Diagnostic and therapeutic arthroscopy of the knee joint is a common day case procedure. Ideally, it demands effective postoperative analgesia of long duration without systemic effects. Intra-articular morphine has been studied and used as a method of postoperative pain relief after knee arthroscopy. Meta-analysis has suggested that intra-articular morphine has a definite but mild analgesic effect, which may be dose dependent. A systematic effect cannot be completely excluded [1,2].
No study has yet compared the conventional intramuscular morphine dose (10 mg) with the same dose given by the intra-articular route. If the effect of intra-articular morphine is dose dependent as has been suggested , then a large dose could potentially give long-lasting analgesia. In addition, if intra-articular morphine is effective locally and leaves the knee slowly as has been suggested [3,4], then analgesia should not only be effective, but safe. If on the other hand, morphine enters the systemic circulation postoperatively through vessels opened in the knee during the operative procedure with the injection acting as a depot rather like an intramuscular injection, then we will be able to assess this by assaying plasma morphine concentrations. No studies have measured plasma morphine concentrations for 24 h after intra-articular administration in patients undergoing knee arthroscopy. This study compared the efficacy and plasma morphine concentrations produced by a large dose (10 mg) of intra-articular and intramuscular morphine up to 24 h after injection in a randomized double-blind fashion.
After obtaining approval from Bro TafLocal Research Ethics Committee, 40 ASA I-II patients between the ages of 18 and 65 yr presenting for day case knee arthroscopy were recruited to the study. Written informed consent was obtained. The study was carried out in the Day Surgery Unit in Llandough Hospital, Cardiff, UK. All patients were required to have an operative intervention and those having simple arthroscopy were not studied. Consequently, randomization of patients was not carried out until the operation was underway. Patients allergic to morphine or propofol, those unable to take non-steroidal anti-inflammatory medication, those who had taken oral analgesics in the 12 h before surgery or those who needed intra-articular drainage after surgery were excluded from the study. Patients with a body mass index over 30 were not included in the study. Patient characteristics data and pain score were collected at the time of admission to the Day Surgery Unit.
In this study, we were interested in the duration of the action of the analgesia so that patient's comfort is improved in the community. The primary outcome variable was the overall pain assessment between discharge (4 h) and 24 h assessed by visual analogue scale (VAS). Joshi's  study found a 24-h mean score of 3.75 (0-10) with a calculated SD of 0.79. To find a clinically important reduction in pain score of about 25%, 30 patients would be necessary for a study with a power of 0.95 showing significance at the 0.05 level. Forty patients were recruited for the study, to allow for losses. For analysing the primary outcome variable, t-test was used and repeated measures analysis of variance (ANOVA) was used to analyse the variables taken at multiple time points, the pain scores and plasma morphine concentrations. U-test was used for non-normally distributed data. χ2-test was used to analyse analgesic consumed or not consumed.
Pain score was assessed in two ways. A VAS designed to enable patients to make an overall assessment of the pain that they had suffered on movement between discharge at 4 and 24 h postoperatively was used as the primary outcome variable for the study. On this scale, patients were asked to assess their pain on movement, where 0 indicates they have been suffering from no pain during the period between discharge and assessment at 24 h and 100 indicates they have been suffering from the worst pain imaginable between discharge until assessment. This assessment was recorded at the patient's home by the visiting researchers. In addition, pain at the time of admission (baseline), and 1, 2, 4 and 24 h time points after surgery was also assessed on a VAS score of 0-100 mm (where 0 is no pain at this time and 100 is worst possible pain at this time). Pain was assessed on movement; however, movement was inevitably minimal until the 4 h assessment when patient mobilization occurred.
Patients were not premedicated and received a standardized general anaesthetic involving preoxygenation and induction with fentanyl 1.5 μg kg−1 and propofol 2-3 mg kg−1 intravenously (i.v.). Ketorolac 10 mg and ondansetron 8 mg were also given at the time of induction, the latter in an attempt to reduce any influence of postoperative nausea and vomiting on the postoperative pain score. Anaesthesia was maintained using sevoflurane in oxygen and nitrous oxide (ratio 1:2) with the patient breathing spontaneously via a laryngeal mask airway. Standard anaesthetic monitoring was employed. No local anaesthetic agents were used during the procedures.
The patients were prospectively allocated randomly into two groups using Lab View 5.1 and a sealed-envelope method. At the end of the operation, the envelope was opened and patients received either 10 mg of preservative-free morphine in 20 mL of normal saline (i.e. 30 mL intra-articular injection) or 10 mg in 1 mL of morphine intramuscularly into the lateral compartment of the thigh on the same side as the operation was performed. The intramuscular group also received 30 mL normal saline intra-articularly at the end of the procedure. All patients received a small dressing where the intramuscular injection would have been, or was actually, placed. The tourniquet was deflated 5 min after injection in all patients. The anaesthetist involved in taking the blood samples and VAS scores was not aware of which treatment the patients had received.
In the recovery room, i.v. meperidine in aliquots of 10-20 mg was available for escape analgesia until the pain score was <40. Thereafter, pain was treated with paracetamol 1 g 6 hourly as required and ibuprofen 400 mg 8 hourly, the latter being added if analgesia was inadequately controlled by paracetamol. Patients were discharged in accordance with day case discharge protocols. The supplementary analgesic requirement (i.e. number of tablets taken in 24 h) pain score at 24 h; and pain experienced between discharge and 24 h were all recorded at 24 h by researchers visiting patients at home.
Venous blood samples were collected at 1, 2, 4, and 24 h post-injection for analysis of plasma morphine concentration. These samples were collected in plain tubes and taken immediately to the laboratory for analysis. The blood samples were centrifuged at 3000 rpm for 15 min, the plasma separated, and frozen and stored at −4°C for analysis of morphine concentration by solid-phase extraction, high-performance liquid chromatography (HPLC) and electrochemical detection. The limit of sensitivity was 1 ng mL−1. Interassay verification was performed. Human serum and Sorensen's buffer spiked with morphine in three concentrations (10, 100 and 200 ng mL−1) were assayed on the same day as the procedure (n = 6). Within run precision (intra-assay coefficients of variation) for these concentrations in serum samples were calculated and were 8.66%, 8.32% and 10.37%, respectively. Results for Sorensen's buffer were similar. Between day data (interassay) for human serum (400 μL) for three concentrations from high-, middle- and low-control samples (10, 100 and 200 ng mL−1) were obtained by performing the normal method of extraction and running them on the HPLC system on different days. Coefficients of variation were also measured and found to be 13.21%, 9.99% and 5.76%, respectively.
Forty patients undergoing day case knee arthroscopy were recruited to the study over a 9-month period. The duration of symptoms in both the groups' patients were chronic and varied from 2 to 30 yr. The symptoms in the intra-articular group were secondary to trauma. In the intramuscular group, it was again largely due to trauma although in three patients the symptoms were secondary to osteoarthritis (Fisher's exact test revealed no significant difference between the groups). One patient was withdrawn from the study due to problems with i.v. access. We were not able to collect the 24-h blood sample in five patients. The patient characteristics data in both groups were similar and are shown in Table 1.
Post discharge data are shown in Table 2. The assessment of pain experienced between discharge (4 h) and 24 h was significantly better in the intra-articular group (n = 20; mean ± SD: 18 ± 20, range: 0-77) than the intramuscular group (n = 19; mean ± SD: 34 ± 20, range: 4-64) (P = 0.027) (Table 2). The number of patients consuming any additional analgesia between discharge and 24 h was significantly different between groups (P = 0.038), with 4 (20%) patients in the intra-articular group and 11 (60%) patients in the intramuscular group consuming supplementary analgesia (Table 2). Four patients in each group required meperidine in the first hour after the procedure.
VAS scores assessing pain at the 1, 2, 4 and 24 h time points after injection are shown in Table 3 and no difference was found between groups at any of the time points.
Plasma morphine concentrations are shown in Table 4. After the completion of 11 patients, on the advice of the laboratory, we started to collect a 15 min blood sample for plasma morphine concentration. The laboratory required an early sample, which would give a potentially high concentration to assist with assay verification. The randomization code was not broken. The 15 min sample was not considered vital to the study as the main emphasis of the study lay in efficacy of post-discharge analgesia, though the results of the 15 min sample are included in Table 4 for completeness. As SDs were large, non-parametric analysis was applied and there was no difference between groups. The plasma morphine concentration in one patient in the intra-articular group was very high (211 ng mL−1) and this could not be explained, unless it was given i.v. in error. Even when this was removed from the analysis, there was no significant difference between the two groups. None of the patients in either group showed any significant side-effects from the morphine administration.
In our study, overall pain experienced between discharge and 24 h, and the need for supplementary analgesia during that time were significantly lower in the intra-articular group than the intramuscular group. The use of more supplementary analgesia in the intramuscular group would explain the similar pain scores found in the pain assessment taken at the 24 h time point. This study suggests that intra-articular morphine has a longer duration of action than systemic morphine.
One weakness of the study may lie in the fact that the VAS scores in the Day Surgery Unit (i.e. in the first 4 h) were taken when patients were inevitably not very mobile, whilst the VAS score taken at 24 h were in their homes by which time they would have been more mobile. It is almost impossible to overcome this problem in study design where patients are making a rapid recovery. This factor, however, would have been the same in both groups and consequently would not have influenced patients' overall assessment of pain since discharge.
Intra-articular morphine has been widely studied as a form of pain relief after knee arthroscopy and meta-analyses has shown intra-articular morphine to produce a mild but definite analgesia compared to placebo, an effect that may be dose dependent [1,2]. The site of action of this analgesic effect is open to debate. Early work by Stein and colleagues  showed that low-dose intra-articular morphine reduced pain after knee surgery when compared with the same dose given i.v. This group showed that this effect could be reversed by intra-articular injection of naloxone and attributed this action to local intra-articular opioid receptors. Similarly, Lawrence and colleagues  also found intra-articular morphine to be effective and provided evidence to suggest that the action of intra-articular morphine is to some extent mediated via a local action within the joint. In this study, synovial tissue samples were successfully assayed for the presence of opioid-binding sites. Animal studies have also shown that when opioids are injected directly into inflamed tissue it produces pain relief that is reversible and dose dependent [6,7]. Keates and colleagues using radio-ligand binding showed an increased density of opioid-binding receptors in inflamed tissues, supporting the clinical use of intra-articular opioids in the treatment of postoperative and chronic inflammatory joint pain . In contrast, other studies have not been able to show any advantage of morphine given by the intra-articular route when compared to placebo or local anaesthetic infiltration [9-13].
A meta-analysis by Møiniche and colleagues  has shown that there is some evidence that intra-articular bupivacaine is effective in reducing mild to moderate postoperative pain in patients undergoing knee arthroscopies, although this is of short duration. However, if bupivacaine is effective, there is evidence to suggest that intra-articular morphine may be more so and work comparing intra-articular morphine and bupivacaine has shown patients receiving morphine to have lower pain scores  and require less supplementary analgesics [15-17]. Cepeda and colleagues  compared intra-articular morphine 10 mg, the same dose given by the subcutaneous route and intra-articular bupivacaine 0.5% 20 mL with epinephrine. They showed that the morphine given by either route provided better and longer-lasting postoperative pain relief than intra-articular bupivacaine with epinephrine. In addition, studies looking at the dose-response relationship have found that increasing doses of intra-articular morphine are associated with improved analgesia [18-21].
Three studies have looked at the plasma concentration of morphine after intra-articular injection. Joshi and colleagues , assessing the analgesic effects of intra-articular morphine 5 mg vs. placebo, measured plasma concentrations up to 4h post-injection and found the concentration measured was always below 10 ng mL−1 during the sampling period. Richardson and colleagues  and Brandsson and colleagues , in their studies comparing i.v. morphine against intra-articular morphine, also measured plasma morphine concentrations. In the former study, the single measurement done after 2 h was again less than 10 ng mL−1 in all the three groups (i.e. 5 mg i.v.), 1 mg intra-articular and 5 mg intra-articular groups. In Brandsson's study, where measurements were done up to 6 h post-injection, the peak plasma morphine concentrations after 5 mg i.v., 1 mg intra-articular and 5 mg intra-articular injections were 67 ng mL−1, 3 ng mL−1 and 20 ng mL−1, respectively. In these studies, the plasma morphine concentrations were thought to be too low to produce any significant systemic analgesic effects.
Studies have shown that the minimum effective plasma morphine concentration varies from 16 to 40 ng mL−1[23,24]. In our study in the intra-articular group, the morphine levels at 15 min and 1 h only just achieved this level. This perhaps suggests that any absorption from the knee through open blood vessels is minimal, but does occur. The effect may be dose dependent as Richardson and colleagues  found the plasma morphine concentrations in patients who were given intra-articular morphine 5 mg were significantly higher than those given intra-articular morphine 1 mg. This early absorption from the knee may contribute to early analgesia (in fact, we showed no difference in morphine concentrations between intra-articular and intramuscular routes), but as plasma morphine concentrations were consistently low, the overall improvement in analgesia observed is probably largely a local, intra-articular phenomenon.
The 10 mg morphine intramuscular injection did not reach above the minimum effective concentration at any time explaining the increased analgesic requirement in this group. Pharmacokinetic studies have shown that after intramuscular injection, peak plasma levels ranging from 50 to 60 ng mL−1 can be achieved within 10-20 min after the injection and elimination half-life to be about 4.5 h . The mean plasma morphine levels in our study at 15 min were found to be lower than this and low levels of morphine were found in the blood at 24h, so that any effect there did not persist.
In summary, this study demonstrates that intra-articular morphine provides better pain relief after knee arthroscopy following discharge from hospital. Although there may be an early minor systemic contribution, a greater peripheral, intra-articular contribution is likely.
This work was kindly supported by a Development Grant from the Royal College of Anaesthetists.
1. Gupta A, Bodin L, Holmstrom B, Berggren L. A systematic review of the peripheral analgesic effects of intraarticular morphine. Anesth Analg
2001; 93: 761-770.
2. Kalso E, Tramer M, Carroll D, McQuay HJ, Moore RA. Pain relief from intra-articular morphine after knee surgery: a qualitative systematic review. Pain
1997; 71: 127-134.
3. Stein C, Comisel K, Haimerl E, et al.
Analgesic effect of intraarticular morphine after arthroscopic knee surgery. N Engl J Med
1991; 325: 1123-1126.
4. Lawrence AJ, Joshi GP, Michalkiewicz A, Blunnie WP, Moriarty DC. Evidence for analgesia mediated by peripheral opioid receptors in inflamed synovial tissue. Eur J Clin Pharmacol
1992; 43: 351-355.
5. Joshi GP, McCarroll SM, Cooney CM, Blunnie WP, O'Brien TM, Lawrence AJ. Intra-articular morphine for pain relief after knee arthroscopy. J Bone Joint Surg Br
1992; 74: 749-751.
6. Stein C, Millan MJ, Yassouridis A, Herz A. Antinociceptive effects of mu- and kappa-agonists in inflammation are enhanced by a peripheral opioid receptor-specific mechanism. Eur J Pharmacol
1988; 155: 255-264.
7. Stein C, Millan MJ, Shippenberg TS, Peter K, Herz A. Peripheral opioid receptors mediating antinociception in inflammation. Evidence for involvement of mu, delta and kappa receptors. J Pharmacol Exp Ther
1989; 248: 1269-1275.
8. Keates HL, Cramond T, Smith M. Intraarticular and periarticular opioid binding in inflamed tissue in experimental canine arthritis. Anesth Analg
1999; 89: 409-415.
9. Heard SO, Edwards WT, Ferrari D, et al.
Analgesic effect of intraarticular bupivacaine or morphine after arthroscopic knee surgery: a randomized, prospective, double-blind study. Anesth Analg
1992; 74: 822-826.
10. Raja SN, Dickstein RE, Johnson CA. Comparison of postoperative analgesic effects of intraarticular bupivacaine and morphine following arthroscopic knee surgery. Anesthesiology
1992; 77: 1143-1147.
11. Laurent SC, Nolan JP, Pozo JL, Jones CJ. Addition of morphine to intra-articular bupivacaine does not improve analgesia after day-case arthroscopy. Br J Anaesth
1994; 72: 170-173.
12. De Andres J, Valia JC, Barrera L, Colomina R. Intra-articular analgesia after arthroscopic knee surgery: comparison of three different regimens. Eur J Anaesthesiol
1998; 15: 10-15.
13. Aasbo V, Raeder JC, Grogaard B, Roise O. No additional analgesic effect of intra-articular morphine or bupivacaine compared with placebo after elective knee arthroscopy. Acta Anaesthesiol Scand
1996; 40: 585-588.
14. Møiniche S, Mikkelsen S, Wettesley J, Dahl JB. A systematic review of intra-articular local anesthesia for postoperative pain relief after arthroscopic knee surgery. Reg Anesth Pain Med
1999; 24: 430-437.
15. Joshi GP, McCarroll SM, O'Brien TM, Lenane P. Intraarticular analgesia following knee arthroscopy. Anesth Analg
1993; 76: 333-336.
16. Heine MF, Tillet ED, Tsueda K, et al.
Intra-articular morphine after arthroscopic knee operation. Br J Anaesth
1994; 73: 413-415.
17. Cepeda MS, Uribe C, Betancourt J, Rugeles J, Carr DB. Pain relief after knee arthroscopy: intra-articular morphine, intra-articular bupivacaine, or subcutaneous morphine? Reg Anesth
1997; 22: 233-238.
18. Likar R, Kapral S, Steinkellner H, Stein C, Schafer M. Dose-dependency of intra-articular morphine analgesia. Br J Anaesth
1999; 83: 241-244.
19. Kanbak M, Akpolat N, Ocal T, Doral MN, Ercan M, Erdem K. Intraarticular morphine administration provides pain relief after knee arthroscopy. Eur J Anaesthesiol
1997; 14: 153-156.
20. Rosseland LA, Stubhaug A, Skoglund A, Breivik H. Intra-articular morphine for pain relief after knee arthroscopy. Acta Anaesthesiol Scand
1999; 43: 252-257.
21. Richardson MD, Bjorksten AR, Hart JA, McCullough K. The efficacy of intra-articular morphine for postoperative knee arthroscopy analgesia. Arthroscopy
1997; 13: 584-589.
22. Brandsson S, Karlsson J, Morberg P, Rydgren B, Eriksson BI, Hedner T. Intra-articular morphine after arthroscopic ACL reconstruction: a double blind placebo controlled study of 40 patients. Acta Orthop Scand
2000; 71: 280-285.
23. Dahlstrom B, Tamsen A, Paalzow L, Hartvig P. Patient-controlled analgesic therapy, Part IV: pharmacokinetics and analgesic plasma concentrations of morphine. Clin Pharmacokin
1982; 7: 266-279.
24. Graves DA, Arrigo JM, Foster TS, Baumann TJ, Batenhorst RL. Relationship between plasma morphine concentrations and pharmacologic effects in postoperative patients using patient-controlled analgesia. Clin Pharm
1985; 4: 41-47.
25. Stanski DR, Greenblatt DJ, Lowenstein E. Kinetics of intravenous and intramuscular morphine. Clin Pharmacol Ther
1978; 24: 52-59.