ORIGINAL ARTICLE
A pilot study of the dose–response of caudalmethylprednisolone with levobupivacaine in chroniclower back pain*
R. A. McCahon,
1
A. Ravenscroft,
2
V. Hodgkinson,
2
R. Evley
3
and J. Hardman
4
1 Consultant Anaesthetist, Department of Anesthesia, Nottingham University Hospitals NHS Trust, Queen’s Medical Centre, Nottingham, UK 2 Consultant Anaesthetist, Department of Anaesthesia, Nottingham University Hospitals NHS, Nottingham CityHospital, Nottingham, UK 3 Research Nurse, 4 Associate Professor & Honorary Consultant, Division of Anesthesia & Intensive Care, University of  Nottingham, Queen’s Medical Centre, Nottingham, UK 
Summary
The question as to what constitutes the ideal epidural steroid injection remains unresolved.We performed a prospective, randomised, double-blind, AB
  ⁄  
BA 2
 ·
 2 crossover study of caudal 40 vs 80 mg methylprednisolone acetate (in 20 ml levobupivacaine 0.125%) in outpatientswith chronic low back pain. Data from 33 participants were analysed. The Oswestry DisabilityIndex improved in both dose groups over time following injection. However, a statisticallysignificant improvement was only observed in the 40 mg methylprednisolone acetate group(40 mg: p < 0.001; 80 mg: p = 0.33). There was no statistically significant difference between thedose groups in change in the Oswestry Disability Index with respect to time. Methylprednisoloneacetate 40 mg appears to be as effective as 80 mg in improving disability associated with chroniclow back pain, and should be considered in preference to the 80 mg dose for outpatients withchronic low back pain attending for repeat caudal steroid injection.
........................................................................................................
Correspondence to: Dr R. A. McCahonEmail: rob.mccahon@nottingham.ac.uk*Presented in part at the Annual Scientific Meeting of the EuropeanSociety of Regional Anaesthesia, Valencia, Spain, September 2007. Accepted: 28 March 2011
Back pain is a major problem in the United Kingdomwith an estimated 11 million working days per annumlost to industry [1]. Epidural steroid injection is afrequent intervention in acute and chronic lower backpain, with 45 948 epidural steroid injections performedin the National Health Service (NHS) during2002
  ⁄  
2003 [2].It has been demonstrated that epidural glucocorti-coids depress the hypothalamo-pituitary axis andimpair glucose tolerance; therefore, it is important thatthe usefulness of epidural steroid is quantified and anydose–response relationship identified [3, 4]. Only onerecently published study has attempted to determinethis relationship [5]. However, this study had severalmethodological flaws including inadequate randomisa-tion and absent blinding.In this pragmatic, crossover study, we tested thehypothesis that caudal epidural methylprednisoloneacetate 40 mg with 25 mg levobupivacaine was not aseffective as caudal epidural methylprednisolone acetate80 mg with 25 mg levobupivacaine in improvingdisability in patients with chronic lower back pain.The primary outcome following epidural steroid injec-tion was assessed using the Oswestry Low Back PainDisability Index (ODI, version 2.1), a validated out-come measure of chronic lower back pain disability [6].
Anaesthesia, 2011,
 66
, pages 595603 doi:10.1111/j.1365-2044.2011.06764.x
.....................................................................................................................................................................................................................
 2011 The AuthorsAnaesthesia
 
 2011 The Association of Anaesthetists of Great Britain and Ireland
 595
 
Methods
This study was approved by the Institutional ReviewBoard, Research Ethics Committee, and the Medicines& Healthcare Regulatory Authority (MHRA), UKbefore commencement.Members of the study team obtained writteninformed consent from all study participants beforerandomisation in accordance with the ethical principlesthat have their srcin in the Declaration of Helsinkiand the ICH (International Conference on Harmon-isation) Guideline for Good Clinical Practice.All patients scheduled to have an epidural steroidinjection at a hospital Day Case Unit over a 6-monthperiod were sent a participant information sheet andstudy invitation letter with their usual appointmentletter. On the day of their treatment, nursing staff asked patients attending for an epidural steroid injec-tion if they would like to participate in the study.Following an explanation of the ODI by a member of the study team, interested patients completed a back-ground questionnaire, an ODI questionnaire, a backand leg pain VAS and a Hospital Anxiety & DepressionScale (HADS). Patients were considered eligible for study inclusion if they satisfied all of the followingcriteria:
1
 Patient scheduled for caudal epidural steroid injection.
2
 At least two epidural steroid injections in the previous12 months.
3
 Back and leg pain of any cause.
4
 ODI score > 20% (i.e.moderate disabilityorworse) [7].
5
 Back or leg VAS > 30 mm (i.e. moderate and abovepain severity) [8].
6
 No anticoagulant therapy, no bleeding diathesis and nolocal
  ⁄  
systemic sepsis.
The study was a standard AB
  ⁄  
BA 2
 ·
 2 crossover design. Each participant was given one 40 mg and one80 mg methylprednisolone acetate epidural injectionseparated by at least 12 weeks.All injections and drug preparations were performedby AR and VH, who are consultants in chronic painmanagement. The assigned dose of methylprednisoloneacetate (40 mg.ml
)
1
) was drawn up using an aseptictechnique into a 20-ml syringe and mixed with 10 mllevobupivacaine 0.25%. Sterile 0.9% saline was thenadded to make total volume 20 ml. Drug preparationtook place immediately before injection and out of view of the study participants. A loss-of-resistancetechnique was used to identify the epidural space via acaudal approach. Fluoroscopic guidance was not usedas previous studies have reported that 91% and 93% of blind caudal injections are accurately placed, especiallywith experienced operators [9, 10].Study participants were given a questionnaire book-let before discharge from the Day Case Unit. Thebooklet comprised 12 ODI questionnaires. A member of study staff explained the ODI questionnaire againand asked the participant to complete one ODI per week on the anniversary of their epidural steroidinjection. Participants were permitted to continue allconcurrent analgesic medications and were asked torecord their weekly analgesic intake for the duration of the study. Between each epidural steroid injection,participants received a weekly telephone reminder tocomplete the questionnaires. Participants returned thebooklets on their next visit to the Day Case Unit.Study participation was complete 12 weeks after thesecond epidural steroid injection.The primary study outcome was the change in ODIfrom baseline over a 12-week period. The ODIquestionnaire comprises 10 domains, each consisting of six simply worded comments scored from 0 to 5, withlarger scoresindicating greater disability(seeAppendix).Patients were instructed to tick one comment pedomain that best described them on that day. The indexis normally calculated and reported in the literature as apercentage(seebelow)[6].TheODIisahighlycitedandutilised assessment of disability in patients with chroniclow back pain. As such, it has been reported todemonstrate high levels of validity, internal consistencyand reliability,anditisabletosignalmeaningful changesin a patient’s functional ability [6].
ODI
¼
Total score5
Number of questions answered
100%
The secondary study outcome was the change inweekly analgesic consumption.The local Clinical Trials Support Unit created asimple web-based, computer-generated randomisationschedule for the study. Treatment sequence alloca-tion took place immediately before the first epidu-ral steroid injection and was performed by VH or AR. The treatment allocation list was kept on apassword-protected webpage of the Clinical TrialsSupport Unit website.The participants were blinded to the allocatedtreatment. All outcome measures were self-assessedby the participants. Both AR and VH, who performedthe epidural steroid injections, were blinded to theparticipant assessment data until after all patients had
R. A. McCahon et al.
Æ
 Dose-response of caudal methylprednisolone Anaesthesia, 2011,
 66
, pages 595–603
......................................................................................................................................................................................................................
 2011 The Authors
596
 Anaesthesia
 
 2011 The Association of Anaesthetists of Great Britain and Ireland
 
completed the study, and they did not take part indata collection or analysis. Statistical analysis wasperformed by JH who was blinded to the treatmentallocation.To achieve a 90% probability that the study woulddetect a clinically significant treatment difference of 10on the ODI at a two-sided 5% significance level, asample size of 32 participants was required. A differ-ence of 10 in the ODI was chosen as previous work hasdemonstrated that changes of 8–12 percentage points inthe ODI score are associated with return to workand
  ⁄  
or resumption of full duties [11, 12]. A within-subject standard deviation of 10 in the ODI differencesover a 12-week period was assumed [13]. To allow for a drop-out rate of 15%, we planned to recruit a total of 38 participants. Analysis was performed using Micro-soft Excel 2007 with Analysis Toolpak (MicrosoftCorporation, Redmond, WA, USA). Oswestry Dis-ability Index scores have been treated previously asnormally distributed [13, 14]; aggregated ODI scores inour study were tested using the Shapiro–Wilk test for normality of distribution. The change in ODI frombaseline (pre-epidural steroid injection) over time wascalculated for weeks 1–12; this value was assessed withrespect to time for the 40 mg and the 80 mg groupsusing ANOVA. The differences in change in ODIfrom baseline for each patient at each time-point for the two doses of methylprednisolone acetate werecalculated, and this value was examined with respect totime using ANOVA. The change in number of analgesics taken per week from baseline was assessedwith respect to time for the 40 mg and the 80 mggroups using ANOVA. Subgroup analysis on the causeof back pain was not performed, as the study was notadequately powered for this.
Results
The study profile is summarised in Fig. 1. There wereno protocol deviations during the study.A total of 33 patients (of 37 recruited) were includedin the final analysis. Failure to complete the question-naires in the second period or failure to return themprecluded analysis of three participantsdata. Oneparticipant withdrew early in the first period. Thus,data from 20 females and 13 males were analysed. Theyhad a mean (SD) age of 56.5 (9.8) years and reported amean lower back pain duration of 18.9 (9.3) years.Baseline measurements before each injection are sum-marised in Table 1.All participants complained of lower back and legpain. Their diagnosis was recorded in the medical notesas follows: 24 out of 33 degenerative disc disease; 2 outof 33 postback surgery syndrome; 8 out of 33 other (i.e. mechanical pain or no diagnosis found in medicalnotes).In the 12 months before entry to the study,participants had a median (IQR [range]) of 3(3–3 [2–3]) epidural steroid injections. Participantshad not received any epidural steroid for a minimum of 12 weeks before recruitment as per the local clinicalprotocol; the usual epidural steroid injection wasmethylprednisolone acetate 80 mg with levobupiva-caine 25 mg. The median (IQR [range]) time betweenstudy epidural steroid injections was 16 (14–19 [12– 23]) weeks. The randomly assigned dose sequence wassuch that 16 (48%) patients had the 40 mg dose firstand 17 (52%) had the 80 mg dose first.Shapiro–Wilk testing indicated that ODI scores wereacceptably close to normal distribution, so they wereanalysed using ANOVA. Oswestry Disability Indeximproved in both dosage groups over time followinginjection,butastatisticallysignificantimprovementonlyoccurred in the 40 mg dosage group (40 mg:p < 0.0001; 80 mg: p = 0.33); see Fig. 2. There wasno statistically significant difference between the 40 and80 mg groups in change in ODI over time (p = 0.95);seeFig. 3.TransformingODItoapercentageofbaselinestill revealed no statistically significant change over time(p = 0.9); see Fig. 3.Each participant was asked to complete 13 ODIquestionnaires per study period. Each questionnairecomprises 10 items (see Appendix) giving a total of 130items per respondent per study period. Thus, there wasa total of 4290 items (33
 ·
 130) to be answered in eachperiod.In period 1, 3.8% (164 out of 4290) items were notanswered by participants. The mean number of omitted items per respondent per week in the firstperiod was 0.38. In period 2, 3.6% (154 out of 4290)items were omitted. The mean number of omitteditems per respondent per week in the second periodwas 0.36. The distribution of omitted items issummarised in Table 2.Therewasnostatisticallysignificantchangeinweeklyanalgesic intake (number of tablets) following epiduralsteroid injection with 40 mg (p = 0.69) or 80 mgmethylprednisolone acetate (p = 0.99); see Fig. 4.There were no adverse events reported by partici-pants during the study.
Anaesthesia, 2011,
 66
, pages 595–603
 R. A. McCahon et al.
Æ
 Dose-response of caudal methylprednisolone
......................................................................................................................................................................................................................
 2011 The AuthorsAnaesthesia
 
 2011 The Association of Anaesthetists of Great Britain and Ireland
 597
 
Discussion
It is important to consider that a pragmatic study aimsto measure the benefit of a treatment in routine clinicalpractice. In contrast, explanatory studies aim tomeasure the benefit a treatment under ideal conditionsusing strictly defined inclusion criteria. Thus, in apragmatic study, the difference between two treat-ments is the sum of the treatment and any placeboeffect, as this best reflects the likely response in clinicalpractice [15]. In this pragmatic, crossover study, wehave demonstrated that caudal epidural administrationof methylprednisolone acetate 40 + 25 mg levobupi-vacaine is as effective as caudal epidural administrationof methylprednisolone acetate 80 + 25 mg levobupi-vacaine in reducing disability scores in patients with abroad diagnosis of chronic lower back pain who attendfor repeat epidural steroid injection.We believe that the lack of a statistically significantreduction in ODI after epidural steroid injection usingmethylprednisolone acetate 80 mg is most likely to be atype-2 error and we do not take it to mean that a40 mg dose is effective, whereas an 80 mg dose is not.Indeed, there was no statistically significant differencebetween the 40 and 80 mg groups with respect to
Table 1
 Baseline measurements before epidural steroidinjection. Values are mean (SD).
Methylprednisoloneacetate 40 mg(n = 33)Methylprednisoloneacetate 80 mg(n = 33)
ODI (%) 54.0 (13.9) 54.7 (15)HADSAnxiety 9.0 (4.2) 8.2 (4.6)Depression 7.4 (3.3) 7.1 (3.2)Pain VAS; mmBack 66.1 (19.6) 66.7 (17.6)Leg 54.2 (28.4) 57.0 (29.2)
ODI, Oswestry Disability Index; HADS, Hospital Anxiety & DepressionScale; VAS, Visual Analogue Score.
Figure 1
 Study profile. ODI, Oswestry Disability Index.
R. A. McCahon et al.
Æ
 Dose-response of caudal methylprednisolone Anaesthesia, 2011,
 66
, pages 595–603
......................................................................................................................................................................................................................
 2011 The Authors
598
 Anaesthesia
 
 2011 The Association of Anaesthetists of Great Britain and Ireland
 
change in ODI over time. Furthermore, visual inspec-tion of the change in ODI over time following 40 and80 mg epidural steroid injection shows treatmenteffects that are very similar between the groups.Although the use of levobupivacaine as a carrier for methylprednisolone acetate may have caused somerelief of symptoms (potentially concealing differencesbetween steroid doses), this was identical in bothgroups. Furthermore, use of local anaesthetic as acarrier reflects common clinical practice. In a previousrandomised, controlled trial of epidural methylpred-nisolone acetate (without local anaesthetic) vs salineplacebo in chronic lower back pain, Carette et al.demonstrated a significant improvement in ODI scoresthat lasted 6 weeks in patients treated with methyl-prednisolone acetate [16]. In the current study, theonly treatment difference between groups was the doseof methylprednisolone acetate, and so differences infunctional state are attributable to differences in dosage.Only one previous study has examined the effect of epidural steroid dose (methylprednisolone acetate spe-cifically) in patients with back pain [5]; this studydemonstrated that a 40 mg dose of methylprednisoloneacetate was equivalent to an 80 mg dose, 3 monthsafter injection. However, the authors used a bluntoutcome measure in that a participant’s treatment wasdeemed successful if they described a 2-point improve-ment in their 10-point pain VAS following epiduralsteroid injection. Thus, there was no means of trulydifferentiating between doses of methylprednisoloneacetate. Furthermore, there was no randomisation,blinding or power calculation. Such methodological
Figure 2
 Change in mean Oswestry Disability Index (ODI)scores from before epidural steroid injection (baseline) for 12 weeks following 40 mg methylprednisolone acetate(dashed line; p < 0.0001) and 80 mg methylprednisoloneacetate (solid line; p = 0.33). Error bars represent 1 SD (upgoing bars relate to 40 mg, and down going bars relate to80 mg dose).
Table 2
 Omitted Oswestry Disability Index (ODI) items bystudy period.
ODI itemNumber of omissionsStudy period 1 Study period 2
Pain intensity 5 3Personal care 10 6Lifting 7 4Walking 9 5Sitting 6 3Standing 5 3Sleeping 5 3Sex life 105 120Social life 5 3Travelling 7 4Total 164 154
Figure 3
 Difference in mean Oswestry Disability Index (ODI) scores following epidural 80 mg methylprednisolone acetateand 40 mg methylprednisolone acetate over 12 weeks, expressed as percentage change (dashed line) and absolute difference(solid line) (p = 0.95). Error bars represent 1 standard deviation (up going bars relate to % change in ODI, and down going barsrelate to absolute change in ODI).
Anaesthesia, 2011,
 66
, pages 595–603
 R. A. McCahon et al.
Æ
 Dose-response of caudal methylprednisolone
......................................................................................................................................................................................................................
 2011 The AuthorsAnaesthesia
 
 2011 The Association of Anaesthetists of Great Britain and Ireland
 599
of 9