The analgesic efficacy and pharmacokinetics of epidural oxycodone after gynaecological laparotomy: a randomized, double‐blind, double‐dummy comparison with intravenous administration

Abstract

Aim

The aim of the present study was to compare the analgesic efficacy of epidural and intravenous (i.v.) oxycodone at the same dose.

Methods

In this randomized, double‐blind, double‐dummy clinical trial, 30 women, aged 24–67 years, undergoing elective gynaecological laparotomy, were administrated either i.v. saline and epidural oxycodone 0.1 mg·kg⁻¹ (EPI group; n = 15) or i.v. oxycodone 0.1 mg·kg⁻¹ and epidural saline (IV group; n = 15). For multimodal analgesia, patients received i.v. paracetamol and dexketoprofen, and a triple‐mixture epidural infusion after the first 4 h postoperatively. The primary outcome was the total dose of i.v. fentanyl for rescue analgesia during the first 4 h postoperatively.

Results

All patients required fentanyl during the first 4 h. The median number of fentanyl doses were three (quartiles 1, 8) in the EPI group and seven (6, 9) in the IV group (mean difference 3.1; 95% confidence interval 0.9, 5.2; P = 0.01). After the first 4 h, the two groups needed a similar total dose of epidural infusion. Patient satisfaction was similarly high in both groups, and both administration routes were well tolerated.

Conclusions

The data support the superiority of epidural oxycodone compared with that of i.v. administration in pain management after laparotomy.

What is Already Known about this Subject

• Oxycodone is an increasingly used semisynthetic opioid agonist in acute and chronic pain management orally and intravenously.

• Epidural oxycodone provides enhanced central nervous system pharmacokinetics than the same dose given intravenously.

• There are sparse data on the efficacy of epidural oxycodone in postoperative pain.

What this Study Adds

• Epidural oxycodone seems to provide better postoperative analgesia than the same dose given intravenously after gynaecological laparotomy.

• Further studies are needed to assess the optimal dose and dose interval for repeated dosing and for continuous infusion of epidural oxycodone.

Introduction

http://guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=7093 is a semisynthetic opioid agonist which is used increasingly in acute and chronic pain management 1. In clinical practice, oxycodone is usually administered by mouth, intravenously, subcutaneously or transmucosally 1, 2.

Epidural administration of oxycodone may be a feasible route as the main site of analgesic action of opioids is in the central nervous system (CNS) 3. In a pharmacokinetic study we carried out previously, it appeared that epidural oxycodone may provide enhanced CNS pharmacokinetics than the same dose given intravenously. The peak cerebrospinal fluid (CSF) concentration and area under the CSF concentration‐time curve concentration–time curve for epidural oxycodone were 320‐ and 120‐fold higher, respectively, than with intravenous (i.v.) administration 4.

Oxycodone is a highly potent opioid for visceral pain and is therefore appropriate for pain management in laparotomy 5. In addition, enhanced CNS pharmacokinetics suggest a high analgesic efficacy of epidural administration because both the binding affinity of oxycodone to http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=50 and receptor activation are lower than those of morphine 6, 7.

The aims of the present study were, firstly, to compare the analgesic efficacy of epidural and i.v. oxycodone at the same dose and, secondly, to evaluate the pharmacokinetics of epidural and i.v. oxycodone in postoperative pain management after gynaecological laparotomy. Our study hypothesis was that epidural administration should be superior compared with i.v. administration.

Methods

The study protocol was approved by the Research Ethics Committee of the Hospital District of Northern Savo, Kuopio, Finland (ref: 83//2014), registered with EudraCT (ref: 2014–004313‐82) and conducted in accordance with the Declaration of Helsinki. The Finnish Medicines Agency was also notified (ref: 115/2014). The study was carried out between May 2015 and December 2016 at Kuopio University Hospital and had institutional approval. The study design was a prospective, randomized, double‐blind, double‐dummy clinical trial with two parallel groups.

We enrolled patients aged 18–75 years scheduled for elective gynaecological laparotomy with planned postoperative epidural analgesia. We did not enrol patients: (i) undergoing major oncological surgery; (ii) with an allergy/hypersensitivity to oxycodone or its excipients; (iii) with reduced respiratory function; (iv) with a bleeding disorder; (v) who had participated in a drug trial; or (vi) who had used oxycodone or monoamine oxidase inhibitors, or cytochrome P450 (CYP) http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=263‐ or http://guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1329 inhibitors during the previous month.

A total of 43 patients were invited and 33 agreed to participate (see Figure 1 for study flow chart). Reasons for declining included: patients did not wish to undergo any additional procedures (n = 3); fear of postpuncture headache (n = 2), fear of needle puncture (n = 1), a history of nausea with oxycodone (n = 1), too many confusing words in the patient information sheet (n = 1), a challenging personal situation (n = 1), no specific reason (n = 1).

Figure 1.

Study flow chart. i.v., intravenous

Three patients who agreed to participate had to be excluded before study drug administration because of a rash on the back that contraindicated epidural analgesia (n = 1), a major surgical complication and intensive care treatment (n = 1) and extensive oncological surgery (n = 1).

After giving written informed consent, 30 patients were randomized, using a randomization plan generator (http://www.randomization.com), into two parallel groups and were administered either i.v. saline and epidural oxycodone 0.1 mg·kg⁻¹ (Oxanest® 10 mg·ml⁻¹; Takeda, Helsinki, Finland) (EPI group) or i.v. oxycodone 0.1 mg·kg⁻¹ and epidural saline (IV group) in the post‐anaesthesia care unit (PACU) after baseline pain assessment. Concealment was achieved using the closed, opaque envelope method. The oxycodone and saline‐containing syringes were prepared by a study nurse who did not participate in the study or patient care.

Dose selection

The dose of oxycodone was based on data indicating that the minimum effective plasma concentration (MEC) of oxycodone is approximately 21 [95% confidence interval (CI) 13, 29] ng·ml⁻¹, and the minimum effective analgesic concentration (MEAC) 55 (19–91) ng·ml⁻¹ for postoperative pain 8. In our previous pharmacokinetic study 4, plasma oxycodone concentrations were in these ranges after epidural and i.v administration of 0.1 mg·kg⁻¹, with a median of 29 (minimum, maximum 14, 77) ng·ml⁻¹ after epidural and 58 (35, 135) ng·ml⁻¹ after i.v. administration. Based on these data, it was anticipated that at this dose, the MEC and MEAC would be achieved in both study groups.

Anaesthesia and pain management

A standardized general anaesthesia was used. For premedication, the patients were given 10 mg diazepam and 2 g paracetamol orally 60 min prior to surgery. An epidural catheter was placed at the Th10–Th12 interspace before anaesthesia induction and tested for i.v. or spinal misplacement with 5 ml of lidocaine (10 mg·ml⁻¹) with epinephrine (10 μg·ml⁻¹). Anaesthesia was induced with http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=5464; rocuronium 0.5 mg·kg⁻¹ was administered to facilitate orotracheal intubation; and a http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=7292 infusion was used for intraoperative analgesia. The infusion rate of remifentanil was adjusted to maintain the Surgical Pleth Index at 20–50. The propofol infusion with a sevoflurane end‐tidal concentration of 0.6–0.8% was adjusted to maintain the response entropy at 40–50 (Carescape™ B650; GE Healthcare, Helsinki, Finland). Muscle relaxation was monitored and the post‐tetanic count was kept at 2–5 with rocuronium. At the end of the anaesthesia, the anaesthetics were discontinued and muscle relaxation was reversed with sugammadex 1–2 mg·kg⁻¹ (train‐of‐four ratio 0.9 or higher). Remifentanil was continued at 0.1 mg·h⁻¹ until the study drug administration.

In the study drug injection, the nominal dose of oxycodone hydrochloride trihydrate was 0.1 mg·kg⁻¹, diluted to 10 ml with saline. Saline was used as placebo. Both study drugs were clear, colourless liquids, ensuring blinding. They were given simultaneously as 5‐min infusions after the patient had arrived in the PACU, had emerged from anaesthesia to respond to verbal comments and had evaluated pain on an 11‐point numerical rating scale (NRS, 0 = no pain, 10 = most pain) at rest and while coughing/during wound compression. The NRS is a commonly used, sensitive and reliable method of evaluating pain in studies concerning different pain therapies 9. The wound area was compressed with a 20 N force (2 kg pressure with three fingers for a 10 cm² area) 8.

For multimodal analgesia, all patients were given i.v. paracetamol 1 g, starting 15 min, and i.v. dexketoprofen 50 mg three daily doses starting 60 min, after administration of the study drugs. For breakthrough pain, patients were given i.v. http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1626 50 μg when pain at rest was ≥3/10 and/or while coughing/during wound compression ≥5/10. The minimum fentanyl dosing interval was 10 min.

Pain was assessed continuously and recorded at 30‐min intervals during the first 4 h, and then every 6 h for the next 20 h. Arterial blood pressure, heart rate, respiratory rate, peripheral capillary oxygen saturation, exhaled carbon dioxide and sedation score, using the 10‐point Richmond Agitation–Sedation scale (−5 = unarousable, to +4 = combative), were monitored for the first 24 h.

After the first 4 h, patients were admitted to a postoperative ward and postoperative analgesia was continued with an epidural infusion of levobupivacaine (0.6 μg·ml⁻¹), fentanyl (4 μg·ml⁻¹) and epinephrine (2 μg·ml⁻¹), which was the standard treatment of the hospital. An infusion rate of 2–8 ml·h⁻¹ and 2 ml boluses were administered as needed, to keep the pain scores <3/10 at rest and <5/10 while coughing and during wound compression. No further oxycodone was given to patients until 24 h after operation. After the 24‐h study period, postoperative pain was treated according to the standard protocol of the hospital (Table 1). Thus, most patients were likely to have received oxycodone i.v. or orally after the first 24 h postoperatively. Patients' satisfaction with the analgesia was assessed at 24 h with an 11‐point NRS (0 = completely dissatisfied, 10 = completely satisfied).

Table 1.

Comparison of the study protocol with the standard care of treatment in Kuopio University Hospital

	Study protocol	Standard care
Intraoperative care
Remifentanil infusion adjustment	Surgical Pleth Index 20–50	Surgical Pleth Index 20–50
Rocuronium administration during surgery	Post tetanic count 2–5	Post‐tetanic count 2–5
Rocuronium administration at the end of surgery	Train‐of‐four stimulation ≥2 twitches	Train‐of‐four stimulation ≥2 twitches
Sevoflurane inhalation	End‐tidal concentration 0.6–0.8%	End‐tidal concentration 0.6–0.8%
Propofol infusion	Response and state entropy 40–60	Response and state entropy 40–60
Sugammadex	1–2 mg·kg−1, train‐of‐four ratio ≥0.9	1–2 mg·kg−1, train‐of‐four ratio ≥0.9
Postoperative care
Paracetamol	Three daily i.v. doses of 1 g. First 15‐min infusion 15 min after administration of study compounds	Three daily i.v. doses of 1 g. First 15‐min infusion after patient has arrived in the PACU if not hypotensive
Dexketoprofen	Three daily i.v. doses of 50 mg. First 5‐min injection 1 h after administration of study compounds	First of the two or three daily i.v. doses of 25–50 mg 1–2 h after surgery if not contraindicated
Opioid analgesics	Epidural or i.v. injection of oxycodone 0.1 mg·kg−1. Fentanyl 50 μg i.v. every 10 min for rescue analgesia if pain at rest ≥3/10 and dynamic pain ≥5/10	During the first postoperative hours, oxycodone 2–3 mg i.v. or fentanyl 25–50 μg i.v. every 10 min if pain at rest ≥3/10 and dynamic pain ≥5/10
Epidural levobupivacaine–fentanyl–epinephrine infusion	Started 4 h after administration of study compounds at dose of 2–8 ml·h−1	Started within the first hour after surgery at dose of 2–8 ml·h−1

i.v., intravenous; PACU, post‐anaesthesia care unit

Efficacy and safety outcomes

The primary outcome was the analgesic efficacy of oxycodone, measured as the dose of fentanyl given during the first 4 h postoperatively. The secondary outcomes were the time from the study drug administration to the first dose of fentanyl, pain scores and the incidence of adverse effects during the first 24 h postoperatively. Adverse effects were monitored and recorded at each time of pain evaluation. All data for the first 4 h and at 24 h were collected by the investigators, and in the 5–23 h postoperatively by study nurses.

Pharmacokinetic outcomes

A paired blood (5 ml) and CSF sample (1 ml) was collected from the last 18 patients 1–4 h after the study drug injection for analysis of oxycodone and metabolite concentrations, to verify the data from our previous study 4. In that study, all patients had both a spinal and epidural catheter. Thus, there was a theoretical potential for epidural oxycodone to diffuse into the subarachnoid space through the breach in the dura 10. The samples were collected at a random time after the injection of study compounds – no earlier than 1 h after the epidural injection, in order to allow the injection to spread and absorb from the epidural space, and at the latest 4 h after the study compound injection as epidural infusion was started. For CSF collection, a lumbar puncture was performed at L4–L5 with a 27G pencil‐point needle, and a blood sample was collected from the contralateral arm, to study compound administration. Blood and CSF samples were centrifuged and 1200 g at 21°C for 10 min and the separated plasma and CSF were stored at −70°C until analysis. The oxycodone and its metabolites oxymorphone, noroxycodone and noroxymorphone concentrations in the plasma and CSF were determined using an ultra‐performance liquid chromatography mass spectrometry system described previously 4. The lower limit of quantification was 0.1 ng·ml⁻¹, the accuracy of the assay 80–120% and the coefficient of variation below 20%.

Statistical analysis

The sample size calculation was based on our pilot study, where the mean [standard deviation (SD)] need for i.v. fentanyl with i.v. oxycodone was 0.36 (0.08) mg during the first 4 h after gynaecological laparotomy 4. In order to show a decrease of 50% (i.e. 0.18 mg) in the dose of i.v. fentanyl, 12 patients per group would be needed to achieve a power of 0.8 at an alpha of 0.05 (two‐sided test). To allow for withdrawals, 15 patients per group were recruited.

The data were recorded and analysed using the Statistical Package for Social Sciences software (IBM SPSS Statistics 23, IBM Corporation, Armonk, NY, USA). The distribution of continuous data was checked visually, and the normal distribution assumption was checked using the Shapiro–Wilk test. Normally distributed continuous data were analysed using a two‐sample t‐test assuming equal variances. Equality of variances was tested using Levene's test. The Mann–Whitney U‐test was used if continuous data were not normally distributed. For multiple comparisons, the Bonferroni correction was applied. Categorical data were analysed using the chi‐square test, and correlations using the Pearson correlation coefficients with two‐tailed significance testing. Data are presented as number of cases, median, quartiles, minimum, maximum, mean, SD and 95% CI, as appropriate. A P‐value of less than 0.05 was considered statistically significant.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org,the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 11 and are permanently archived in the Concise Guide to PHARMACOLOGY 12, 13.

Results

Patient characteristics are presented in Table 2. Before the injection of the study compounds, the pain scores were similar in the two groups; in the EPI group, the median pain score at rest was 6 (quartiles 6, 8), during wound compression was 7 (5, 8) and while coughing was 6 (5, 8), and in the IV group these values were 7 (4, 8), 7 (4, 8) and 7.5 (5, 8), respectively. All patients completed the 24‐h follow‐up and there were no missing data. There was one major protocol violation. One patient in the EPI group was given a single dose of fentanyl for mild pain (NRS = 1) at rest during the first 4 h in the PACU, with no further doses needed. Her data were included in the analysis.

Table 2.

Patient characteristics

Variable	EPI group (n = 15)	IV group (n = 15)	P‐value
Age, years	48 (14)	52 (13)	0.40
Weight, kg	62 (11)	70 (11)	0.047
Height, cm	161 (5)	166 (5)	0.014
BMI, kg·m −2	24 (3)	25 (5)	0.35
ASA, I/II	7/8	8/7	1.00
Duration of surgery, min	212 (101)	179 (75)	0.32
Intraoperative bleeding, ml	807 (673)	567 (421)	0.25

Data are presented as the mean (standard deviation) and number of cases. body mass index

All patients were given fentanyl during the first 4 h postoperatively. The median number of fentanyl doses was significantly lower after epidural oxycodone [three (1, 8) doses] than after i.v. oxycodone [seven (6, 8) doses (mean difference 3.1; 95% Cl: 0.9, 5.2; P = 0.01)] (Figure 2). The duration of surgery did not correlate with the need for fentanyl; in the EPI group, the Pearson correlation coefficient was 0.058 (P = 0.838), and in the IV group −0.108 (P = 0.702).

Figure 2.

Fentanyl doses during the first 4 h. The EPI group received intravenous saline and epidural oxycodone; the IV group received intravenous oxycodone and epidural saline

In support of the superiority of epidural administration, the pain scores at rest, while coughing and during wound compression at 30–60 min after administration of the study drug were lower in the EPI group than in the IV group (Figure 3).

Figure 3.

Pain scores at rest (A), while coughing (B) and during wound compression (B). These were lower in the EPI group (who received intravenous saline and epidural oxycodone) than in the IV group (who received intravenous oxycodone and epidural saline) at 30 min (P = 0.026, P = 0.021 and P = 0.029, respectively), and at 60 min (P = 0.003, P = 0.013 and P = 0.023, respectively) after study drug administration. NRS 0–10, 11‐point numerical rating scale* P < 0.05

The mean time from administration of the study drug to the first dose of fentanyl was similar in the two groups: 35 (SD 60) min in the EPI group and 19 (8) min in the IV group; P = 0.15. The epidural infusion requirement 5–24 h postoperatively was similar in the two groups: 86 (22) ml in the EPI group and 86 (20) ml in the IV group; P = 0.93.

Patient satisfaction with postoperative analgesia was similarly high in both groups: 9.0 (1.4) in the EPI group and 9.4 (0.8) in the IV group; P = 0.48.

In the EPI group (n = 11), plasma oxycodone ranged between 7.9 ng·ml⁻¹ and 26 ng·ml⁻¹, and in the IV‐group (n = 7) between 6.3 ng·ml⁻¹ and 14 ng·ml⁻¹. CSF oxycodone concentrations were 100‐fold higher in the EPI group (210–1960 ng·ml⁻¹) than in the IV group (6.5–14 ng·ml⁻¹) (Figure 4). The active metabolite of oxycodone, http://guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=7094, was detected in 11/11 patients in the EPI group and 3/7 patients in the IV group. CSF oxymorphone levels ranged between 0.12 ng·ml⁻¹ and 0.41 ng·ml⁻¹ in the EPI group and between 0 and 0.42 ng·ml⁻¹ in the IV group. The other active metabolite, noroxymorphone, was detected in the CSF of one patient in the IV group (0.31 ng·ml⁻¹ – i.e. below the lower limit of quantification). Interestingly, plasma noroxycodone levels were similar between the two groups but CSF noroxycodone levels were 5–10‐times higher after epidural (2.2–8.2 ng·ml⁻¹) than after i.v. (0.27–1.5 ng·ml⁻¹) administration.

Figure 4.

Plasma (A) and cerebrospinal fluid (B) oxycodone concentrations after the administration of an epidural or intravenous injection of 0.1 mg kg⁻¹ to patients undergoing gynaecological laparotomy in the EPI group (who received intravenous saline and epidural oxycodone; n = 11) and the IV group (who received intravenous oxycodone and epidural saline; n = 7) Oxycodone concentrations in both plasma and cerebrospinal fluid are presented in panel C (C)

There were no differences between the two groups in adverse effects during the first 24 h postoperatively. The cumulative incidences of adverse effects during the first 24 h are presented in Table 3.

Table 3.

Adverse effects during the first 24 h postoperatively

	EPI group (n = 15)	IV group (n = 15)
Patients with adverse effects	12	12
Total number of adverse effects	19	18
Postoperative nausea and vomiting	8	9
Pruritus	7	2
Respiratory rate <10/min	2	2
Headache	–	2
Dizziness	1	1
Hypotension	–	1
Motor block	–	1
Major surgical bleeding	1	–

Data represent the number of patients with adverse effects in the first row and the number of adverse effects in the second row

One patient in the EPI group had a major surgical complication. During the first 4 h postoperatively, the patient bled 2800 ml into the abdominal cavity, although intraoperative blood loss had been minimal, at 50 ml. The postoperative bleed was noticed 3.5 h after study drug administration, and the patient completed the first 4‐h follow‐up. However, the bleed may have affected the patient's pain level and need for analgesics. She received 0.45 mg fentanyl during the first 4 h. A paired blood and CSF sample was collected from this patient at 80 min after injection of the study compounds. Her oxycodone concentration in the plasma was 10 ng·ml⁻¹ and in the CSF 983 ng·ml⁻¹. This patient's data for the first 24 h were included in the analyses.

Discussion

In the present study, epidural oxycodone provided superior analgesia to i.v. administration during the first 4 h after laparotomy. The need for fentanyl was lower and pain scores at 30–60 min were lower in the EPI group compared with the IV group. These efficacy data support our preliminary data 4 on the superiority of epidural over i.v. oxycodone. Plasma oxycodone concentrations were similar after epidural and i.v. administration but CSF concentrations were significantly higher in the EPI group. These pharmacokinetic data are consistent with those found in our earlier study 4.

In the present study, CSF oxycodone levels in the EPI group were approximately 100‐fold higher than in the IV group. However, the observed peak concentrations in the present study were approximately one‐quarter of those in our previous study 4. The most likely explanation for this is that, in the present study, CSF samples were collected at the L4–5 level by a single puncture, whereas in the previous study a spinal catheter was used for CSF sampling, with the tip higher in the subarachnoid space. Our aim in that study was to insert the spinal catheter tip close to the epidural catheter in the lower thoracic spinal canal. We assume that in the CSF aliquots in the present study, obtained in the spinal canal 10–15 cm caudal to the level of epidural injection, oxycodone should have been more diluted than in the previous study, where the aliquots were obtained adjacent to the epidural oxycodone injection site. Another explanation could be that in our previous study some of the oxycodone injected into the epidural space may have passed into the subarachnoid space via a dural hole around the spinal catheter 10.

Epidural administration of oxycodone has been evaluated in few previous clinical trials 4, 14, 15, 16, 17. In the study by Bäcklund et al. 14, epidural oxycodone had no benefit over i.v. oxycodone after major laparotomy. However, the comparison was not double blinded, making the results inconclusive. In another study, epidural oxycodone 0.5 mg·kg⁻¹·h⁻¹ after a 4 mg bolus was as effective as epidural morphine 0.25 mg·kg⁻¹·h⁻¹ after a 2 mg bolus for the first 72 h after gynaecological surgery. Pain relief with a smaller dose of epidural oxycodone, 0.25 mg·kg⁻¹·h⁻¹ after a 2 mg bolus, was a little lower during the first 6 h, but thereafter was similar with a higher dose of oxycodone and with morphine 10. Inconsistent with our data on oxycodone, epidural fentanyl appeared to be superior to i.v. administration. After thoracotomy, the need for fentanyl has been found to be significantly lower in epidural infusion (1.0 μg·kg⁻¹·h⁻¹) than that in i.v. (1.7 μg·kg⁻¹·h⁻¹) administration 18.

Although epidural oxycodone was superior to i.v. oxycodone for analgesia after gynaecological laparotomy, in the present clinical trial all patients needed rescue fentanyl. Thus, the appropriate dose of oxycodone for pain relief after laparotomy should be evaluated in further studies. Most of the fentanyl doses in the EPI group were given during the first 30 min after the epidural oxycodone, and there was a second peak at 3 h after the oxycodone dose. Based on these data, both the onset and duration of the analgesic action of epidural oxycodone should be considered. Population pharmacokinetic modelling of these and earlier published data on the disposition characteristics of oxycodone can be used to support the evaluation of the dose rationale in future clinical trials 6.

Oxycodone was well tolerated in the present study and there were no differences in adverse events between the two groups. However, the small sample size was unlikely to detect any differences in adverse effects between the groups. The most common adverse event was postoperative nausea and vomiting (PONV) in both groups. However, there is a significant incidence of PONV in gynaecological laparotomy patients, regardless of the pain management used 19. Pruritus was more common in the EPI group, and this was probably associated with oxycodone. Intrathecal opioids cause a higher incidence of pruritus compared with other administration routes, and this should be taken in account when using epidural oxycodone. The incidence of pruritus seems to be similar after epidural oxycodone and epidural morphine 16.

Oxycodone has active circulating metabolites, and their role in the analgesia achieved with the parent compound has been under debate. In experimental studies, Lemberg et al. 20, 21 showed that noroxymorphone administered spinally, and oxymorphone systemically and spinally are more potent analgesics than oxycodone. Our earlier data and the present study showed that CSF concentrations of both noroxymorphone and oxymorphone are very low after epidural oxycodone. This finding suggests that the circulating metabolites of oxycodone may have some role in analgesic efficacy after systemic but not spinal administration of oxycodone 4. It remains unclear whether i.v. and epidural administration of oxycodone may produce sufficient concentrations of noroxymorphone and/or oxymorphone in the supraspinal pain‐modulating regions of the CNS in humans.

Another interesting finding was that the CSF concentrations of the inactive metabolite, noroxycodone, in the EPI group were 5–10 times higher than in the IV group. Plasma noroxycodone was similar between groups. The explanation for this requires further evaluation.

One of the strengths of the study was that intraoperative care was standardized, with patients receiving a standardized multimodal postoperative analgesic regimen consisting of paracetamol and dexketoprofen. The main limitation of the study was the relatively small sample size. The study power was sufficient to support our hypothesis that the analgesic efficacy of epidural oxycodone is superior to i.v. administration, but no conclusions can be drawn on the safety of this approach based on this small number of patients. Respiratory depression and neurotoxicity are the main concerns with any compounds injected epidurally. In both our previous pharmacokinetic study 4 and the present study, no signs or symptoms of neural irritation were noted. An in vitro study that we conducted in human neuroblastoma and mouse motoneuronal cell cultures suggested that the neuronal toxicity of oxycodone is similar to or less than that of morphine 22. However, carefully designed experimental and behavioural studies are needed to clarify this issue. Another of the limitations of the present study was that no pain scores were recorded in the database just prior to the first rescue dose of fentanyl. Thus, we were unable to evaluate whether the pain scores correlated with oxycodone concentrations. However, oxycodone CNS penetration is associated with a time to reach peak concentration in the CSF of approximately 0.6 h after epidural administration, and over 1 h after i.v. administration 6. Surgical technique may affect postoperative pain in gynaecological surgery, and optimal epidural oxycodone dosing should be assessed after different procedures 1, 23.

Conclusions

Epidural oxycodone provided better postoperative analgesia than i.v. oxycodone after gynaecological laparotomy. Further research is needed to assess the optimal dose of epidural oxycodone and to compare the efficacy of epidural oxycodone with that of other opioids.

Competing Interests

There are no competing interests to declare.

This work was supported by the Finnish Cultural Foundation, Governmental VTR‐funding (507A009) of Kuopio University Hospital and Olvi‐Säätiö (201710334). The study was not financially supported by any external sources.

Contributors

P.P. and M.K. carried out data collection. P.P., H.K., V‐P.R. and M.K. analysed and interpreted the data. P.P., H.K., H.H., V‐P.R. and M.K. contributed to the writing of the manuscript. P.P., H.K. H.H., and V‐P.R. were responsible for the critical evaluation of the intelligent content of the final version of the manuscript. H.K. and M.K. designed the study. H.H. measured drug concentrations. M.K. gave final approval of the version of the submitted manuscript.

Piirainen, P. , Kokki, H. , Hautajärvi, H. , Ranta, V.‐P. , and Kokki, M. (2018) The analgesic efficacy and pharmacokinetics of epidural oxycodone after gynaecological laparotomy: a randomized, double‐blind, double‐dummy comparison with intravenous administration. Br J Clin Pharmacol, 84: 2088–2096. 10.1111/bcp.13643.