Printed on: December 22nd, 2024

Please note, once printed or downloaded, articles cannot be updated. By bookmarking the article and continuing to access it online, you can be sure you are reading the most up-to-date information and that you are not in breach of copyright laws.

Acute cough in children

Acute cough in children

A clinical review in the BMJ considered the very common problem of assessing childhood cough (BMJ 2012;344:e1177). We see this frequently and it is a source of distress and uncertainty to children and parents. The review is divided into two sections:

• Acute cough = cough less than 3 weeks in duration.
• Chronic cough = cough greater than 3 weeks in duration.

It focuses on the conundrum of diagnosis, rather than the treatment of all causes of cough.

If you are seeing a child with a chronic cough, you may find the Chronic cough in children article helpful. You may also find the separate articles on Pneumonia in children, Bronchiolitis and Croup of interest. Clearly, the BMJ review was written before COVID-19 so does not consider the need for PCR testing in a symptomatic child.

This article was updated in May 2024.

The vast majority of children seen in general practice will have a self-limiting viral upper respiratory tract infection. However, if we don't think of the other stuff, we will miss it!

Assessment involves a careful history and examination:

• Are there typical features of an URTI (coryza, sneezing, temperature)?
• Was the cough sudden in onset?
• Is there breathlessness?
• Are there features of atopy or a strong family history?
• Is there a background of an unwell child or faltering growth?
• Take pulse rate, respiratory rate, temperature and oxygen saturations as part of assessment (do you have a saturation monitor for paediatric use?).
• Listen for focal chest signs.

This flowchart offers an approach to diagnosis, though the division between URTI and LRTI may be oversimplified as you will see when you read on (BMJ 2012;344:e1177):

No! Supportive treatment, time and safety-netting!

• Antibiotics do not alter the clinical course.
• There is no good evidence for OTC cough remedies, and there may be harms, resulting in many being withdrawn for under-6-year-olds (Cochrane 2008:CD001831).

We are familiar with tools such as FeverPAIN when assessing sore throat.

STARWAVe was a similar model proposed for assessment of acute cough in children. It included criteria such as short illness duration, high temperature, age <2y, respiratory signs such as wheeze, increased respiratory effort, known asthma and vomiting. Internal validation was good (Lancet Resp Med 2016;4:902).

Subsequent external validation found it was less successful than a simple 3-item checklist when it came to identifying the most unwell children. A cohort study in the BJGP used respiratory rate, oxygen saturation and the presence of sputum or rattly chest to risk stratify (BJGP 2023;73:e885). A robust, fully validated scoring system hasn’t materialised yet so we’ll continue to use our clinical judgement for now!

In the UK, at least 40% of children classed as having a ‘chest infection’ are prescribed antibiotics. This translates to 2 million prescriptions per year and costs the NHS about £30 million, as well as impacting on medicalisation of self-limiting illness and reconsultation (BJGP 2018;68(675):e682-e693).

A UK population-based cohort study tried to assess whether antibiotics given for URTI-associated cough in children reduced the risk of subsequent admission or deterioration. However, despite starting with 8000 children, only 65 actually needed later admission, numbers too small to draw any meaningful conclusions from (except that children with URTI-associated cough rarely get admitted!) (BJGP 2018;68(675):e682-e693).

The ARTIC PC study was a small, well-designed, UK primary care-based, double-blind RCT (Lancet 2021;398:1417).

Just over 400 children aged 6 months to 12 years with an uncomplicated acute ‘lower respiratory tract infection’, where acute cough was the dominant symptom but there was no suspicion of pneumonia (so, children we might code as chest infection or bronchitis), were randomised to receive either placebo or amoxicillin for 7 days.

Children with other symptoms localising the infection to the lower respiratory tract, e.g. shortness of breath, sputum and pain, were included in the trial.

Exclusions: immunocompromised children, those who had symptoms for >21 days and those who had had an antibiotic prescription in the past 30 days were excluded. Children for whom the clinician judged pneumonia was likely or who were judged ‘severely ill’ were also excluded.

It found no statistically significant difference in:

• The duration of ‘moderately bad or worse’ symptoms (5 days in antibiotic group vs. 6 days in placebo group).
• Median duration of symptoms until rated absent or very little problem.
• Complications and need for assessment in, or admission to, hospital (which was rare, occurring in about 2% of both groups – though it is important to note that the study was not powered to be definitive about this).
• No differences in outcomes for pre-specified subgroups, including those with chest signs (though, unfortunately, this study was not adequately powered to be definitive about this because of early termination due to COVID-19), or in groups identified in post-hoc analysis, e.g. lower oxygen saturations.

There was a small statistically significant difference in favour of antibiotic prescribing in self-reported symptom severity between days 2–4, which is of questionable clinical significance as it equates to less than 1 in 3 children rating their symptoms as a ‘slight problem’ vs. ‘very little problem!’

A subsequent prospective cohort study looked at the same question. This used the cohort of children from the ARTIC PC trial and an associated wider observational group who were not included in the placebo-controlled arm of the trial, but where the data was collected (BJGP 2023;73(728):e156). Even in this wider group, similar results were shown:

• A non-significant reduction in symptom duration of 1 day for symptoms rated ‘moderately bad or worse’.
• No difference in hospitalisation rates.

This study is highly relevant and applicable to our patients in UK general practice. It is the largest of its kind that includes young children.

There is no clinically significant benefit in prescribing antibiotics to children presenting in primary care with uncomplicated LRTI (i.e. where we don't clinically judge the child to have pneumonia or be seriously ill). There is also no additional risk of complications or admission where antibiotics are not prescribed. The authors conclude:

“Unless pneumonia is suspected, clinicians should provide safety-netting advice but not prescribe antibiotics for most children presenting with chest infections.”

In some children, we will suspect pneumonia and we will prescribe antibiotics – but at what dose and for how long?

A further study looked at the dose and duration of oral amoxicillin for children diagnosed with community-acquired pneumonia and seen but discharged in the emergency department, observational unit or from an inpatient ward (in under 48 hours). So, this study is helpful, but not directly applicable to our patients (JAMA 2021; 326(17):1713,  BMJ 2022;378:o1767, NIHR (National Institute for Health and Care Research) alert).

It randomised more than 800 children who were felt to need antibiotics in a 2x2 design to either low-dose (35–50mg/kg) or high-dose (70–90mg/kg) oral amoxicillin given for either 3 or 7 days.

In summary, it found that low-dose was non-inferior to high-dose, and that 3 days was non-inferior to 7 days with respect to later need for antibiotic retreatment within 28 days.

So, this raises the question, what makes us judge that a child is ‘likely to have pneumonia’?

The British Thoracic Society produced guidance on community-acquired pneumonia (Thorax 2011;66:ii1-ii23). It states that we should consider it in children with:

• Parent-reported cough in the past 96 hours AND
• Temperature ≥38°C AND
• Signs of laboured/difficult breathing or focal chest signs.

This issue was discussed as part of the BMJ Uncertainties series. It reviewed all the evidence of what features were most useful in primary care at predicting which children with acute cough are more likely to have pneumonia, and may therefore benefit from antibiotics or admission (BMJ 2012;345:e6212).

The two most useful features to rule out pneumonia in a primary care setting were:

• Absence of difficult or laboured breathing.
• Absence of GP’s subjective assessment that the child is unwell.

No signs or symptoms were sufficiently useful or validated in a primary care setting to devise a clinical prediction rule to rule in pneumonia.

Based on emergency department studies, the authors suggest that GPs, particularly in the out-of-hours setting, should consider the presence of an unwell child with fever, chest signs and raised respiratory rate as suggestive of pneumonia (hmmm…think most of us might have spotted that one!).

Raised respiratory rate is consistently the best predictor of serious lower respiratory tract infection in children. However, it has been accepted wisdom that raised temperature raises respiratory rate. To make best use of paediatric vital signs, we need cut-off values which take account of a child's temperature.

This prospective study based in emergency departments in Holland attempted to look at this issue. It studied 1500 children presenting to the emergency department and paediatric assessment unit – so a secondary care population (BMJ 2012;345:e4224).

First, it determined respiratory rate, temperature, age and presence of LRTI, and used this to produce the centile charts. It then used this prospectively to see how useful it was to predict which children have a LRTI:

• Respiratory rate increased by 2.2 for each 1°C increase in temperature.
• It found that using the 90th and 97th percentile cut-off for respiratory rate according to temperature was more effective at accurately identifying children with LRTI than the established thresholds, e.g. Advanced Paediatric Life Support.
• The 50th centile was not sensitive or specific enough to rule out LRTI.
• It was not good at differentiating between viral, bacterial and atypical infections.

• This is a useful reminder of the importance of measuring and documenting vital signs in children.
• However, these values are based on a secondary care population (i.e. those children already referred) so is not generalisable to our patients.
• The authors suggest it should be validated in a primary care population before we use it.

In this UK primary care-based study, parents of more than 8000 children presenting with acute cough or respiratory tract infection were asked about their perceptions of how unwell their child was, and what signs they used to assess this. This was then compared with the signs clinicians used to make their assessments of how unwell a child might be (BJGP 2019;69:680).

In general, and perhaps unsurprisingly, parents tended to perceive their child as being more unwell than their clinician did.

• Parents used different signs from clinicians to assess severity.
• Some symptoms that parents felt were very important were not used by clinical teams as part of their assessment (severity of cough, disturbed sleep).
• Parents tended to underestimate the importance of drowsiness/irritability in their assessment when compared with clinicians.
• However, there were important areas of agreement around symptoms such as high fever and rapid breathing which both groups thought were important.

This study highlights the differences in perception between patients and clinicians. We can use presentations in minor illness as an opportunity to educate families about which signs are more useful when assessing illness severity in the future.

Acute cough in children Acute cough lasts <3–4w. It is most commonly caused by self-limiting URTI and there is no evidence that any treatments are beneficial in managing the cough. Tachypnoea, increased work of breathing, fever and focal chest signs are suggestive of pneumonia. Consider an inhaled foreign body if the cough is sudden in onset without other features. There is no evidence that inhalers, cough mixtures or antibiotics improve the prognosis of a viral URTI cough. In a child with a LRTI where we don’t suspect pneumonia and they are not severely unwell, we should safety-net and not prescribe antibiotics.