Combined Analgesia to Control Pain in Children Seen in Emergency Department (ED) for a Trauma of a Limb

Children's pain in Emergency Departments (ED) is poorly managed even though pain is one of the main symptoms of children consulting to the ED (Alexander et al., 2003; O'Donnell et al., 2002). Literature has repeatedly demonstrated the provision of analgesia to be suboptimal in the pediatric population in general. Friedland \& Kulick (1994) reviewed medical records of 99 children with acute fractures assumed to be painful and found that only 53% received analgesia. Southall et al. (1993) showed that of 181 invasive procedures performed on children receiving intensive care, only 50 (28%) received additional analgesia or sedation and during 36 of these procedures the child grimaced or cried (In Maurice et al., 2002). Pain management practices by nurses and doctors in the emergency department (ED) are poor. Selbst \& Clark (1990) noted that children (< 19 years) presenting with painful conditions in ED were much less likely to receive pain medications than adults (p = .001). Moreover, children were also less likely than adults to receive analgesics at discharge (p < .001). The authors concluded by mentioning that pediatricians and emergency physicians are reluctant to use analgesics for children in pain and needed additional education about management of acute pain. Recently, Brown et al. (2003) obtained similar results in a comparative chart review of adults and children presenting to the ED with a diagnostic of fracture. According to their results, children were least likely to receive any analgesics especially opioids. Compared with adults, a lower proportion (54%) of children (< 15 years) received an opioid (p < .001).

Moreover, results of the PAMPER study (Le May et al., 2005; Le May et al., 2007) revealed that on 150 children who presented themselves in ED for either a fracture, a severe sprain, a burn, an acute abdomen or deep lacerations, nurses and physicians only documented pain on 58% (87/150) of these children, even though these diagnoses are known to generate moderate to severe pain (> 4/10). Treatment of children presenting to the ED with a simple fracture is also poorly managed. Results from the same study showed that only 4 children out of 28 (14%) who presented to the ED with a simple fracture received an analgesic.

Efficacy of several medications has been verified to relieve pain of limb injuries in children consulting to the ED. Results of those studies show that ibuprofen was effective to relieve pain. But, most of these studies had a small sample size which affects their power to generalize their results (Koller et al., 2007; Tanabe et al. 2002). Only the RCT performed by Clark et al. (2007) demonstrated that ibuprofen (10 mg/kg) was more effective, at 60-minute post-administration, than both acetaminophen 15 mg/kg) and codeine (1 mg/kg) alone, to relieve children's pain following a musculoskeletal trauma. This trial had a total sample of 300 patients under 18-year with 100 patients per group. Nevertheless, only 52% of the group of children who received ibuprofen considered that their pain was well managed. On the other hand, it is important to specify that we are aware that codeine is pro-drug and necessitates the co-enzyme 2D6 of the CYP 450 to release morphine the analgesic component. Close to 12% of the Caucasian population do not have this co-enzyme, so codeine is not effective to relieve their pain (Williams et al., 2002). Nevertheless, since it is not part of the emergency physicians' practices at our ED to prescribe an opioid for a limb trauma, we needed evidence-based data to prove that a combination of medication would help decrease pain intensity. Further, this study is the first done at our ED using a combination of analgesics implying an opioid, so we needed to start with a weaker and safer opioid such as codeine.

Kennedy (2004) and Mehlisc (2002) propose the use of a combination of analgesics (opioid and non-steroidal inflammatory drug-NSAID) to better relieve children's pain generated by a musculoskeletal trauma. Their rationale is based on the potential of the NSAID to block pain impulses peripherically and decrease edema and the potential for the opioid to block pain at the level of the nervous central system. Finally, very few studies verified the efficacy of a combination of analgesics to relieve the pain generated by a limb trauma. An exhausted review of the literature did not retrieve any study comparing the efficacy of a combination of codeine and ibuprofen with ibuprofen alone to relieve musculoskeletal pain of children consulting to the ED.

Design: Randomized controlled trial (double blind) with the experimental group receiving codeine + ibuprofen and the control group receiving ibuprofen + placebo. Single site study.

Instruments :Pain will be measured with a Verbal Analogue Scale (VAS). The VAS is a sliding scale divided from 0 mm (absence of pain) to 100 mm (worst pain experienced). The VAS was used in several clinical trials with children and is considered valid for children over 6 years (Clark et al., 2007; Giannoni et al., 2002; Warnock \& Lander, 1998; Elhakim et al., 2003; Tyler et al., 1993).

A checklist will be used to evaluate the incidence of common side effects that might be experienced following the administration of an opioid.

Sample: Randomized sample of 200 children presenting to the ED with an injured limb.

Sample size calculation: The sample size was estimated according to the minimal difference required on the VAS for a significant clinical value. A difference of 15 mm with a standard deviation of 25 mm will be considered clinically significant (Clark et al., 2007). Since we have two groups and four different measuring times, we chose an alpha of 0,01 (bilateral). A sample of 67 children in both experimental and control groups was considered sufficient to detect a difference of 15 mm between each group, with a power of 80%. Our final sample size will be 100 children per group considering a dropping rate of 40%. Since Ste-Justine's ED handles close to 5000 cases of limb trauma per year, it is reasonable to estimate that we will be able to recruit 200 patients over a period of 12 months during the day and evening shifts and on week days only.

Inclusion criteria: 1) 6 to 18 y.o., 2) pain intensity > 30 mm, 3) understand, read and/or speak French or English, 4) injury within the last 24 hrs, 5) trauma to a limb without significant angulation (< 30 degrees).

Exclusion criteria: 1) allergies to opioids, ibuprofen and food coloring, 2) children suffering from asthma, 3) children presenting in a critical state, 4) social and physical abuse cases, 5) children with cognitive deficits (autism), 6) children with multiple traumas, 7) children suffering from chronic pain, 8) Children who received an analgesic within the past six hours.

Setting: The ED of Ste. Justine's University Health Centre in Montreal.

Interventions and double blind procedure: The randomized sequence will be generated by a computer program managed by a pharmacist independent of the study. Children allocated to the experimental group will receive a syrup of codeine and chewable tablets of ibuprofen. The children in the control group will receive a syrup containing a placebo and chewable tablets of ibuprofen. Pain will be measured at triage (before medication administration)-T1, at 60 minutes post-administration-T2, at 90 minutes-T3, and at 120 minutes-T4. Side effects and vital signs including saturation will be measured by the research nurse at each measuring time of the study. For security purposes, a standardized protocol will be elaborated if ever a child presents severe sedation, respiratory depression, nausea/vomiting following the administration of the medication.

Data collection and representativeness of the sample: Research nurses associated with this project will collect data from 12h00 to 20h00 on week days with a possibility of collecting data one week end per month during the same period of time. Children presenting after 10h00 in the ED will be considered for the study. Each recruited participant will be registered on a log sheet (screening log) where all the required information will be registered and the patient will receive a code for confidential purposes. Also, patients refusing to participate or patients presenting outside of the hours allocated for data collection will be registered in a different log (Log RMO- Refusal, Missed, Other) where we will register the chart number, the name of the patient, sex, age, date of visit, reason for consultation and reason why the patient refused or missed.

Statistical analyses: Descriptive statistics will be presented on sociodemographic variables for each group. A covariant model with repeated measures will be used to verify the variation of pain intensity among the groups and over the times of the study. The model will include a Time variable (T-1 to T-4), a Group variable (experimental and control), initial pain intensity at triage (T-1), presence of parents, and use of non-pharmacological methods. Comparisons will be done with an alpha of 0,01. Incidence of side effects among the groups will be compared using chi-squared tests. Finally, intention to treat analyses will be performed on participants randomized, who received the medication, and left the study afterwards.

The following outcomes will be used to measure the achievement of the objectives:

Main outcomes: Pain intensity at 60 min (T-2) and 90 min. (T-3) following administration of the medication compared to pain intensity at triage (T-1).

Secondary outcomes: Incidence of side effects at 60, 90 and 120 min. (T-4) following medication administration.