- Email: [email protected]
Three-in-one femoral nerve block as analgesia for fractured neck of femur in the emergency department: A randomized, controlled trial
PAIN MANAGEMENT/ORIGINAL RESEARCH
Three-in-One Femoral Nerve Block as Analgesia for Fractured Neck of Femur in the Emergency Department: A Randomized, Controlled Trial
Alan K. Fletcher, MRCP(UK) Alan S. Rigby, MSc Francis L. P. Heyes, FRCS From the Department of Emergency Medicine, Rotherham General Hospital, Rotherham, United Kingdom (Fletcher, Heyes), and the Department of Statistics, University of Sheffield, Sheffield, United Kingdom (Rigby).
Study objective: We determine whether 3-in-1 femoral nerve block is effective as analgesia for fractured neck of femur when administered by emergency physicians. Methods: This was a prospective, randomized controlled trial with blinded assessors conducted in a district general hospital emergency department in the United Kingdom. Over a 6-month period, all patients with fractured neck of femur were considered for study. Patients were randomly assigned to receive 3-in-1 nerve block with bupivacaine plus intravenous morphine or intravenous morphine. An accreditation package for all ED medical staff was devised to ensure competence in the technique of 3-in-1 nerve block. Pain scores were recorded on arrival and at intervals up to 24 hours after admission. Morphine consumption in the first 24 hours was recorded. Results: Ninety-four patients sustained fractured neck of femur during the study period; 50 were studied. Of 44 not studied, 42 were confused, 1 did not consent, and 1 was overlooked. Patients receiving 3-in-1 nerve blocks recorded a faster time to reach the lowest pain score: 2.88 hours for patients with nerve block and 5.81 hours for control patients (mean difference –2.93 h; 95% confidence interval [CI] –5.48 to –0.38 h). Nerve block recipients required significantly less morphine per hour than control patients (mean of 0.49 mg/h versus 1.17 mg/h; mean difference –0.68 mg/h; 95% CI –1.23 to –0.12 mg/h). Conclusion: Three-in-one femoral nerve block is an effective method of providing analgesia to patients with fractured neck of femur in the ED. All grades of medical staff were able to apply and consolidate this skill. [Ann Emerg Med. 2003;41:227-233.]
Copyright © 2003 by the American College of Emergency Physicians. 0196-0644/2003/$30.00 + 0 doi:10.1067/mem.2003.51
FEBRUARY 2003
41:2
ANNALS OF EMERGENCY MEDICINE
2 2 7
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
INTRODUCTION
Fractured neck of femur causes significant mortality and morbidity in elderly patients. Even with optimum so-called fast tracking of these patients through hospital admission, surgery, and rehabilitation, mortality rates of 18% at 90 days can be anticipated.1 Patients with fractured neck of femur are often in considerable pain, and minimizing this has important implications for patient comfort and reduction of morbidity and mortality.2 The 3-in-1 femoral nerve block is a safe and widely practiced local anesthetic technique used to supplement anesthesia and provide postoperative analgesia after hip surgery.3 Local anesthetic infiltrates the femoral nerve sheath and tracks cranially and laterally to anesthetize the femoral and obturator nerves, the lower cords of the lumbar plexus, and the lateral cutaneous nerve of the thigh (Figure 1). Two previous studies4,5 have shown that femoral nerve block provides effective analgesia for patients with fractured neck of femur in the emergency department, but each relied on one practitioner’s expertise for block administration and might consequently have been relatively selective in patients recruited. One of these studies5 did not have a control group. Another recent study6 has successfully used 3-in-1 femoral nerve block for fractures of the upper femur, but the study population was not clearly described, had no control group, and once again relied on expert block administration. This study was conducted to assess whether 3-in-1 femoral nerve block is effective when taught to and implemented by ED medical staff. Our study was powered to detect a 40% reduction in pain scores in those given a nerve block. The applicability of the technique has been evaluated in a United Kingdom district general hospital ED setting.
obtained. Patients were recruited during a 6-month period, from February until August, coinciding with senior house officer (intern) posts. Two consultants, 4 middle-grade physicians, and 7 senior house officers staff our ED. An accreditation package was developed for all ED medical staff to ensure competence in the technique of 3-in-1 femoral nerve block before the study commenced. The accreditation package comprised a half-hour structured session of instruction in the technique, including anatomy, pharmacology, and local anesthetic toxicity, followed by supervised practice on a resuscitation mannequin. A nerve block performed on a patient with fractured neck of femur was also supervised by the same observer (AKF). These patients were not included in the study. Competence was considered achieved when supervised practice
Figure 1.
Injection technique of 3-in-1 femoral nerve block. Inguinal ligament Nerve Artery Injection of Vein 20 mL of 0.5% plain bupivacaine into the neural sheath
30°
Digital pressure distal to injection site during and 5 minutes after injection
M AT E R I A L S A N D M E T H O D S
The study was conducted in the ED of our mediumsized district general hospital in the United Kingdom. The ED handles 65,000 new attendances per year. Local research ethical committee approval for the study was
2 2 8
ANNALS OF EMERGENCY MEDICINE
41:2
FEBRUARY 2003
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
showed that key safety and technical points had been achieved and when the supervised patient nerve block was considered safe and satisfactory. As all physicians achieved competence, it was possible to consider all patients with all types of fractured neck of femur for this study. Patients who were confused (and therefore unable to give informed consent), had a bleeding diathesis or were taking warfarin, had local or systemic infection, or had previous hypersensitivity to local anesthetics were excluded (Figure 2). Informed consent was obtained once fracture was confirmed with radiography, and patients were randomly assigned to 1 of 2 groups. Patients in the first group were control patients and received analgesia in the conventional fashion for our district general hospital (ie, 5 to 10 mg of intravenous morphine, available
Figure 2.
Study flow diagram.
90 patients with fractured neck of femur 42 excluded because of confusion 1 patient refused consent 1 patient overlooked 50 patients included
24 randomized to receive IV morphine
26 randomized to receive 3-in-1 femoral nerve block with 20 mL of 0.5% bupivacaine
Pain scores at 0, 1, 4, 8, 12, 16, and 24 h Morphine consumption at 24 h
Complications and mortality recorded after case note review at 6 mo
FEBRUARY 2003
41:2
ANNALS OF EMERGENCY MEDICINE
hourly until analgesia was achieved). This prescription continued until surgery. Patients in the second group received a 3-in-1 femoral nerve block with 20 mL of 0.5% plain bupivacaine. This was performed by using the technique described by Winnie et al7; a 21-gauge needle without a peripheral nerve stimulator was used instead of a nerve block needle for convenience and practicality. Aseptic technique was observed, and patients were monitored with ECG and basic observations. An intravenous cannula was sited before nerve block in all cases, and its patency was confirmed. The femoral pulse was palpated, and the nerve block was inserted 1 cm lateral to the pulse once paresthesias were elicited to localize the nerve. The local anesthetic was injected in a cranial direction and with pressure distal to the needle during and shortly after injection to encourage the local anesthetic to track cephalad. If patients required analgesia before radiography, they received incremental doses of intravenous morphine. Sample size was estimated by using the method described by Altman8 from data of previously published studies and from 10 pilot patients in whom a pain score reduction of 50% was observed. We calculated that a total study size of 50 patients would be required to achieve 80% power to observe a 40% pain reduction score difference at the 0.05 significance level. A pilot study indicated that approximately half of all patients with fractured neck of femur would be ineligible because of confusion, and this was allowed for in assessment of study feasibility. The data from the pilot study are not included in the results. The randomization sequence was derived from a random number generator, and allocation concealment was achieved by means of the sealed opaque envelope method. Pain on movement was assessed by using a numeric rating scale that has been adopted as a hospital standard for measuring pain, ensuring that scores were reproducible. This pain scale has been validated before.9,10 We decided to adopt a numeric rating scale instead of a visual analog scale partly because of the difficulties we encountered with patient comprehension of visual analog scales in the pilot study. Pain was graded as 0, 1, 2, or 3 corresponding to no pain, mild pain, moderate pain,
2 2 9
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
or severe pain, respectively. Pain on movement was assessed on arrival (ie, before randomization) and at 1, 4, 8, 12, 16, and 24 hours after randomization. These assessments were made by ward nursing staff blinded to the intervention and formed part of regular nursing observations undertaken at these times according to the hospital’s fractured neck of femur protocol. With these, records were also made of blood pressure, pulse rate, respiratory rate, and oxygen saturation. If measurements did not exactly coincide with time intervals, they were assigned to the closest standard time. The admitting orthopedic senior house officer was also unaware of study intervention, and therefore, analgesic prescription (although standard at Rotherham) was not influenced by patient allocation. Patients were not blinded to group allocation because our research ethics committee considered placebo injection unacceptable. The same blinded observer (AKF) abstracted all data. The total dose of morphine in the first 24 hours after admission or until surgery (if this was sooner), the type of fracture, and the time to surgery were recorded. This dose included morphine given to patients before radiography. Hospital notes were examined again at 6 months when the incidence of postoperative complications, time to discharge, and death rate were recorded. All data were compiled into a Microsoft Excel (Microsoft Corporation, Redmond, WA) spreadsheet before the study code was deciphered. The statistical analysis is concerned with the measurement of serial data (ie, repeat observations made at baseline and 1, 4, 8, 12, 16, and 24 hours after randomization). A potential problem with serial data is correlation between subsequent measurements. The statistical methods adopted to overcome this problem are outlined by Mathews et al,11 who suggest that serial data should be analyzed with the “method of summary measures.” The method considers the individual patient as the basic unit of assessment and constructs 1 or 2 numeric values that best summarize the patients’ response curve. Thus, for each variable, 1 or 2 important summary features are calculated per patient, and these summary features are then analyzed as if they were the raw data. Summary measures do not have the
2 3 0
problem of correlation over time and also allow for missing values (at one or more time points) to be taken into consideration. Thus, the data on pain scores, pulse rate, oxygen saturation, and respiratory rate were analyzed by calculating the area under the curve (essentially a weighted mean) after adjusting for baseline values in an analysis of covariance. However, much of the data at 24 hours were not recorded, and therefore, this time point was excluded from our analyses. Missing data at earlier time points were minimal and were incorporated as truncated means. For the pain score data, another summary measure (time to best response; ie, time at which the lowest pain score was reached) was also calculated. Results are presented as mean differences with 95% confidence intervals (CIs) for the differences. The Generalized Linear Interactive Modeling (GLIM; Royal Statistical Society, London, United Kingdom) statistical computer package was used for the statistical calculations. A nominal level of 5% statistical significance was used throughout. R E S U LT S
The characteristics of the 2 patient groups are given in Table 1. All but 1 (consultant) of the 14 ED medical staff performed nerve blocks. No one physician or grade performed more than 5 nerve blocks, and no clinically important variations were seen in nerve block efficacy between physicians. Among the study patients, none experienced adverse effects as a result of nerve block
Table 1.
Patient characteristics.
Variable Mean age, y (SD) Female sex, No. (%) Intertrochanteric, No. (%) Subcapital-transcervical, No. (%) Mean time to surgery, h (SD) Mean pain score on arrival (SD)
Study Patients (N=24)
Control Group (N=26)
76 (13) 17 (71) 15 (63) 9 (37) 29.3 (20.8) 2.8 (0.4)
80 (9) 18 (69) 15 (58) 11 (42) 27.4 (16.5) 2.7 (0.6)
ANNALS OF EMERGENCY MEDICINE
41:2
FEBRUARY 2003
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
administration. Patients receiving 3-in-1 nerve blocks recorded a faster time to reach the lowest pain score (Figure 3), and nerve block recipients required significantly less morphine per hour than control patients (Figure 4). Analgesia was required initially by virtually all patients, and if this was given before radiography, it has been included in the total dose. All patients were prescribed morphine for pain; one patient initially received diamorphine, and this dose was translated to an equivalent morphine dose. Only 2 patients received additional analgesia in the form of nonsteroidal anti-
Figure 3.
Mean pain score: time to best response (mean difference [95% CI]: –2.93 h [–5.48 to –0.38 h]).
7
inflammatory drugs: one in the study group and the other in the control group. No other forms of analgesia (eg, mechanical traction or anxiolytic drugs) were used. Morphine dose is presented as dose per hour because some patients underwent surgery less than 24 hours after admission. Morphine consumption was only included up to 24 hours after admission for those who underwent surgery later because we believed the nerve block could realistically only be expected to have an effect up to 24 hours after administration. There were no clinically important differences between the groups in respect to pulse rate, oxygen saturation, or respiratory rate at any time interval (Table 2). At 6 months’ follow-up, 3 study patients and 3 control patients had died. At 6 months, note review identified 2 lower respiratory tract infections in study patients and 4 in control patients. One patient in each group had a deep vein thrombosis on the same side as the fracture (and nerve block).
6 Time (h)
5
DISCUSSION
4 3 2 1 0
No block
3-in-1 nerve block
We show that 3-in-1 femoral nerve block is effective in producing analgesia in the ED for patients with fractured femoral neck. Analgesia is achieved substantially quicker in patients receiving the nerve blocks, and these patients require less morphine, reproducing the
Table 2.
Figure 4.
Mean morphine dose per hour (mean difference [95% CI]: –0.68 mg/h [–1.23 to –0.12 mg/h]).
Three-in-one femoral nerve block randomized controlled trial: Analysis of serial measurements.*
Morphine dose/h (mg)
1.4 1.2
Variable
No Block Block (Mean) (Mean)
Mean Difference (95% CI), Unadjusted
Mean Difference (95% CI), Adjusted
1 Pain score, 1.34 0-3 scale Pulse, beats/min 82.99 Oxygen saturation, 94.06 % Respiratory rate, 14.38 breaths/min
0.8 0.6 0.4 0.2 0
FEBRUARY 2003
No block
41:2
3-in-1 nerve block
ANNALS OF EMERGENCY MEDICINE
*Statistical
0.57 –0.77 (–1.10 to –0.53) –0.78 (–1.02 to –0.54) 83.01 94.87
0.02 (–6.80 to 6.83) 0.81 (–0.87 to 2.50)
–2.27 (–7.82 to 3.28) 0.11 (–1.45 to 1.68)
13.72 –0.66 (–1.97 to 0.55)
–0.53 (–1.58 to 0.52)
note: Adjusting for baseline values (age and sex) in an analysis of covariance.
2 3 1
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
findings of other studies. This has many potential advantages for patient care, and above all, our data show that the 3-in-1 nerve block technique can be generally applicable. The frequency of admissions of patients with fractured neck of femur was sufficiently high to ensure that the practical skills acquired were used enough to permit continued familiarity and confidence among junior physicians. Patients with fractured neck of femur are often in considerable pain, which, if inadequately controlled, might contribute to increased morbidity by reducing mobility and increasing anxiety and confusion. In the ED, the management of pain in these patients centers on the administration of parenteral morphine, which itself has attendant disadvantages, such as respiratory depression and increased confusion, in particularly susceptible elderly patients. Apart from obvious humanitarian reasons for seeking adequate analgesia, there are therefore sound clinical reasons why effective and safe analgesia for fractured neck of femur is in our patients’ best interest. A best practice review of the care of patients with fractured neck of femur included a femoral nerve block as analgesia in the ED.12 This is not common practice, and evidence in support of its use is sparse. Three published studies have described the successful use of femoral nerve block or 3-in-1 femoral nerve block in this situation.4-6 All studies relied on experts to administer the block, 2 had no control patients,5,6 and 1 included some young patients (mean age, 68 years) in a study population that was not clearly defined.7 Our study examines the applicability of 3-in-1 femoral nerve blocks in a district general ED in which admitted patients with fractured neck of femur might present at any time and on any day and are assessed and initially managed by junior physicians with hitherto little experience of regional anesthesia. Study limitations include the fact that patients were not blind to group allocation for ethical considerations. We believe, however, that the placebo effect alone could not account for such marked differences between the groups when the distraction of pain and circumstance was considerable. It is conceivable that nurse assessors
2 3 2
on the orthopedic ward might have been biased if patients revealed their group allocation to them. We believe this is unlikely because the recorded observations form part of normal assessment and patients would not draw attention to group allocation in the ED to either the patient or the nurse. It is possible that the nerve blocks worked earlier than the times given because we did not take more frequent pain measurements during the early period after randomization. Although orthopedic house staff might have identified patients with nerve block after neurologic examination on the ward, they were not responsible for recording pain scores and were bound by a protocol morphine prescription. We would emphasize that in undertaking this pragmatic study, we set out to establish whether this local anesthetic technique would be feasible and effective in a district hospital ED. Studies have shown numeric rating scales to be on par with visual analog scales, with target scores of 0 or 1 corresponding to no pain or mild pain. All our patients achieved either of these targets eventually (their best pain score). A peripheral nerve stimulator was not used because our aim was to apply the nerve block technique to a typically busy ED, where such items of equipment frequently are misplaced and are time-consuming to use. Despite improvement in pain scores and morphine consumption, there were no significant effects on physiologic parameters or complications after nerve block. Such parameters are influenced by many factors, and differences were not expected in this study, which was powered to examine the effect on pain scores and morphine consumption. Similarly, this study was not powered to examine safety of the block, and although no adverse events occurred, it is important that clinicians are aware of potential problems, such as inadvertent intravascular injection, infection, intraneural injection, and masking compartment syndrome. To our knowledge, the latter has only been described very rarely and confined to epidural anesthesia and lower leg fractures.13,14 It is important to be aware that injection of a large volume of bupivacaine in the close proximity of blood vessels is not without risk of serious adverse effects,
ANNALS OF EMERGENCY MEDICINE
41:2
FEBRUARY 2003
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
and therefore, it is important to achieve competence before 3-in-1 nerve blocks are performed and to have adequate monitoring and resuscitation facilities. Those patients who could potentially benefit most from this technique, the confused elderly, were excluded from this study for ethical reasons. If the practice of 3-in-1 femoral nerve block were adopted as part of a hospital’s fast-track package, there might be additional benefits for these especially vulnerable patients. Author contributions: AKF conceived the original idea, collected the data, and wrote the paper. ASR performed the statistical analysis and edited the paper. FLPH discussed core ideas and edited the paper. AKF takes responsibility for the paper as a whole. Received for publication February 27, 2002. Revision received August 18, 2002. Accepted for publication August 26, 2002. Address for reprints: Alan Fletcher, MRCP(UK), Department of Emergency Medicine, Northern General Hospital, Herries Road, Sheffield, United Kingdom S5 7AU; +44 0114 2714741, fax +44 0114 2560472; E-mail [email protected].
REFERENCES 1. Todd CJ, Freeman CJ, Camilleri-Ferrante C, et al. Differences in mortality after fracture of the hip: the East Anglian audit. BMJ. 1995;310:905-908. 2. Bonica JJ. Importance of effective pain control. Acta Anaesthesiol Scand. 1987;31(suppl 85):1-16. 3. Fournier R, Van Gessel E, Gaggerro G, et al. Postoperative analgesia with “3-in-1” femoral nerve block after prosthetic hip surgery. Can J Anesth. 1998;45:35-38. 4. Haddad FS, Williams RL. Femoral nerve block in extracapsular femoral neck fractures. J Bone Joint Surg. 1995;77:922-923. 5. Finlayson BJ, Underhill TJ. Femoral nerve block for analgesia in fractures of the femoral neck. Arch Emerg Med. 1988;5:173-176. 6. Van Leeuwen FL, Bronselaer K, Gilles M, et al. The “three in one” block as locoregional analgesia in the emergency department. Eur J Emerg Med. 2000;7:35-38. 7. Winnie AP, Rammamurthy S, Durrani Z. The inguinal paravascular technique of lumbar plexus anesthesia—the “3-in-1 block.” Anesth Analg. 1973;52:989-996. 8. Altman DG. Statistics and ethics in medical research. III. How large a sample? BMJ. 1980;281:1336-1338. 9.
Scott J, Huskisson E. Graphic representation of pain. Pain. 1976;2:175-184.
10. Gould T, Crosby D, Harmer M, et al. Policy for controlling pain after surgery: effect of sequential changes in management. BMJ. 1992;305:1187-1193. 11. Mathews JNS, Altman DG, Campbell MJ, et al. Analysis of serial measurements in medical research. BMJ. 1990;300:230-235. 12. Parker MJ. Managing an elderly patient with a fractured femur. BMJ. 2000;320:102103. 13. Price C, Ribeiro J, Kinnebrewt T. Compartment syndrome associated with postoperative epidural analgesia: a case report. J Bone Joint Surg. 1996;78:597-599. 14. Hyder N, Kessler S, Jennings AG, et al. Compartment syndrome in tibial shaft fracture missed because of a local nerve block. J Bone Joint Surg. 1996;78:499-500.
FEBRUARY 2003
41:2
ANNALS OF EMERGENCY MEDICINE
2 3 3
Three-in-One Femoral Nerve Block as Analgesia for Fractured Neck of Femur in the Emergency Department: A Randomized, Controlled Trial
Alan K. Fletcher, MRCP(UK) Alan S. Rigby, MSc Francis L. P. Heyes, FRCS From the Department of Emergency Medicine, Rotherham General Hospital, Rotherham, United Kingdom (Fletcher, Heyes), and the Department of Statistics, University of Sheffield, Sheffield, United Kingdom (Rigby).
Study objective: We determine whether 3-in-1 femoral nerve block is effective as analgesia for fractured neck of femur when administered by emergency physicians. Methods: This was a prospective, randomized controlled trial with blinded assessors conducted in a district general hospital emergency department in the United Kingdom. Over a 6-month period, all patients with fractured neck of femur were considered for study. Patients were randomly assigned to receive 3-in-1 nerve block with bupivacaine plus intravenous morphine or intravenous morphine. An accreditation package for all ED medical staff was devised to ensure competence in the technique of 3-in-1 nerve block. Pain scores were recorded on arrival and at intervals up to 24 hours after admission. Morphine consumption in the first 24 hours was recorded. Results: Ninety-four patients sustained fractured neck of femur during the study period; 50 were studied. Of 44 not studied, 42 were confused, 1 did not consent, and 1 was overlooked. Patients receiving 3-in-1 nerve blocks recorded a faster time to reach the lowest pain score: 2.88 hours for patients with nerve block and 5.81 hours for control patients (mean difference –2.93 h; 95% confidence interval [CI] –5.48 to –0.38 h). Nerve block recipients required significantly less morphine per hour than control patients (mean of 0.49 mg/h versus 1.17 mg/h; mean difference –0.68 mg/h; 95% CI –1.23 to –0.12 mg/h). Conclusion: Three-in-one femoral nerve block is an effective method of providing analgesia to patients with fractured neck of femur in the ED. All grades of medical staff were able to apply and consolidate this skill. [Ann Emerg Med. 2003;41:227-233.]
Copyright © 2003 by the American College of Emergency Physicians. 0196-0644/2003/$30.00 + 0 doi:10.1067/mem.2003.51
FEBRUARY 2003
41:2
ANNALS OF EMERGENCY MEDICINE
2 2 7
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
INTRODUCTION
Fractured neck of femur causes significant mortality and morbidity in elderly patients. Even with optimum so-called fast tracking of these patients through hospital admission, surgery, and rehabilitation, mortality rates of 18% at 90 days can be anticipated.1 Patients with fractured neck of femur are often in considerable pain, and minimizing this has important implications for patient comfort and reduction of morbidity and mortality.2 The 3-in-1 femoral nerve block is a safe and widely practiced local anesthetic technique used to supplement anesthesia and provide postoperative analgesia after hip surgery.3 Local anesthetic infiltrates the femoral nerve sheath and tracks cranially and laterally to anesthetize the femoral and obturator nerves, the lower cords of the lumbar plexus, and the lateral cutaneous nerve of the thigh (Figure 1). Two previous studies4,5 have shown that femoral nerve block provides effective analgesia for patients with fractured neck of femur in the emergency department, but each relied on one practitioner’s expertise for block administration and might consequently have been relatively selective in patients recruited. One of these studies5 did not have a control group. Another recent study6 has successfully used 3-in-1 femoral nerve block for fractures of the upper femur, but the study population was not clearly described, had no control group, and once again relied on expert block administration. This study was conducted to assess whether 3-in-1 femoral nerve block is effective when taught to and implemented by ED medical staff. Our study was powered to detect a 40% reduction in pain scores in those given a nerve block. The applicability of the technique has been evaluated in a United Kingdom district general hospital ED setting.
obtained. Patients were recruited during a 6-month period, from February until August, coinciding with senior house officer (intern) posts. Two consultants, 4 middle-grade physicians, and 7 senior house officers staff our ED. An accreditation package was developed for all ED medical staff to ensure competence in the technique of 3-in-1 femoral nerve block before the study commenced. The accreditation package comprised a half-hour structured session of instruction in the technique, including anatomy, pharmacology, and local anesthetic toxicity, followed by supervised practice on a resuscitation mannequin. A nerve block performed on a patient with fractured neck of femur was also supervised by the same observer (AKF). These patients were not included in the study. Competence was considered achieved when supervised practice
Figure 1.
Injection technique of 3-in-1 femoral nerve block. Inguinal ligament Nerve Artery Injection of Vein 20 mL of 0.5% plain bupivacaine into the neural sheath
30°
Digital pressure distal to injection site during and 5 minutes after injection
M AT E R I A L S A N D M E T H O D S
The study was conducted in the ED of our mediumsized district general hospital in the United Kingdom. The ED handles 65,000 new attendances per year. Local research ethical committee approval for the study was
2 2 8
ANNALS OF EMERGENCY MEDICINE
41:2
FEBRUARY 2003
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
showed that key safety and technical points had been achieved and when the supervised patient nerve block was considered safe and satisfactory. As all physicians achieved competence, it was possible to consider all patients with all types of fractured neck of femur for this study. Patients who were confused (and therefore unable to give informed consent), had a bleeding diathesis or were taking warfarin, had local or systemic infection, or had previous hypersensitivity to local anesthetics were excluded (Figure 2). Informed consent was obtained once fracture was confirmed with radiography, and patients were randomly assigned to 1 of 2 groups. Patients in the first group were control patients and received analgesia in the conventional fashion for our district general hospital (ie, 5 to 10 mg of intravenous morphine, available
Figure 2.
Study flow diagram.
90 patients with fractured neck of femur 42 excluded because of confusion 1 patient refused consent 1 patient overlooked 50 patients included
24 randomized to receive IV morphine
26 randomized to receive 3-in-1 femoral nerve block with 20 mL of 0.5% bupivacaine
Pain scores at 0, 1, 4, 8, 12, 16, and 24 h Morphine consumption at 24 h
Complications and mortality recorded after case note review at 6 mo
FEBRUARY 2003
41:2
ANNALS OF EMERGENCY MEDICINE
hourly until analgesia was achieved). This prescription continued until surgery. Patients in the second group received a 3-in-1 femoral nerve block with 20 mL of 0.5% plain bupivacaine. This was performed by using the technique described by Winnie et al7; a 21-gauge needle without a peripheral nerve stimulator was used instead of a nerve block needle for convenience and practicality. Aseptic technique was observed, and patients were monitored with ECG and basic observations. An intravenous cannula was sited before nerve block in all cases, and its patency was confirmed. The femoral pulse was palpated, and the nerve block was inserted 1 cm lateral to the pulse once paresthesias were elicited to localize the nerve. The local anesthetic was injected in a cranial direction and with pressure distal to the needle during and shortly after injection to encourage the local anesthetic to track cephalad. If patients required analgesia before radiography, they received incremental doses of intravenous morphine. Sample size was estimated by using the method described by Altman8 from data of previously published studies and from 10 pilot patients in whom a pain score reduction of 50% was observed. We calculated that a total study size of 50 patients would be required to achieve 80% power to observe a 40% pain reduction score difference at the 0.05 significance level. A pilot study indicated that approximately half of all patients with fractured neck of femur would be ineligible because of confusion, and this was allowed for in assessment of study feasibility. The data from the pilot study are not included in the results. The randomization sequence was derived from a random number generator, and allocation concealment was achieved by means of the sealed opaque envelope method. Pain on movement was assessed by using a numeric rating scale that has been adopted as a hospital standard for measuring pain, ensuring that scores were reproducible. This pain scale has been validated before.9,10 We decided to adopt a numeric rating scale instead of a visual analog scale partly because of the difficulties we encountered with patient comprehension of visual analog scales in the pilot study. Pain was graded as 0, 1, 2, or 3 corresponding to no pain, mild pain, moderate pain,
2 2 9
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
or severe pain, respectively. Pain on movement was assessed on arrival (ie, before randomization) and at 1, 4, 8, 12, 16, and 24 hours after randomization. These assessments were made by ward nursing staff blinded to the intervention and formed part of regular nursing observations undertaken at these times according to the hospital’s fractured neck of femur protocol. With these, records were also made of blood pressure, pulse rate, respiratory rate, and oxygen saturation. If measurements did not exactly coincide with time intervals, they were assigned to the closest standard time. The admitting orthopedic senior house officer was also unaware of study intervention, and therefore, analgesic prescription (although standard at Rotherham) was not influenced by patient allocation. Patients were not blinded to group allocation because our research ethics committee considered placebo injection unacceptable. The same blinded observer (AKF) abstracted all data. The total dose of morphine in the first 24 hours after admission or until surgery (if this was sooner), the type of fracture, and the time to surgery were recorded. This dose included morphine given to patients before radiography. Hospital notes were examined again at 6 months when the incidence of postoperative complications, time to discharge, and death rate were recorded. All data were compiled into a Microsoft Excel (Microsoft Corporation, Redmond, WA) spreadsheet before the study code was deciphered. The statistical analysis is concerned with the measurement of serial data (ie, repeat observations made at baseline and 1, 4, 8, 12, 16, and 24 hours after randomization). A potential problem with serial data is correlation between subsequent measurements. The statistical methods adopted to overcome this problem are outlined by Mathews et al,11 who suggest that serial data should be analyzed with the “method of summary measures.” The method considers the individual patient as the basic unit of assessment and constructs 1 or 2 numeric values that best summarize the patients’ response curve. Thus, for each variable, 1 or 2 important summary features are calculated per patient, and these summary features are then analyzed as if they were the raw data. Summary measures do not have the
2 3 0
problem of correlation over time and also allow for missing values (at one or more time points) to be taken into consideration. Thus, the data on pain scores, pulse rate, oxygen saturation, and respiratory rate were analyzed by calculating the area under the curve (essentially a weighted mean) after adjusting for baseline values in an analysis of covariance. However, much of the data at 24 hours were not recorded, and therefore, this time point was excluded from our analyses. Missing data at earlier time points were minimal and were incorporated as truncated means. For the pain score data, another summary measure (time to best response; ie, time at which the lowest pain score was reached) was also calculated. Results are presented as mean differences with 95% confidence intervals (CIs) for the differences. The Generalized Linear Interactive Modeling (GLIM; Royal Statistical Society, London, United Kingdom) statistical computer package was used for the statistical calculations. A nominal level of 5% statistical significance was used throughout. R E S U LT S
The characteristics of the 2 patient groups are given in Table 1. All but 1 (consultant) of the 14 ED medical staff performed nerve blocks. No one physician or grade performed more than 5 nerve blocks, and no clinically important variations were seen in nerve block efficacy between physicians. Among the study patients, none experienced adverse effects as a result of nerve block
Table 1.
Patient characteristics.
Variable Mean age, y (SD) Female sex, No. (%) Intertrochanteric, No. (%) Subcapital-transcervical, No. (%) Mean time to surgery, h (SD) Mean pain score on arrival (SD)
Study Patients (N=24)
Control Group (N=26)
76 (13) 17 (71) 15 (63) 9 (37) 29.3 (20.8) 2.8 (0.4)
80 (9) 18 (69) 15 (58) 11 (42) 27.4 (16.5) 2.7 (0.6)
ANNALS OF EMERGENCY MEDICINE
41:2
FEBRUARY 2003
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
administration. Patients receiving 3-in-1 nerve blocks recorded a faster time to reach the lowest pain score (Figure 3), and nerve block recipients required significantly less morphine per hour than control patients (Figure 4). Analgesia was required initially by virtually all patients, and if this was given before radiography, it has been included in the total dose. All patients were prescribed morphine for pain; one patient initially received diamorphine, and this dose was translated to an equivalent morphine dose. Only 2 patients received additional analgesia in the form of nonsteroidal anti-
Figure 3.
Mean pain score: time to best response (mean difference [95% CI]: –2.93 h [–5.48 to –0.38 h]).
7
inflammatory drugs: one in the study group and the other in the control group. No other forms of analgesia (eg, mechanical traction or anxiolytic drugs) were used. Morphine dose is presented as dose per hour because some patients underwent surgery less than 24 hours after admission. Morphine consumption was only included up to 24 hours after admission for those who underwent surgery later because we believed the nerve block could realistically only be expected to have an effect up to 24 hours after administration. There were no clinically important differences between the groups in respect to pulse rate, oxygen saturation, or respiratory rate at any time interval (Table 2). At 6 months’ follow-up, 3 study patients and 3 control patients had died. At 6 months, note review identified 2 lower respiratory tract infections in study patients and 4 in control patients. One patient in each group had a deep vein thrombosis on the same side as the fracture (and nerve block).
6 Time (h)
5
DISCUSSION
4 3 2 1 0
No block
3-in-1 nerve block
We show that 3-in-1 femoral nerve block is effective in producing analgesia in the ED for patients with fractured femoral neck. Analgesia is achieved substantially quicker in patients receiving the nerve blocks, and these patients require less morphine, reproducing the
Table 2.
Figure 4.
Mean morphine dose per hour (mean difference [95% CI]: –0.68 mg/h [–1.23 to –0.12 mg/h]).
Three-in-one femoral nerve block randomized controlled trial: Analysis of serial measurements.*
Morphine dose/h (mg)
1.4 1.2
Variable
No Block Block (Mean) (Mean)
Mean Difference (95% CI), Unadjusted
Mean Difference (95% CI), Adjusted
1 Pain score, 1.34 0-3 scale Pulse, beats/min 82.99 Oxygen saturation, 94.06 % Respiratory rate, 14.38 breaths/min
0.8 0.6 0.4 0.2 0
FEBRUARY 2003
No block
41:2
3-in-1 nerve block
ANNALS OF EMERGENCY MEDICINE
*Statistical
0.57 –0.77 (–1.10 to –0.53) –0.78 (–1.02 to –0.54) 83.01 94.87
0.02 (–6.80 to 6.83) 0.81 (–0.87 to 2.50)
–2.27 (–7.82 to 3.28) 0.11 (–1.45 to 1.68)
13.72 –0.66 (–1.97 to 0.55)
–0.53 (–1.58 to 0.52)
note: Adjusting for baseline values (age and sex) in an analysis of covariance.
2 3 1
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
findings of other studies. This has many potential advantages for patient care, and above all, our data show that the 3-in-1 nerve block technique can be generally applicable. The frequency of admissions of patients with fractured neck of femur was sufficiently high to ensure that the practical skills acquired were used enough to permit continued familiarity and confidence among junior physicians. Patients with fractured neck of femur are often in considerable pain, which, if inadequately controlled, might contribute to increased morbidity by reducing mobility and increasing anxiety and confusion. In the ED, the management of pain in these patients centers on the administration of parenteral morphine, which itself has attendant disadvantages, such as respiratory depression and increased confusion, in particularly susceptible elderly patients. Apart from obvious humanitarian reasons for seeking adequate analgesia, there are therefore sound clinical reasons why effective and safe analgesia for fractured neck of femur is in our patients’ best interest. A best practice review of the care of patients with fractured neck of femur included a femoral nerve block as analgesia in the ED.12 This is not common practice, and evidence in support of its use is sparse. Three published studies have described the successful use of femoral nerve block or 3-in-1 femoral nerve block in this situation.4-6 All studies relied on experts to administer the block, 2 had no control patients,5,6 and 1 included some young patients (mean age, 68 years) in a study population that was not clearly defined.7 Our study examines the applicability of 3-in-1 femoral nerve blocks in a district general ED in which admitted patients with fractured neck of femur might present at any time and on any day and are assessed and initially managed by junior physicians with hitherto little experience of regional anesthesia. Study limitations include the fact that patients were not blind to group allocation for ethical considerations. We believe, however, that the placebo effect alone could not account for such marked differences between the groups when the distraction of pain and circumstance was considerable. It is conceivable that nurse assessors
2 3 2
on the orthopedic ward might have been biased if patients revealed their group allocation to them. We believe this is unlikely because the recorded observations form part of normal assessment and patients would not draw attention to group allocation in the ED to either the patient or the nurse. It is possible that the nerve blocks worked earlier than the times given because we did not take more frequent pain measurements during the early period after randomization. Although orthopedic house staff might have identified patients with nerve block after neurologic examination on the ward, they were not responsible for recording pain scores and were bound by a protocol morphine prescription. We would emphasize that in undertaking this pragmatic study, we set out to establish whether this local anesthetic technique would be feasible and effective in a district hospital ED. Studies have shown numeric rating scales to be on par with visual analog scales, with target scores of 0 or 1 corresponding to no pain or mild pain. All our patients achieved either of these targets eventually (their best pain score). A peripheral nerve stimulator was not used because our aim was to apply the nerve block technique to a typically busy ED, where such items of equipment frequently are misplaced and are time-consuming to use. Despite improvement in pain scores and morphine consumption, there were no significant effects on physiologic parameters or complications after nerve block. Such parameters are influenced by many factors, and differences were not expected in this study, which was powered to examine the effect on pain scores and morphine consumption. Similarly, this study was not powered to examine safety of the block, and although no adverse events occurred, it is important that clinicians are aware of potential problems, such as inadvertent intravascular injection, infection, intraneural injection, and masking compartment syndrome. To our knowledge, the latter has only been described very rarely and confined to epidural anesthesia and lower leg fractures.13,14 It is important to be aware that injection of a large volume of bupivacaine in the close proximity of blood vessels is not without risk of serious adverse effects,
ANNALS OF EMERGENCY MEDICINE
41:2
FEBRUARY 2003
FEMORAL NERVE BLOCK AS AN ANALGESIA Fletcher, Rigby & Heyes
and therefore, it is important to achieve competence before 3-in-1 nerve blocks are performed and to have adequate monitoring and resuscitation facilities. Those patients who could potentially benefit most from this technique, the confused elderly, were excluded from this study for ethical reasons. If the practice of 3-in-1 femoral nerve block were adopted as part of a hospital’s fast-track package, there might be additional benefits for these especially vulnerable patients. Author contributions: AKF conceived the original idea, collected the data, and wrote the paper. ASR performed the statistical analysis and edited the paper. FLPH discussed core ideas and edited the paper. AKF takes responsibility for the paper as a whole. Received for publication February 27, 2002. Revision received August 18, 2002. Accepted for publication August 26, 2002. Address for reprints: Alan Fletcher, MRCP(UK), Department of Emergency Medicine, Northern General Hospital, Herries Road, Sheffield, United Kingdom S5 7AU; +44 0114 2714741, fax +44 0114 2560472; E-mail [email protected].
REFERENCES 1. Todd CJ, Freeman CJ, Camilleri-Ferrante C, et al. Differences in mortality after fracture of the hip: the East Anglian audit. BMJ. 1995;310:905-908. 2. Bonica JJ. Importance of effective pain control. Acta Anaesthesiol Scand. 1987;31(suppl 85):1-16. 3. Fournier R, Van Gessel E, Gaggerro G, et al. Postoperative analgesia with “3-in-1” femoral nerve block after prosthetic hip surgery. Can J Anesth. 1998;45:35-38. 4. Haddad FS, Williams RL. Femoral nerve block in extracapsular femoral neck fractures. J Bone Joint Surg. 1995;77:922-923. 5. Finlayson BJ, Underhill TJ. Femoral nerve block for analgesia in fractures of the femoral neck. Arch Emerg Med. 1988;5:173-176. 6. Van Leeuwen FL, Bronselaer K, Gilles M, et al. The “three in one” block as locoregional analgesia in the emergency department. Eur J Emerg Med. 2000;7:35-38. 7. Winnie AP, Rammamurthy S, Durrani Z. The inguinal paravascular technique of lumbar plexus anesthesia—the “3-in-1 block.” Anesth Analg. 1973;52:989-996. 8. Altman DG. Statistics and ethics in medical research. III. How large a sample? BMJ. 1980;281:1336-1338. 9.
Scott J, Huskisson E. Graphic representation of pain. Pain. 1976;2:175-184.
10. Gould T, Crosby D, Harmer M, et al. Policy for controlling pain after surgery: effect of sequential changes in management. BMJ. 1992;305:1187-1193. 11. Mathews JNS, Altman DG, Campbell MJ, et al. Analysis of serial measurements in medical research. BMJ. 1990;300:230-235. 12. Parker MJ. Managing an elderly patient with a fractured femur. BMJ. 2000;320:102103. 13. Price C, Ribeiro J, Kinnebrewt T. Compartment syndrome associated with postoperative epidural analgesia: a case report. J Bone Joint Surg. 1996;78:597-599. 14. Hyder N, Kessler S, Jennings AG, et al. Compartment syndrome in tibial shaft fracture missed because of a local nerve block. J Bone Joint Surg. 1996;78:499-500.
FEBRUARY 2003
41:2
ANNALS OF EMERGENCY MEDICINE
2 3 3