1428
`
`Comparative Clinical Efficacy and Safety of a Novel
`Controlled-Release Oxycodone Formulation and
`Controlled-Release Hydromorphone in the Treatment
`of Cancer Pain
`
`Neil A. Hagen, M.D., F.R.C.P.1
`Najib Babul, Pharm.D.2
`
`1 Department of Medicine, Tom Baker Cancer
`Centre, and Clinical Neurosciences, The Univer-
`sity of Calgary, Alberta, Canada.
`
`2 Department of Scientific Affairs, Purdue Fred-
`erick, Pickering, Ontario, Canada.
`
`Presented in part at the 8th World Congress
`on Pain, Vancouver, British Columbia, Canada,
`August 18, 1996.
`
`The authors thank Drs. Helen Hays, H. S. Dhali-
`wal, Dwight Moulin, Louise Yelle, Oscar Wong,
`and Deborah Dudgeon for patient recruitment;
`Ms. Iva Ford for study monitoring; and Mr. Zol-
`tan Harsanyi for data analysis.
`
`reprints: Neil A. Hagen, M.D.
`Address for
`F.R.C.P.C., Department of Medicine, Tom Baker
`Cancer Centre, 1331 29 St. NW, Calgary, Al-
`berta, Canada T2N 4N2.
`
`Received September 17, 1996; revision received
`December 2, 1996; accepted December 2, 1996.
`
`q 1997 American Cancer Society
`
`BACKGROUND. The use of oxycodone to treat chronic cancer pain has been ham-
`pered by its short elimination half-life, which necessitates administration every 4
`hours. This study compared the clinical efficacy and safety of a novel oxycodone
`formulation with that of hydromorphone in the treatment of cancer pain.
`METHODS. In a double-blind crossover study, 44 patients with stable cancer pain
`were randomized to controlled-release oxycodone or controlled-release hydromor-
`phone, each given every 12 hours for 7 days. Pain intensity, nausea, and sedation
`were assessed by patients four times daily, and breakthrough analgesia was re-
`corded.
`RESULTS. Thirty-one patients completed the study (18 women, 13 men; mean age,
`56 { 3 years) and received a final controlled-release oxycodone dose of 124 { 22
`mg per day and a final controlled-release hydromorphone dose of 30 { 6 mg per
`day. There were no significant differences between treatments in overall Visual
`Analogue Scale (VAS) pain intensity (VAS 28 { 4 mm vs. 31 { 4 mm), categorical
`pain intensity (1.4 { 0.1 vs. 1.5 { 0.1), daily rescue analgesic consumption (1.4 {
`0.3 vs. 1.6 { 0.3), sedation scores (24 { 4 mm vs. 18 { 3 mm), nausea scores
`(15 { 3 mm vs. 13 { 3 mm), or patient preference. Two patients experienced
`hallucinations on controlled-release hydromorphone, but none did while receiving
`controlled-release oxycodone.
`CONCLUSIONS. Controlled-release oxycodone demonstrated excellent pharmaco-
`dynamic characteristics, analgesic efficacy, and safety as compared with con-
`trolled-release hydromorphone and represents an important new therapeutic op-
`tion for cancer pain management. Cancer 1997;79:1428–37.
`q 1997 American Cancer Society.
`
`KEYWORDS: oxycodone, hydromorphone, controlled release, cancer pain, drug treat-
`ment, drug safety, clinical trial.
`
`O
`
`xycodone is a semisynthetic opioid analgesic with a high oral-to-
`parenteral bioavailability1,2 and a twofold greater oral potency
`than oral morphine.3–6 There are anecdotal data suggesting a lower
`incidence of side effects, such as hallucinations, for oxycodone rela-
`tive to morphine2 and relative to other opioids.7 Use of oxycodone
`for cancer pain has been hampered by its short elimination half-
`life, which necessitates dosing every 4 hours in order to maintain an
`optimal level of analgesia.
`
`/ 7b52$$0984
`
`03-07-97 11:35:20
`
`canas
`
`W: Cancer
`
`1
`
`

`

`In contrast, the use of morphine has gained wide-
`spread acceptance due to its global availability, sig-
`nificant pharmacokinetic and pharmacodynamic data,
`and the development of a controlled-release formula-
`tion.8–13 Despite its extensive use, the occurrence of
`adverse effects necessitates discontinuation in some
`patients. Clinicians treating cancer pain with opioids
`report significant variability among patients in efficacy
`and side effects with available opioid analgesics. Pa-
`tients who have poor analgesic efficacy or safety out-
`comes with one opioid will frequently tolerate another
`opioid well.11,13–20 The clinical observation of a differ-
`ential pharmacologic response has led to the growing
`acceptance and clinical practice of opioid rota-
`tion.11,13–20 The recent availability of controlled-release
`formulations of other opioids, such as hydromor-
`phone, provides clinicians with therapeutic options for
`cancer pain management.9,10,21,22 Recently, a con-
`trolled-release oxycodone formulation has been devel-
`oped that provides a biphasic absorption pattern,
`characterized by an initial rapid onset followed by a
`prolonged phase.5,23,24 We compared the clinical effi-
`cacy and safety of this formulation of oxycodone with
`a previously evaluated controlled-release formulation
`of hydromorphone9,10,25 available in Canada and other
`countries.
`
`MATERIALS AND METHODS
`Subjects
`Forty-four patients with chronic cancer pain and
`stable analgesic requirements were selected for par-
`ticipation. Criteria for exclusion included known hy-
`persensitivity to opioid analgesics, intolerance of ox-
`ycodone or hydromorphone, presence of a medical
`or surgical condition likely to interfere with drug
`absorption in the gastrointestinal tract, concurrent
`use of other opioid analgesics during the study pe-
`riod, and presence of intractable nausea or vomiting.
`Patients who had undergone or were expected to
`undergo therapeutic procedures likely to influence
`their pain during the study period were also ex-
`cluded. The study protocol and informed consent
`form received scientific and ethical approval, and
`patients gave written informed consent before par-
`ticipating in the study.
`
`Medications
`Controlled-release oxycodone (Oxycontin; Purdue
`Frederick, Pickering, Ontario, Canada) and controlled-
`release hydromorphone (Hydromorph Contin; Purdue
`Frederick, Pickering, Ontario, Canada) were each ad-
`ministered every 12 hours for 7 days. Further blind-
`label dose changes were permitted during the study;
`
`Oxycodone for Cancer Pain/Hagen and Babul
`
`1429
`
`and in the event of a dose change, the rescue analgesic
`dose was modified correspondingly. Blinding was
`maintained by the double-dummy technique, which
`involved matching placebos. In the active treatment
`phases, patients received either active controlled-re-
`lease oxycodone and placebos matching controlled-
`release hydromorphone or active controlled-release
`hydromorphone and placebos matching controlled-
`release oxycodone.
`Incident and nonincident breakthrough pain was
`treated with immediate-release oxycodone and hydro-
`morphone matching the active opioid analgesic at a
`dosage of approximately 10% of the daily scheduled
`opioid dose. No opioids other than the test and rescue
`analgesic medications were permitted during the
`course of the study. Nonopioid analgesics, such as
`nonsteroidal anti-inflammatory drugs and acetamino-
`phen, and adjuvant drugs, such as antidepressants,
`anticonvulsants,
`corticosteroids, bisphosphonates,
`and psychostimulants, that had been part of the pa-
`tient’s therapy were continued at the same dose level
`throughout the study period.
`
`Study Design
`The study was a double-blind, randomized, two-way
`crossover evaluation of the clinical efficacy and
`safety of controlled-release oxycodone and con-
`trolled-release hydromorphone. Pain was character-
`ized at the start of the study using the Edmonton
`staging system according to pathophysiologic mech-
`anisms (visceral, bone/soft
`tissue, neuropathic,
`mixed, or unknown) and pain characteristics (inci-
`dent or nonincident pain).26 Patients with 3 days of
`stable analgesic requirements on a prestudy opioid
`were randomized to controlled-release oxycodone
`or controlled-release hydromorphone. Stable anal-
`gesia was defined as 2 or fewer rescue doses of opioid
`analgesic per 24-hour period, calculated over 3 or
`more days. Once controlled-release oxycodone and
`controlled-release hydromorphone treatment were
`initiated, further dose adjustments were permitted
`throughout the study period, and breakthrough pain
`was treated with rescue doses of the matching im-
`mediate-release opioid. At the end of Phase I, pa-
`tients were crossed over to the alternative treatment
`in Phase II without an intervening washout period.
`Pain intensity was assessed by the patient 4
`times a day (at 8 a.m., 12 noon, 4 p.m., and 8 p.m.)
`on a 100 mm Visual Analogue Scale (VAS) (anchors:
`no pain to excruciating pain) and on a 5-point cate-
`gorical scale (0 (cid:129) none, 1 (cid:129) mild, 2 (cid:129) moderate, 3
`(cid:129) severe, 4 (cid:129) excruciating). Nausea and sedation
`were also assessed on a 100 mm VAS at the same
`
`/ 7b52$$0984
`
`03-07-97 11:35:20
`
`canas
`
`W: Cancer
`
`2
`
`

`

`Characteristic
`
`Men/Women
`Mean age (yrs), (mean { SE)
`Primary tumor
`Breast
`Colorectal
`Lung
`Urologic/prostate
`CNS
`Unknown primary site
`Other
`Total
`
`CNS: central nervous system; SE: standard error.
`
`No. of patients
`
`13/18
`56 { 3
`
`7
`5
`1
`5
`4
`2
`7
`31
`
`1430
`
`CANCER April 1, 1997 / Volume 79 / Number 7
`
`TABLE 1
`Characteristics of Study Patients (n (cid:129) 31)
`
`times (nausea anchors: no nausea to severe nausea;
`sedation anchors: no sedation to extreme sedation).
`Spontaneously reported, investigator-observed, and
`elicited adverse events were recorded at the end of
`each phase, and patient and investigator treatment
`preferences were recorded on completion of both
`treatments. At the end of the double-blind study,
`patients who requested continued treatment with
`controlled-release oxycodone received the drug in
`an open-label, longitudinal evaluation.
`The primary measures of efficacy were the 100
`mm pain intensity VAS and the 5-point categorical
`intensity assessment. Previous studies of cancer
`pain have suggested that standard deviations of 17
`mm (pain VAS)27 and 0.6 units (categorical)28 could
`be expected, and we adopted these values for sample
`size estimation. With each patient serving as his or
`her own control, and assuming low correlation be-
`tween responses in a single subject in the 2 treat-
`was determined, and the average number of doses
`ment periods, 24 completed patients (12 patients per
`per time segment was computed. Multivariate re-
`sequence) were calculated to provide 80% power to
`peated measure analysis was used to test for effects
`detect a difference of 15 mm pain intensity VAS and
`of drug, time segment, and their interaction. Data
`0.6 units categorical pain intensity at a statistical
`on rescue consumption from 2 12-hour periods be-
`significance of 0.05.
`ginning at 8 a.m. and 8 p.m. were combined to ob-
`Analysis of treatment sequence revealed no sig-
`tain an overall estimate of rescue use during succes-
`nificant carryover effects. Therefore, differences in
`sive 4-hour intervals during a 12-hour period corre-
`treatment effects for a two-period crossover design
`sponding to the dosing interval of controlled-release
`were evaluated. The pain intensity (VAS and categor-
`oxycodone and controlled-release hydromorphone.
`ical), nausea (VAS), and sedation (VAS) scores were Multivariate repeated measures analysis was used to
`summarized across days and times as well as overall.
`test for the effects of drug, time segment, and their
`interaction.
`The scores for each patient were averaged as follows:
`(1) over 4 timepoints for each of 7 days, (2) over 7
`In addition to analysis of nausea and sedation
`scores, adverse events were analyzed in terms of the
`days for each of 4 timepoints, and (3) over all days
`and times. Three-way analysis of variance was used
`number of patients reporting each event. Fisher’s
`exact test was used to determine the significance of
`to test for the effect of drug and treatment sequence
`and for phase, using data from the overall mean
`differences in the frequency of side effects. Differ-
`ences in patients’ and investigators’ assessment of
`scores. Multivariate repeated measures analysis was
`used to test for effect of drug, day, time of day, and
`treatment effectiveness were compared using the chi
`square test. Treatment preferences were compared
`their interactions. Two separate repeated measures
`analyses were performed: (1) test for drug, day, and
`using normal approximation to the binomial.
`drug 1 day interaction and (2) test for drug, time of
`Statistical significance was defined as P (cid:155) 0.05
`day, and drug 1 time-of-day interaction.
`for a two-tailed hypothesis. All means are presented
`with the corresponding standard error.
`Rescue analgesic use was determined for each
`patient as the total number of rescue doses per day
`of study. Three-way analysis of variance was used to
`test for the effect of drug, treatment sequence, and
`phase, using the data from the overall mean scores.
`Rescue use by time of day was computed by catego-
`rizing time of rescue into 1 of 6 4-hour segments.
`Because the first dose of each phase began at 8 p.m.,
`the interval from 8 p.m. to 12 midnight was desig-
`nated the first 4-hour segment. The number of doses
`for each patient during each of these time segments
`
`RESULTS
`Forty-four patients were enrolled and 31 patients
`completed the study (18 women and 13 men; mean
`age, 56 { 3 years). Reasons for premature withdrawal
`from the study included adverse events (8 patients),
`inadequate pain control (3 patients), intercurrent ill-
`ness (1 patient), and voluntary withdrawal (1 pa-
`tient). Failure to complete both phases of the study
`did not appear to be related to toxicity of one of the
`
`/ 7b52$$0984
`
`03-07-97 11:35:20
`
`canas
`
`W: Cancer
`
`3
`
`

`

`Oxycodone for Cancer Pain/Hagen and Babul
`
`1431
`
`TABLE 2
`Comparison of Clinical Efficacy and Safety Variables by Day of Treatment after Drug Administrationa
`
`CR oxycodone
`
`CR hydromorphone
`
`Variable
`
`Day 1
`
`Day 2
`
`Day 3
`
`Day 4
`
`Day 5
`
`Day 6
`
`Day 7
`
`Overall
`
`Day 1
`
`Day 2
`
`Day 3
`
`Day 4
`
`Day 5
`
`Day 6
`
`Day 7
`
`Overall
`
`Pain intensity VAS
`(0–100 mm)
`Pain intensity ordinal
`(0–5)
`Rescue analgesic use
`(doses/day)
`Sedation VAS (0–100 mm)
`Nausea VAS (0–100 mm)
`
`28.4
`
`1.4
`
`1.3
`23.4
`16.1
`
`27.5
`
`1.3
`
`1.2
`23.5
`15.4
`
`29.7
`
`1.5
`
`1.5
`25.4
`14.3
`
`29.3
`
`1.4
`
`1.6
`24.4
`17.1
`
`28.6
`
`1.4
`
`1.5
`24.1
`15.5
`
`27.4
`
`1.3
`
`1.1
`22.4
`15.7
`
`25.2
`
`1.3
`
`1.3
`22.4
`14.1
`
`28.0
`
`1.4
`
`1.4
`23.6
`15.5
`
`29.7
`
`1.5
`
`2.0
`18.0
`11.8
`
`29.5
`
`1.6
`
`1.5
`18.1
`13.8
`
`32.1
`
`1.5
`
`1.5
`18.0
`12.8
`
`30.6
`
`1.5
`
`1.6
`18.3
`11.8
`
`33.3
`
`1.6
`
`1.5
`16.6
`15.5
`
`31.3
`
`1.5
`
`1.4
`20.6
`14.4
`
`27.7
`
`1.3
`
`1.5
`17.6
`12.2
`
`30.6
`
`1.5
`
`1.6
`18.2
`13.1
`
`CR: controlled-release; VAS: Visual Analogue Scale.
`a Data represent mean daily scores over 4 assessment periods (8 a.m., 12 noon, 4 p.m., and 8 p.m.); overall score represents mean over all days and times.
`
`TABLE 3
`Comparison of Clinical Efficacy and Safety Variables by Time of Day after Drug Administrationa
`
`CR oxycodone
`
`CR hydromorphone
`
`8:00
`a.m.
`
`12:00
`a.m.
`
`4:00
`p.m.
`
`8:00
`p.m.
`
`Overallb
`
`8:00
`a.m.
`
`12:00
`a.m.
`
`4:00
`p.m.
`
`8:00
`p.m.
`
`Overallb
`
`30.7
`
`23.8
`
`27.3
`
`30.3
`
`1.5
`
`0.2
`
`22.9
`
`16.0
`
`1.2
`
`0.4
`
`23.8
`
`15.2
`
`1.4
`
`0.3
`
`24.1
`
`14.7
`
`1.5
`
`0.1
`
`23.9
`
`15.9
`
`28.0
`
`1.4
`
`23.6
`
`15.5
`
`32.9
`
`27.4
`
`29.6
`
`32.6
`
`1.6
`
`0.3
`
`19.8
`
`14.5
`
`1.4
`
`0.4
`
`17.9
`
`13.0
`
`1.5
`
`0.3
`
`16.8
`
`12.3
`
`1.6
`
`0.2
`
`18.2
`
`13.0
`
`30.6
`
`1.5
`
`18.2
`
`13.2
`
`Variable
`
`Pain intensity VAS
`(0–100 mm)
`Pain intensity ordinal
`(0–5)
`Rescue analgesic use
`(dose/4 hrs)
`Sedation VAS
`(0–100 mm)
`Nausea VAS
`(0–100 mm)
`
`CR: controlled-released; VAS: Visual Analogue Scale.
`a Data represent mean score by time of day over 7 days.
`b Overall score represents mean over all days and times.
`
`study drugs over another. The analyses of all efficacy
`outcome variables, including VAS and categorical
`pain intensity, sedation VAS, and nausea VAS were
`restricted to patients completing both phases of the
`study. Spontaneously reported safety variables were
`analyzed for all patients enrolled. The mean daily
`initial controlled-release oxycodone and controlled-
`release hydromorphone doses were 120 { 22 mg and
`24 { 4 mg, respectively. The mean final controlled-
`release oxycodone and controlled-release hydro-
`morphone doses were 124 { 22 and 30 { 6 mg, re-
`spectively. There were 0.65 dose changes during the
`controlled-release oxycodone phase and 0.81 dose
`changes during the controlled-release hydromor-
`phone phase. Sixty-one percent of patients had bone
`
`pain, 29% had soft tissue pain, 23% had visceral pain,
`and 45% had neuropathic pain. Sixteen percent de-
`scribed pain as lancinating. Sixty-one percent of pa-
`tients had steady pain and 52% experienced incident
`pain with or without steady pain. Table 1 provides
`characteristics of evaluable patients.
`The smallest treatment differences that could be
`detected for overall pain intensity using the VAS and
`categorical scales were estimated assuming a type 1
`error rate (alpha) of 0.05 and type 2 error rate (beta)
`of 0.2. Using the sample size of 31, intrasubject cor-
`relations for pain intensity, and the variance esti-
`mates from the data, differences of 4.4 mm on the
`100 mm VAS and 0.2 points on the 5-point categori-
`cal scale were detectable with 80% power.
`
`/ 7b52$$0984
`
`03-07-97 11:35:20
`
`canas
`
`W: Cancer
`
`4
`
`

`

`1432
`
`CANCER April 1, 1997 / Volume 79 / Number 7
`
`FIGURE 1. Mean pain intensity Visual Analogue Scale (VAS) by time of
`day, after administration of controlled-release (CR) oxycodone and CR
`hydromorphone; h (cid:129) hours.
`
`FIGURE 3. Mean rescue analgesic use at successive 4-hour intervals,
`after administration of controlled-release (CR) oxycodone and CR hydro-
`morphone; h (cid:129) hours (time of day).
`
`treatments in pain intensity on the categorical scale
`when tested by day of treatment (P (cid:129) 0.1009) and
`by time of day (P (cid:129) 0.0960). The overall mean cate-
`gorical pain intensity score across all days was 1.4
`{ 0.1 for controlled-release oxycodone and 1.5 { 0.1
`for controlled-release hydromorphone (P (cid:129) 0.0960).
`For both treatment groups, there were significant
`changes in pain intensity VAS (P (cid:129) 0.0018) and cate-
`gorical (P (cid:129) 0.0001) over the course of a day, with
`less pain during the night.
`The frequency of rescue analgesic use is shown in
`6 4-hour time segments in Figure 3 and by day of
`treatment and time of day Tables 2 and 3, respectively.
`There were no significant differences in daily rescue
`analgesic use between controlled-release oxycodone
`and controlled-release hydromorphone (1.4 { 0.3 vs.
`1.6 { 0.3, P (cid:129) 0.1906). Repeated measures analysis
`of variance by time of day indicated no significant
`differences between the treatments in the pattern of
`rescue analgesic use over 6 4-hour time segments (P
`(cid:129) 0.2258), although for both formulations there was a
`significant difference (P (cid:129) 0.0066) in the pattern of
`rescue analgesic use over each day, with fewer doses
`taken during the night. The percentage of rescue anal-
`gesic use in the first, second, and third 4-hour period
`representing the combined 12-hour dosing frequency
`of controlled-release oxycodone (8 a.m. to 8 p.m. and
`8 p.m. to 8 a.m.) showed no significant differences
`between treatments (P (cid:129) 0.3929) and was 23.8%,
`42.8%, and 33.3% for controlled-release oxycodone
`and 33.3%, 37.5%, and 29.2% for controlled-release
`hydromorphone.
`At the completion of the study, patients and inves-
`tigators were asked to identify their treatment prefer-
`ence while blinded to the treatment assignment.
`
`FIGURE 2. Mean categorical pain intensity by day of treatment, after
`administration of controlled-release (CR) oxycodone and CR hydromor-
`phone.
`
`The mean pain intensity VAS scores for con-
`trolled-release oxycodone and controlled-release
`hydromorphone are given in Table 2 by day of treat-
`ment and in Table 3 and Figure 1 by time of day.
`There were no significant differences between treat-
`ments in pain intensity VAS when tested by day of
`treatment (P (cid:129) 0.1091) and by time of day (P (cid:129)
`0.1091). The overall mean pain intensity VAS across
`all days was 28 { 4 mm for controlled-release oxyco-
`done and 31 { 4 mm for controlled-release hydro-
`morphone (P (cid:129) 0.1119). The mean pain intensity
`scores on the categorical scale after administration
`of controlled-release oxycodone and controlled-re-
`lease hydromorphone are given in Table 2 and Fig-
`ure 2 by day of treatment and in Table 3 by time of
`day. There were no significant differences between
`
`/ 7b52$$0984
`
`03-07-97 11:35:20
`
`canas
`
`W: Cancer
`
`5
`
`

`

`Oxycodone for Cancer Pain/Hagen and Babul
`
`1433
`
`FIGURE 4. Mean sedation Visual Analogue Scale (VAS) score by time
`of day, after administration of controlled-release (CR) oxycodone and CR
`hydromorphone, is given; h (cid:129) hours.
`
`FIGURE 5. Mean nausea Visual Analogue Scale (VAS) score by time of
`day, after administration of controlled-release (CR) oxycodone and CR
`hydromorphone, is given; h (cid:129) hours.
`
`Among patients, 25.8% indicated no preference, 35.5%
`preferred controlled-release oxycodone, and 38.7%
`preferred controlled-release hydromorphone. Among
`investigators, 50% indicated no preference, 26.7% pre-
`ferred controlled-release oxycodone, and 23.3% pre-
`ferred controlled-release hydromorphone. Among pa-
`tients and investigators, the proportions expressing a
`preference for controlled-release oxycodone and hy-
`dromorphone were not significantly different (P (cid:129)
`0.2980 and P (cid:129) 0.6290, respectively). There were no
`significant differences in overall effectiveness ratings
`by patients (2.1 { 0.1 vs. 2.1 { 0.1, P (cid:129) 0.9275) and
`by investigators (2.0 { 0.2 vs. 1.9 { 0.1, P (cid:129) 0.9510).
`The mean sedation VAS scores for each treatment
`are given in Table 2 by day of treatment and in Table 3
`and Figure 4 by time of day. There were no significant
`differences between treatments in sedation VAS when
`tested by day of treatment (P (cid:129) 0.1114) and by time
`of day (P (cid:129) 0.1114). The overall mean sedation VAS
`scores across all days was 24 { 4 mm for controlled-
`release oxycodone and 18 { 3 mm for controlled-re-
`lease hydromorphone (P (cid:129) 0.0843).
`Nausea VAS scores for each treatment are given
`in Table 2 by day of treatment and in Table 3 and
`Figure 5 by time of day. There were no significant
`differences between treatments in nausea scores when
`tested by day of treatment (P (cid:129) 0.3807) and by time
`of day (P (cid:129) 0.3807). The overall mean nausea VAS
`score across all days was 15 { 3 mm for controlled-
`release oxycodone and 13 { 3 mm for controlled-re-
`lease hydromorphone (P (cid:129) 0.1958). The overall fre-
`quency of adverse events with controlled-release oxy-
`codone and controlled-release hydromorphone, re-
`corded from spontaneous reports and through the use
`of a nondirected adverse events questionnaire, were
`
`not significantly different (P (cid:129) 0.8250) and generally
`consistent with the use of opioid analgesics to treat
`patients with advanced cancer. Among patients com-
`pleting the study, there were no significant differences
`in the frequency of elicited adverse events, except with
`regard to drowsiness, which was reported more fre-
`quently with oxycodone than with hydromorphone (28
`vs. 19 patients, P (cid:129) 0.0160). Two patients developed
`hallucinations with controlled-release hydromor-
`phone, but none did with controlled-release oxyco-
`done.
`At the completion of the double-blind evaluation,
`13 patients asked to participate in long term, open-
`label treatment with controlled-release oxycodone.
`The starting dose of controlled-release oxycodone was
`111 { 27 mg (range, 20–360 mg), and the daily final
`dose was 149 { 27 mg (range, 20–360), taken for a
`mean duration of 149 days (range, 3–373 days). Pain
`was generally well controlled, although most patients
`required dose escalation at some point in the trajec-
`tory of their illness.
`
`DISCUSSION
`This is the first randomized clinical trial to compare
`the efficacy and safety of hydromorphone and oxy-
`codone, two potent opioid agonists, in the treatment
`of chronic severe cancer pain.
`Oral morphine is widely utilized for the manage-
`ment of chronic severe cancer pain, and it has been
`considered the opioid of choice by many clinicians.
`Unfortunately, the occurrence of unmanageable ad-
`verse effects requires its discontinuation in some pa-
`tients. Recent clinical experience suggests that pa-
`tients who have failed to obtain adequate analgesia
`
`/ 7b52$$0984
`
`03-07-97 11:35:20
`
`canas
`
`W: Cancer
`
`6
`
`

`

`1434
`
`CANCER April 1, 1997 / Volume 79 / Number 7
`
`due to intolerable and unmanageable side effects
`while taking one opioid may benefit from switching
`to an alternative opioid.11,13 – 20 Maddocks et al. have
`shown that in patients with morphine-induced delir-
`ium, the substitution of morphine with oxycodone
`results in significant improvement in mental status,
`nausea, and vomiting.20
`There is extensive clinical experience with the
`use of hydromorphone as an alternative to mor-
`phine. The recent availability of a controlled-release
`formulation of hydromorphone9,10,25 has provided an
`opportunity for an expanded role for this semisyn-
`thetic congener of morphine.
`In contrast, oxycodone is an opioid analgesic
`that has traditionally been utilized in the treatment
`of moderate pain. Oxycodone was first introduced
`in North America as a fixed dose opioid/non-opioid
`combination, and it has generally been used for the
`same indications as codeine or propoxyphene com-
`binations. Although there is convincing evidence
`that oral oxycodone is twice as potent as oral mor-
`phine3 – 6 and equally effective,1,2,29 its widespread use
`has been hampered in part by a need for dosing
`every 4 hours. Ferrell et al. have demonstrated that
`compliance with opioid analgesics increases as the
`required dosing frequency decreases and that non-
`compliance results in suboptimal pain control and
`poor quality of life.30 Therefore, availability of a con-
`trolled-release formulation of oxycodone could en-
`hance its utility as a strong opioid alternative to oral
`morphine for severe cancer pain.
`Opioid rotation has been widely utilized as a
`strategic intervention for dose-limiting toxicities to
`opioids, and it has been recommended in clinical
`practice guidelines for cancer pain management.11
`There is published evidence from case series that
`switching to hydromorphone, oxycodone, or other
`morphine alternatives can be an effective strategy
`for the analgesic management of patients who de-
`velop encephalopathy, sedation, myoclonus, nausea
`and vomiting, hyperalgesia, and analgesic fail-
`ure.11,14 – 20 Variations
`in pharmacodynamic
`re-
`sponses to different opioids have been observed dur-
`ing treatment with drugs that appear to have phar-
`macologically similar sites of activity.14 – 20 These
`reports contradict
`the hypothesis that any one
`strong opioid is uniquely more efficacious than an-
`other. Recent evidence suggests that in addition to m
`receptor – mediated analgesia, oxycodone possesses
`intrinsic activity at the k receptor,31 providing fur-
`ther rationale for its use in treating patients unable
`to obtain an optimal analgesic or side effect profile
`with morphine or hydromorphone. Clinicians can
`
`exploit this variability in drug response by empiri-
`cally offering sequential trials of different opioids
`in order to optimize analgesia and minimize side
`effects.
`In selecting opioid analgesics for cancer pain,
`clinicians need to be mindful of the high prevalence
`of organ impairment and substantial accumulation
`of opioid metabolites, such as morphine-3-glucuro-
`nide and morphine-6-glucuronide.32,33 Morphine-3-
`glucuronide accumulation has been implicated in
`hyperalgesia, respiratory stimulation, and behav-
`ioral excitatory properties through nonopioid recep-
`tor mechanisms.34 – 39 Morphine-6-glucuronide accu-
`mulation has been implicated in increasing levels of
`nausea and sedation in patients with renal impair-
`ment.40 Furthermore, morphine and its principle
`metabolites may also accumulate in patients with
`hepatic dysfunction.41 In contrast, the major metab-
`olite of oxycodone, noroxycodone, is less than 1/
`100th as potent as the parent compound;42 oxyco-
`done administered to patients with renal or hepatic
`impairment only slightly alters the pharmacokinet-
`ics of oxycodone and its metabolites.43,44 Oxymor-
`phone, another metabolite of oxycodone, is pharma-
`cologically active but is produced in such small con-
`centrations that it is unlikely to be of consequence
`in most clinical settings.44
`The results of our study demonstrate that
`chronic severe cancer pain can be well controlled
`with a regular regimen of controlled-release oxyco-
`done given every 12 hours and that the efficacy of
`this regimen is at least equal to that of controlled-
`release hydromorphone.9,10,25 The duration of anal-
`gesia provided by controlled-release oxycodone and
`controlled-release hydromorphone was compared
`through analysis of pain intensity scores and rescue
`analgesic use during three equal 4-hour periods con-
`stituting the 12-hour dosing interval. Analgesics last-
`ing less than 12 hours would be expected to result
`in an increase in pain intensity or rescue analgesic
`use during the period immediately before the next
`dose. The lack of such an end of dose effect in the
`current study provides pharmacodynamic confir-
`mation of the previously documented sustained-re-
`lease pharmacokinetic profile of controlled-release
`oxycodone.45,46
`There were no significant differences in the fre-
`quency of spontaneously reported and elicited ad-
`verse events in the current study, with the exception
`of drowsiness, which occurred more frequently dur-
`ing treatment with oxycodone. However, this finding
`was not consistent with the results of more system-
`
`/ 7b52$$0984
`
`03-07-97 11:35:20
`
`canas
`
`W: Cancer
`
`7
`
`

`

`atic evaluation of sedation using a VAS four times a
`day.
`The mean final daily controlled-release oxyco-
`done dose of 124 mg in our study (equivalent to
`approximately 250 mg of oral morphine) is compara-
`ble to the dose of controlled-release morphine used
`in previous randomized trials in cancer pain.27,28 The
`dose of oral oxycodone required to provide optimal
`analgesia without intolerable side effects ranged
`from 20 – 550 mg per day. This wide variability
`among patients is consistent with the results of pre-
`vious studies with controlled-release morphine27,28
`and controlled-release hydromorphone9,10,25 and is
`well within the range of oxycodone doses used in
`the management of cancer pain.1,2 Differences in an-
`algesic requirements may result from the pharmaco-
`kinetic variability observed after oral administration
`of oxycodone.1,45 However, variability in pharmaco-
`kinetics is probably not sufficient to explain the wide
`range of oxycodone doses required in the current
`study. A variety of factors may be important determi-
`nants of opioid requirement, including prior opioid
`exposure, pain mechanism, predisposition to side
`effects, and psychologic distress.47
`Breakthrough pain is reported by about two-
`thirds of patients with well-controlled cancer pain
`and is generally managed by titration of the sched-
`uled dose of controlled release opioids to manage
`baseline pain, together with administration of im-
`mediate-release opioids as required. It is notewor-
`thy that the pattern of steady pain and break-
`through pain in our patients was dynamic despite
`relatively stable analgesic requirements. When we
`evaluated pain intensity on VAS and categorical
`scales by time of day, it was apparent that pain
`was at its worst for both treatments during daytime
`hours. This period of increased pain intensity also
`coincided with a rise in rescue analgesic use. In our
`study, the most likely cause of circadian worsening
`of pain appeared to be due to an increase of patient
`activity in the daytime (incident pain).
`At the start of this study, 45% of our patients had
`at least some neuropathic component to their pain.
`Although the efficacy of opioid analgesics in neuro-
`pathic pain syndromes remains a polarized area of
`clinical practice, data from well-controlled clinical tri-
`als indicate that pain of neuropathic origin can be at
`least partially responsive to opioid analgesics.48–50 The
`mean daily oxycodone dose given to patients with
`neuropathic pain was only slightly higher than that
`given to patients with nonneuropathic pain (136 { 36
`mg vs. 114 { 26 mg, respectively).
`
`Oxycodone for Cancer Pain/Hagen and Babul
`
`1435
`
`CONCLUSIONS
`We conclude that the recently available controlled-
`release oxycodone formulation, with its biphasic ab-
`sorption profile, is highly effective in the management
`of chronic severe cancer pain when administered ev-
`ery 12 hours. Our experience suggests that oxycodone
`can be successfully used for the treatment of severe
`cancer pain, adding to extensive experience of others
`in the use of oxycodone for moderate cancer pain.
`
`REFERENCES
`1.
`Po¨yho¨a R, Vainio A, Kalso E. A review of oxycodone’s clinical
`pharmacokinetics and pharmacodynamics. J Pain Symptom
`Manage 1993;8:63–7.
`2. Kalso E, Vainio A. Morphine and oxycodone hydrochloride
`in the management of cancer pain. Clin Pharmacol Ther
`1990;47:639–46.
`3. Beaver WT, Wallenstein SL, Rogers A, Houde RW. A