Pharmacodynamics of Orally Administered Sustained- release Hydromorphone in Humans

This report summarizes analgesic data obtained with an experimental pain model in human volunteers after oral administration of hydromorphone. Data were collected during five different study sessions at least 5 days apart, i.e. , after administration of a single dose of immediate-release hydromorphone (session 1) followed in randomized sequence by three different doses of hydromorphone OROS and placebo, respectively (sessions 2–5). Hydromorphone plasma concentrations are reported only in the context of the pharmacodynamic analysis. A detailed pharmacokinetic analysis including data from intravenous administration of hydromorphone will be the subject of a subsequent manuscript.

The study was approved by the Institutional Review Board of Stanford University. Twelve healthy subjects (six men and six women) with a mean age of 27 yr (range, 21–34 yr) and a mean body weight of 76 kg (range, 60–88 kg) for men and 64 kg (range, 59–70 kg) for women gave written informed consent. All subjects had a normal physical examination, electrocardiogram, routine laboratory profile, negative drug screen, and pregnancy test (women) before enrollment. No over-the-counter medication was allowed 48 h before each study day. Intake of prescription drugs or chronic medication was prohibited during and 14 days before the study except for the use of oral contraceptives. Before each study session subjects fasted overnight. On arrival at the study center, a catheter was inserted in one arm vein, and recording of vital signs was started (electrocardiogram, noninvasive arterial blood pressure, hemoglobin oxygen saturation, and respiratory rate).

Subjects received a single dose of 8 mg, open-label, immediate-release hydromorphone with 240 ml of water (Dilaudid, Knoll Pharmaceutical Company, Mount Olive, NJ). No other water intake was allowed from 1 h before dosing through 1 h after dosing. A meal was taken 3 h after drug intake.

Venous blood samples were obtained and vital signs recorded before and 15, 30, 45, 60, 90, 120, 180, 240, 300, and 360 min after drug intake. Experimental pain testing was performed in duplicate before and 30, 60, 120, 240, and 360 min after drug intake. Occurrence and severity of adverse events were recorded (mild = no limitation of usual activity, slight discomfort; moderate = some limitation of usual activity or significant discomfort; severe = inability to conduct usual activity or intolerable discomfort).

Using a double-blind design, subjects were randomly allocated to receive a single dose of 8, 16, and 32 mg hydromorphone OROS (Hydromorphone SR, Knoll Pharmaceutical Company), and placebo. All subjects received all treatments but on different occasions. Dose selection was based on developing a single daily dose formulation matching the doses of the immediate-release formulation (2, 4, and 8 mg) administered every 4–6 h.

At the beginning of each study session, subjects swallowed three tablets, all containing no drug or two containing no drug and one containing 8, 16, or 32 mg hydromorphone OROS. Water intake was restricted as outlined for study session 1. Meals were taken at 3, 9, 24, and 28 h after drug intake. Venous blood samples were obtained before and 1, 2, 3, 6, 9, 12, 15, 18, 21, 24, and 30 h after drug intake. Experimental pain testing was performed in duplicate before and 3, 6, 9, 12, and 30 h after drug intake. Vital signs were recorded hourly. Adverse events were recorded as outline for study session 1.

Seven milliliters of venous blood was drawn into heparinized glass tubes, centrifuged, frozen within 1 h, and stored at −20°C. Plasma and the internal standard tri-deuterated hydromorphone were injected onto a SCIEX API III-Plus LC-MS-MS (PE Biosystems, Foster City, CA). The assay was linear over a range of 0.05–10.0 ng/ml. The between-day coefficient of variation ranged from 1.4 % to 11.2%. The lower limit of quantification was 0.05 ng/ml.

A constant current device (Neurometer, Neurotron Inc., Baltimore, MD) with a maximum output of 20 mA and delivering 5 Hz sine wave pulses of 3-s duration was used to administer nociceptive electrical stimuli. An aluminum–gold ring electrode was attached to the surface of the skin at the right lateral upper arm as previously described. 13 

Before study participation, subjects were familiarized with the test procedure. During study sessions, subjects completed experimental pain test cycles of approximately 10-min duration at specified times. Standardized sentences emphasizing the test procedure were read to subjects before starting a test cycle. A test cycle consisted of two distinct series of electrical stimuli.

A first series was administered using an ascending staircase design. The intensity of the first stimulus varied randomly between 1.2 and 2.4 mA. The intensity of subsequent stimuli increased by 30% steps until a subject reported pain, and then by 15% steps until a subject indicated that he or she was unable to tolerate the next stronger stimulus. The interstimulus interval was 15 s. The first series allowed estimating the pain threshold (average of highest electrical current not evoking pain and lowest electrical current evoking pain) and the pain tolerance (current evoking most painful perception tolerated).

The second series was composed of 16 stimuli evenly spaced starting 30% less than the estimated pain threshold and ending at the estimated pain tolerance. Each stimulus was administered once in random sequence. Subjects rated the magnitude of pain evoked by each stimulus on a 100-mm visual analog scale (VAS) anchored by the words “no pain” and “most intense pain tolerable.” A scatter plot depicting the 16 stimulus intensities (mA) versus  pain (VAS) resulted for each test cycle (fig. 2, plot on the left).

To compare the stimulus intensity (mA) versus  pain (VAS) relation obtained at different times, parameters describing the relation were derived. A line was fitted to the data using the simple linear model,

where x is the electrical current, y is the VAS-score, a is the slope of the relation, and b is the x-intercept (fig. 2, middle graph). A maximum likelihood approach was used for parameter estimation. A power model was also explored but failed to improve the fit as judged by visual inspection and the likelihood ratio test.

The fitting procedure assumed that errors (observed − predicted value) were normally distributed around the predicted value. Fitting a line to a VAS-score between 1 and 99 maximized the likelihood of observed error, i.e. , minimized the distance parallel to the y-axis between the fitted line and observed data point (inset graph A of the middle graph of fig. 2). Fitting a line to a VAS score equal to 0 or 100 maximized the probability of observed error, i.e. , maximized the distance parallel to the y-axis between the fitted line and observed data point (inset graph B of the middle graph of fig. 2). The second fitting procedure accounted for the fact that VAS scores of 0 and 100 are censored data (0 indicates a true value of approximately 0 or lower but not higher, 100 indicates a true value of approximately 100 or higher but not lower).

One analgesic efficacy variable was the pain threshold (x-intercept of the linear model denoted “b”) and the pain tolerance calculated with aid of the linear model and by setting y = 100 (fig. 2, plot on the right).

The pain threshold and pain tolerance measured after drug administration was expressed as the percentage change from baseline (before drug). The baseline value represented 100%. Inspection of individual plots depicting plasma concentration versus  pain threshold and pain tolerance suggested counterclockwise hysteresis for data collected after immediate-release hydromorphone but not after hydromorphone OROS administration. Therefore, individual effect site concentrations were used for further exploration of the concentration–effect relation after administration of immediate-release hydromorphone, but plasma concentrations were used in case of hydromorphone OROS. Effect site concentration and equilibration rate constant were determined using a nonparametric approach and the equilibration half-life, i.e. , the time necessary to achieve 50% of the plasma concentration at the effect site was calculated using standard formulas. 14 

Data obtained after administration of immediate-release hydromorphone and hydromorphone OROS were pooled separately for further evaluation of a pharmacodynamic model describing the effect site and plasma concentration versus  effect relation, respectively. NONMEM 15was used to explore pharmacodynamic models nested in a power model of the general form,

where y is the analgesic efficacy variable, x is the plasma or effect site concentration, c is the y-intercept or the effect at baseline, d is the slope of the relation, and λ is a variable exponent. 15Model selection was based on the likelihood ratio test and visual assessment of the goodness of fit. A sigmoidal model was not explored because effects close to the maximum were not observed.

The relation between hydromorphone plasma concentration and heart rate, arterial blood pressure, hemoglobin oxygen saturation, and respiratory rate was explored as outlined for the analgesic efficacy variables.

Results of summary statistics are expressed as mean and SD for data passing the normality test (Kolmogorov-Smirnov) or alternatively as median and interquartile range. Parameters of linear equations are expressed as mean and corresponding 95% confidence interval.

To compare the magnitude of observed analgesic effect after administration of 8, 16, and 32 mg hydromorphone OROS and placebo, individual area under the curves (AUCs) depicting the time versus  the percentage change of the pain tolerance were calculated. Four AUCs resulted per subject, one for each treatment. A one-way repeated-measure analysis of variance (ANOVA) tested for significant differences among treatments. For treatments significantly different from placebo, a one-way repeated-measure ANOVA tested if analgesic effects measured at different times after drug administration differed significantly from baseline. The Student-Newman-Keuls test was used for post hoc  analysis. Outlined procedure rather than a two-way repeated-measure ANOVA was used because raw data did not pass the normality and equal variance test. AUCs were only determined for pain tolerance data because no significant relation existed between the hydromorphone plasma concentration and the pain threshold after hydromorphone OROS administration.

Areas under the curve depicting the electrical pain tolerance versus  time for sessions 2–5 were compared among each other to test for the presence of an order effect (ANOVA on ranks). During each session, an equal number of subjects received each treatment, i.e. , the total dose of hydromorphone administered during each session remained constant. An order effect should manifest itself by AUCs that are increasing or decreasing with subsequent study sessions.

The electrical pain tolerance measured before drug administration was compared among study sessions to test for the presence of changing baseline conditions (one-way repeated-measure ANOVA on ranks). The Student-Newman-Keuls test was used for post hoc  analysis.

To explore the fact that a significant relation between hydromorphone plasma concentration and pain tolerance but not pain threshold was detected, omega squared statistics were used. 16Omega squared (ω²) estimates the fraction of the overall variability of the effect measure that can be attributed to the fact that subjects received different treatments. It has been suggested that 0.01 < ω²< 0.06, 0.06 < ω²< 0.15, and ω²> 0.15 indicate a small, medium, and large treatment effect, respectively. 16 

To compare plasma concentration versus  time profiles observed after administration of immediate-release hydromorphone and hydromorphone OROS, the following parameters were determined: peak plasma concentration, time to achieve peak plasma concentration, first and last time plasma concentration was greater than 50% of measured peak concentration, and the length of time the plasma concentration remained at greater than 50% of the peak concentration. To compare analgesic effect versus  time profiles observed after administration of immediate-release hydromorphone and hydromorphone OROS, the following parameters were determined: peak pain tolerance, time to achieve peak pain tolerance, first and last time pain tolerance was greater than 50% of peak pain tolerance, and the length of time the pain tolerance remained at greater than 50% of the peak pain tolerance. Linear interpolation was used to determine first and last time 50% of a peak value was observed. A 50% level of peak plasma concentration and peak analgesic effect was used to obtain a normalized, i.e. , dose-independent measure quantifying how sustained the plasma concentration and the analgesic effect were after administration of the immediate or sustained-release formulation. The 50% level does not relate to the adequacy of the plasma concentration or measured analgesic effect, but to the performance of examined sustained-release formulation. Treatments were compared using one-way repeated-measure ANOVA on ranks and Student-Newman-Keuls post hoc  test.

One-way repeated-measure ANOVA on ranks and Dunnett’s post hoc  analysis tested for a significant difference in the incidence of side effects between active treatments and placebo. A P  value less than 0.05 was considered statistically significant.

All 12 subjects completed the investigation according to the protocol. Oral contraceptives taken by five women were the only ongoing drug therapy during the study.

The parameters of the linear model describing the stimulus intensity (mA) versus  pain (VAS) relation at baseline were 2.8 ± 1.3 mA for the x-intercept and 16.2 (interquartile range, 11.2–25.3) for the slope. Baseline pain tolerance did not change significantly over the course of the study.

Counterclockwise hysteresis was present between the hydromorphone plasma concentration and the pain threshold and the pain tolerance, respectively (fig. 3). The median equilibration half-life between plasma and the effect site was 18 min (1–42 min) and 38 min (8–66 min) for the pain threshold and the pain tolerance, respectively. The following significant linear relation was found between the hydromorphone effect site concentration (Ce [ng/ml]) and the percentage increase of the pain threshold (Pthr [%]) from baseline (corresponding to 100%),

Numbers in parentheses indicate corresponding 95% confidence intervals. Similarly, a significant linear relation was found between the hydromorphone effect site concentration (Ce) and the percentage increase of the pain tolerance (Ptol) from baseline,

The average analgesic efficacy of hydromorphone OROS did not change over the course of the different study sessions. There was no significant relation between the hydromorphone plasma concentration and the percentage increase of the pain threshold from baseline (fig. 4, upper graph). However, a significant linear relation existed between the hydromorphone plasma concentration (Cp) and the percentage increase of the pain tolerance (Ptol) from baseline (fig. 4, lower graph),

Numbers in parentheses indicate corresponding 95% confidence intervals. Pharmacodynamic models with a variable exponent did not improve the goodness of fit. Despite a lacking relation between the pain threshold and the hydromorphone plasma concentration, the percentage increase in pain threshold and pain tolerance correlated significantly (Pearson correlation coefficient = 0.53;P < 0.001).

Omega squared (ω²), i.e. , the fraction of the overall variability of the pain threshold and the pain tolerance that could be attributed to the fact that subjects received different treatments (three doses of sustained-release hydromorphone and placebo) was 0.07 and 0.15, respectively. In other words, the magnitude of the analgesic treatment effect was small to medium and medium to large when assessing the pain threshold and pain tolerance, respectively.

There was a significant relation between increasing hydromorphone plasma concentration and decreasing hemoglobin oxygen saturation. The lowest hemoglobin oxygen saturation measured was 95%. No significant relation was detected between the hydromorphone plasma concentration and heart rate, arterial blood pressure, or respiratory rate. The lowest respiratory rate was 8–10 breaths/min and was measured in one subject 0.5–2 h after administration of immediate-release hydromorphone. All other measurements were at least 12 breaths/min.

There was no significant relation between the hydromorphone plasma concentration and heart rate, arterial blood pressure, hemoglobin oxygen saturation (no supplemental oxygen), or respiratory rate. The lowest hemoglobin oxygen saturation measured on one occasion 22 h after administration of 32 mg sustained-release hydromorphone was 91%. All other measurements were at least 93%. The lowest respiratory rate was 8–10 breaths/min, measured in one subject 8–14 h after administration of 16 mg sustained-release hydromorphone. All other measurements were at least 12 breaths/min.

Table 1summarizes peak plasma concentration, the time to peak plasma concentration, the first and the last time 50% of peak concentration was present in plasma, and the time span that plasma concentration remained at greater than 50% of the peak concentration. Peak plasma concentration was significantly higher after administration of 8 mg immediate-release hydromorphone than after any dose of hydromorphone OROS. However, accounting for observed equilibration delay between plasma and effect site after administration of immediate-release hydromorphone, the peak effect site rather than the peak plasma concentration is relevant to relate peak concentration to peak effect. The peak effect site concentration after administration of immediate-release hydromorphone was 2.24 ± 1.19 ng/ml. This was similar to the peak concentration measured after 32 mg hydromorphone OROS but significantly higher than that measured after 8 and 16 mg hydromorphone OROS. Statistically, peak concentrations after different doses of hydromorphone OROS were dose proportional and differed significantly among each other. The time to peak plasma concentration, the first and last time 50% of peak concentration was present in plasma, and the time span that plasma concentration remained at greater than 50% of the peak concentration were all significantly shorter after immediate-release hydromorphone than after any dose of hydromorphone OROS. This was also true if considering the effect site rather than the plasma concentration in case of immediate-release hydromorphone. The time to peak effect site concentration was 2 h (1–2 h). The first and last time 50% of peak effect site concentration was present was 0.7 ± 0.4 h and 4.1 ± 1.6 h, respectively. The time span that effect site concentration remained at greater than 50% of the peak concentration was 3.4 ± 1.4 h. Different doses of hydromorphone OROS did not result in a different time to peak plasma concentration, a different first and last time 50% of the peak concentration was present in plasma, or a different time span plasma concentration was at greater than 50% of the peak concentration. Similarly, different doses of hydromorphone OROS did not result in significantly different intersubject variability of the plasma concentration. The average coefficient of variation of the plasma concentration was 43%, 43%, and 41% after administration of 8, 16, and 32 mg sustained-release hydromorphone, respectively. The bottom graphs of fig. 5illustrate the plasma concentration versus  time profile after administration of 8, 16, and 32 mg hydromorphone OROS, and the effect site concentration versus  time profile after administration of 8 mg immediate-release hydromorphone, respectively.

Table 2summarizes peak pain tolerance, the time to peak pain tolerance, the first and last time 50% of peak pain tolerance was achieved, and the time span that pain tolerance remained at greater than 50% of the peak pain tolerance. Because the AUC depicting time versus  pain tolerance was significantly different from placebo after administration of 16 and 32 mg, but not after administration of 8 mg hydromorphone OROS, only the 16- and 32-mg doses are discussed (fig. 6). Peak percentage increase in pain tolerance was not significantly different after administration of immediate-release hydromorphone or the two doses of hydromorphone OROS. However, compared with hydromorphone OROS administration, immediate-release hydromorphone resulted in a significantly shorter time to peak pain tolerance, first and last time 50% of peak pain tolerance was achieved, and time span that the pain tolerance remained at greater than 50% of the peak pain tolerance. Pain tolerance peaked approximately six times faster after administration of immediate-release hydromorphone. Comparable analgesic effects (pain tolerance at greater than 50% of the peak pain tolerance) were prolonged by a factor of approximately 4 after administration of hydromorphone OROS. The top graphs of figure 5depict the pain tolerance versus  time profile after administration of 8, 16, and 32 mg hydromorphone OROS, 8 mg immediate-release hydromorphone, and placebo, respectively.

Compared with baseline, a significant increase in pain tolerance was present 0.5, 1, 2, 4, and 6 h after 8 mg immediate-release hydromorphone. Pain tolerance was significantly increased from baseline at 6, 9, 12, and 30 h after 16 and 32 mg hydromorphone OROS, respectively.

Table 3summarizes the incidence and severity of various adverse events. Compared with placebo, the incidence of pruritus and nausea was significantly greater after administration of 32 mg hydromorphone OROS. No serious adverse events were observed during the study. All adverse events were of mild or moderate severity, and most resolved spontaneously. However, nausea and vomiting made intravenous administration of 4–8 mg ondansetron necessary on 3, 1, 5, and 7 occasions after administration of 8 mg immediate-release hydromorphone and 8, 16, and 32 mg hydromorphone OROS, respectively. Pruritus was treated with 25–50 mg diphenhydramine administered intravenously on two occasions each after administration of 16 and 32 mg hydromorphone OROS. Finally, headache was treated with 325–650 mg acetaminophen orally on 2, 1, 2, and 3 occasions after administration of 8 mg immediate-release hydromorphone and 8, 16, and 32 mg hydromorphone OROS, respectively. With one exception, acetaminophen was given after completion of all experimental pain tests.

A new sustained-release hydromorphone formulation using oral osmotic pump system technology has been developed for the treatment of chronic pain with a single daily dose. Hydromorphone OROS evoked a significant dose-dependent analgesic response beginning 6 h and lasting more than 24 h of its administration. The observed analgesic response was substantially longer than that observed after a similarly effective dose of immediate-release hydromorphone.

Oral Osmotic pump systems have been successfully used for the sustained release of a variety of drugs that are quickly eliminated from the body and therefore require dosing of the conventional tablet several times a day. Pharmacokinetics reported for hydromorphone OROS are consistent with previous reports of other compounds using oral osmotic pump systems. 10,11Plasma concentrations peaked significantly later after administration of hydromorphone OROS compared with conventional tablets, but were sustained over the 24-h interval after drug administration.

Several studies have reported equivalent or superior efficacy for compounds administered with oral osmotic pump systems if compared with corresponding conventional tablet. 10–12Experimental pain testing reported here suggests that a single daily dose of hydromorphone OROS provides sustained analgesia closely related to measured plasma concentration. Because previous studies have demonstrated similar analgesic efficacy for equivalent doses of various immediate and controlled-release opioids in the setting of clinical pain, it is conceivable that once an effective plasma concentration is achieved, a single daily dose of hydromorphone OROS will be as efficacious as multiple daily doses of the conventional tablet. 5,17Based on our data, a 6-h period lacking sufficient plasma concentrations to produce significant analgesia has to be expected after an initial dose of hydromorphone OROS. A possible solution to prevent a 6-h delay in the onset of analgesia may be to coadminister immediate-release hydromorphone with the first dose of hydromorphone OROS. However, with repetitive dosing of hydromorphone OROS, sustained plasma concentrations throughout the 24-h dosing interval may be expected.

Peak analgesic effects to suprathreshold stimuli lagged behind peak plasma concentrations by approximately 40 min after administration of immediate-release hydromorphone. In monkeys, maximum brain uptake of ¹¹C-hydromorphone was delayed by 10–15 min. 18A previous study in humans reported a delay between peak plasma hydromorphone concentration and peak analgesia of 10–20 min. 18The 40-min delay reported here must be interpreted as estimation because collection of analgesic data over time was relatively sparse. Nevertheless, the observed delay points to a relatively slow transit of hydromorphone from plasma to the effect site, i.e. , the central nervous system, and is similar to that described for morphine. 19,20This is in agreement with the fact that the polarity of hydromorphone and morphine is similar at physiologic pH. 21Generally, drug uptake into the central nervous system becomes slower with increased polarity of the compound. 22 

Administration of hydromorphone OROS resulted in a significant relation between plasma concentration and measured analgesic effect for the pain tolerance (noxious input causing maximum tolerable pain) but not the pain threshold (noxious input causing equally likely a painful or a nonpainful perception). Several studies using various experimental pain modalities and different classes of drugs support the view that measuring the pain tolerance is more suitable to detect analgesic effects than measuring the pain threshold. 13,23,24omega-squared statistics reported here and for a previous study implies that the signal (difference between treatments) to noise (variance unrelated to treatments) ratio is superior if measuring the pain tolerance. 13This may be explained, in part, by a superior test–retest reliability if measuring the pain tolerance rather than the pain threshold. This view is supported by our study revealing a smaller within-day (9 vs.  15%) and between-day (17 vs.  23%) coefficient of variation for measurements of the pain tolerance rather than the pain threshold at baseline.

The experimental pain model used in our study did not detect a significantly different analgesic efficacy after administration of 16 and 32 mg hydromorphone OROS. This raises the question how the precision and sensitivity of the pain model used compares with other models testing analgesic efficacy. Postoperative pain models have often been used to compare the analgesic efficacy of different doses of an orally administered opioid. However, despite the fact that these studies enrolled a larger number of patients than that enrolled in our study, they could not consistently demonstrate increased analgesic efficacy if doubling or even tripling the dose. 25–27This suggests that the precision and sensitivity of our experimental pain model is at least equivalent to that of a postoperative pain model.

The difficulty of differentiating the analgesic efficacy among various doses of an orally administered opioid may result from significant intersubject pharmacokinetic and pharmacodynamic variability, less than desirable precision inherent to any pain assay, and the fact that many doses studied may constitute the low or the high end of the dose versus  effect relation. The latter seems relevant because the dose versus  effect relation is S-shaped for most drugs. Therefore, the gain in effect is much smaller and harder to detect if a low rather than a midsized dose is doubled.

Currently, a sustained-release hydromorphone formulation different from hydromorphone OROS is commercially available. This formulation has a pharmacokinetic and pharmacodynamic profile that allows safe and effective pain control if administered every 12 h. 6,28However, the plasma concentration after administration of hydromorphone OROS was sustained approximately three times longer if compared with the alternative formulation. 28Hydromorphone OROS likely needs less frequent dosing than the alternative formulation.

There are a variety of other opioids including morphine, oxycodone, and tramadol that are available as sustained-release formulations. 17,29,30Most of these formulations allow safe and effective pain control if administered every 12 h. One formulation containing morphine was given once a day to patients with cancer pain and provided adequate pain control without the need for rescue medication in approximately 50% of the patients. 31This formulation had pronounced sustained-release characteristics, with plasma concentrations remaining above 50% of the peak concentration for approximately 12 h. 31By comparison, administration of hydromorphone OROS resulted in plasma concentrations that were similarly sustained for approximately 24 h. Contrasting with most sustained-release opioid formulations, a single daily dose of hydromorphone OROS may allow safe and adequate pain control and a decreased need for rescue medication.

The incidence of side effects after administration of hydromorphone OROS was dose-related. During clinical analgesic trials of opioids, a fraction of patients commonly withdrew because of disturbing opioid side effects. 5,6,32However, studies comparing sustained- and immediate-release formulations of various opioids or among different release formulations of the same opioid did not find significant differences in the severity and incidence of side effects. 5,6,32,33One study reported an increased therapeutic window in patients controlling the rate of a continuous intravenous infusion rather than the number of intravenous bolus injections of morphine. 34More data are needed to establish if the therapeutic window between long-term administration of immediate- and sustained-release opioids differs in a relevant way.

Compliance with taking a medication is inversely related to the dosing frequency, particularly if three or more daily doses are required. 8This was even found to be true in patients suffering from seizure disorders or chronic pain, conditions where noncompliance may result in immediate harm or increased suffering. 7,9Cancer pain patients on controlled release morphine were more compliant, experienced less pain, and rated their quality of life higher than patients on immediate-release tablets. 9 

The development of tolerance, i.e. , a need to increase the dose to maintain an effect, may be different if a drug is administered as an immediate- or a sustained-release formulation. 35,36One study suggested that more pronounced tolerance to the analgesic effects of morphine developed in patients receiving a continuous infusion rather than intermittent intramuscular bolus injections. 37However, another study reported that, although analgesic effects were preserved, nausea and constipation became less prominent in the course of a treatment with controlled-release morphine. 9Further investigation is needed to clarify if administration of immediate- and sustained-release opioids produce differential tolerance to analgesic and nonanalgesic effects.

Variation in gastrointestinal transit time could alter the release and absorption profile of hydromorphone OROS. No data regarding hydromorphone OROS are yet available. However, food intake and altered gastrointestinal transit time did not affect the release rate of radio-labeled markers from an oral osmotic pump system. 38The release and absorption rate of metoprolol and oxprenolol using oral osmotic pump system technology were approximately constant along the entire gastrointestinal tract. 39,40However, an inverse relation between the residual amount of drug recovered from excreted oral osmotic pump system and the gastrointestinal transit time has been reported. 41Reduced systemic drug uptake can result if the gastrointestinal transit time is considerably shorter than the duration of drug release from an oral osmotic pump system.

In summary, hydromorphone OROS may become a valuable addition to other sustained-release formulations containing opioids. The presented data obtained in human volunteers using experimental pain provide strong evidence that a single daily dose of hydromorphone OROS may be effective in the treatment of opioid responsive chronic pain. Clinical studies are needed to substantiate and expand the findings presented here.

The authors thank Lawrence J. Saidman, M.D. (Professor, Department of Anesthesia, Stanford University School of Medicine, Stanford, California), for helpful comments on the manuscript; and Frances A. Monroe, B.A. (Life Science Research Assistant II, Department of Anesthesia, Stanford University School of Medicine, Stanford, California), for graphical assistance.