Where can I find this paper?

http://www.jem-journal.com/article/S0736-4679(12)01136-5/abstract

What is this paper about (what is the research question)?

Is codeine in combination with ibuprofen better at improving paediatric pain scores for children presenting to the ED with limb injuries than ibuprofen alone?

Summary of the Paper

Design: single centre, randomised, double-blinded placebo-controlled trial

Objective: to compare the efficacy of a combination of codeine with ibuprofen to codeine alone for the management of pain in children with limb injuries presenting to the ED.

Computer generated block randomisation with double blinding was used.

Intervention: 1mg/kg codeine syrup plus 10mg/kg ibuprofen caplet

Comparison: placebo syrup plus 10mg/kg ibuprofen caplet

Primary outcome: difference in mean pain score between initial score at triage (T1) and 90mins after administration of medication (T3). Validated visual analogue scale (VAS) was used for pain scoring.

Population: children aged 7-18 years presenting to a tertiary paediatric ED between Feb 2008-Nov 2009, between 10am-4pm, with musculoskeletal limb injury in preceding 72h

• Inclusion: pain score >3 on 0-10 VAS at triage, French/English speaking, limb with bony tenderness, swelling, limited range-of-movement or angulation <30deg.
• Exclusion: allergy to opioids, NSAIDs or food colouring, critical illness, injury requiring orthopaedic reduction, trauma to >1 limb, social/physical abuse, cognitive deficit, chronic pain, analgesia in last 6h

Results: 83 patients enrolled from 279 eligible patients, 81 analysed; 40 in experimental and 41 in control arms. 47% had presented with fractures

Mean pain score at T1: experimental 5.9 +/- 1.7, control 5.7 +/- 1.3

Mean pain score at T3: experimental 4.0 +/- 2.4, control 4.1 +/- 2.0

Mean difference in pain scores between T1 & T3: -0.1 (95% CI -1.3 to 1.0), p=0.55

Authors’ Conclusions:

The addition of codeine to ibuprofen did not significantly improve pain management of children with a musculoskeletal trauma. Pain control provided by the medications remained suboptimal for most patients.

On the study design

Despite being a convenience sample, there are plenty of good methodological examples to be found in this paper. The use of allocation concealment (randomisation by computer generated blocks) and placebo, and careful preservation of blinding until after completion of the study are good examples of the role of study designers in reducing human influence on the results. The paper gives details as to how randomisation was carried out, which increases our faith in the strength of the methodology.

Arguably, a convenience sample reduces our ability to apply this data to our general PED population, however it could be countered that the majority of paediatric injury presentations occur between 09:00 and 21:00. In my clinical experience, the weighting does fall towards the evening, so perhaps an adjustment of the research nurse’s hours to reflect this might have been sensible.

There are a great number of inclusion and exclusion criteria, but these seem straightforward and sensible. The key to measuring the outcome (pain) is an ability to understand and communicate a pain score as per VAS, and many of the criteria are designed around ensuring that outcomes can be measured accurately in the sample. After all, for severely injured children we are likely to consider stronger opioids/opiates (such as diamorphine or fentanyl via the intranasal route), so if there is an analgesic role for codeine in PED it is likely to be in those children with mild-moderate pain.

It is worth noting that children requiring plaster immobilisation had to wait until the full 2hr study period had elapsed before their plaster was applied. I can understand that  this might cause a confounding effect, but I wonder if further studies might help us to quantify the benefit of splintage? Anecdotally from working with adults, plastering a fractured limb does provide relief. With increasing use of ultrasound to identify fractures at the bedside, should our emphasis be towards plastering earlier in the clinical encounter (and perhaps x-ray after plaster instead of before)?

What were the results and what does this mean?

The sample size calculation undertaken prior to the study required 28 children to be analysed in each arm at T3 in order to achieve power to detect a difference in means of 1.5. We can see this was achieved (31 patients remained in each arm at T3).

There was no statistically significant difference in mean pain scores between the two treatment arms at T3, which suggests no benefit in adding codeine to ibuprofen for these patients. Of note, one patient in the experimental arm complained of nausea. It would seem sensible to consider an alternative analgesic for co-administration.

What can we take from this paper into clinical practice?

Codeine is infrequently suggested for paediatric analgesia in my experience. As discussed above, if there is a role for its use it would likely be in those patients with mild-moderate pain. This is not supported by the findings of this study and as such it seems unlikely that anyone will be advocating the use of codeine for paediatric pain as yet.

More questions to ask

• How can we best provide analgesia for children with limb injuries?
• Does the role of robust splintage (e.g. plaster) compare favourably with analgesic agents?

Follow us on twitter: @PEMLit