Assessment of non-inferiority of OROS hydromorphone compared with sustained release (SR) oxycodone with regard to pain control by measuring the change from baseline in pain severity, using BPI item 6 "pain right now" score at week 24. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "pain right now".

PP population (all randomized subjects who took the study medication at least once, who had post-baseline efficacy data, and who were without major protocol violation)

Change from baseline to week 24 in BPI pain severity, pain at its worst (BPI item 3) assessed using the BPI questionnaire. Score values ranges from 0 (no pain) to 10 (pain as bad as you can imagine). Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "pain at its worst".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Assessment of non-inferiority of OROS hydromorphone compared with SR oxycodone with regard to pain control by measuring the change from baseline in pain severity, using BPI item 6 "pain right now" score at week 24. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "pain right now".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality was assessed using the Medical Outcomes Study (MOS) questionnaire at week 24, specifically the sleep subscale index I. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improved sleep quality.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline to week 24 in subject diary evening mean pain score "pain right now". Subjects rated the severity of "pain right now" on a 10 point numeric scale, with 0 being the least pain and 10 being the most pain. Negative change from baseline scores indicate improvement in subject diary evening mean pain score "pain right now".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline to week 24 in subject diary morning mean pain score "pain right now". Subjects rated the severity of "pain right now" on a 10 point numeric scale, with 0 being the least pain and 10 being the most pain. Negative change from baseline scores indicate improvement in subject diary evening mean pain score "pain right now".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Number of subjects with dose increase in study medication.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically "pain right now" (BPI item 6) at week 4. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "pain right now".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically "pain at its least" (BPI item 4) at week 4. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "pain at its least".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically "pain at its worst" (BPI item 3) at week 4. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "pain at its worst".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically "average pain" (BPI item 5) at week 4. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "average pain".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically pain relief (BPI item 8) at week 4. Scores could have ranged from 0 to 100, where 0 = no relief and 100 = complete relief. Positive change from baseline scores indicate improvement in pain relief.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in BPI pain severity was assessed using the BPI questionnaire (mean of BPI items 3 to 6) at week 4. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in pain severity.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically "pain at its least" (BPI item 4) at week 24. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "pain at its least".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically "average pain" (BPI item 5) at week 24. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative change from baseline scores indicate improvement in "average pain".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity was assessed using the BPI questionnaire, specifically pain relief (BPI item 8) at week 24. Scores could have ranged from 0 to 100, where 0 = no relief and 100 = complete relief. Positive change from baseline scores indicate improvement in pain relief.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change in pain severity was assessed using the BPI questionnaire, specifically average (mean) score of BPI items 3 to 6 (worst pain, least pain, average pain, and pain right now) at week 24. Scores could have ranged from 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine. Negative scores indicate improvement in pain severity.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in interference of pain was assessed using the BPI questionnaire, specifically BPI item 9a "pain interfered with general activity" at week 4. Scores could have ranged from 0 to 10, where 0 = does not interfere and 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with general activity".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9b "pain interfered with mood" at week 4. Scores could have ranged from 0 to 10, where 0 = does not interfere and 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with mood".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9c "pain interfered with walking ability" at week 4. Scores could have ranged from 0 to 10, where 0 = does not interfere and 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with walking ability".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9d "pain interfered with normal work" at week 4. Scores could have ranged from 0 to 10, where 0 = does not interfere and 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with normal work".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using BPI questionnaire, specifically BPI item 9e "pain interfered with relations with other people" at week 4. Scores could have ranged from 0 to 10, where 0 = does not interfere and 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with relations with other people".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using BPI questionnaire, specifically BPI item 9f "pain interfered with sleep" at week 4. Scores could have ranged from 0 to 10, where 0 = does not interfere and 10 - completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with sleep".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9g "pain interfered with enjoyment of life" at week 4. Scores could have ranged from 0 to 10, where 0 = does not interfere and 10 = interferes completely. Negative change from baseline scores indicate improvement in "pain interfered with enjoyment of life".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9a "pain interfered with general activity" at week 24. Scores could have ranged from 0 to 10, where 0 = does not interfere to 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with general activity".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9b "pain interfered with mood" at week 24. Scores could have ranged from 0 to 10, where 0 = does not interfere to 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with mood".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9c "pain interfered with walking ability" at week 24. Scores could have ranged from 0 to 10, where 0 = does not interfere to 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with walking ability".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9d "pain interfered with normal work" at week 24. Scores could have ranged from 0 to 10, where 0 = does not interfere to 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with normal work".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9e "pain interfered with relations with other people" at week 24. Scores could have ranged from 0 to 10, where 0 = does not interfere to 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with relations with other people".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9f "pain interfered with sleep" at week 24. Scores could have ranged from 0 to 10, where 0 = does not interfere to 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with sleep".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain interference was assessed using the BPI questionnaire, specifically BPI item 9g "pain interfered with enjoyment of life" at week 24. Scores could have ranged from 0 to 10, where 0 = does not interfere to 10 = completely interferes. Negative change from baseline scores indicate improvement in "pain interfered with enjoyment of life".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in pain severity, pain relief, and pain interference was assessed using the BPI questionnaire at week 52. BPI items 3 to 6, score range 0 to 10, where 0 = no pain and 10 = pain as bad as you can imagine; BPI items 9a to 9g, score range from 0 to 10, where 0 = does not interfere and 10 = completely interferes. Negative change from baseline scores indicate improvement in pain severity and pain interference. BPI item 8, score range from 0 to 100, where 0 = no relief and 100 = complete relief. Positive change from baseline scores indicate improvement in pain relief.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality was assessed using the sleep subscales of the MOS questionnaire, which consists of 12 items; MOS sleep scale index I (average of item 1, 3, 7, 8, 9, and 12) was assessed at week 4. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improvement in sleep quality.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality was assessed using the sleep subscales of the MOS questionnaire, which consists of 12 items. MOS index II (average of items 1, 3, 4, 5, 6, 7, 8, 9, and 12) was assessed at week 4. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improvement in sleep quality.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality was assessed using the sleep subscales of the MOS questionnaire, which consists of 12 items. MOS index II (average of items 1, 3, 4, 5, 6, 7, 8, 9, and 12) was assessed at week 24. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improvement in sleep quality.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality (sleep disturbance) was assessed using the MOS questionnaire at week 24. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improvement in sleep disturbance.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality (snoring) was assessed using the MOS questionnaire at week 24. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improvement in snoring.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality (sleep shortness of breath or headache) was assessed using the MOS questionnaire at week 24. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improvement in sleep shortness of breath or headache.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality (sleep adequacy) was assessed using the MOS questionnaire at week 24. Score range 0 to 100, where 0 = worst sleep quality and 100 = best sleep quality. Positive change from baseline scores indicate improvement in sleep adequacy.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality (sleep somnolence) was assessed using the MOS questionnaire at week 24. Score range 0 to 100, where 0 = best sleep quality and 100 = worst sleep quality. Negative change from baseline scores indicate improvement in sleep somnolence.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality (sleep quantity) was assessed using the MOS questionnaire at week 24. Score range 0 to 100, where 0 = worst sleep quality and 100 = best sleep quality. Positive change from baseline scores indicate improvement in sleep quantity.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Number of subjects indicating that they had optimal sleep was assessed based on the number of hours of sleep reported on the MOS questionnaire at week 24. Optimal sleep was defined as 7 to 8 hours sleep per night.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in sleep quality was assessed using the MOS questionnaire at week 52. Score range 0 to 100. For disturbance, snoring, shortness of breath or headache, and somnolence, 0 = best sleep quality and 100 = worst sleep quality; negative change from baseline scores indicate improvement in sleep quality for these measures. For adequacy and quantity, 0 = worst sleep quality and 100 = best sleep quality; positive change from baseline scores indicate improvement in sleep quality for these measures.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Number of subjects who experienced optimal sleep was assessed based on the number of hours of sleep reported on the MOS questionnaire at week 52. Optimal sleep was defined as 7-8 hours sleep per night.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline to week 24 in subject diary evening, morning and all day mean pain scores for pain right now, at its worst, at its least, and average. Subjects rated the severity of pain on a 10 point numeric scale, with 0 being the least pain and 10 being the most pain. Negative change from baseline scores indicate improvement in subject diary mean pain scores.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in subject diary mean pain score "pain at its worst" from morning to evening at weeks 4, 8, 12, 16, 20, and 24. Subjects rated the severity of "pain right now" on a 10 point numeric scale, with 0 being the least pain and 10 being the most pain. Negative change from baseline scores indicate improvement in subject diary mean pain score "pain at its worst".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

The number of subjects with dose increase in study medication was assessed at week 4.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

The number of subjects with dose increase in study medication was assessed at week 24.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the Short Form (SF)-36 QoL questionnaire, specifically the SF-36 bodily pain index. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in bodily pain.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 general health perceptions score at week 4. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in general health perceptions.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 health transition score at week 4. Scores could range from 0 to 100, with higher scores indicating a better QoL. Positive change from baseline scores indicate improvement in health transition.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 mental health score at week 4. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in mental health score.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 physical functioning score at week 4. Scores could range from 0 to 100, with high scores indicating a better QoL. Positive change from baseline scores indicate improvement in physical functioning.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 "role emotional" score at week 4. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in "role emotional".

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 role physical score at week 4. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in role physical.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 social functioning score at week 4. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in social functioning.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 vitality score at week 4. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in vitality.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 bodily pain index score at week 24. Score could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in bodily pain.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 general health perceptions at week 24. Scores could range from 0 to 100 with a higher score indicating a better QoL. Positive change from baseline scores indicate improvement in health perceptions.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 health transition score at week 24. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in health transition.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 mental health score at week 24. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline score indicates improvement in mental health.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 physical functioning score at week 24. Scores could range from 0 to 100, with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in physical functioning.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 "role emotional" score at week 24. Scores could range from 0 to 100 with a higher score indicating a better QoL. Positive change from baseline scores indicate improvement in "role emotional."

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 role physical score at week 24. Scores could range from 0 to 100 with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in role physical.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 social functioning score at week 24. Scores could range from 0 to 100 with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in social functioning.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire, specifically SF-36 vitality score at week 24. Scores could range from 0 to 100 with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in vitality.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Change from baseline in QoL was assessed using the SF-36 QoL questionnaire at week 52. Scores could range from 0 to 100 with a high score indicating a better QoL. Positive change from baseline scores indicate improvement in QoL.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Overall clinical efficacy was assessed by the Investigator using the following global ratings: very good, good, moderate, poor, or very poor, at weeks 4, 24, and 52.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Number of subjects with change in dose of study treatment was assessed at weeks 4, 24, and 52.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Number of subejcts with change in dose of study treatment was assessed and stratified by time on study, at least 4 weeks versus dropped out at highest dose before week 4, at weeks 4 and 24.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Number of drop-outs according to reasons for drop-out and due to inefficacy at maximal dosage was assessed at weeks 24 and 52.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Number of days with add-on pain medication during the first 24 weeks of the study was assessed at week 24.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Total amount of add-on pain medication (paracetamol) for the first 24 weeks was assessed at week 24.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Subjects filled out a questionnaire based on the mode and convenience of drug intake and could rate their responses as very convenient, convenient, neither convenient or inconvenient, inconvenient, and very inconvenient.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Resource utilization was defined as the number of additional visits including additional telephone visits during the treatment period. This was assessed at week 24.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

If non-inferiority of OROS hydromorphone was established, the daily dose of OROS hydromorphone and SR oxycodone that induced the same pain control was to be calculated (average dose used at week 24). Relative equi-analgesic dose was defined as mean dose/allowed maximum dose*100. Allowed maximum doses were 32mg OROS hydromorphone and 80mg SR oxycodone respectively.

PP population (all subjects who took the study medication at least once, who had post-baseline efficacy data, and who were without major protocol violation)

If non-inferiority of OROS hydromorphone was established, the daily dose of OROS hydromorphone and SR oxycodone that induced the same pain control was to be calculated (average dose used at week 24). Relative equi-analgesic dose was defined as mean dose/allowed maximum dose*100. Allowed maximum doses were 32mg OROS hydromorphone and 80mg SR oxycodone respectively.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)

Dose of OROS hydromorphone and SR oxycodone that induced the same pain control at steady state, defined as the mean dose from week 4 to week 24.

PP population (all subjects who took the study medication at least once, who had post-baseline efficacy data, and who were without major protocol violation)

Dose of OROS hydromorphone and SR oxycodone that induced the same pain control at steady state, defined as the mean dose from week 4 to week 24.

ITT population (all randomized subjects who took the study medication at least once, excluding subjects who had no post-baseline efficacy data)