- Email: [email protected]
Efficacy and respiratory effects of low-dose spinal morphine for postoperative analgesia following knee arthroplasty
British Journal of Anaesthesia 85 (2): 233±7 (2000)
Ef®cacy and respiratory effects of low-dose spinal morphine for postoperative analgesia following knee arthroplasty P. J. Cole*, D. A. Craske and R. G. Wheatley The Acute Pain Management Unit, York District Hospital, Wigginton Road, York YO3 7HE, UK *Corresponding author A randomized, double-blind study of 38 patients undergoing total knee replacement was undertaken to compare the ef®cacy and respiratory effects of low-dose spinal morphine and patientcontrolled i.v. morphine against patient-controlled i.v. morphine alone. Patients received either morphine 0.3 mg or saline 0.3 ml with 0.5% heavy spinal bupivacaine 2±2.5 ml. Respiratory effects were measured continuously for 14 h postoperatively with an Edentec 3711 respiratory monitor. There was an improvement in pain relief in the intrathecal morphine group, with signi®cantly lower median VAS pain scores on movement at 4 h (0 (median 0±1.5) vs 5 (1.25± 7.75) P<0.01), 12 h (2 (1±5) vs 6 (3±8) P<0.01) and 24 h (3 (1±5) vs 5 (3±7) P<0.05) postoperatively, despite using signi®cantly less patient-controlled morphine (20 mg (10.25±26.25) vs 38.5 mg (27±51) P<0.01) in the ®rst 24 h. There was a small but statistically signi®cant reduction in the median oxygen saturation (SpO2) in the intrathecal morphine group 97 (95±99)% compared with the placebo group 99 (97±99)% (P<0.05). Although marked disturbances in respiratory pattern were observed in both groups, none of the patients in the study had severe hypoxaemia (SpO2 <85% >6 min h±1) and there was no signi®cant difference in the incidence of mild (SpO2 <94% >12 min h±1) or moderate (SpO2 <90% >12 min h±1) hypoxaemia or in the incidence of episodes of apnoea or hypopnoea in the two groups. Br J Anaesth 2000; 85: 233±7 Keywords: anaesthetic techniques, subarachnoid; analgesics, opioid; analgesics, morphine; surgery, orthopaedic Accepted for publication: February 2, 2000
Intrathecal opioids have been shown to provide effective analgesia in a variety of surgical settings1 2 since the introduction of the technique into clinical practice in 1979.3 The advantage of spinally administered opioids is that prolonged analgesia can be provided using a single injection at the time of surgery without the need for cumbersome and expensive pumps in the postoperative period. However, the use of this technique has been limited by a high incidence of opioid-related side effects including nausea, pruritus and urinary retention, and the fear of respiratory depression, which may be delayed in onset. In an attempt to limit major and minor opioid side effects, the use of low-dose spinal opioids has been advocated.4±6 When using low doses there are concerns about the appropriate choice of rescue analgesia if patients have ineffective pain relief. Although studies using intrathecal opioids have demonstrated improved analgesia, there have been a limited number of studies that assess their respiratory effects.7 The aim of this study was to investigate whether the combination of low-dose (0.3 mg) intrathecal morphine and
patient-controlled i.v. morphine provides more effective analgesia than patient-controlled i.v. morphine alone in patients undergoing knee arthroplasty. A secondary aim was to investigate the respiratory effects of the combination of intrathecal morphine and patient-controlled i.v. morphine as rescue analgesia.
Methods After obtaining Hospital Ethics Committee approval and informed patient consent, we studied 38 ASA 1 or 2 patients scheduled for unilateral knee arthroplasty, in a prospective, randomized, double-blind, placebo-controlled study. Patients undergoing knee arthroplasty were chosen for this study because audit from our acute pain team has consistently shown that these patients have high analgesic requirements and require early mobilization. Patients in whom either non-steroidal anti-in¯ammatory drugs or spinal anaesthesia were contraindicated were excluded from the study.
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000
Cole et al.
The anaesthetic management of all patients was standardized. Spinal anaesthesia was undertaken using a mixture of 0.5% heavy bupivacaine 2±2.5 ml plus a pharmacyprepared solution of morphine or saline, which was allocated on a double-blind randomized basis by pharmacy, using a computer-generated random number system. The randomization was strati®ed to ensure that morning and afternoon cases were divided equally between the two groups. Patients received either morphine 0.3 mg (0.3 ml from a 2-ml ampoule of 1 mg ml±1 preservative-free morphine), or normal saline 0.3 ml (0.3 ml from a 2 ml 0.9% sodium chloride solution) in addition to the bupivacaine in their spinal anaesthetic. General anaesthesia was induced in unpremedicated patients with propofol (1±3 mg kg±1) and fentanyl (0±1 mg kg±1) at the anaesthetist's discretion. Maintenance of anaesthesia was provided with oxygen, nitrous oxide and en¯urane in patients breathing spontaneously through a laryngeal mask airway. In addition to the spinal injection, postoperative analgesia consisted of diclofenac 50 mg every 8 h (PO or PR). Patients were allowed to self-administer i.v. morphine via a patientcontrolled analgesia system (Graseby 9300 model) in 1 mg boluses with a 5-min lockout period. Metoclopramide was routinely prescribed for postoperative nausea and vomiting. All patients were given oxygen by nasal cannulae at 2 litres min±1, if at any time their oxygen saturation was noted to be 94% or below. The use of a visual analogue scale was described at the preoperative visit and the effectiveness of analgesia was measured by visual analogue pain scores on movement (VASm 0±10) at 4, 12 and 24 h. Patients were also asked to give a verbal rating score for satisfaction with the quality of analgesia (poor, fair, adequate, good, excellent), and pain severity (absent, mild, discomforting, distressing, excruciating) for the 24-h period. In addition, the PCA 24 h morphine consumption was recorded. Respiration was monitored continuously using the Edentec model 3711 digital recorder (Edentec 3711, Nellcor, Eden Prairie, MN, USA) for 1 h preoperatively for a baseline reading, and then for 13±14 h after the patient had returned to the orthopaedic ward postoperatively. This digital recorder comprises a lightweight portable recorder, with non-invasive probes attached to the subject, continuously measuring nasal air¯ow, chest impedance and pulse oximetry. Chest impedance and heart rate were recorded by attaching two electrodes to each side of the chest (mid axillary line at T4±5). Chest wall movement was measured by passing a constant current through the chest giving a measure of impedance. Nasal air¯ow was measured by a thermistor attached between the upper lip and nose. All Edentec respiratory data were downloaded onto a desktop computer. Each tracing was assessed in 10-min epochs by two of the authors (P.C., D.C.) in a blinded fashion. These data were edited to exclude obvious artefact recordings, which included non-physiological hypoxaemia,
interference due to patient movement and/or disconnection from the recorder. The edited data were then analysed using an ETS 2.0 software program. An apnoea event starts when the air¯ow stays between +5% and ±5% of baseline for at least 10 s, and ends when the air¯ow is greater than 5% or less than ±5%. A hypopnoea event starts when the trace amplitude is less than 50% of the average amplitude over the previous 2 min and stays at or below this level for at least 10 s. The event ends as soon as the amplitude increases above 50% of the average amplitude.8 Frequencies of events were calculated per hour of recording. Studies with less than 4 h of recorded data were regarded as technically unsatisfactory and rejected. The respiratory distress index (RDI) is the sum of apnoea and hypopnoea events per hour. An RDI greater than 30 is said to be diagnostic of sleep apnoea/hypopnoea syndrome requiring treatment.8 Hypoxaemia was classi®ed by previously de®ned criteria9 as mild when the SpO2 was <94% >12 min h±1, moderate when the SpO2 <90% >12 min h±1 and severe when the SpO2 <85% >6 min h±1. The mean saturations and the percentage time spent below a saturation of 95% during the study period for each patient were also recorded. Measurement of oxygen saturation, sedation scoring (0=none, patient alert; 1=mild, occasionally drowsy, easy to arouse; 2=moderate, frequently drowsy, easy to arouse; 3=severe; somnolent, dif®cult to arouse), respiratory rate and visual analogue score for pain were recorded on the ward at hourly intervals for the ®rst 4 h and then every 4 h thereafter. The incidence of nausea or pruritus requiring treatment was noted. Previously collected audit data on patients who had received a spinal anaesthetic with bupivacaine alone and a morphine PCA following knee arthroplasty had shown that the mean morphine consumption was 40 (SD 16) mg. Assuming a clinically signi®cant reduction in morphine consumption of 20 mg, it was calculated10 that 30 patients (15 pairs) would be required to have a 90% power of detecting a 50% reduction in 24 h morphine consumption at a signi®cance level of 0.05. Data were analysed by using the SPSS statistics package (version 8). Mann±Whitney and chi-squared tests were used for the non-parametric data, and an unpaired Student's t-test for the patient characteristics. A value of P<0.05 was regarded as statistically signi®cant.
Results Thirty-eight patients completed the study. Two patients were excluded as an incorrect dose of intrathecal drug was administered to one and premedication was given to the second. Due to software problems with the Edentec monitor and patient non-compliance, a further four patients had incomplete respiratory data for analysis, but their data on morphine consumption were included.
234
Effects of low-dose spinal morphine
Fig 1 Box plot of visual analogue score on movement (VASm) scores at 4, 12 and 24 h postsurgery. Boxes represent interquartile range, horizontal lines in the boxes represent median values, and error bars the range. The VASm scores were signi®cantly reduced at 4 (P<0.01), 12 (P=0.01) and 24 (P<0.05) h in the intrathecal morphine group when compared with the intrathecal saline group. Open inverted triangle represents an outlying value.
Patient characteristics were similar between the two groups. Mean (SD) age, body weight and height were 70.3 (9.9) yr, 82.3 (15.7) kg and 1.7 (0.1) m, and 71.3 (7.6) yr, 80.5 (14.3) kg and 1.7 (0.1) m in the intrathecal morphine and intrathecal saline groups, respectively. The preoperative baseline respiratory monitoring showed median saturations (interquartile range) of 98.5 (96±99)% in the morphine group and 98.5 (97±99)% in the control group. Patients in the intrathecal morphine group had signi®cantly lower VASm pain scores on movement at 4, 12 and 24 h (Fig. 1), and used a signi®cantly lower dose of patientcontrolled i.v. morphine in the ®rst 24 h after surgery (Fig. 2). Although more patients in the intrathecal morphine group scored the severity of their pain as mild or absent and more patients in this group required treatment for emesis, these differences were not statistically signi®cant. There was no signi®cant difference between the two groups in the incidence of pruritus or their verbal rating of satisfaction (Table 1). When comparing the apnoea and hypopnoea episodes, there was no signi®cant difference between the two groups in the median RDI or in the incidence of those with a respiratory distress index >30 (Table 2). Patients in both groups had disordered respiratory patterns with episodes of central and obstructive apnoea. The median values of the mean saturations in the postoperative period between the two groups were compared and were signi®cantly lower in the intrathecal morphine group compared with the intrathecal saline group.
Fig 2 Box plot of i.v. PCA morphine consumption in the ®rst 24 h postsurgery in the intrathecal morphine and intrathecal saline (placebo) groups. Boxes represent interquartile range, horizontal lines in the boxes represent median values, and error bars the range. There is a signi®cant difference between the intrathecal morphine and intrathecal placebo groups (P<0.01). Closed diamond and closed square represent outlying values.
Using previously de®ned criteria,9 5/17 patients in the intrathecal morphine group could be classi®ed as mildly hypoxaemic and 1/17 as moderately hypoxaemic during the study period. In the saline group, 3/15 subjects showed mild hypoxaemia. However, this difference was not statistically signi®cant. No patients had severe hypoxaemia using these criteria. Although patients in the intrathecal morphine group spent a median of 10% of their time with a saturation below 95% compared with a median of 3% in the placebo group, this was not statistically signi®cant.
Discussion This study has achieved the primary aim of demonstrating that, following knee arthroplasty, intrathecal morphine 0.3 mg followed by patient-controlled i.v. morphine, improves postoperative analgesia and reduces i.v. morphine consumption compared with patient-controlled i.v. morphine alone. This con®rms earlier studies on the quality of analgesia after Caesarean section11±14 and hip surgery.15 Whilst this combination of intrathecal and patientcontrolled i.v. morphine provides effective analgesia, there are concerns in clinical practice about the safety of intrathecal opioids particularly if systemic opioids are used for rescue analgesia. A 17-nation European survey revealed that, except for the UK, morphine was the most commonly used of the long-acting opioids.16 This initial study used morphine because it has been extensively used, and whatever degree of respiratory depression was seen with intrathecal morphine, respiratory depression would be less
235
Cole et al. Table 1 Highest sedation score, minor side effects and patient satisfaction/severity during the ®rst 24 h. There was no difference in the median highest sedation scores in the ®rst 24 h between the two groups. Median values and interquartile ranges are shown. The number of patients requiring antiemetic treatment and reporting pruritis was higher in the intrathecal morphine group, but this was not statistically signi®cantly. Patients reported their analgesia as good or excellent, and their pain severity as absent or mild more frequently in the intrathecal morphine group compared with the saline group
Morphine Placebo P-value
Sedation score
Emesis (treated) (n=31)
Pruritus (n=29)
Satisfaction (good/excellent) (n=32)
Severity (absent/mild) (n=31)
1 [0±2] 1 [0±1.5] NA
Nine out of 15 (60%) Six out of 16 (38%) 0.07
Four out of 15 (27%) Two out of 14 (14%) 0.65
Nine out of 15 (60%) Nine out of 17 (53%) 0.96
12 out of 14 (86%) Nine out of 17 (53%) 0.07
Table 2 Postoperative respiratory data. Median (interquartile range) respiratory distress indices (RDI) and number of patients who had an RDI count >30 are shown, there was no signi®cant difference within these groups. The median values of the mean saturations (interquartile range) (%) show a statistical difference between the two groups*. There were more episodes of hypoxaemia in the intrathecal morphine group compared with the saline group but this did not reach statistical signi®cance. *P<0.05
RDI RDI count >30 Median SpO2 (%) Incidence of mild/moderate hypoxaemia
Morphine group (n=17)
Placebo group (n=15)
P-value
19 (4±35.5) Six out of 17 97 (95±99) Six out of 17
13 (3±34) Five out of 15 99 (97±99) Three out of 15
0.53 0.91 0.03* 0.36
likely to occur when low doses of more lipid-soluble opioids are administered by the intrathecal route.7 17 Due to the dose-dependent nature of respiratory depression,5 a dose of morphine 0.3 mg was used, based on an earlier dose±response study that showed that this is an effective dose with the fewest side effects.4 However, more recent studies have shown that lower doses (0.1 mg) are effective for analgesia after Caesarean section12 and hip replacement surgery.14 There have been many reports of respiratory depression following spinal opioids.7 Previous studies looking at the respiratory effects of intrathecal opioids have used a combination of pulse oximetry, sedation scoring, respiratory rate and invasive arterial blood±gas analysis.5 7 18 The availability of relatively simple to use, bedside respiratory monitors allows us to continuously study the ventilatory pattern and oxygenation of patients receiving intrathecal opioids. Although it is often assumed that bedside monitors such as the Edentec are less accurate than conventional sleep studies with polysomnography, this equipment has recently been used to validate the detection of respiratory disturbances in patients with sleep apnoea/ hypopnoea syndrome8 and to investigate patients recovering from major abdominal surgery.19 Our study demonstrated that respiratory disturbances occurred in both groups. The pattern of respiratory dysfunction was similar to that previously described following the use of i.v. morphine,20 21 and included obstructive and central apnoeas. Although patients in the morphine group spent three times as long with a saturation below 95% and twice as many were classi®ed as `hypoxaemic patients'; that is they spent greater than 20% of their time with mild or moderate
hypoxaemia (85±94%),9 these differences were not statistically signi®cant. The only statistical difference was between the median oxygen saturations in the two groups. In view of the risks of postoperative hypoxaemia22 and its potential effects on postoperative morbidity, it was deemed necessary to give patients in this study supplementary oxygen if their oxygen saturations fell below 94%. Because oxygen was not administered to all the study patients, the small but statistically signi®cant difference between the mean saturations (97% and 99%) may not be of clinical signi®cance. None of the patients in the study met the pre-de®ned criteria9 for severe hypoxaemia, that is, spending more than 10% of their time with a saturation below 85%. The only patient in the study who approached this level was a patient in the spinal morphine group who spent 90% of their time with saturation below 95% including 4% of their time below 85%. This would be regarded by most clinicians as a signi®cant level of hypoxaemia. Because of the low incidence of respiratory depression, a much larger study would be needed to detect or exclude a signi®cantly increased risk with low-dose spinal morphine. Further studies are required into the ef®cacy and duration of action of more lipid-soluble opioids which are likely to be associated with a lower incidence of respiratory depression.
Acknowledgements We thank Mrs E. McManus, senior pharmacist at York District Hospital, acute pain nurses D. Hunter and R. Clayton, and Miss V. Algar, statistician at St James's University Teaching Hospital, Leeds for their help in this study.
236
Effects of low-dose spinal morphine
References 1 Rawal N. Epidural and spinal agents for postoperative analgesia. Surg Clinics North Am 1999; 79: 313±44 2 Carr DB, Cousins MJ. Spinal route of analgesia: opioids and future options. In: Cousins MJ, Bridenbaugh PO, eds, Neural Blockade in Clinical Anaesthesia and Management of Pain, 3rd edn. Philadelphia: Lippincott-Raven, 1998 3 Wang JK, Nauss LA, Thomas JE. Pain relief by intrathecally applied morphine in man. Anaesthesiology 1979; 50: 149±51 4 Sarma VJ, Bostrom UV. Intrathecal morphine for the relief of post hysterectomy pain ± a double blind dose response study. Acta Anaesthesiol Scand 1993; 37: 223±7 5 Bailey PL, Rhondeau S, Schafer PG et al. Dose±response pharmacology of intrathecal morphine in human volunteers. Anaesthesiology 1993; 79: 49±59 6 Jacobson L, Chabal C, Brody MC. A dose±response study of intrathecal morphine: ef®cacy, duration, optimum dose, and side effects. Anaesth Analg 1988; 67: 1082±8 7 Etches RC, Sandler AN, Daley MD. Respiratory depression and spinal opioids. Can J Anaesth 1989; 36: 165±85 8 Whittle AT, Finch SP, Mortimer IL, MacKay TW, Douglas NJ. Use of home sleep studies for diagnosis of sleep apnoea/ hypopnea syndrome. Thorax 1997; 52: 1068±73 9 Wheatley RG, Shepherd D, Jackson IJB, Madej TH, Hunter D. Hypoxaemia and pain relief after upper abdominal surgery: comparison of I.M. and patient controlled analgesia. Br J Anaesth 1992; 69: 558±61 10 Altman DG. How large a sample? Statistics and ethics in medical research. BMJ 1980; 281: 1336±8 11 Swart M, Sewell J, Thomas D. Intrathecal morphine for Caesarean section: an assessment of pain relief, satisfaction and side effects. Anaesthesia 1997; 52: 373±7 12 Palmer CM, Emerson S, Volgoropolous D, Alves D. Dose± response relationship of intrathecal morphine for postcesarean analgesia. Anesthesiology 1999; 90: 437±44
237
13 Husaini SW, Russell IF. Intrathecal diamorphine compared with morphine for postoperative analgesia after Caesarean section under spinal anaesthesia. Br J Anaesth 1998; 81: 135±9 14 Gerancher JC, Floyd H, Eisanach J. Determination of an effective dose of intrathecal morphine for pain relief after cesarean delivery. Anesth Analg 1999; 88: 346±51 15 Slappendel R, Weber EW, Dirksen R, Gielen MJ, Van Limbeek J. Optimisation of the dose of intrathecal morphine in total hip surgery: a dose ®nding study. Anesth Analg 1999; 88: 822±6 16 Rawal N, Allvin R. Euro-pain study on acute pain. Epidural and intrathecal opioids for postoperative pain management in Europe ± a 17 nation questionnaire study of selected hospitals. Acta Anaesthesiol Scand 1996; 40: 1119±26 17 Hamber EA, Viscomi CM. Intrathecal lipophilic opioids as adjuncts to surgical spinal analgesia. Reg Anaesth Pain Medicine 1999; 24: 255±63 18 Kelly MC, Carabine UA, Mirakhur RK. Intrathecal diamorphine for analgesia after Caesarean section. A dose ®nding study and assessment of side effects. Anaesthesia 1998; 53: 231±7 19 Rosenberg J, Rasmussen GI, Wojdemann KR, Kirkeby LT, Jorgensen LN, Kehlet H. Ventilatory pattern and associated episodic hypoxaemia in the late postoperative period in the general surgical ward. Anaesthesia 1999; 54: 323±28 20 Jones JG, Jordan C, Scudder C, Rocke DA, Barrowcliffe M. Episodic postoperative oxygen saturation: the value of added oxygen. J Roy Soc Med 1985; 78: 1019±22 21 Catley DM, Thornton C, Jordan C, Lehane JR, Royston D, Jones JG. Pronounced, episodic oxygen desaturation in the postoperative period; its association with ventilatory pattern and analgesic regime. Anaesthesiology 1985; 63: 20±8 22 Reeder MK, Muir AD, Foex P, Goldman MD, Loh L, Smart D. Postoperative myocardial ischaemia: temporal association with nocturnal hypoxaemia. Br J Anaesthesia 1991; 67: 626±31
Ef®cacy and respiratory effects of low-dose spinal morphine for postoperative analgesia following knee arthroplasty P. J. Cole*, D. A. Craske and R. G. Wheatley The Acute Pain Management Unit, York District Hospital, Wigginton Road, York YO3 7HE, UK *Corresponding author A randomized, double-blind study of 38 patients undergoing total knee replacement was undertaken to compare the ef®cacy and respiratory effects of low-dose spinal morphine and patientcontrolled i.v. morphine against patient-controlled i.v. morphine alone. Patients received either morphine 0.3 mg or saline 0.3 ml with 0.5% heavy spinal bupivacaine 2±2.5 ml. Respiratory effects were measured continuously for 14 h postoperatively with an Edentec 3711 respiratory monitor. There was an improvement in pain relief in the intrathecal morphine group, with signi®cantly lower median VAS pain scores on movement at 4 h (0 (median 0±1.5) vs 5 (1.25± 7.75) P<0.01), 12 h (2 (1±5) vs 6 (3±8) P<0.01) and 24 h (3 (1±5) vs 5 (3±7) P<0.05) postoperatively, despite using signi®cantly less patient-controlled morphine (20 mg (10.25±26.25) vs 38.5 mg (27±51) P<0.01) in the ®rst 24 h. There was a small but statistically signi®cant reduction in the median oxygen saturation (SpO2) in the intrathecal morphine group 97 (95±99)% compared with the placebo group 99 (97±99)% (P<0.05). Although marked disturbances in respiratory pattern were observed in both groups, none of the patients in the study had severe hypoxaemia (SpO2 <85% >6 min h±1) and there was no signi®cant difference in the incidence of mild (SpO2 <94% >12 min h±1) or moderate (SpO2 <90% >12 min h±1) hypoxaemia or in the incidence of episodes of apnoea or hypopnoea in the two groups. Br J Anaesth 2000; 85: 233±7 Keywords: anaesthetic techniques, subarachnoid; analgesics, opioid; analgesics, morphine; surgery, orthopaedic Accepted for publication: February 2, 2000
Intrathecal opioids have been shown to provide effective analgesia in a variety of surgical settings1 2 since the introduction of the technique into clinical practice in 1979.3 The advantage of spinally administered opioids is that prolonged analgesia can be provided using a single injection at the time of surgery without the need for cumbersome and expensive pumps in the postoperative period. However, the use of this technique has been limited by a high incidence of opioid-related side effects including nausea, pruritus and urinary retention, and the fear of respiratory depression, which may be delayed in onset. In an attempt to limit major and minor opioid side effects, the use of low-dose spinal opioids has been advocated.4±6 When using low doses there are concerns about the appropriate choice of rescue analgesia if patients have ineffective pain relief. Although studies using intrathecal opioids have demonstrated improved analgesia, there have been a limited number of studies that assess their respiratory effects.7 The aim of this study was to investigate whether the combination of low-dose (0.3 mg) intrathecal morphine and
patient-controlled i.v. morphine provides more effective analgesia than patient-controlled i.v. morphine alone in patients undergoing knee arthroplasty. A secondary aim was to investigate the respiratory effects of the combination of intrathecal morphine and patient-controlled i.v. morphine as rescue analgesia.
Methods After obtaining Hospital Ethics Committee approval and informed patient consent, we studied 38 ASA 1 or 2 patients scheduled for unilateral knee arthroplasty, in a prospective, randomized, double-blind, placebo-controlled study. Patients undergoing knee arthroplasty were chosen for this study because audit from our acute pain team has consistently shown that these patients have high analgesic requirements and require early mobilization. Patients in whom either non-steroidal anti-in¯ammatory drugs or spinal anaesthesia were contraindicated were excluded from the study.
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000
Cole et al.
The anaesthetic management of all patients was standardized. Spinal anaesthesia was undertaken using a mixture of 0.5% heavy bupivacaine 2±2.5 ml plus a pharmacyprepared solution of morphine or saline, which was allocated on a double-blind randomized basis by pharmacy, using a computer-generated random number system. The randomization was strati®ed to ensure that morning and afternoon cases were divided equally between the two groups. Patients received either morphine 0.3 mg (0.3 ml from a 2-ml ampoule of 1 mg ml±1 preservative-free morphine), or normal saline 0.3 ml (0.3 ml from a 2 ml 0.9% sodium chloride solution) in addition to the bupivacaine in their spinal anaesthetic. General anaesthesia was induced in unpremedicated patients with propofol (1±3 mg kg±1) and fentanyl (0±1 mg kg±1) at the anaesthetist's discretion. Maintenance of anaesthesia was provided with oxygen, nitrous oxide and en¯urane in patients breathing spontaneously through a laryngeal mask airway. In addition to the spinal injection, postoperative analgesia consisted of diclofenac 50 mg every 8 h (PO or PR). Patients were allowed to self-administer i.v. morphine via a patientcontrolled analgesia system (Graseby 9300 model) in 1 mg boluses with a 5-min lockout period. Metoclopramide was routinely prescribed for postoperative nausea and vomiting. All patients were given oxygen by nasal cannulae at 2 litres min±1, if at any time their oxygen saturation was noted to be 94% or below. The use of a visual analogue scale was described at the preoperative visit and the effectiveness of analgesia was measured by visual analogue pain scores on movement (VASm 0±10) at 4, 12 and 24 h. Patients were also asked to give a verbal rating score for satisfaction with the quality of analgesia (poor, fair, adequate, good, excellent), and pain severity (absent, mild, discomforting, distressing, excruciating) for the 24-h period. In addition, the PCA 24 h morphine consumption was recorded. Respiration was monitored continuously using the Edentec model 3711 digital recorder (Edentec 3711, Nellcor, Eden Prairie, MN, USA) for 1 h preoperatively for a baseline reading, and then for 13±14 h after the patient had returned to the orthopaedic ward postoperatively. This digital recorder comprises a lightweight portable recorder, with non-invasive probes attached to the subject, continuously measuring nasal air¯ow, chest impedance and pulse oximetry. Chest impedance and heart rate were recorded by attaching two electrodes to each side of the chest (mid axillary line at T4±5). Chest wall movement was measured by passing a constant current through the chest giving a measure of impedance. Nasal air¯ow was measured by a thermistor attached between the upper lip and nose. All Edentec respiratory data were downloaded onto a desktop computer. Each tracing was assessed in 10-min epochs by two of the authors (P.C., D.C.) in a blinded fashion. These data were edited to exclude obvious artefact recordings, which included non-physiological hypoxaemia,
interference due to patient movement and/or disconnection from the recorder. The edited data were then analysed using an ETS 2.0 software program. An apnoea event starts when the air¯ow stays between +5% and ±5% of baseline for at least 10 s, and ends when the air¯ow is greater than 5% or less than ±5%. A hypopnoea event starts when the trace amplitude is less than 50% of the average amplitude over the previous 2 min and stays at or below this level for at least 10 s. The event ends as soon as the amplitude increases above 50% of the average amplitude.8 Frequencies of events were calculated per hour of recording. Studies with less than 4 h of recorded data were regarded as technically unsatisfactory and rejected. The respiratory distress index (RDI) is the sum of apnoea and hypopnoea events per hour. An RDI greater than 30 is said to be diagnostic of sleep apnoea/hypopnoea syndrome requiring treatment.8 Hypoxaemia was classi®ed by previously de®ned criteria9 as mild when the SpO2 was <94% >12 min h±1, moderate when the SpO2 <90% >12 min h±1 and severe when the SpO2 <85% >6 min h±1. The mean saturations and the percentage time spent below a saturation of 95% during the study period for each patient were also recorded. Measurement of oxygen saturation, sedation scoring (0=none, patient alert; 1=mild, occasionally drowsy, easy to arouse; 2=moderate, frequently drowsy, easy to arouse; 3=severe; somnolent, dif®cult to arouse), respiratory rate and visual analogue score for pain were recorded on the ward at hourly intervals for the ®rst 4 h and then every 4 h thereafter. The incidence of nausea or pruritus requiring treatment was noted. Previously collected audit data on patients who had received a spinal anaesthetic with bupivacaine alone and a morphine PCA following knee arthroplasty had shown that the mean morphine consumption was 40 (SD 16) mg. Assuming a clinically signi®cant reduction in morphine consumption of 20 mg, it was calculated10 that 30 patients (15 pairs) would be required to have a 90% power of detecting a 50% reduction in 24 h morphine consumption at a signi®cance level of 0.05. Data were analysed by using the SPSS statistics package (version 8). Mann±Whitney and chi-squared tests were used for the non-parametric data, and an unpaired Student's t-test for the patient characteristics. A value of P<0.05 was regarded as statistically signi®cant.
Results Thirty-eight patients completed the study. Two patients were excluded as an incorrect dose of intrathecal drug was administered to one and premedication was given to the second. Due to software problems with the Edentec monitor and patient non-compliance, a further four patients had incomplete respiratory data for analysis, but their data on morphine consumption were included.
234
Effects of low-dose spinal morphine
Fig 1 Box plot of visual analogue score on movement (VASm) scores at 4, 12 and 24 h postsurgery. Boxes represent interquartile range, horizontal lines in the boxes represent median values, and error bars the range. The VASm scores were signi®cantly reduced at 4 (P<0.01), 12 (P=0.01) and 24 (P<0.05) h in the intrathecal morphine group when compared with the intrathecal saline group. Open inverted triangle represents an outlying value.
Patient characteristics were similar between the two groups. Mean (SD) age, body weight and height were 70.3 (9.9) yr, 82.3 (15.7) kg and 1.7 (0.1) m, and 71.3 (7.6) yr, 80.5 (14.3) kg and 1.7 (0.1) m in the intrathecal morphine and intrathecal saline groups, respectively. The preoperative baseline respiratory monitoring showed median saturations (interquartile range) of 98.5 (96±99)% in the morphine group and 98.5 (97±99)% in the control group. Patients in the intrathecal morphine group had signi®cantly lower VASm pain scores on movement at 4, 12 and 24 h (Fig. 1), and used a signi®cantly lower dose of patientcontrolled i.v. morphine in the ®rst 24 h after surgery (Fig. 2). Although more patients in the intrathecal morphine group scored the severity of their pain as mild or absent and more patients in this group required treatment for emesis, these differences were not statistically signi®cant. There was no signi®cant difference between the two groups in the incidence of pruritus or their verbal rating of satisfaction (Table 1). When comparing the apnoea and hypopnoea episodes, there was no signi®cant difference between the two groups in the median RDI or in the incidence of those with a respiratory distress index >30 (Table 2). Patients in both groups had disordered respiratory patterns with episodes of central and obstructive apnoea. The median values of the mean saturations in the postoperative period between the two groups were compared and were signi®cantly lower in the intrathecal morphine group compared with the intrathecal saline group.
Fig 2 Box plot of i.v. PCA morphine consumption in the ®rst 24 h postsurgery in the intrathecal morphine and intrathecal saline (placebo) groups. Boxes represent interquartile range, horizontal lines in the boxes represent median values, and error bars the range. There is a signi®cant difference between the intrathecal morphine and intrathecal placebo groups (P<0.01). Closed diamond and closed square represent outlying values.
Using previously de®ned criteria,9 5/17 patients in the intrathecal morphine group could be classi®ed as mildly hypoxaemic and 1/17 as moderately hypoxaemic during the study period. In the saline group, 3/15 subjects showed mild hypoxaemia. However, this difference was not statistically signi®cant. No patients had severe hypoxaemia using these criteria. Although patients in the intrathecal morphine group spent a median of 10% of their time with a saturation below 95% compared with a median of 3% in the placebo group, this was not statistically signi®cant.
Discussion This study has achieved the primary aim of demonstrating that, following knee arthroplasty, intrathecal morphine 0.3 mg followed by patient-controlled i.v. morphine, improves postoperative analgesia and reduces i.v. morphine consumption compared with patient-controlled i.v. morphine alone. This con®rms earlier studies on the quality of analgesia after Caesarean section11±14 and hip surgery.15 Whilst this combination of intrathecal and patientcontrolled i.v. morphine provides effective analgesia, there are concerns in clinical practice about the safety of intrathecal opioids particularly if systemic opioids are used for rescue analgesia. A 17-nation European survey revealed that, except for the UK, morphine was the most commonly used of the long-acting opioids.16 This initial study used morphine because it has been extensively used, and whatever degree of respiratory depression was seen with intrathecal morphine, respiratory depression would be less
235
Cole et al. Table 1 Highest sedation score, minor side effects and patient satisfaction/severity during the ®rst 24 h. There was no difference in the median highest sedation scores in the ®rst 24 h between the two groups. Median values and interquartile ranges are shown. The number of patients requiring antiemetic treatment and reporting pruritis was higher in the intrathecal morphine group, but this was not statistically signi®cantly. Patients reported their analgesia as good or excellent, and their pain severity as absent or mild more frequently in the intrathecal morphine group compared with the saline group
Morphine Placebo P-value
Sedation score
Emesis (treated) (n=31)
Pruritus (n=29)
Satisfaction (good/excellent) (n=32)
Severity (absent/mild) (n=31)
1 [0±2] 1 [0±1.5] NA
Nine out of 15 (60%) Six out of 16 (38%) 0.07
Four out of 15 (27%) Two out of 14 (14%) 0.65
Nine out of 15 (60%) Nine out of 17 (53%) 0.96
12 out of 14 (86%) Nine out of 17 (53%) 0.07
Table 2 Postoperative respiratory data. Median (interquartile range) respiratory distress indices (RDI) and number of patients who had an RDI count >30 are shown, there was no signi®cant difference within these groups. The median values of the mean saturations (interquartile range) (%) show a statistical difference between the two groups*. There were more episodes of hypoxaemia in the intrathecal morphine group compared with the saline group but this did not reach statistical signi®cance. *P<0.05
RDI RDI count >30 Median SpO2 (%) Incidence of mild/moderate hypoxaemia
Morphine group (n=17)
Placebo group (n=15)
P-value
19 (4±35.5) Six out of 17 97 (95±99) Six out of 17
13 (3±34) Five out of 15 99 (97±99) Three out of 15
0.53 0.91 0.03* 0.36
likely to occur when low doses of more lipid-soluble opioids are administered by the intrathecal route.7 17 Due to the dose-dependent nature of respiratory depression,5 a dose of morphine 0.3 mg was used, based on an earlier dose±response study that showed that this is an effective dose with the fewest side effects.4 However, more recent studies have shown that lower doses (0.1 mg) are effective for analgesia after Caesarean section12 and hip replacement surgery.14 There have been many reports of respiratory depression following spinal opioids.7 Previous studies looking at the respiratory effects of intrathecal opioids have used a combination of pulse oximetry, sedation scoring, respiratory rate and invasive arterial blood±gas analysis.5 7 18 The availability of relatively simple to use, bedside respiratory monitors allows us to continuously study the ventilatory pattern and oxygenation of patients receiving intrathecal opioids. Although it is often assumed that bedside monitors such as the Edentec are less accurate than conventional sleep studies with polysomnography, this equipment has recently been used to validate the detection of respiratory disturbances in patients with sleep apnoea/ hypopnoea syndrome8 and to investigate patients recovering from major abdominal surgery.19 Our study demonstrated that respiratory disturbances occurred in both groups. The pattern of respiratory dysfunction was similar to that previously described following the use of i.v. morphine,20 21 and included obstructive and central apnoeas. Although patients in the morphine group spent three times as long with a saturation below 95% and twice as many were classi®ed as `hypoxaemic patients'; that is they spent greater than 20% of their time with mild or moderate
hypoxaemia (85±94%),9 these differences were not statistically signi®cant. The only statistical difference was between the median oxygen saturations in the two groups. In view of the risks of postoperative hypoxaemia22 and its potential effects on postoperative morbidity, it was deemed necessary to give patients in this study supplementary oxygen if their oxygen saturations fell below 94%. Because oxygen was not administered to all the study patients, the small but statistically signi®cant difference between the mean saturations (97% and 99%) may not be of clinical signi®cance. None of the patients in the study met the pre-de®ned criteria9 for severe hypoxaemia, that is, spending more than 10% of their time with a saturation below 85%. The only patient in the study who approached this level was a patient in the spinal morphine group who spent 90% of their time with saturation below 95% including 4% of their time below 85%. This would be regarded by most clinicians as a signi®cant level of hypoxaemia. Because of the low incidence of respiratory depression, a much larger study would be needed to detect or exclude a signi®cantly increased risk with low-dose spinal morphine. Further studies are required into the ef®cacy and duration of action of more lipid-soluble opioids which are likely to be associated with a lower incidence of respiratory depression.
Acknowledgements We thank Mrs E. McManus, senior pharmacist at York District Hospital, acute pain nurses D. Hunter and R. Clayton, and Miss V. Algar, statistician at St James's University Teaching Hospital, Leeds for their help in this study.
236
Effects of low-dose spinal morphine
References 1 Rawal N. Epidural and spinal agents for postoperative analgesia. Surg Clinics North Am 1999; 79: 313±44 2 Carr DB, Cousins MJ. Spinal route of analgesia: opioids and future options. In: Cousins MJ, Bridenbaugh PO, eds, Neural Blockade in Clinical Anaesthesia and Management of Pain, 3rd edn. Philadelphia: Lippincott-Raven, 1998 3 Wang JK, Nauss LA, Thomas JE. Pain relief by intrathecally applied morphine in man. Anaesthesiology 1979; 50: 149±51 4 Sarma VJ, Bostrom UV. Intrathecal morphine for the relief of post hysterectomy pain ± a double blind dose response study. Acta Anaesthesiol Scand 1993; 37: 223±7 5 Bailey PL, Rhondeau S, Schafer PG et al. Dose±response pharmacology of intrathecal morphine in human volunteers. Anaesthesiology 1993; 79: 49±59 6 Jacobson L, Chabal C, Brody MC. A dose±response study of intrathecal morphine: ef®cacy, duration, optimum dose, and side effects. Anaesth Analg 1988; 67: 1082±8 7 Etches RC, Sandler AN, Daley MD. Respiratory depression and spinal opioids. Can J Anaesth 1989; 36: 165±85 8 Whittle AT, Finch SP, Mortimer IL, MacKay TW, Douglas NJ. Use of home sleep studies for diagnosis of sleep apnoea/ hypopnea syndrome. Thorax 1997; 52: 1068±73 9 Wheatley RG, Shepherd D, Jackson IJB, Madej TH, Hunter D. Hypoxaemia and pain relief after upper abdominal surgery: comparison of I.M. and patient controlled analgesia. Br J Anaesth 1992; 69: 558±61 10 Altman DG. How large a sample? Statistics and ethics in medical research. BMJ 1980; 281: 1336±8 11 Swart M, Sewell J, Thomas D. Intrathecal morphine for Caesarean section: an assessment of pain relief, satisfaction and side effects. Anaesthesia 1997; 52: 373±7 12 Palmer CM, Emerson S, Volgoropolous D, Alves D. Dose± response relationship of intrathecal morphine for postcesarean analgesia. Anesthesiology 1999; 90: 437±44
237
13 Husaini SW, Russell IF. Intrathecal diamorphine compared with morphine for postoperative analgesia after Caesarean section under spinal anaesthesia. Br J Anaesth 1998; 81: 135±9 14 Gerancher JC, Floyd H, Eisanach J. Determination of an effective dose of intrathecal morphine for pain relief after cesarean delivery. Anesth Analg 1999; 88: 346±51 15 Slappendel R, Weber EW, Dirksen R, Gielen MJ, Van Limbeek J. Optimisation of the dose of intrathecal morphine in total hip surgery: a dose ®nding study. Anesth Analg 1999; 88: 822±6 16 Rawal N, Allvin R. Euro-pain study on acute pain. Epidural and intrathecal opioids for postoperative pain management in Europe ± a 17 nation questionnaire study of selected hospitals. Acta Anaesthesiol Scand 1996; 40: 1119±26 17 Hamber EA, Viscomi CM. Intrathecal lipophilic opioids as adjuncts to surgical spinal analgesia. Reg Anaesth Pain Medicine 1999; 24: 255±63 18 Kelly MC, Carabine UA, Mirakhur RK. Intrathecal diamorphine for analgesia after Caesarean section. A dose ®nding study and assessment of side effects. Anaesthesia 1998; 53: 231±7 19 Rosenberg J, Rasmussen GI, Wojdemann KR, Kirkeby LT, Jorgensen LN, Kehlet H. Ventilatory pattern and associated episodic hypoxaemia in the late postoperative period in the general surgical ward. Anaesthesia 1999; 54: 323±28 20 Jones JG, Jordan C, Scudder C, Rocke DA, Barrowcliffe M. Episodic postoperative oxygen saturation: the value of added oxygen. J Roy Soc Med 1985; 78: 1019±22 21 Catley DM, Thornton C, Jordan C, Lehane JR, Royston D, Jones JG. Pronounced, episodic oxygen desaturation in the postoperative period; its association with ventilatory pattern and analgesic regime. Anaesthesiology 1985; 63: 20±8 22 Reeder MK, Muir AD, Foex P, Goldman MD, Loh L, Smart D. Postoperative myocardial ischaemia: temporal association with nocturnal hypoxaemia. Br J Anaesthesia 1991; 67: 626±31