A randomised comparison of regular oral oxycodone and intrathecal morphine for post-caesarean analgesia

Introduction 

Opioids, administered by the intravenous or neuraxial route, form the foundation of most multimodal analgesic regimens after caesarean delivery. The side effects of opioid use vary with the dose administered and the route of administration. Intravenous opioids are associated with a high incidence of nausea, vomiting and sedation whereas neuraxial opioids are more likely to cause pruritus and urinary retention.1 The subarachnoid administration of morphine is popular and provides prolonged post-caesarean analgesia, especially when combined with non-steroidal anti-inflammatory drugs (NSAIDs).2

Oral opioids are commonly employed as ‘step down’ analgesics but there is insufficient evidence supporting oral regimens as the primary post-caesarean analgesic method on the first postoperative day.3, 4, 5, 6 The potential benefits of such an approach include ease of administration, high maternal acceptability, fewer opioid-related side effects than neuraxial opioids, and the avoidance of complications associated with epidural analgesia. Published audit data show promising results, especially in regards to patient satisfaction.6, 7 A randomised controlled trial found superior pain relief and fewer opioid-related side effects associated with an oral oxycodone regimen compared with morphine patient controlled intravenous analgesia.3

The aim of this study was to compare the quality of postoperative pain relief provided by a multimodal analgesic regimen based on intrathecal morphine with that provided by one based on regular oral administration of oxycodone, among women undergoing caesarean section under spinal anaesthesia. Secondary objectives included an evaluation of the side effects and acceptability of a multimodal analgesic regimen based on regular oral oxycodone. The null hypothesis was that there would be no difference between the intrathecal morphine and oral oxycodone regimens with respect to early postoperative analgesia, as assessed by the area under the curve (AUC) for dynamic pain scores in the first 24h.

Methods 

After obtaining Institutional Ethics Committee approval and written informed consent, 120 women scheduled for elective primary or repeat caesarean section were recruited to this randomised, double-blind, double-dummy, placebo-controlled trial. Women of American Society of Anesthesiologists (ASA) physical status 1 or 2 who were aged 18 years and over and who consented to combined spinal-epidural anaesthesia were eligible for recruitment. Exclusion criteria were concurrent opioid therapy, contraindications to combined spinal-epidural anaesthesia, a contraindication to any of the study medications, a history of preoperative pruritus or nausea, failure to identify the subarachnoid space at the time of anaesthesia, inadvertent dural puncture with the epidural needle and conversion of spinal to general anaesthesia.

Preoperatively, subjects were randomised into one of two groups using a random number generator prepared by the hospital pharmacy department. Patients in group O were assigned to postoperative analgesia using regular, fixed-dose oral oxycodone. They received a 0.2-mL placebo intrathecal injection of sterile saline at the time of spinal anaesthesia and 20mg of sustained-release oral oxycodone immediately on arrival in the recovery room. This was followed by immediate-release oral oxycodone 10mg every 6h postoperatively, in conjunction with oral paracetamol 1g. Patients assigned to group I received preservative-free intrathecal morphine 100μg (0.2mL) at the time of spinal anaesthesia and matching oral placebo medication, instead of oxycodone, at the same time intervals as group O (Table 1). The hospital pharmacy provided individual sealed packs containing 1mL of the intrathecal study solution (either intrathecal morphine 500μg or sterile saline) and either active or dummy oral oxycodone tablets of identical appearance. All investigators, patients and staff involved in the care of the patients (anaesthetists, midwives, nursing staff and obstetricians) were blinded to group allocation.

Table 1. Intra- and postoperative analgesia

		Treatment regimen
	Subarachnoid medication	Hyperbaric bupivacaine 11-12.5mg
		Fentanyl 15μg
		Group I: morphine 100μg; Group O: saline placebo
	Intraoperative	Parecoxib 40mg i.v.
	Recovery room	Paracetamol 1g orally
		Group I: placebo orally; Group O oxycodone SR 20mg orally
	Postoperative	Paracetamol 1g orally
	6, 12, 18 and 24h post recovery room medication	Group I: placebo orally; Group O: oxycodone IR 10mg orally
	8-hourly starting at least 12h postoperatively	Diclofenac 50mg orally
	Rescue analgesia	Oxycodone IR 10-15mg orally 2-hourly as required
		Tramadol 100mg orally 4-hourly as required

SR = sustained release; IR = immediate release. Group I = intrathecal morphine; Group O = oral oxycodone.

All patients received aspiration prophylaxis with oral ranitidine 300mg on the morning of surgery, were fasted from solid food from midnight and allowed clear oral fluids up until 3h before surgery. Patients started free oral fluids after discharge from the recovery area and thereafter were permitted solid food at the discretion of the attending midwife.

In both groups spinal anaesthesia was achieved using hyperbaric bupivacaine 11-12.5mg (according to the preference of the attending anaesthetist) and fentanyl 15μg, plus 0.2mL of the intrathecal study solution. The woman was in the seated position and a combined spinal-epidural technique (16-gauge Tuohy needle and 27-gauge pencil-point spinal needle) was performed at a low lumbar interspace. Intravenous dexamethasone 4mg was given as antiemetic prophylaxis immediately after delivery of the newborn. The epidural catheter was used only if surgery was delayed or intraoperative pain experienced, in which case epidural lidocaine with adrenaline 1: 200000 was given, and catheter was removed at the conclusion of surgery.

As part of the multimodal analgesic regimen, all patients received i.v. parecoxib 40mg at the conclusion of surgery and 8-hourly oral diclofenac 50mg starting 12h postoperatively. Supplemental postoperative analgesia was available as immediate release oral oxycodone, 10-15mg 2-hourly on demand. If this proved insufficient (0-10 dynamic pain numerical rating score [NRS] > 6) then oral tramadol 100mg was administered. Postoperative pruritus was treated with i.v. ondansetron 4mg 6-hourly on demand or if this was ineffective with hourly i.v. naloxone 50μg. Postoperative nausea and/or vomiting were treated with i.v. ondansetron or metoclopramide.

Maternal age, weight, parity and gestation, the number of previous caesarean sections and a preoperative quality of recovery (QoR) score were recorded. The QoR is a reliable, validated instrument for assessing quality of recovery after anaesthesia and surgery, being based on patient responses to questions about their well-being and functional status.8 The duration of surgery and need for intraoperative analgesia were noted. NRS pain scores (rest pain and dynamic or movement pain) and sedation scores were assessed at 6, 12, 18 and 24h postoperatively. At 24h postoperatively a modified brief pain inventory –short form (MBPI-SF) questionnaire was completed. This instrument is a stable and valid tool in the postoperative period that was designed to assess how much pain interferes with an increasing number of functions over time.9 Supplementary analgesic requirement and dosage, a maternal NRS for satisfaction, an opioid symptom distress score and global nausea, sedation and pruritus scores, were recorded at 24h. At 48h the presence of urinary retention or a respiratory complication during the preceding 48-hour period was assessed and maternal satisfaction score repeated. All women were monitored for respiratory depression by means of hourly assessments of respiratory rate and sedation during the first 12h postoperatively and then 4-hourly until 48h. Respiratory depression was defined as a respiratory rate of less than 8 or a sedation score of 3 or more (0: none, 1: drowsy. 2: asleep, rousing to command, 3: asleep, rousing on physical contact, 4: difficult to rouse).

Results 

One hundred and twenty patients were recruited to the study and nine withdrew after recruitment and were excluded from analyses (Fig. 1). Of the nine patients excluded from the analyses, three withdrew consent for the study shortly after receiving the intrathecal study drug (group O n=2, group I n=1) and one patient required conversion to general anaesthesia (group O). Four patients (group I n=3, group O n=1) withdrew their consent because of uncontrolled pain in the recovery room and one patient (Group O) withdrew within the first six hours secondary to pain. One hundred and eleven patients with complete data (group O n=55, group I n=56) were included in the analysis. The patients excluded did not differ significantly from the analysed study patients with respect to age, weight, height, QoR score, parity, ASA status or group randomisation. However, the pre-operative symptom distress score of all subjects who withdrew from the study was significantly higher than for those who remained in the study (median score 6 [IQR 4-10] vs. 3 (IQR 1-5]; P =0.019).

View Large Image Download to PowerPoint

Fig. 1. Participant flow through each stage of the trial.

The characteristics of the two patient groups were similar (Table 2). Thirteen patients in group I (23%) required intraoperative supplementation versus nine in group O (16%) (P =0.37). The 0 to 24h AUC pain scores at rest and with movement did not differ significantly between groups (Table 3, Fig 2). The AUC pain scores also did not differ significantly with simultaneous adjustment for the use of postoperative oxycodone (P values 0.47 and 0.53 respectively). The median time to first request for supplemental analgesia did not differ between groups (160 [92-230] min vs. 144 [62-218] min for groups I and O respectively). Pain scores at rest were low in both groups at all time points (Table 3), although at 12h median pain score at rest in group O was higher than that in group I (2 [1-3] vs. 1 [0-2]; P =0.03). There were no differences in pain scores with movement, although additional analgesia was more likely to be requested by group O (46/55 [82%] versus 35/56 [63%] in group I; P =0.034). Total scores for the MBPI did not differ between the groups. More patients in group O than in group I rated their worst pain experienced during the first 24h postoperatively as 4–10 (87% vs. 64%; P=0.007) (Table 4).

Table 2. Patient characteristics

		Group I (n=56)	Group O (n=56)	P
	Age (years)	32 [28-37; 20-44]	31 [26-36; 20-42]	0.311
	Weight (kg)	86 [73-107; 58-153]	87 [76-107; 51-157]	0.864
	Height (cm)	165 [162-170; 152-183]	165 [158-169; 152-183]	0.666
	Parity	2 [1-2; 0-4]	1 [1-2; 0-4]	0.345
	ASA 2	21 (38%)	24 (43%)	0.561
	Previous caesarean section	1 [1-2; 0-4]	1 [0-1; 0-3]	0.194
	Pre-op QoR score	17 [15-18; 10-18]	16 [15-17; 9-18]	0.621
	(range 0-18)
	Pre-op SDQ score	3 (2-5; 0-16]	3 (1-5; 0-14]	0.684
	(range 0-36)

N (%) or median [IQR; Range] are shown as appropriate. QoR= Quality of Recovery; SDQ= Symptom distress questionnaire.

Table 3. Analgesia

		N	Group I	N	Group O	P
	Pain scores (0-10)
	At rest
	6h	53	2 [0-3; 0-7]	53	2 [1-3; 0-10]	0.178
	12h	55	1 [0-2; 0-7]	53	2 [1-3; 0-7]	0.030
	18h	56	1 [0-2; 0-6]	54	1 [0-2; 0-5]	0.589
	24h	56	1 [0-2; 0-8]	54	1 [0-2; 0-8]	0.688
	With movement
	6h	53	4 [3-6; 0-10]	51	5 [3-7;1-10]	0.062
	12h	55	3 [2-5; 0-8]	53	4 [3-6; 0-8]	0.100
	18h	56	3 [2-6; 0-8]	53	3 [2-5; 0-8]	0.852
	24h	56	3 [2-5; 0-9]	54	3 [2-5; 0-9]	0.792
	AUC to 24h
	rest	56	21 [6-35; 0-117]	54	24 [12-39; 0-84]	0.306
	movement	56	63 [37-96; 3-132]	54	71 [42-99; 18-135]	0.319
	Time to first on-demand oxycodone request (min)	35	160 [92-230; 23-1419]	46	144 [64-218; 20-1440]	0.597
	Additional analgesia requested	56	35 (63%)	55	46 (82%)	0.034
	Additional oxycodone (mg)	32	23 [10-35; 5-70]	44	25 [10-44; 5-85]	0.953
	Additional tramadol (mg)	19	100 [100-100; 50-300]	29	100 [100-200; 50-500]	0.299

Values are median, [IQR; R] or n (%). AUC = area under the curve.

View Large Image Download to PowerPoint

Fig. 2. Area under the curve for scores of rest pain, movement pain and sedation in the first 24h. No significant differences between groups. Group O: Oral oxycodone; group I: intrathecal morphine.

Table 4. Modified brief pain inventory-short form (MBPI – SF) scores

		MPBI-SF	Group I (n=56)	Group O (n=55)	P
	Any pain last 24h		33 (59%)	39 (70%)	0.234
	Worst pain 24h	0-3	20 (36%)	7 (13%)	0.007
		4-10	36 (64%)	48 (87%)
	Least pain 24h	0-3	52 (93%)	53 (96%)	0.679
		4-10	4 (7%)	2 (4%)
	Average pain 24h	0-3	38 (68%)	40 (73%)	0.679
		4-10	18 (32%)	15 (27%)
	Pain now	0-3	48 (86%)	50 (91%)	0.557
		4-10	8 (14%)	5 (9%)
	Relief after medication (%)	0–30%	0	3 (6%)	0.112
		40–100%	56 (100%)	50 (94%)

Values are n (%).

There was a significantly higher incidence of pruritus in group I (P =0.001) and this was also more severe (Table 5). There were no significant differences between groups for the incidence of nausea, epigastric pain, urinary retention or sedation scores. No respiratory depression (based on clinical observation of conscious state and ventilation) was observed in either group. At the 24-h review patients in group I had higher satisfaction scores but this advantage was no longer present by 48h.

Table 5. Recovery outcomes and treatment side effects

		Group I (n=56)	Group O (n=55)	P
	24h post-op
	Nausea	26 (46%)	19 (34%)	0.247
	Epigastric pain	6 (11%)	3 (5%)	0.490
	Satisfaction with analgesia (0-10)	10 [8-10; 3-10]	8 [7-10; 2-10]	0.010
	Postoperative pruritus (any)	48 (87%)	31 (56%)	0.001
	Global severity score (0-10)
	Nausea	1 [0-3; 0-8]	0 [0-4; 0-9]	0.469
	Sedation	2 [0-5; 0-9]	3 [0-6; 0-10]	0.141
	Pruritus	4 [2-5; 0-9]	1 [0-4; 0-9]	0.001
	SDQ score (0-36)	7 [3-10; 0-19]	6 [3-12; 0-18]	0.806
	48h post-op:
	Satisfaction with analgesia (0-10)	9 [8-10; 2-10]	9 [8-9; 0-10]	0.887
	Urinary retention	0	1 (2%)	1.000
	Respiratory depression	0	0

N (%) or median [IQR; Range] are shown as appropriate. Respiratory depression defined as a respiratory rate < 8 breaths/min or a sedation score of 3 (0-4). SDQ=Symptom distress questionnaire.

Discussion 

In this randomised controlled trial we confirmed that a multimodal regimen consisting of regular oral non-opioids and oxycodone provided satisfactory analgesia after caesarean delivery, resulting in early postoperative analgesia of similar quality to that from a regimen based on intrathecal morphine. Based on a post hoc power analysis, the study had 95% power to detect a 30% difference in 24-h AUC pain scores between groups and 86% power to show a 25% difference between groups. However, the intrathecal morphine regimen was associated with higher maternal satisfaction, despite a higher incidence of pruritus. We speculate that this may have been a consequence of factors such as less severe episodic pain and a lesser requirement for rescue analgesics.

Multimodal analgesic techniques are commonly recommended for post-caesarean analgesia but the oral route of opioid administration has received little attention other than as a means of transition from intravenous, subarachnoid or epidural opioid analgesia.10, 11 Cohort studies suggest that oral opioid regimens are convenient for nursing staff, of low cost and result in less severe opioid-related side effects,4, 5, 6, 7 but to our knowledge only one randomised controlled trial has been performed.3 Despite concerns about decreased bowel motility and increased postoperative nausea and vomiting,4 early oral intake after caesarean section is acceptable and is associated with shorter lengths of hospital stay and earlier recovery of bowel function.12 Oral opioid administration avoids issues associated with rectal administration, including unreliable drug absorption and low patient acceptability, and with intravenous administration, such as the risk of device programming errors, device malfunction and rapid fluctuation in plasma concentrations.13, 14 We chose oral oxycodone rather than oral morphine, because it is the most popular oral opioid in Australia in this setting.15 The cost of these two drug formulations in this country is similar, while oxycodone has more rapid absorption and reliable kinetics and an established role in postoperative pain management,16 and is associated with minimal risk to the neonate.15 Limited experience with patient-controlled oral morphine after caesarean delivery has been encouraging, local audit data from the United Kingdom suggesting good patient acceptability.5, 6, 7 However, this approach poses a number of administrative obstacles before effective implementation, including the provision of secure storage areas in patient rooms.5 In addition, oral morphine does not appear to have been adequately investigated or compared with alternatives, and its effect on the neonate has not been evaluated. The breast milk transfer of oxycodone is low and considered compatible with breast feeding,15 especially during the first day or two after delivery, when milk production is low. Nevertheless, maternal plasma concentrations of opioid are likely to be significantly higher than with neuraxial opioid administration, which may need consideration in individual cases, especially when doses of more than 90mg per 24h are used for several days.15 We administered parecoxib on the basis of its prolonged analgesic effect (12-24h), proven benefit when combined with intrathecal morphine17 and reduced risk of exacerbating post-partum blood loss or hypertension in the parturient compared with traditional NSAIDs. Intrathecal morphine 100μg causes postoperative nausea and vomiting in approximately a third of women, justifying antiemetic prophylaxis. Dexamethasone is effective against both nausea and vomiting,18 although arguably not as a sole agent in the first 24h after caesarean section.19

Primary analgesic methods based on oral opioid and non-opioids given on request have been evaluated in open non-randomised studies. Jakobi et al. suggested pain relief was satisfactory,4 and Monagle et al. reported a low incidence of side effects and good satisfaction.20 Jakobi et al. compared an on-demand approach with predetermined regular dosing and noted lower pain scores and enhanced maternal satisfaction with the fixed dosing method.21 Recently, Davis et al. compared intravenous morphine patient-controlled analgesia (PCA) with a fixed dose combination of oxycodone/acetaminophen (5mg/325mg tablets, two tablets 3-4-hourly, maximum 12 tablets in the first 24h) in an open-label randomised trial.3 The regimen included intravenous ketorolac 30mg and unrelieved pain was treated with intramuscular pethidine. Significantly lower pain scores were found at 6 and 24h in the oral analgesia group (VAS 3.2 [1.8] vs. 4.1 [2.5], P=0.04 and 2.9 [1.7] vs. 4.1 [2.1], P=0.004 respectively) and nausea and sedation were less common. However, the differences in scores do not appear clinically important and it is unclear whether dynamic pain was assessed. Furthermore, greater drowsiness and nausea experienced by the morphine PCA group may have been a consequence of using a continuous infusion in addition to demand boluses.

In our study both regimens produced satisfactory pain management, confirming the analgesic efficacy of the regular oral oxycodone-based multimodal regimen. However, at 24h maternal satisfaction was higher after intrathecal morphine, despite more pruritus with this method. Some secondary study endpoints indicate that analgesia was more consistent in the intrathecal morphine group, with women being more likely to report their worst pain as being of lower intensity. This might suggest that post-caesarean patients consider consistently good pain relief more important than side effects such as pruritus, although other explanations are possible.

The 100-μg dose of intrathecal morphine is supported by good levels of evidence, higher doses resulting in little improvement in the quality of analgesia and more side effects.22 Lower doses combined with NSAIDs may show efficacy,2 but even after a 100-μg dose we found that 63% of women receiving this combination requested additional opioid analgesia within 24h. This finding is consistent with previous studies2, 21 and audits23 and highlights the need to provide patients receiving neuraxial opioids with ready access to supplemental analgesia. We based the fixed oral dosing regimen on the work of Davis et al.,3 but oxycodone is available in immediate-release and sustained-release formulations, so alternative regimens may have increased efficacy. Further research is required to determine optimal doses, dose intervals and combinations of formulations.

Opioid-related side effects are partly determined by the dose and route of administration. Pruritus results from activation of μ-opioid and 5-hydroxytryptamine3 receptors and non-nociceptive neurones in the medulla and dorsal horn of the spinal cord, particularly in trigeminal-nerve distribution. Pruritus is the most frequent side effect of intrathecal morphine and the incidence in this study (87%) was consistent with previous research.24 As expected, pruritus was less common and severe in the oral oxycodone group, possibly because of lower opioid concentrations in relevant central nervous areas following systemic administration.

This study has a number of limitations. As noted, it is possible that the fixed dose regimen of oxycodone chosen empirically was not optimal and that different doses or dose intervals or combinations of immediate and sustained release oxycodone might prove more effective or reduce side effects. On the other hand, opioid-induced side effects may be increased if total opioid use were to be increased, so other regimens would need to be investigated. It is also possible that a lower dose of intrathecal morphine might have been associated with reduced side effects, although the high rate of supplementation might mitigate such a benefit. A future study should explore the reasons for greater maternal dissatisfaction with an oral opioid regimen, because this information could provide insights into the balance between opioid efficacy and side effects from the patient’s perspective, as well as examine the economic implications of such a regimen. Finally, we did not evaluate the neonate or the establishment of breastfeeding in this study.

Conclusion 

In conclusion, this study shows that a multimodal analgesic regimen based on oral oxycodone confers generally comparable and satisfactory clinical analgesic efficacy and a favourable side-effect profile compared with a multimodal regimen based on intrathecal morphine. An intrathecal morphine-based multimodal regimen appeared to have some small efficacy advantages and was associated with reduced need for rescue analgesia and higher maternal satisfaction scores. We consider that a multimodal oral analgesic regimen including oxycodone, paracetamol, parecoxib and diclofenac is an acceptable approach to post-caesarean analgesia. Further refinements to this technique may improve efficacy and maternal satisfaction. Cost-effectiveness evaluation appears warranted.