IV Dexamethasone for Back Pain and Radiculopathy in the ED
IV Dexamethasone for Back Pain and Radiculopathy in the ED
This was a double-blind randomised controlled trial approved by the Health District's Human Research and Ethics Committee and registered prior commencement with the Australia and New Zealand Clinical Trials Register (no.: 12611001020976) without any protocol changes. We adhered to the CONSORT statement (http://www.consort-statement.org).
The trial was conducted between November 2011 and November 2012 in the EDs of the two public hospitals within the same Health District in Southeast Queensland, Australia. The main campus is a 570-bed major metropolitan teaching hospital and the second campus (located 12 km from the main campus) is an urban district hospital with 200 beds. The ED census in 2012 was 67 000 and 50 000, respectively.
Patients who presented with low back pain and leg radiation at triage were identified as potential participants in this trial. Further inclusion criteria were age between 18 and 55 years, a positive SLR test and difficulty mobilising. Exclusion criteria were likely alternative diagnosis or 'red flags' (fever, recent trauma, history of malignancy), pregnancy, known allergy to dexamethasone, current use of glucocorticoids, history of lower back surgery and inability to provide consent. Eligible patients were recruited by the triage nurse, treating staff or research nurse according to these inclusion and exclusion criteria. Written informed consent from eligible participants was obtained before randomisation.
An inclusion criterion was to have a positive SLR test, as assessed by either a doctor or a physiotherapist. An SLR test was positive if pain was reproduced by passive movement of the hip joint between 0° and 70° as measured with a goniometer. The SLR test is highly sensitive for detecting LBPR and assessing the effect of treatment. The average angle of passive hip flexion from three attempts was calculated and documented at initial presentation and again upon discharge.
After being consented, eligible participants were randomised to either treatment (8 mg intravenous dexamethasone in 2 mL) or control (2 mL intravenous 0.9% sodium chloride) group. Syringes for both treatment and control were identical except for unique sequence numbers. The allocated treatment was in addition to current routine care for LBPR. Both groups received a standardised regimen of regular analgesia, physiotherapy referral and education. Standardised analgesia included regular oral acetaminophen/codeine, ibuprofen and oral oxycodone as required.
The randomisation was overseen by an independent statistician using computer-generated block randomisation with blocks of 10 for each hospital. Allocation concealment was maintained using individual sequentially numbered opaque sealed envelopes. These envelopes were opened after the patient had provided written consent.
A study nurse (not involved with recruitment, consenting or the participant's care) opened the next available sealed envelope (containing the patient allocation) and would ask the treating nurse to administer the allocated treatment from the syringe matching the sequence number in the envelope. Nurses and doctors involved with patient care and data collection, participants and the analysing statistician were blinded to the group allocation.
The primary outcome measure was the change in level of pain, documented using a VAS. A 10 cm line ranging from 'no pain' to 'worst pain ever' was presented. Patients were instructed to mark an 'X' on any point of the line that best described their current level of pain. The VAS is a validated, reliable and easy means of measuring pain in the acute medicine setting with a change of between 1.3 and 2.0 points generally accepted as the minimum clinically meaningful effect. We did not prespecify patient position for VAS measurement. The VAS pain score was collected at four distinct points: on presentation, on discharge, 24 h after randomisation and 6 weeks postinitial presentation to ED.
Secondary outcome measures included (1) EDLOS (defined as triage time to time of discharge), (2) time to return to normal activities (in days), (3) SLR range of motion (degrees) as described above, (4) analgesia requirements and proportion of adverse effects and (5) the Oswestry Disability Index (ODI). This index is a validated outcome measure in assessing a patient's level of disability status as a result of spinal disorders. The questionnaire consists of 10 multiple choice questions worth up to 5 marks each, giving a scoring range of 0–50, with a lower score representing a lower degree of disability. The score is then used to calculate (in percentage) the level of a patient's disability. Participants' scores were collected at three times: on presentation, at 24 h and 6 weeks.
Data sheets were deposited in a locked box. A single research nurse collected these and was responsible for data entry.
The treating staff (medical and physiotherapy) discharged patients from ED based on their clinical judgement, with ability to mobilise safely a compulsory component. Participants were given follow-up questionnaires (including VAS and ODI) and instructions to self-complete and return them by reply-paid envelopes at 24 h and 6 weeks.
A power calculation was performed based on the assumption that in the placebo group (routine management only) the VAS pain score would improve by 1.5 points. This was based on local audit data of all patients in 2010 with LBPR, where VAS scores decreased from 7.5 to 6 with routine management. We hypothesised that the addition of dexamethasone would decrease the VAS pain score by a further 1.5 points (7.5 to 4.5), which we deemed clinically significant. All VAS scores were assumed to have an SD of 2 points. We calculated a required number of 44 patients (22 per group) to have 80% power to detect this difference (α 0.05, two-sided). To account for loss to follow-up and to have complete follow-up for the secondary endpoint of VAS at 6 weeks, we aimed to enrol 70–100 patients. Analysis was by intention to treat. No interim analyses were performed.
Data taken from completed questionnaires were collated using Microsoft Excel spreadsheet software and then coded prior to transfer to the Statistical Package for the Social Sciences (V.17.0, SPSS, Chicago, Illinois, USA) for statistical analysis. Before analysis, all variables were reviewed for accuracy of data entry, missing values and outliers using SPSS. No modifications were made for missing data. For continuous variables, we used an independent t test or Mann–Whitney U test (if variables were not normally distributed) to compare treatment groups. We conducted a regression analysis as a secondary analysis to account for baseline imbalance, with change in VAS as dependent variable and baseline VAS, amount of oxycodone and intervention (placebo/dexamethasone) as independent variables. For categorical variables, the χ test was used. An α of 0.05 was statistically significant.
Methods
Study Design
This was a double-blind randomised controlled trial approved by the Health District's Human Research and Ethics Committee and registered prior commencement with the Australia and New Zealand Clinical Trials Register (no.: 12611001020976) without any protocol changes. We adhered to the CONSORT statement (http://www.consort-statement.org).
Setting
The trial was conducted between November 2011 and November 2012 in the EDs of the two public hospitals within the same Health District in Southeast Queensland, Australia. The main campus is a 570-bed major metropolitan teaching hospital and the second campus (located 12 km from the main campus) is an urban district hospital with 200 beds. The ED census in 2012 was 67 000 and 50 000, respectively.
Selection of Participants
Patients who presented with low back pain and leg radiation at triage were identified as potential participants in this trial. Further inclusion criteria were age between 18 and 55 years, a positive SLR test and difficulty mobilising. Exclusion criteria were likely alternative diagnosis or 'red flags' (fever, recent trauma, history of malignancy), pregnancy, known allergy to dexamethasone, current use of glucocorticoids, history of lower back surgery and inability to provide consent. Eligible patients were recruited by the triage nurse, treating staff or research nurse according to these inclusion and exclusion criteria. Written informed consent from eligible participants was obtained before randomisation.
Measurements: Straight Leg Raise Test
An inclusion criterion was to have a positive SLR test, as assessed by either a doctor or a physiotherapist. An SLR test was positive if pain was reproduced by passive movement of the hip joint between 0° and 70° as measured with a goniometer. The SLR test is highly sensitive for detecting LBPR and assessing the effect of treatment. The average angle of passive hip flexion from three attempts was calculated and documented at initial presentation and again upon discharge.
Interventions
After being consented, eligible participants were randomised to either treatment (8 mg intravenous dexamethasone in 2 mL) or control (2 mL intravenous 0.9% sodium chloride) group. Syringes for both treatment and control were identical except for unique sequence numbers. The allocated treatment was in addition to current routine care for LBPR. Both groups received a standardised regimen of regular analgesia, physiotherapy referral and education. Standardised analgesia included regular oral acetaminophen/codeine, ibuprofen and oral oxycodone as required.
Randomisation and Allocation Concealment
The randomisation was overseen by an independent statistician using computer-generated block randomisation with blocks of 10 for each hospital. Allocation concealment was maintained using individual sequentially numbered opaque sealed envelopes. These envelopes were opened after the patient had provided written consent.
A study nurse (not involved with recruitment, consenting or the participant's care) opened the next available sealed envelope (containing the patient allocation) and would ask the treating nurse to administer the allocated treatment from the syringe matching the sequence number in the envelope. Nurses and doctors involved with patient care and data collection, participants and the analysing statistician were blinded to the group allocation.
Outcome Measures and Data Collection
The primary outcome measure was the change in level of pain, documented using a VAS. A 10 cm line ranging from 'no pain' to 'worst pain ever' was presented. Patients were instructed to mark an 'X' on any point of the line that best described their current level of pain. The VAS is a validated, reliable and easy means of measuring pain in the acute medicine setting with a change of between 1.3 and 2.0 points generally accepted as the minimum clinically meaningful effect. We did not prespecify patient position for VAS measurement. The VAS pain score was collected at four distinct points: on presentation, on discharge, 24 h after randomisation and 6 weeks postinitial presentation to ED.
Secondary outcome measures included (1) EDLOS (defined as triage time to time of discharge), (2) time to return to normal activities (in days), (3) SLR range of motion (degrees) as described above, (4) analgesia requirements and proportion of adverse effects and (5) the Oswestry Disability Index (ODI). This index is a validated outcome measure in assessing a patient's level of disability status as a result of spinal disorders. The questionnaire consists of 10 multiple choice questions worth up to 5 marks each, giving a scoring range of 0–50, with a lower score representing a lower degree of disability. The score is then used to calculate (in percentage) the level of a patient's disability. Participants' scores were collected at three times: on presentation, at 24 h and 6 weeks.
Data sheets were deposited in a locked box. A single research nurse collected these and was responsible for data entry.
Follow-up Procedure
The treating staff (medical and physiotherapy) discharged patients from ED based on their clinical judgement, with ability to mobilise safely a compulsory component. Participants were given follow-up questionnaires (including VAS and ODI) and instructions to self-complete and return them by reply-paid envelopes at 24 h and 6 weeks.
Primary Data Analysis
A power calculation was performed based on the assumption that in the placebo group (routine management only) the VAS pain score would improve by 1.5 points. This was based on local audit data of all patients in 2010 with LBPR, where VAS scores decreased from 7.5 to 6 with routine management. We hypothesised that the addition of dexamethasone would decrease the VAS pain score by a further 1.5 points (7.5 to 4.5), which we deemed clinically significant. All VAS scores were assumed to have an SD of 2 points. We calculated a required number of 44 patients (22 per group) to have 80% power to detect this difference (α 0.05, two-sided). To account for loss to follow-up and to have complete follow-up for the secondary endpoint of VAS at 6 weeks, we aimed to enrol 70–100 patients. Analysis was by intention to treat. No interim analyses were performed.
Data taken from completed questionnaires were collated using Microsoft Excel spreadsheet software and then coded prior to transfer to the Statistical Package for the Social Sciences (V.17.0, SPSS, Chicago, Illinois, USA) for statistical analysis. Before analysis, all variables were reviewed for accuracy of data entry, missing values and outliers using SPSS. No modifications were made for missing data. For continuous variables, we used an independent t test or Mann–Whitney U test (if variables were not normally distributed) to compare treatment groups. We conducted a regression analysis as a secondary analysis to account for baseline imbalance, with change in VAS as dependent variable and baseline VAS, amount of oxycodone and intervention (placebo/dexamethasone) as independent variables. For categorical variables, the χ test was used. An α of 0.05 was statistically significant.
Source...