Total Hip Arthroplasty 2010

Total Hip Arthroplasty-Specific Evidence

Arguments for…

• One study (n=217) assessed the effect of valdecoxib (20 mg or 40 mg orally) administered pre-operatively and then postoperatively every 12 h, compared with placebo, on postoperative pain scores and on the requirement for supplementary analgesia 48 h (Camu 2002).
  • Oral valdecoxib was effective in reducing postoperative pain scores at two doses (20 mg or 40 mg) compared with placebo when administered pre-operatively and then postoperatively (p<0.05 for 6, 12 and 18 h) (Camu 2002)
  • Oral valdecoxib was effective in reducing the requirement for supplementary analgesia compared with placebo (p<0.001) (Camu 2002)
  • Oral valdecoxib was not associated with a significant increase in the incidence of nausea or vomiting compared with placebo (Camu 2002)

Transferable Evidence from Other Procedures

Arguments for…

• COX-2-selective inhibitors significantly reduced postoperative pain and the requirement for supplementary analgesia compared with placebo following knee arthroplasty
  • Parecoxib sodium (20 or 40 mg intravenously) significantly reduced postoperative pain scores and the requirement for supplementary analgesia compared with placebo following total knee arthroplasty (two studies, n=403) (Hubbard 2003, Rasmussen 2002).
  • Valdecoxib (40 or 80 mg orally) was superior to placebo for postoperative pain scores (p<0.05) and for the requirement for supplementary analgesia (p<0.05) compared with placebo following total knee arthroplasty (n=209) (Reynolds 2003)
• Parecoxib (20 or 40 mg) was as effective as ketorolac (30 mg) administered intravenously for postoperative pain scores (for level, onset and duration of analgesia) following total knee arthroplasty (n=208) (Rasmussen 2002)
• Randomised endoscopic trials in healthy volunteers have shown that COX-2-selective inhibitors are associated with a lower incidence of upper gastrointestinal ulceration compared with conventional NSAIDs for short-term use (Harris 2001, Stoltz 2002)
• Studies in healthy volunteers demonstrated that COX-2-selective inhibitors had no effect on platelet aggregation (Greenberg 2000, Leese 2000, Leese 2002, Noveck 2001, Noveck 2004, Wilner 2002) or bleeding time (Leese 2000, Leese 2002, Noveck 2001) compared with placebo
• Clinical studies investigating the response to oral challenge with COX-2-selective inhibitors in patients with aspirin-induced asthma have demonstrated that COX-2-selective inhibitors do not have an effect on respiratory function (Bavbek 2004, Gyllfors 2003, Martin-Garcia 2003, Sanchez Borges 2005, West 2003, Woessner 2004)
• A study to assess the safety of the COX-2-selective inhibitors parecoxib and valdecoxib following non-cardiac general surgery (including gastrointestinal, orthopaedic, gynaecological, urological, and thoracic surgeries) showed no difference in the incidence of cardiovascular thromboembolic events, renal dysfunction/failure, gastrointestinal ulcer complications, and surgical wound-healing complications, compared with placebo (n=1062) (in press, Anesthesiology)

Arguments against…

• Two clinical trials showed that in patients who had undergone CABG surgery, COX-2-selective inhibitors (valdecoxib and parecoxib) were associated with a higher rate of serious cardiovascular thromboembolic events (including myocardial infarction) compared with placebo (Nussmeier 2005, Ott 2003)
• Hypersensitivity reactions and serious skin reactions (e.g. toxic epidermal necrolysis, Stevens-Johnson syndrome, and erythema multiforme) can occur with all COX-2-selective inhibitors. Serious skin reactions have been reported in association with valdecoxib at a higher rate than with other COX-2-selective agents (EPAR Bextra 2004, EPAR Dynastat 2004, CHMP Updated Questions & Answers on Cox 2 Inhibitors 2005)
• Conventional NSAIDs and COX-2-selective inhibitors have been associated with an increased risk of transient hepatotoxicity. Cases of acute hepatic failure have also been reported. Elderly females, with autoimmune disease, and taking other potentially hepatotoxic drugs, may be most susceptible (O’Connor 2003, Teoh 2003)
• Short-term use of COX-2-selective inhibitors or conventional NSAIDs can increase the risk of transient renal impairment, especially in patients with existing renal dysfunction, and both types of agent have been associated with cases of acute renal failure in high-risk patients (Cheng 2004, Harris 2002, Perazella 2002)
• COX-2-selective inhibitors and conventional NSAIDs may elevate blood pressure, particularly in hypertensive patients receiving antihypertensive medication (Cheng 2004, Harris 2002, Perazella 2002)
• Although there is some concern that COX-2-selective inhibitors may impair wound healing, evidence from animal and clinical studies is conflicting (Blomme 2003, Cahill 2004, Nussmeier 2005, Ott 2003)
• Conventional NSAIDs and COX-2-selective inhibitors have been associated with impaired bone healing in various animal models and in vitro studies (Gajraj 2003, Gerstenfeld 2004, Harder 2003, Seidenberg 2004) but clinical studies to evaluate the effects of conventional NSAIDs and COX-2-specific inhibitors on bone healing are lacking

Clinical Practice

• Rofecoxib has been withdrawn from the market, and marketing of valdecoxib has been suspended in most countries.
• There is no difference in the analgesic efficacy between the different COX2-specific inhibitors.

PROSPECT Recommendations

• COX-2-selective inhibitors are recommended for their analgesic and opioid-sparing effect (grade A). They should be given in combination with strong opioids for high-intensity pain (grade A), or with weak opioids for moderate- or low-intensity pain (grade D). Voting agreement: 8/8
• COX-2-selective inhibitors may be preferred to conventional NSAIDs in the peri-operative setting, in patients who have an increased risk of bleeding, or who are at risk of gastroduodenal ulcer/erosion (grade B).
• The use of COX-2-selective inhibitors should depend upon assessment of individual patient risks (cardiovascular morbidity, renal function and hepatic function [grade B]. Voting agreement: 8/8
• COX-2-selective inhibitors should be used in patients with aspirin-sensitive asthma in preference to NSAIDs (Grade B). Voting agreement: 8/8
• Although there is concern about impairment of bone-healing with COX-2-selective inhibitors, there are no conclusive clinical studies to show that they have detrimental effects

Total Hip Arthroplasty-Specific Evidence

Study details: Seven studies (n=380) have examined the effect of postoperative conventional NSAIDs compared with placebo on postoperative pain in total hip arthroplasty (Dahl 1995, Fletcher 1995, Fogarty 1995b, Iohom 2002, Laitinen 1992, Segstro 1991, Serpell 1989). One study compared the following conventional NSAIDs postoperatively, when the spinal anaesthesia had worn off: ketorolac 30 mg IV loading dose plus 90 mg IV infusion over 15.5 h, diclofenac 75 mg IV loading dose plus 75 mg IV over 15.5 h, and ketoprofen 100 mg IV loading dose plus 100 mg IV over 15.5 h (Kostamovaara 1998).

Arguments for…

• Conventional NSAIDs were effective in reducing pain scores compared with placebo up to 32 h following operation
  • Qualitative analysis: of seven studies, six demonstrated superiority for conventional NSAIDs compared with placebo for VAS scores: oral ibuprofen 800 mg (Dahl 1995); intravenous ketorolac 60 mg (Fletcher 1995); intramuscular ketorolac 30 mg (Fogarty 1995b); diclofenac 75 mg intravenous loading dose with 5 mg/h infusion (Laitinen 1992); rectal indomethacin 100 mg (Segstro 1991); oral dexketoprofen 25 mg (Iohom 2002). One study showed no significant benefit for oral piroxicam 40 mg versus placebo (Serpell 1989).
  • Quantitative analysis: conventional NSAIDs were superior for VAS scores for the 0–8 h grouping (3 studies, WMD -9.48 [-18.06, -0.90], p=0.03), the 8–16 h grouping (2 studies WMD -4.12 [-7.56, -0.68], p=0.02) and the 16–32 h grouping versus placebo (2 studies WMD -8.46 [-14.88, -2.05], p=0.01).
• Conventional NSAIDs were effective for reducing the requirement for supplementary analgesia compared with placebo
  • Qualitative analysis: of seven studies, six showed significant reductions in supplementary analgesia use compared with placebo, and one arm of one study showed no significant benefit (intravenous ketorolac 60 mg at skin closure) (Fletcher 1995).
  • Quantitative analysis: conventional NSAIDs significantly reduced 24-h postoperative morphine consumption (WMD -9.07 mg [-14.16, -3.97], p=0.0005)
• Similar incidence of nausea and vomiting compared with placebo
  • Qualitative analysis: of seven studies, one study showed superiority for conventional NSAIDs in reducing nausea (oral dexketoprofen 25 mg) (Iohom 2002) and six showed no significant difference for the incidence of nausea or vomiting, when compared with placebo.
  • Quantitative analysis: conventional NSAIDs had no significant effect on the incidence of nausea (OR 0.74 [0.29, 1.91], p=0.5) or vomiting (OR 1.20 [0.60, 2.39], p=0.6).
• Postoperative ketorolac, diclofenac and ketoprofen were similar for pain scores and use of supplementary PCA fentanyl (Kostamovaara 1998)

Arguments against…

• Inconsistent evidence for the effect of conventional NSAIDs on the time to first analgesic request
  • The time to first analgesic request was recorded in two studies: compared with placebo, oral dexketoprofen 25 mg was superior for this outcome (p=0.03; n=30) (Iohom 2002) and ketorolac 30 mg intramuscular had no significant effect compared with placebo (n=60) (Fogarty 1995b).

Transferable Evidence from Other Procedures

Arguments for…

• Ketorolac (10 mg orally or 30 mg intravenously) was significantly more effective than placebo for reducing pain scores as shown in a number of studies of orthopaedic surgery (including knee, hip and spinal procedures) (Kinsella 1992, Maslanka 1994, Rasmussen 2002)
• Ketoprofen 100 mg orally significantly reduced VAS scores compared with placebo following spinal fusion (p<0.001; n=50) (Aubrun 2000)
• Oral keterolac 10 mg was as effective as intramuscular morphine 5 mg or 10 mg for reducing postoperative pain scores after orthopaedic procedures (DeAndrade 1994, Maslanka 1994).
• Intravenous ketorolac 30 mg was more effective than intravenous morphine 4 mg for reducing postoperative pain scores in one study of total knee replacement surgery (n=196) (Rasmussen 2002)
• Ketoprofen significantly reduced supplementary analgesia requirements compared with placebo following spinal fusion (p=0.002; n=50) (Aubrun 2000)
• Ketorolac (30 mg bolus, then 5 mg/h infusion for 24 h) reduced the duration of myocardial ischaemic episodes – compared with placebo – for 24 h following the operation in patients undergoing total hip or knee arthroplasty (Beattie 1997)
• One randomised trial in patients undergoing transurethral prostatectomy showed that diclofenac did not affect total blood loss compared with placebo (Bricker 1987)

Arguments against…

• Randomised endoscopic trials in healthy volunteers have shown that COX-2-selective inhibitors are associated with a lower incidence of upper gastrointestinal ulceration compared with conventional NSAIDs for short-term use (Harris 2001, Stoltz 2002)
• Randomised trials in healthy volunteers showed that conventional NSAIDs (ketorolac, naproxen or diclofenac) reduced the platelet aggregation response compared with placebo; ketorolac and naproxen also prolonged bleeding time compared with placebo (Greer 1999, Leese 2000, Leese 2002)
• Meta-analyses of randomised, controlled trials showed that peri-operative conventional NSAIDs increased the risk of postoperative bleeding requiring treatment and/or the risk of re-operation for haemostasis after tonsillectomy compared with controls (Marret 2003, Möiniche 2003)
• A randomised trial in healthy volunteers showed that three conventional NSAIDs (diclofenac, ketorolac and ketoprofen) caused a reversible platelet dysfunction (Niemi 1997)
• A randomised controlled trial showed that three conventional NSAIDs (diclofenac, ketorolac and ketoprofen) were associated with a similar incidence of surgical site bleeding after elective surgery (Forrest 2002)
• Aspirin and conventional NSAIDs can induce asthma attacks in patients with aspirin-exacerbated respiratory disease (Stevenson 2004). Aspirin-induced asthma occurs in approximately 4–10% of the adult asthmatic population (Hamad 2004)
• Conventional NSAIDs and COX-2-selective inhibitors have been associated with an increased risk of transient hepatotoxicity. Cases of acute hepatic failure have also been reported. Elderly females, with autoimmune disease, and taking other potentially hepatotoxic drugs, may be most susceptible (O’Connor 2003, Teoh 2003)
• Short-term use of conventional NSAIDs or COX-2-selective inhibitors can increase the risk of transient renal impairment, especially in patients with existing renal dysfunction, and both types of agent have been associated with cases of acute renal failure in high-risk patients (Cheng 2004, Harris 2002, Perazella 2002)
• Conventional NSAIDs and COX-2-selective inhibitors may elevate blood pressure, particularly in hypertensive patients receiving antihypertensive medication (Cheng 2004, Harris 2002, Perazella 2002)
• Conventional NSAIDs and COX-2-selective inhibitors have been associated with impaired bone healing in various animal models and in vitro studies (Gajraj 2003, Gerstenfeld 2004, Harder 2003, Seidenberg 2004) but clinical studies to evaluate the effects of conventional NSAIDs and COX-2-specific inhibitors on bone healing are lacking

Clinical Practice

[None cited]

PROSPECT Recommendations

• Conventional NSAIDs are recommended for their analgesic and opioid-sparing effect (grade A). They should be given in combination with strong opioids for high-intensity pain (grade A), or with weak opioids for moderate- or low-intensity pain (grade D). Voting agreement: 8/8
• Conventional NSAIDs are not recommended in patients who have an increased risk of bleeding, aspirin-sensitive asthma or who are at risk of gastroduodenal ulcer/erosion (grade A). Voting agreement: 8/8
• The use of conventional NSAIDs should depend upon assessment of individual patient risks (cardiovascular morbidity, renal function and hepatic function) (grade B). Voting agreement: 8/8
• Although there is concern about impairment of bone-healing with conventional NSAIDs, there are no conclusive clinical studies to show that they have detrimental effects

Total Hip Arthroplasty-Specific Evidence

[None cited]

Transferable Evidence from Other Procedures

• Studies of ketamine in abdominal, orthopaedic, gastric, hepatic and renal surgery showed a reduction in postoperative pain and opioid use when used as an adjuvant to morphine, either epidurally or intravenously (Subramaniam 2004)

Clinical Practice

[None cited]

PROSPECT Recommendations

• Ketamine is not recommended at this time due to limited procedure-specific evidence, although there are some positive analgesic data from other procedures and in opiod-tolerant patients (grade B). Voting agreement: 8/8

Total Hip Arthroplasty-Specific Evidence

Study details: No total hip arthroplasty-specific studies were found within the parameters of the systematic review that compared the analgesic effects of postoperative strong opioids with placebo. The studies found have compared different strong opioids and various postoperative dosing regimens (Fee 1989, Frater 1989, Keita 2003, McCormack 1993, Robinson 1991).

One study compared oral morphine sulphate sustained-release tablet, 20 mg with premedication and every 12 h thereafter for 48 h, with IM morphine 10 mg after regression of the spinal motor block and at 6 h intervals thereafter (Bourke 2000)

Arguments for…

• Morphine (0.15 mg/kg intramuscularly) was superior to nalbuphine (0.3 mg/kg intramuscularly) for postoperative pain scores (p<0.02) and for the use of supplementary analgesia (p<0.05; n=80) (Fee 1989)
• Morphine (2 mg bolus) was superior to meptazinol (20 mg bolus) – each given by PCA – for 8 h postoperative pain scores (p<0.05), and for the incidence of nausea and vomiting (p<0.05; n=49) (Frater 1989)
• Morphine 2 mg and diamorphine 1 mg – given by PCA – were equally effective for reducing postoperative pain scores (n=40) (Robinson 1991)
• IV PCA morphine 1 mg 8-min lockout was superior to 4-hourly intramuscular morphine 0.1 mg/kg for postoperative pain scores at rest (p<0.01) and on movement (p<0.05) at 24 and 48 h. There was no significant difference between groups for PONV (n=40) (Keita 2003)
• Oral morphine (20 mg every 4 h) was superior to IM morphine (5–10 mg on demand) for postoperative pain scores (p<0.05; n=47) (McCormack 1993)
• Oral sustained-release morphine 20 mg every 12 h, starting at premedication, and IM morphine 10 mg every 6 h postoperatively, were similar for VAS pain scores (Bourke 2000)

Transferable Evidence from Other Procedures

Arguments for…

• Meperidine significantly reduced postoperative pain scores compared with placebo following hip or knee arthroplasty
  • In one study of a group of total hip-arthroplasty and knee-arthroplasty patients (n=48), the use of meperidine (100 mg bolus intravenously on demand) resulted in a significant reduction in pain scores within 30 min of administration compared with placebo (p<0.01) (Tarradell 1996). At 30–240 min following the treatment, the benefit was not significant.
• Morphine 4 mg was superior for reducing pain scores and reducing the time to onset of analgesia compared with placebo following total knee arthroplasty (Rasmussen 2002)
• Oral flupirtine maleate 100–200 mg and pentazocine 50–100 mg had similar effects on pain scores and on the requirement for supplementary analgesia after total hip or knee arthroplasty (n=66) (Galasko 1985)
• Morphine administered by PCA device had a similar effect to continuous infusion on postoperative pain scores following total hip or knee arthroplasty (n=32) (Smythe 1996)
• Opioids administered by PCA improved analgesia and decreased the risk of pulmonary complications, and patients preferred them compared with conventional intramuscular, intravenous or subcutaneous opioid treatment, as determined in a quantitative systematic review of randomised trials of postoperative pain management following various surgical procedures (Walder 2001)

Arguments against…

• Meperidine induced significantly higher sedation and respiratory depression compared with tramadol (both 100 mg intravenously) (n=48) (Tarradell 1996)

Clinical Practice

• Strong opioids in equipotent doses are considered to give a similar level of analgesia
• Intramuscular administration of strong opioids is considered to be more painful than intravenous administration. However, the size and speed of intravenous dose should be assessed to minimise the risk of respiratory depression

PROSPECT Recommendations

• Pre-operative strong opioids are not recommended because of inconsistent evidence of their postoperative analgesic efficacy (grade A). Voting agreement: 8/8
• As with all analgesia, strong opioids should be administered in time to provide sufficient analgesia in the early postoperative recovery period (grade D). Voting agreement: 8/8
• Strong opioids are recommended in combination with non-opioid analgesia for managing high-intensity pain following total hip arthroplasty (grade A). Voting agreement: 8/8
• Intravenous patient-controlled analgesia (grade A) or intravenous administration titrated for pain intensity (grade D), is recommended over ‘on-demand’ administration. Voting agreement: 8/8
• Intramuscular administration of strong opioids is not recommended (grade A). Voting agreement: 8/8
• Transdermal fentanyl patches are contra-indicated for the postoperative pain due to regulatory reasons (grade A). Voting agreement: 8/8

Total Hip Arthroplasty-Specific Evidence

Arguments against…

• Tramadol (50 or 100 mg orally) provided no benefit over placebo for postoperative pain scores up to 5 h after total hip arthroplasty (n=137) (Stubhaug 1995)
• Codeine (60 mg orally) provided no benefit over placebo for postoperative pain scores up to 6 h after total hip arthroplasty (n=121) (Dahl 1995)

Transferable Evidence from Other Procedures

Arguments against…

• Tramadol (100 mg bolus intravenously) provided no benefit over placebo for postoperative pain scores after total hip or knee arthroplasty (n=48) (Tarradell 1996)

Clinical Practice

• It is considered that weak opioids are ineffective as a single therapy for postoperative pain following total hip arthroplasty

PROSPECT Recommendations

• Weak opioids are not recommended for controlling pain <6 h following total hip arthroplasty (grade A). Voting agreement: 8/8
• Weak opioids are recommended for moderate- or low-intensity pain after 6 h if conventional NSAIDs or COX-2-selective inhibitors, are insufficient or are contraindicated (grade D). Voting agreement: 8/8

Total Hip Arthroplasty-Specific Evidence

Arguments for…

• Parenteral propacetamol 2 g (intravenous infusion) had a significant benefit over placebo in reducing postoperative morphine consumption (p<0.001; n=97) (Peduto 1998)
• Oral paracetamol 1 g plus codeine 60 mg was superior to placebo and to tramadol (50 or 100 mg) for reducing postoperative pain scores at 2–6 h (p<0.01, all comparisons) and supplementary analgesic use (p<0.01, all comparisons) (n=137) (Stubhaug 1995)

Arguments against…

• Parenteral propacetamol 2 g (intravenous infusion) had no benefit over placebo for reducing postoperative pain scores after total hip arthroplasty (n=97) (Peduto 1998)

Transferable Evidence from Other Procedures

Arguments for…

• Paracetamol reduced supplementary analgesic consumption compared with placebo following orthopaedic surgery
  • In a qualitative review of 28 paracetamol studies, five out of seven studies of orthopaedic procedures reported a significant reduction in analgesic consumption with paracetamol compared with placebo (Römsing 2002).

Arguments against…

• Lack of consistent evidence for the benefit of paracetamol to reduce postoperative pain scores following orthopaedic surgery
  • In a qualitative review of 28 paracetamol studies, rectal or parenteral paracetamol was effective compared with placebo for reducing postoperative pain in a variety of surgical procedures (Römsing 2002). However, of the nine studies examining orthopaedic procedures, six showed no significant benefit compared with placebo for postoperative pain scores; the orthopaedic procedures included arthroscopy, surgery for hallux valgus, hip arthroplasty, knee ligamentoplasty, limb surgery and lumbar discectomy and the studies examined propacetamol 1–2 g administered intravenously or intramuscularly, or paracetamol 14 mg/kg administered rectally four times daily. An additional study, not included in this analysis, did show a benefit of intravenous propacetamol 2 g for postoperative pain scores in patients undergoing spinal fusion (n=42) (Hernandez-Palazon 2001).
• Lack of consistent evidence for the benefit of adding paracetamol to NSAIDs to reduce postoperative pain following orthopaedic surgery
  • In a qualitative review of paracetamol studies, six studies assessed the effects on postoperative pain of adding paracetamol to NSAIDs as compared with NSAIDs alone in a variety of surgical procedures (including only two orthopaedic procedures) (Römsing 2002). Only two out of the six studies (one in lumbar discectomy and the other in dental surgery) demonstrated a reduction in postoperative pain scores for combined ketoprofen (50 mg intravenously four times daily) and propacetamol (2 g intravenously four times daily), compared with ketoprofen alone.

Clinical Practice

[None cited]

PROSPECT Recommendations

• Paracetamol is recommended for all pain intensities because it reduces supplementary analgesic requirements immediately following total hip arthroplasty (grade A). Voting agreement: 8/8
• Paracetamol is recommended in combination with conventional NSAIDs or COX-2-selective inhibitors, with or without rescue opioids (grade B). Voting agreement: 8/8

Total Hip Arthroplasty-Specific Evidence

Arguments for…

• Two studies showed that continuous epidural anaesthesia/analgesia was superior to general anaesthetic plus intravenous morphine analgesia for reducing postoperative pain scores
  • Continuous epidural infusion of bupivacaine 0.625 mg/ml plus morphine 0.05 mg/ml at 4 ml/h for 48 h and piroxicam 40 mg orally (evening before surgery and 1 h before surgery) then piroxicam 20 mg/day, was superior to general anaesthetic followed by intravenous morphine 5 mg and intramuscular morphine 0.125 mg/kg with paracetamol orally on demand for postoperative pain scores on movement, during (p=0.0002) and after (p=0.02) the epidural phase (n=20) (Möiniche 1994).
  • Infused epidural ropivacaine (2 mg/ml at 4–6 ml/h) was superior to general anaesthetic plus postoperative intravenous bolus and on-demand morphine (10 mg and 1.0–1.5 mg via a patient-controlled administration device) for 10- and 24-h pain scores (p<0.001; n=90) (Wulf 1999).
• Epidural bolus (pethidine hydrochloride 60 mg) was superior to IM bolus (pethidine hydrochloride 1 mg/kg) administered postoperatively for pain scores at 0.5 and 1 h (p<0.05), but non-significant for 2–4 h pain scores and supplementary analgesic consumption (n=14) (Gustafsson 1986)
• Postoperative epidural infusions of bupivacaine, ropivacaine or levobupivacaine were similar for postoperative pain scores, supplementary analgesia and PONV in two studies
  • Two studies were identified: one compared epidural ropivacaine 2 mg/ml with bupivacaine 2 mg/ml, both administered as a constant infusion of 6 ml/h supplemented by PCEA bolus doses of 2 ml. Both groups received rescue IV PCA morphine on demand (n=51) (Bertini 2001). The second study assessed levobupivacaine (0.5% intra-operative epidural block and 0.125% postoperative analgesia) versus bupivacaine (0.5% intra-operative epidural block and 0.125% postoperative analgesia) versus ropivacaine (0.5% intra-operative epidural block and 0.2% postoperative analgesia). Each group also received the study drug postoperatively using a PCEA infusion at 5 ml/h with an incremental bolus of 2 ml and lockout time of 20 min. IV ketoprofen was used for supplementary analgesia (n=45) (Casati 2003)
  • Postoperative VAS scores: Qualitative analysis: Bupivacaine and ropivacaine were equivalent for postoperative VAS pain scores in two studies (0–24 h and 0–12 h respectively) (Bertini 2001, Casati 2003) and in one study bupivacaine, ropivacaine and levobupivacaine provided similar postoperative analgesia (Casati 2003)
  • Postoperative VAS scores: Quantitative outcomes: Bupivacaine and ropivacaine were not significantly different for VAS scores at 6 h (2 studies, WMD 2.37 [-5.78, 10.52], p=0.57) and at 12 h (2 studies, WMD 5.20 [-11.71, 22.12], p=0.55)
  • Supplementary analgesia use: Qualitative analysis: Bupivacaine and ropivacaine were not significantly different for postoperative supplementary analgesia use (Bertini 2001, Casati 2003) and in one study bupivacaine, ropivacaine and levobupivacaine were not significantly different (Casati 2003)
  • PONV: Qualitative analysis: Bupivacaine and ropivacaine were not significantly different for PONV (Bertini 2001, Casati 2003) and in one study bupivacaine and levobupivacaine were not significantly different (Casati 2003)
  • PONV: Quantitative analysis: Bupivacaine and ropivacaine not significantly different for the incidence of PONV (two studies, OR 1.48 [0.27, 7.96], p=0.65)
• Epidural infusion of bupivacaine (0.125 mg/ml) combined with morphine (0.05 mg/ml) was similar to combination with fentanyl (0.005 mg/ml) for postoperative pain scores after total hip arthroplasty (n=30) (Berti 1998)
• Different epidural sufentanil doses (0.5 µg/ml, 0.75 µg/ml or 1.0 µg/ml) administered with ropivacaine 0.1% by continuous epidural infusion, were similar for VAS scores, supplementary analgesic consumption, pain-mobility-and-ability-to-walk scores and adverse events within 0–44 h (n=32) (Kampe 2003)
• Epidural clonidine was superior to epidural local anaesthetic or morphine for a number of postoperative analgesia outcomes
  • A bolus of 150 µg of clonidine was superior to a bolus of bupivacaine (10 ml 0.25%) for 60 min postoperative pain scores, time to first analgesic request and use of supplementary analgesia (p<0.05 for both; n=90) (Carabine 1992a).
  • Infused clonidine 8.3 µg/ml (6 ml/h) was superior to 0.125% levobupivacaine (6 ml/h) for time to first analgesic request and use of supplementary analgesia (p<0.05 for both; n=86) (Milligan 2000).
  • Bolus and infused clonidine at higher (150 µg and 50 µg/ml at 1 ml/h) and lower (150 µg and 25 µg/ml at 1 ml/h) doses were superior to bolus and infused morphine (1 mg plus 0.1 mg/ml at 1 ml/h) for postoperative pain scores and for the proportion of patients experiencing moderate pain. Higher dose clonidine was also superior to morphine for time to first analgesia request and for the use of supplementary analgesia (p<0.05 for both comparisons; n=90) (Carabine 1992b).
• Addition of clonidine to local anaesthetic or opioid was superior to local anaesthetic or opioid alone for a variety of postoperative analgesia outcomes
  • Bolus bupivacaine (10 ml 0.25%) administered with clonidine 150 µg was superior to bupivacaine alone for postoperative pain scores, time to first analgesia request and use of supplementary analgesia (p<0.05 for both; n=90) (Carabine 1992a).
  • The addition of a clonidine infusion (8.3 µg/ml at 6 ml/h) to a levobupivacaine infusion (0.125% at 6 ml/h) was superior to levobupivacaine alone for time to first analgesia request and use of supplementary analgesia (p<0.01 for both; n=86) (Milligan 2000).
  • The combination of clonidine bolus 150 µg with morphine infusion 0.1 mg/ml at 1 ml/h was superior to morphine alone for postoperative pain scores, time to first analgesia request and use of supplementary analgesia (p<0.05 for both; n=40) (Carabine 1992b).

Arguments against…

• The addition of ropivacaine (1 mg/ml, 3 ml/h) to an infusion of fentanyl (10 µg/ml, 3 ml/h) conferred no benefit over fentanyl alone for postoperative pain scores (n=39) (Kostamovaara 2001)
• Epidural clonidine reduced mean arterial pressure and heart rate compared with epidural morphine but did not increase the incidence of sedation and did not consistently reduce arterial pressure compared with epidural local anaesthetic
  • In one study, epidural clonidine (150 µg bolus plus infusion 25 µg/ml or 50 µg/ml) significantly reduced mean arterial pressure compared with epidural morphine (1 mg bolus plus infusion 0.1 mg/ml) during 30 min–2 h (p<0.01) and at 18 and 24 h (p<0.05), and the epidural combination of clonidine (150 µg) and morphine (1 mg bolus plus infusion 0.1 mg/ml) significantly reduced mean arterial pressure compared with either agent alone during 5–20 min after injection (p<0.05) (n=100) (Carabine 1992b).
  • Epidural clonidine significantly decreased heart rates compared with epidural morphine at 15 and 20 min after injection, and during 6–24 h after operation (p<0.05).
  • Epidural morphine significantly increased the incidence of sedation compared with epidural clonidine at 30 min (p<0.05) but not at 60 min.
  • Epidural clonidine (150 µg) and epidural bupivacaine (10 ml 0.25%) were similar for mean arterial pressure, heart rate and the incidence of sedation (n=60) (Carabine 1992a).
  • Epidural infusion at 6 ml/h of clonidine (8.3 µg/ml) or clonidine (8.3 µg/ml) plus levobupivacaine (0.125%) significantly reduced mean systolic pressures compared with levobupivacaine (0.125%) during 4–26 h (p<0.05), but there was no significant difference in the incidence of hypotension or cardiac arrhythmias (n=90) (Milligan 2000).

Transferable Evidence from Other Procedures

Arguments for…

• Epidural analgesia was superior to systemic analgesia for postoperative pain outcomes
  • Bolus and infused epidural analgesia with bupivacaine (13 ml 0.25% and 0.125% at 10 ml/h) and sufentanil (10 mg/ml and 0.1 mg/ml at 10 ml/h) was significantly more effective (VAS at rest: 4 h p<0.001; 24 h p=0.04 and 48 h p=0.03; VAS on movement: 4 h p<0.001; 24 h p=0.01; 48 h p=0.06) than patient-controlled intravenous morphine (2 mg/ml, bolus of 1.5 mg) for reducing postoperative pain scores after total knee arthroplasty (n=45) (Singelyn 1998).
  • Infused and on-demand epidural clonidine (8 µg/kg over 30 min and 30 µg boluses) were similar to intravenous clonidine (8 µg/kg over 30 min and 30 µg boluses) for reducing postoperative pain scores, but were superior (p<0.01) for the requirement of supplementary patient-controlled clonidine, in patients undergoing scoliosis correction (n=24) (Bernard 1995).
  • Epidural analgesia significantly reduced postoperative VAS scores at rest compared with systemic analgesia in the first 4–6 h after surgery (standardised mean differences -0.77; 95% CI -1.24 to -0.31) following hip or knee replacement (7 studies, n=236). There was no significant difference at 18–24 h postoperatively (standardised mean differences -0.29; 95% CI -0.73 to 0.16) (6 studies, n=182) (Choi 2003).
• Epidural analgesia was associated with a lower incidence of pulmonary complications, sedation, urinary retention, pruritis and hypotension compared with systemic analgesia
  • In a systematic review of studies conducted in a variety of procedures, epidural opioids were superior to systemic opioids for decreasing the incidence of atelectasis (risk ratio [RR] 0.53, 95% confidence interval [CI] 0.33 to 0.85), and provided non-significant decreases in the incidence of pulmonary infections (RR 0.53, 95% CI 0.18 to 1.53) and pulmonary complications (RR 0.51, 95% CI 0.20 to 1.33). In addition, epidural local anaesthetics were superior to systemic opioids for decreasing the incidence of pulmonary infections (RR 0.36, 95% CI 0.21 to 0.65) and overall pulmonary complications (RR 0.58, 95% CI 0.42–0.80) (Ballantyne 1998).
  • In a review, epidural analgesia was associated with a lower incidence of sedation (95% CI 3.5 to 42.0) (3 studies, n=116), urinary retention (OR 3.50; 95% CI 1.63 to 7.51) (5 studies, n=166), pruritus (OR 4.74; 95% CI 1.76 to 12.78) (6 studies, n=209) and hypotension (OR 2.78; 95% CI 1.15 to 6.72) (4 studies, n=137) compared with systemic analgesia (Choi 2003). The groups were not significantly different for the frequency of postoperative nausea and vomiting (OR 0.95; 95% CI 0.6 to 1.49) (9 studies, n=371) or the frequency of respiratory depression (OR 1.07; 95% CI 0.45 to 2.54) (5 studies, n=204) (Choi 2003).
• Ropivacaine epidural infusion was superior to placebo (no epidural infusion) for reducing postoperative pain scores (all groups received rescue IV morphine)
  • One study assessed the effect of a 0.2% ropivacaine infusion compared with control over a range of infusion rates in a group of patients undergoing total hip or knee arthroplasty (n=135). Infusion of ropivacaine at 8 ml/h was superior to control for 4-h postoperative pain scores (p<0.01). Ropivacaine at 10, 12 or 14 ml/h was superior to placebo for 8-h postoperative pain scores and for reducing the requirement for supplementary analgesia up to 21 h following the operation (p<0.01 for all). Ropivacaine at 12 ml/h was superior to 8 ml/h for 8-h postoperative pain scores; and ropivacaine at 14 ml/h was superior to 8 ml/h for 21-h postoperative pain scores (p<0.01 for both) (Turner 1996).
• A combination of epidural local anaesthetic plus opioid produced superior analgesic benefits compared with either drug administered alone
  • Epidural infusion of ropivacaine (2 mg/ml) plus sufentanil (1 µg/ml) was significantly more effective than ropivacaine alone, for reducing postoperative pain scores following major knee surgery (p=0.009; n=115) (Lorenzini 2002).
  • Combined epidural levobupivacaine (20 ml 0.75% bolus and 0.125% at 4 ml/h) and fentanyl (4 µg/ml at 4 ml/h) was superior to either drug administered alone for reducing postoperative pain scores and for increasing time to request of analgesia after total hip or knee arthroplasty. Resting and dynamic VAS pain scores were lower in the combination group than in the plain fentanyl group at 6 h (p=0.022 and 0.036) and 12 h (p=0.002 and 0.001). The combination of levobupivacaine and fentanyl increased the time to first PCEA request compared with either drug alone (p=0.007 combination versus fentanyl and p=0.006 combination versus levobupivacaine). There was no significant difference among the groups for the incidence of postoperative nausea, pruritus, hypotension or sedation (n=65) (Kopacz 1999).
  • Epidural infusion of bupivacaine and meperidine (1 mg/ml) had a significantly slower regression of sensory anaesthesia and slower development of pain, in contrast to infusions of bupivacaine alone (control) or bupivacaine and fentanyl (3 µg/ml) following total knee arthroplasty (p<0.05; n=48) (Ferrante 1993).
• Epidural morphine (bolus 4 mg, infusion 12.5 mg/h over 13 h) was similar to epidural ketoprofen (200 mg over 30 min, 12.5 mg/h over 13 h) for postoperative pain scores following total hip or knee arthroplasty (Hommeril 1994)
• The addition of epidural clonidine (bolus of 150 µg) to epidural bupivacaine (bolus of 50 mg) significantly prolonged the duration of analgesia in patients undergoing hip surgery for traumatic fracture (p<0.05; n=40) (Klimscha 1995)

Arguments against…

• Combined spinal epidural block or spinal block were superior to epidural block (0.5% bupivacaine plus 0.2 or 0.4 mg morphine for spinal, or 0.5% bupivacaine plus 4 mg morphine for epidural) for surgical analgesia and for reducing consumption of perioperative sedatives and analgesics in major orthopaedic surgery (Holmstrom 1993)
• There was no significant difference between epidural fentanyl (3 µg/ml) plus bupivacaine and bupivacaine alone in the rate of regression of sensory anaesthesia or the development of postoperative pain (n=48) (Ferrante 1993)
• Epidural infusion of ropivacaine (2 mg/ml) plus sufentanil (1 µg/ml) produced a significantly higher incidence of adverse events (pruritus, nausea and vomiting) than ropivacaine alone following major knee surgery (n=115) (Lorenzini 2002)
• Neuraxial and parenteral opioids are associated with a greater risk of side-effects compared with continuous plexus and peripheral neural blockades after major orthopaedic surgery (Sinatra 2002)

Clinical Practice

• The protective benefits of epidural analgesia are less marked when compared with ivPCA in more recent perioperative care studies, compared to the data from older studies. (Popping, D and Tramer, M 2008)

PROSPECT Recommendations

• Epidural infusion with local anaesthetic plus opioid is recommended for cardiopulmonary risk patients (grade A) because of the reduction in cardiopulmonary morbidity associated with epidural analgesia. Voting agreement: 8/8
• Despite the analgesic benefits of epidural clonidine, it is not recommended because of the risk of bradycardia, hypotension and sedation (grade A). Voting agreement: 8/8
• Epidural magnesium sulphate is not recommended because of limited data and lack of safety data (grade B). Voting agreement: 8/8
• Epidural extended release morphine is not recommended because of limited procedure specific efficacy data and high incidence of serious adverse effects. (Grade B). Voting agreement: 8/8

Total Hip Arthroplasty-Specific Evidence

Arguments for…

• Posterior lumbar plexus block ‘single shot’ given pre- or postoperatively was superior to placebo for reducing postoperative pain scores and supplementary analgesic consumption
  • In two studies, ‘single shot’ posterior lumbar plexus block was superior to placebo as follows: lumbar plexus block using 0.4 ml/kg bupivacaine and epinephrine after induction of general anaesthetic was effective in reducing pain scores (p=0.007) and reducing the need for supplementary analgesia (p<0.0001) compared with placebo up to 6 h (n=60) (Stevens 2000); and lumbar plexus block using 2 mg/kg 0.375% bupivacaine plus 2 µg/kg of clonidine at the end of surgery was superior to control (no nerve block) for reducing pain scores at 0–4 h (p=0.001) and opioid use at 0–12 h after extubation (p=0.002) (n=45) (Biboulet 2004).
  • Femoral nerve block (‘single shot’ 40 ml bupivacaine 0.5% plus epinephrine after induction of general anaesthesia) increased the time to first analgesic request by approximately 4 h compared with placebo (p<0.05) (n=40) (Fournier 1998)
• Postoperative lumbar plexus block was superior to femoral nerve block for reducing postoperative pain scores at rest and supplementary analgesic consumption
  • In one study, lumbar plexus block was superior to femoral nerve block (each using 2 mg/kg 0.375% bupivacaine plus 2 µg/kg of clonidine ‘single shot’) for reducing pain scores at 0–4 h (p=0.001) and initial opioid use (p=0.004) (n=45) (Biboulet 2004).

Arguments against…

• Femoral nerve block did not significantly reduce postoperative pain scores, and there was inconclusive evidence for the effect on supplementary analgesic consumption, compared with placebo
  • In two studies, femoral nerve block showed no significant benefit over placebo for reducing postoperative pain scores as follows: ‘single shot’ 40 ml bupivacaine 0.5% plus epinephrine after induction of general anaesthesia did not reduce postoperative pain scores for 24 h (n=40) (Fournier 1998); and ‘single shot’ 2 mg/kg 0.375% bupivacaine plus 2 µg/kg of clonidine after surgery did not reduce postoperative pain scores at rest or movement for 48 h (n=45) (Biboulet 2004). However, femoral nerve block was superior to control for reducing initial opioid use in one study (p=0.004) (n=45) (Biboulet 2004), but not in the second study at any time (n=40) (Fournier 1998).
• Posterior lumbar plexus block provided no significant benefit over femoral nerve block or control (no nerve block) for postoperative pain scores on movement, PONV or articular mobility during rehabilitation (n=45) (Biboulet 2004)
• Posterior lumbar plexus block using ropivacaine 0.475% was less effective than spinal morphine 0.1 mg for reducing pain scores during 6–18 h (p<0.05) and supplementary analgesic consumption during 48 h, and there was no significant difference in the incidence of nausea, vomiting or pruritis (n=53) (Souron 2003)
• Femoral nerve block delivered by PCA provided no benefit over delivery by continuous infusion for postoperative pain scores (n=45) (Singelyn 2001)

Transferable Evidence from Other Procedures

Arguments for…

• In a systematic review of seven randomised trials in hip fracture, nerve blocks administered pre- or peri-operatively resulted in a reduction in pain score and supplementary analgesic requirement compared with control (n=269) (Parker 2001a)
• ‘Single shot’ femoral nerve block reduced pain scores for up to 8 h and reduced morphine consumption following total knee arthroplasty (Allen 1998)
• ‘Single shot’ or continuous peripheral nerve block was significantly more effective than placebo for reducing the requirement for supplementary analgesia following total knee or hip arthroplasty (n=242) (Allen 1998, Bogoch 2002, Edwards 1992, Serpell 1989, Wang 2002)
• The posterior approach to the lumbar plexus block produces more reliable analgesia to the hip joint than the distal approach (femoral nerve or 3-in-1 blocks); however, the lumbar plexus block has the potential for more serious complications than the femoral nerve block (Auroy 2002)
• Continuous plexus and peripheral neural blocks are associated with a reduced risk of side-effects compared with neuraxial and parenteral opioids, in pain management after major orthopaedic surgery (Sinatra 2002)
• Femoral nerve block is associated with a lower risk of serious complications than spinal anaesthesia (using bupivacaine or lidocaine) (Auroy 2002)

Arguments against…

• Addition of epinephrine did not alter the duration of analgesia with a ‘single shot’ 3-in-1 femoral nerve block (20 ml ropivacaine 0.5% or 0.2%) following total knee arthroplasty (n=41) (Weber 2001)
• Continuous 3-in-1 femoral block 10 ml/h (0.125% bupivacaine plus sufentanil 0.1 µg/ml and clonidine 1 µg/ml) provided no benefit over IV PCA morphine for the incidence of nausea and vomiting although a reduction in catheter problems was noted (Singelyn 1998)
• As with all regional anaesthetic techniques, peripheral neural blockade has a failure rate and can cause neural injury and local anaesthetic toxicity (Cox 2003)

Clinical Practice

• Catheter techniques are recommended over a ‘single shot’ approach as they provide an extended height of block and a greater duration of analgesia
• Continuous infusion, patient-controlled or ‘on-demand’ femoral nerve blocks are recommended over a ‘single shot’ approach as they provide an extended height of block and a greater duration of analgesia

PROSPECT Recommendations

• Posterior lumbar plexus blocks (psoas sheath blocks) have a greater efficacy than distal lumbar plexus blocks (femoral nerve blocks) in total hip replacement, and are recommended (grade A). However, they have a potential for more serious complications than the femoral block (plus supplementary obturator and lateral cutaneous nerve of thigh blocks) (grade B).This recommendation must be balanced against risks of motor blockade and falls and the efficacy of systemic multimodal analgesia. Voting agreement: 8/8

Total Hip Arthroplasty-Specific Evidence

Arguments for…

• Spinal morphine 0.1 mg was superior to a posterior lumbar plexus block using ropivacaine 0.475% for reducing postoperative pain scores for 6–18 h (p<0.05) and reducing supplementary analgesic consumption for 48 h, but there was no significant difference for the incidence of nausea, vomiting or pruritis (n=53) (Souron 2003)
• Continuous spinal bupivacaine analgesia (and anaesthesia) was superior to IV PCA morphine analgesia (plus ‘single shot’ spinal anaesthesia) for reducing VAS scores for 3–24 h (p<0.05) and the incidence of PONV (n=68) (Maurer 2003)
• Spinal analgesia was superior to epidural analgesia in one study for pain scores
  • Spinal analgesia with bupivacaine (1 ml 0.25% bolus and 10 ml 0.25% over 24 h) was superior to epidural analgesia with bupivacaine (10 ml 0.25% bolus and 2 ml/h) for postoperative pain scores (n=102; p<0.05), and for the use of supplementary analgesia (p<0.05) (Mollmann 1999).
• Strong opioid plus local anaesthetic was superior to local anaesthetic alone for postoperative pain scores, requirement for supplementary analgesia and time to first analgesia request, when administered by the spinal route
  • Postoperative pain scores: Qualitative analysis: of six studies, all six showed that the addition of opioid to local anaesthetic was superior to local anaesthetic alone for postoperative pain scores: intrathecal bolus of fentanyl 25 µg and hyperbaric bupivacaine 12.5 mg (Fernandez-Galinski 1996); intrathecal bolus morphine 1 mg and 2.75 ml bupivacaine 0.5% (Fogarty 1993); intrathecal bolus of morphine 0.5 mg and 13.75 mg isobaric bupivacaine (Grace 1995, Grace 1994); intrathecal bolus of diamorphine 0.75–1 mg and 2.75 ml 0.5% bupivacaine (Milligan 1993); and spinal anaesthesia with hyperbaric bupivacaine 15 mg and spinal doses of morphine 100 µg or 200 µg (Murphy 2003)
  • Postoperative pain scores: Quantitative analysis: combination treatment is superior for VAS scores compared with LA alone for the 0–8 h grouping (WMD -14.27 [-18.01, -10.53], p<0.00001), the 8–16 h grouping (WMD -14.74 [-20.38, -9.09], p<0.00001) but not for the 16–32 h grouping (WMD -4.04 [-9.14, 1.06], p=0.12).
  • Supplementary analgesia: Qualitative analysis: the addition of opioid to local anaesthetic was superior to local anaesthetic alone for supplementary analgesic requirement in all five studies reporting this outcome (Fogarty 1993, Grace 1995, Grace 1994, Milligan 1993, Murphy 2003)
  • Supplementary analgesia: Quantitative analysis: postoperative morphine consumption was significantly lower for the group on combined therapy compared with local anaesthetic alone (WMD -10.94 mg [-15.25, -6.63], p<0.00001).
  • Time to first analgesic request: Qualitative analysis: the addition of opioid to local anaesthetic was superior to local anaesthetic alone for the time to first analgesic request in all five studies reporting this outcome (Fogarty 1993, Grace 1995, Grace 1994, Milligan 1993, Murphy 2003)
  • Time to first analgesic request: Quantitative outcomes: the time to first analgesic request was significantly extended for the group on combined therapy compared with local anaesthetic alone (WMD -499.30 min [-664.25, -334.34], p<0.00001).
• In three of four studies, different intrathecal strong opioids were similar for analgesic outcomes at the doses tested.
  • Intrathecal nalbuphine 400 µg or morphine 160 µg administered as a postoperative bolus dose were similar for postoperative pain scores, but morphine was superior to nalbuphine for time to first analgesia request (p<0.05) and use of supplementary analgesia (p<0.001) (n=24) (Fournier 2000a).
  • Fentanyl 40 µg and sufentanil 7.5 µg administered postoperatively by intrathecal bolus were similar for postoperative pain scores, time to first analgesia request, the requirement for supplementary analgesia, and for nausea and vomiting (n=42) (Fournier 2000b).
  • Morphine-6-glucuronide 100 µg and morphine 0.5 mg bolus doses intrathecally had similar analgesic profiles when administered with intrathecal bupivacaine 13.75 mg (n=75) (Grace 1996).
  • In the remaining study, intrathecal morphine 1 mg was superior to intrathecal diamorphine 0.75 mg for reducing postoperative pain scores at 4 h (p<0.01) and supplementary analgesic requirements (p<0.05) (n=60) (Fogarty 1995a).
• Isobaric bupivacaine 0.5% 3.5 ml and levobupivacaine 0.5% 3.5 ml were similar for postoperative pain scores and for the requirement of supplementary analgesia (n=80) (Glaser 2002)
• Spinal morphine 1 mg was superior to clonidine 75–100 µg in combination with isobaric bupivacaine 2.75 ml 0.5% for pain scores up to and including 10 h after the operation (p<0.01), time to first analgesia request and use of supplementary analgesia (p<0.05; n=90) (Fogarty 1993).
• Patient-controlled spinal analgesia was superior to spinal bolus doses on demand in one study for pain
  • Patient-controlled spinal bupivacaine 0.125% at 0.6 mg/h plus 0.6 mg boluses on demand was superior to spinal bupivacaine 3.75 mg bolus on demand for postoperative pain scores (n=41; p<0.01) and use of study analgesia (p<0.01) (Rundshagen 1997)

Arguments against…

• A combination of bupivacaine and morphine may cause a greater incidence of vomiting compared with bupivacaine alone
  • Qualitative analysis: Of six studies showing postoperative pain benefits with combination therapy, one study demonstrated that a combination of bupivacaine and morphine gave a higher incidence of vomiting than bupivacaine alone (Grace 1995); results of four studies were not significant (Fogarty 1993, Grace 1994, Milligan 1993, Murphy 2003), and one study did not report the incidence of vomiting (Fernandez-Galinski 1996). No significant difference was found for the incidence of nausea.
  • Quantitative outcomes: LA alone was associated with a significantly lower incidence of vomiting compared with combined LA plus strong opioid (OR 1.85 [1.15, 2.99], p=0.01) but no significant difference for the incidence of nausea (OR 1.32 [0.81, 2.15], p=0.26)
• Continuous spinal bupivacaine demonstrated a significant reduction in mean arterial pressure during anaesthetic induction compared with ‘single shot’ spinal bupivacaine (21 ± 11 mmHg versus 29 ± 14; p<0.05) (n=68) (Maurer 2003)
• Spinal clonidine provided no significant benefit over placebo in two of three studies
  • Two out of three studies showed no significant benefit of clonidine compared with placebo in the following comparisons: clonidine hydrochloride 75 µg in combination with morphine and bupivacaine versus morphine and bupivacaine alone (Grace 1995) or clonidine 30 µg in combination with sufentanil versus sufentanil alone (Fournier 2000a). In the remaining study, clonidine 75–100 µg plus bupivacaine was superior to bupivacaine alone for pain scores at 2 and 4 h (n=90; p<0.05) and time to first analgesic request (p<0.05), but was not superior for the use of supplementary analgesia (Fogarty 1993).

Transferable Evidence from Other Procedures

Arguments for…

• Bolus spinal morphine (300 µg) was significantly more effective than saline placebo for reducing postoperative pain scores after total knee arthroplasty (p<0.05; n=60) (Tan 2001)
• Combined spinal epidural block or spinal block were superior to epidural block (0.5% bupivacaine plus 0.2 or 0.4 mg morphine for spinal, or 0.5% bupivacaine plus 4 mg morphine for epidural) for surgical analgesia and for reducing consumption of perioperative sedatives and other analgesics in major orthopaedic surgery (Holmstrom 1993)

Arguments against…

• In a systematic review, spinal morphine in patients undergoing caesarean section was shown to increase the relative risk of postoperative pruritis, nausea and vomiting compared with control; increasing the dose of morphine increased the relative risk of postoperative nausea and vomiting (Dahl 1999)
• Spinal administration of bolus clonidine or morphine produced a high incidence of bladder distension in patients undergoing hip surgery, but there was a greater incidence with spinal morphine than clonidine (p<0.001) (Gentili 1996)
• Neuraxial opioids are associated with a higher risk of side-effects compared with peripheral neural blocks (Sinatra 2002)
• Spinal anaesthesia (using bupivacaine or lidocaine) is associated with a higher risk of serious complications than femoral nerve block (Auroy 2002)

Clinical Practice

• Long-acting opioids, such as morphine, are preferred to short-acting opioids for a long duration of analgesia postoperatively

PROSPECT Recommendations

• Where appropriate, single bolus spinal morphine (0.1–0.2 mg) as a part of spinal anaesthesia is recommended as it provides pain relief for up to 24 h (grade A). However, this should be balanced against the risk of late respiratory depression and monitoring requirements as well as side effects such as postoperative nausea and vomiting and urinary retention. Voting agreement: 8/8
• Spinal clonidine is not recommended because it is less effective than spinal morphine for analgesia (grade A). Voting agreement: 8/8
• Short-acting opioids are not recommended because of their shorter duration of effect compared with morphine at appropriate doses (grade D). Voting agreement: 8/8
• Continuous spinal anaesthesia is not recommended due to limited and inconsistent data (grade B). Continuous spinal administration of morphine is not recommended due to safety concerns (grade D). Voting agreement: 8/8

Total Hip Arthroplasty-Specific Evidence

Arguments for…

• Wound infiltration with lidocaine before drain removal was superior to no wound infiltration for reducing VAS pain scores during drain removal and at 1 h after drain removal (p<0.05, n=82) (Yiannakopoulos 2004)
  • This study compared wound infiltration using 10 ml lidocaine 1% before suction drain removal at 48 h (dispersed around the drain tube to anaesthetise the skin and subcutaneous tissue) with no wound infiltration (n=82) (Yiannakopoulos 2004)

Transferable Evidence from Other Procedures

Study details: Two studies, one in elective hip or knee arthroplasty and the other in spine fusion surgery, compared the same regimens: post-closure continuous wound infusion using a loading dose of ropivacaine 0.5% 200 mg/40 ml followed by ropivacaine 0.2% 5 ml/h through a catheter for 55 h versus control (continuous IV infusion of morphine 0.5 mg/h plus keterolac 3.6 mg/h for 24 h) administered directly into the surgical area –both groups received saline by the alternative route and rescue IV tramadol plus IM diclofenac was available (n=37, n=37) (Bianconi 2004, Bianconi 2003)

Arguments for…

• Post-closure continuous wound infusion with ropivacaine 0.2% was superior for reducing postoperative pain scores at rest and mobilisation when compared with control (IV morphine plus keterolac) after hip or knee arthroplasties (8–72 h; p<0.05) (n=37) (Bianconi 2003) or spinal fusion surgery (4–72 h; p<0.01) (n=37) (Bianconi 2004)
• Post-closure continuous wound infusion with ropivacaine 0.2% was superior to control (IV morphine plus keterolac) for reducing supplementary tramadol and diclofenac consumption after hip or knee arthroplasties (0–24 h, p<0.05; n=37) (Bianconi 2003) and after spinal fusion surgery (0–48 h, p<0.05; n=37) (Bianconi 2004)
• Post-closure continuous wound infusion with ropivacaine 0.2% was superior for reducing length of hospital stay when compared with control (IV morphine plus keterolac) after hip or knee arthroplasties (6.34 versus 8.79 days, p<0.05; n=37) (Bianconi 2003) or spinal fusion surgery (5.11 versus 7.56 days, p<0.05; n=37) (Bianconi 2004)

Arguments against…

• Continuous wound infusion with ropivacaine 0.2% was not significantly different for the incidence of postoperative nausea and vomiting when compared with control (IV morphine plus keterolac) after hip or knee arthroplasties or spinal fusion surgery (Bianconi 2004, Bianconi 2003)

Clinical Practice

• In contrast to positive LIA-data for knee replacement the hip replacement data are inconsistent

PROSPECT Recommendations

• Postoperative wound catheter techniques with local anaesthetic cannot be recommended (grade D) due to lack of procedure-specific evidence, although analgesic data from other procedures are promising. Voting agreement: 8/8
• Wound infiltration at the drain site is not recommended (grade D) because drains are not recommended. Voting agreement: 8/8
• Intraoperative high volume, low concentration wound infiltration (LIA) is not recommended at this time because of inconsistent data (grade A. Voting agreement: 8/8