A 7-year-old boy is admitted to the general paediatric ward with bacterial community acquired pneumonia affecting the right lower lobe. It is suggested on the ward round that we arrange chest physiotherapy to try to reduce the length of his hospital stay. We wonder if there is evidence to support the use of physiotherapy in this case.

Structured clinical question

In a child with community acquired pneumonia [patient], does chest physiotherapy [intervention] reduce the length of hospital admission [outcome]?

Search strategy and outcome

Search date: March 2007.

Cochrane Library using “pneumonia and child* and physiotherapy” revealed 19 results but none of these were relevant.

Pubmed using the same search terms revealed 110 articles. Eleven of these were relevant, but only two were available in English (six in Russian, one in French, one in German and one in Italian). One was a randomised control trial1 and one was a case series.2

PEDro (physiotherapy evidence database). Selecting options of “respiratory therapy” for therapy, “difficulty with sputum clearance” for problem, “chest” for body part and “paediatrics” for subdiscipline revealed 62 articles, none of which was relevant. A second search for articles with “pneumonia” in the title revealed 14 articles; one randomised control trial was relevant.3 See table 3.

Commentary

The three studies failed to show any benefit from chest physiotherapy in the treatment of community acquired pneumonia in terms of: rate of initial improvement of pneumonic infiltrates on chest x ray,1 length of hospital stay,2 3 duration of cough or coryza,2 persistence of wheeze, rhonchi and rales,2 development of dynamic airflow3 or subjective assessment of healing time.3 Two of the studies1 3 also showed that chest physiotherapy was associated with a significant increase in the duration of fever. The types of chest physiotherapy used in the three studies included: intermittent positive pressure breathing,1 postural drainage,1 3 vibration,1 3 percussion1^–3 and external help with breathing.3

Although the results of the three studies seem to be very clear, there are identifiable weaknesses in all of the three included trials. This means the results must be interpreted with a degree of caution.

Inclusion criteria in all three trials are varied and none of the studies directly assesses community acquired pneumonia in children. The study by Levine1 includes 55 children (2–12 years of age) with presumed viral pneumonia, excluding those with bacterial pneumonia. Stapleton2 looked at 55 children with acute uncomplicated respiratory tract infections, which included bronchitis (nine), bronchiolitis (20) and pneumonia (26). Britton et al3 included 171 patients with primary pneumonia, but they were adults (15–75 years).

The study by Britton et al3 is a good quality single blind randomised control trial, but there are methodological flaws in the other two trials. The method of randomisation was poor in the study of Levine.1 Randomisation consisted of numbering the patients consecutively at entry, even-numbered patients then received chest physiotherapy and odd-numbered patients did not. Given this method of randomisation, they failed to explain the difference in the size of the two groups (32 treated, 23 controls). The study by Stapleton2 was an observational study in which 34 patients received chest physiotherapy and 21 did not; this was decided according to the practice of the admitting doctor. Both of these methods allow bias to be introduced into the studies.

The Levine1 study included children with presumed viral pneumonia and therefore the use of antibiotics prior to or during admission was not deemed significant. As they did not confirm the diagnosis of viral pneumonia by viral isolation or serological studies, children with bacterial pneumonia may have been included in the study. If such patients were included, antibiotics are likely to have influenced the pneumonic infiltrates and the duration of fever and so any difference between the treatment and control groups in terms of the use of antibiotics should have been further assessed.

Stapleton2 provided a textual summary of the findings of his study but the only numerical data he gave referred to the mean hospital stay of the two groups. He also supplied no information as to how many children there were with each of the three diagnoses in the two treatment groups and did not mention the age of the children in the study, other than saying that the mean age did not differ between the two groups.