of anti-cholinergics for in children with non-specific cough 229. Use of bronchodilators must be weighed against adverse events (eg tremor, irritability 48 behaviour change, cost).

Only 2 RCTs on ICS for chronic non-specific cough in children have been published and both groups have cautioned against prolonged use of ICS 4647. There is no RCT on oral steroids for non-specific cough in children. In cough associated with pertussis, dexamethasone provides no significant benefit for the symptomatic relief of cough 220. Even in children with wheeze, a RCT found that oral steroids may confer no benefit 230. In contrast to high doses used in adults, low dose ICS has been shown to be effective in the management of the majority of childhood asthma 231232233 and there were reported significant adverse events on high doses 234235. Thus if a trial of asthma therapy is ever warranted, use of a moderate dose (400 mcg/day equivalent of budesonide) is suggested. This practice is however discouraged in most settings. As the earlier studies in adults and children that utilised medications for asthma for the era reported that cough related to asthma completely resolved by 2–7 days 888997236 it is recommended that reassessment is done in 2–3 weeks. Cough unresponsive to ICS should not be treated with increased doses of ICS. Cough that resolves with ICS use may be related to the period effect (spontaneous resolution) 211 or a transient effect responsive to ICS use (ICS may also impact on non-asthmatic airways with pulmonary toxicants 237). Thus clinicians should be cognisant that the child that appears to respond to ICS does not necessarily have asthma and the child should be re-evaluated off asthma treatment.

Cromoglycate and nedocromil reduces cough associated with asthma 238239 and in children born prematurely 240. An open, single arm trial reported significant reduction in cough scores from 30 to 15/week after 2-weeks of treatment with nedocromil (4 mg qid) with no additional benefit in subsequent 4-weeks 241. There are no published RCTs 242. Leukotriene receptor antagonists have been examined in adults for cough 243 but there is no RCT data in children. Theophylline utilized in old studies 97227 may have an effect on cough separate from its 'anti-asthmatic' properties but there are no RCTs in children 244 and theophylline has a narrow therapeutic range. Oral theophylline, but not placebo, induced complete remission in adults with ACE inhibitor related cough 245. There is a need for RCTs examining the effectiveness of theophylline for non-specific cough in children.

The American Academy of Family Physician's guidelines discourages use of antibiotics except when rhinosinusitis and cough are present and not improving after 10 days 246. Meta-analysis on anti-microbials for acute bronchitis (recent onset of productive cough without chronic obstructive pulmonary disease, sinusitis or pneumonia) in older children (aged >8 years) and adults showed a small benefit of 0.58 days but with significantly more adverse events 221. In subacute cough, two paediatric RCT have shown that anti-microbials (amoxycillin/clavulanic acid 247 and erythromycin 248) were more likely to achieve 'clinical cure' and also prevented progression of illness defined by need for antibiotics 249. The quality of cough in both studies was not clearly defined but the secretions in both studies cultured M catarrhalis 247248.

In the management of any child with cough irrespective of the aetiology, attention to exacerbation factors is encouraged. A single report was found on cessation of parental smoking as a successful form of therapy for the children's cough 250. Behavioural counselling for smoking mothers has been shown to reduce young children's ETS exposure in both reported and objective measures of ETS 251.