Morphine produces immunosuppressive effects in nonhuman primates at the proteomic and cellular levels.

Morphine has long been known to have immunosuppressive properties in vivo, but the molecular and immunologic changes induced by it are incompletely understood. To explore how these changes interact with lentiviral infections in vivo, animals from two nonhuman primate species (African green monkeys and pigtailed macaques) were provided morphine and studied using a systems biology approach. Biological specimens were obtained from multiple sources (e.g. lymph node, colon, cerebrospinal fluid, and peripheral blood) before and after the administration of morphine (titrated up to a maximum dose of 5 mg/kg over a period of 20 days). Cellular immune, plasma cytokine, and proteome changes were measured and morphine-induced changes in these parameters were assessed on an interorgan, interindividual, and interspecies basis. In both species, morphine was associated with decreased levels of Ki-67(+) T-cell activation but with only minimal changes in overall T-cell counts, neutrophil counts, and NK cell counts. Although changes in T-cell maturation were observed, these varied across the various tissue/fluid compartments studied. Proteomic analysis revealed a morphine-induced suppressive effect in lymph nodes, with decreased abundance of protein mediators involved in the functional categories of energy metabolism, signaling, and maintenance of cell structure. These findings have direct relevance for understanding the impact of heroin addiction and the opioids used to treat addiction as well as on the potential interplay between opioid abuse and the immunological response to an infective agent.

Characteristics of distribution of morphine and metabolites in cerebrospinal fluid and plasma with chronic intrathecal morphine infusion in humans.

BACKGROUND: Despite widespread use of chronic intrathecal (IT) infusions of morphine, there is little systematic human work evaluating the steady state morphine concentrations or cerebrospinal (CSF) chemistry after long-term IT morphine delivery. We sought to address these issues in patients receiving chronic IT morphine infusion. METHODS: Pain patients with implanted catheters and pumps (range: 127 to 2165 days), receiving a stable dosing (>1 week) of IT morphine by infusion, were entered into the study. The following sequence was performed: (1) estimation of pain score; (2) radiograph localization of catheter tip; (3) percutaneous sampling of lumbar CSF at the L4 to 5 or L5-S1 space. CSF/plasma samples were assayed for chemistry, and morphine and its 3/6 glucuronide metabolites (M3G, M6G) by liquid chromatography mass spectrometry. RESULTS: Nineteen patients were enrolled. CSF samples were obtained from 16 subjects. Three patients were not included in the primary analysis because 1 catheter was epidural, 1 catheter was fractured, and 1 had a granuloma at the catheter tip. Of the 13 sampled patients, the range of daily doses, rates, and concentrations were 1.6 to 25 mg/d and 0.1 to 1 mL/d, 5 to 50 mg/mL, respectively. The principal observations were as follows: (i) morphine, M3G, and M6G were present in the CSF and plasma and showed a significant regression slope when plotted versus daily dose; (ii) in contrast, the regression slope of the group ratio morphine:M3G:M6G plotted versus daily dose in CSF or plasma was not different from zero; (iii) plotting "normalized" CSF analyte concentration (e.g., concentration at site/daily IT morphine dose) against the segmental distance of the sampling site from the catheter tip revealed a significant decline in concentration of morphine, but not of conjugates as a function of distance from the catheter tip; (iv) plotting CSF protein, glucose, and red and white cell counts versus daily morphine dose or morphine concentration at the sampling site revealed no significant regression; and (v) patients with a catheter failure or a granuloma showed reduced concentrations of morphine in their CSF. CONCLUSION: Chronic infusion of morphine shows high concentrations, which correlate with the infusion dose and the proximity of the sampling site to the infusion site with no effects on CSF chemistry.

Morphine reduces spinal c-fos expression dose-dependently during experimental laparotomy in pigs: a combined pharmacokinetic and surgical study.

The pharmacokinetics of intravenous morphine 2.5mg/kg (n=4) and 10mg/kg (n=4) in plasma and cerebrospinal fluid (CSF) of pigs was studied. Plasma half-life was 1.0+/-0.1h and the main metabolite was morphine-3-glucuronide, whereas morphine-6-glucuronide was negligible. CSF morphine concentration peaked after 20-30min (2.5mg/kg) and 60-120min (10mg/kg), and elimination half-life was 3.5+/-0.3h. Subsequently, the effect of morphine on surgery-induced spinal nociception in pigs subjected to unilateral laparotomy was evaluated by stereological quantification of the total number of Fos-like-immunoreactive (Fos-LI) spinal neurons of the dorsal horn. Surgery (n=4) induced 91,680+/-14,974 Fos-LI neurons ipsilaterally and morphine reduced this number to 45,771+/-8755 following the 2.5mg/kg dose (p<0.01; n=6) and 14,981+/-2327 following the 10mg/kg dose (p<0.001; n=6). These results indicate that morphine dose-dependently reduces the number of surgery-induced Fos-LI neurons in the spinal cord. As even a high dose of morphine does not reduce spinal c-fos expression to basal level, it may be appropriate to use other analgesics simultaneously with morphine during surgery.

Prediction of human CNS pharmacokinetics using a physiologically-based pharmacokinetic modeling approach.

Knowledge of drug concentration-time profiles at the central nervous system (CNS) target-site is critically important for rational development of CNS targeted drugs. Our aim was to translate a recently published comprehensive CNS physiologically-based pharmacokinetic (PBPK) model from rat to human, and to predict drug concentration-time profiles in multiple CNS compartments on available human data of four drugs (acetaminophen, oxycodone, morphine and phenytoin). Values of the system-specific parameters in the rat CNS PBPK model were replaced by corresponding human values. The contribution of active transporters for the four selected drugs was scaled based on differences in expression of the pertinent transporters in both species. Model predictions were evaluated with available pharmacokinetic (PK) data in human brain extracellular fluid and/or cerebrospinal fluid, obtained under physiologically healthy CNS conditions (acetaminophen, oxycodone, and morphine) and under pathophysiological CNS conditions where CNS physiology could be affected (acetaminophen, morphine and phenytoin). The human CNS PBPK model could successfully predict their concentration-time profiles in multiple human CNS compartments in physiological CNS conditions within a 1.6-fold error. Furthermore, the model allowed investigation of the potential underlying mechanisms that can explain differences in CNS PK associated with pathophysiological changes. This analysis supports the relevance of the developed model to allow more effective selection of CNS drug candidates since it enables the prediction of CNS target-site concentrations in humans, which are essential for drug development and patient treatment.

Comparative measurement of spinal CSF microdialysate concentrations and concomitant antinociception of morphine and morphine-6beta-glucuronide in rats.

Morphine-6beta-glucuronide (M6G) is well known as a potent active metabolite in humans. To clarify concentration-antinociceptive effect relationships for morphine and M6G, we evaluated comparatively the pharmacokinetics and antinociceptive effects of morphine and M6G. The spinal CSF concentration and antinociception were simultaneously measured by using the combination of a microdialysis method and the formalin test in conscious rats after the s.c. administration of morphine (0.3-3 mg/kg) and M6G (0.1-3 mg/kg). The plasma concentration of M6G after s.c. administration was higher than that of morphine, as shown by the 2.1 times greater value of area under the concentration-time curve (AUC(plasma)). The spinal CSF concentrations of morphine and M6G increased dose-dependently. The AUC(CSF) of M6G was 1.6-1.8 times higher than that of morphine at each dose. Administration of morphine and M6G dose-dependently suppressed the flinching behavior induced by formalin injection. The ED(50) values for M6G were 3 times lower than those of morphine, although the spinal CSF concentration versus antinociceptive effect curves of morphine and M6G were very similar, with similar EC(50) values. These results suggest that the antinociceptive potencies of morphine and M6G, evaluated by simultaneous measurements of spinal CSF drug concentration and antinociception, are equivalent. Simultaneous measurement of spinal CSF concentration and antinociception by using microdialysis should be useful for elucidating the relationship between pharmacokinetics and pharmacodynamics of various opioids.

Effects of morphine and endomorphins on the polysynaptic reflex in the isolated rat spinal cord.

At the spinal level, mu-opioids exert their actions on nociceptive primary afferent neurons both pre- and postsynaptically. In the present study, we used an in vitro isolated neonatal rat (11-15 days old) spinal cord preparation to examine the effects of morphine and the endogenous mu-opioid ligands endomorphin-1 (EM-1) and endomorphin-2 (EM-2) on the polysynaptic reflex (PSR) of dorsal root-ventral root (DR-VR) reflex. The actions of mu-opioids on spinal nociception were investigated by quantification of the firing frequency and the mean amplitude of the PSR evoked by stimuli with 20 x threshold intensity. EM-1 decreased the mean amplitude of PSR, whereas EM-2 and morphine decreased the firing frequency. The pattern of the effects elicited by morphine was the same as that for EM-2, except at high concentration. Naloxonazine, a selective mu(1) opioid receptor antagonist, had no significant effect on PSR by itself, but blocked the inhibition of PSR firing frequency or amplitude induced by EM-1, -2 and morphine. This may suggest that EM-1, EM-2 and morphine modulate spinal nociception differently and act mainly at the mu(1)-opioid receptors. Although they all act via mu(1)-opioid receptors, their different effects on the PSR may suggest the existence of different subtypes of the mu(1)-opioid receptor. The present data is also consistent with a further hypothesis, namely, that morphine and EM-2 activate a subtype of mu(1)-opioid receptor presynaptically, while EM-1 acts mainly through another subtype postsynaptically. However, since other reports indicate that EM-2, but not EM-1, could stimulate the release of enkephalins or dynorphin, presynaptic delta and kappa receptors may be also involved indirectly in the different regulation by mu-opioids at the spinal level.

Disposition of morphine in plasma and cerebrospinal fluid varies during neonatal development in pigs.

The pharmacological effects of morphine are mediated via the central nervous system (CNS) but its clearance from the CNS in neonates has not been investigated. We have proposed that neonatal development of the blood-brain barrier affected CNS clearance mechanisms and CNS exposure to morphine. Male piglets (n = 5) aged one, three and six weeks were given morphine sulfate (0.5 mg kg(-1), i.v.). Serial blood and cerebrospinal fluid (CSF) samples were withdrawn over 360 min after morphine administration. Morphine concentration was measured by radioimmunoassay. A three-compartment model was fitted to individual data. Estimated parameters were reported as median and range. The peak morphine concentrations in plasma were not significantly different in the one-, three- or six-week-old piglets. Plasma clearance at one week (4.5, 3.8-8.6 mL min(-1) kg(-1)) was significantly lower than at three weeks (30.0, 19.1- 39.0 mL min(-1) kg(-1)) and six weeks (37.0, 29.7-82.8 mL min(-1) kg(-1)). The peak morphine concentration in CSF at one week (59.84, 31-67 ng mL(-1)) was higher than at three weeks (18.8, 17.7-25 ng mL(-1)) and six weeks (24.51, 16.5-84 ng mL(-1)), while CSF clearance was lower at one week (1.0, 0.18-9 mL min(-1) kg(-1)) compared with three weeks (6.2, 2.3-9.3 mL min(-1) kg(-1)) and six weeks (3.95, 1.3-85.7 mL min(-1) kg(-1)). Apparent plasma:CSF transfer ratio at one week was greater than at three and six weeks. The reduced plasma and CSF morphine clearance in early infancy resulted in elevated systemic and central morphine exposure in neonatal pigs.

Spinal fluid kinetics of morphine and heroin.

Nine patients undergoing cardiopulmonary bypass surgery were given either 2 mg diamorphine or 2.5 mg morphine by intrathecal injection. Spinal fluid (sf) samples were collected over 25 min and drug concentrations measured by HPLC. Concentrations in sf were about 4000 times as great as after 1 mg/kg IV morphine. The kinetic properties of morphine and heroin in sf differed; diamorphine was removed from sf much more rapidly than morphine. Lipophilic opiates may be safer for intrathecal use because of the shorter life of substantial drug concentrations in the mobile sf phase.

Morphine pharmacokinetics and pain assessment after intracerebroventricular administration in patients with terminal cancer.

Morphine pharmacokinetics and pain relief were evaluated after intracerebroventricular administration of morphine (0.4 +/- 0.11 mg) in seven patients with cancer suffering from intractable pain. Ventricular cerebrospinal fluid (CSF), lumbar CSF, and plasma morphine concentrations were analyzed by a specific morphine radioimmunoassay. A two-compartment model was sufficient to describe the kinetics of morphine in ventricular CSF. Morphine diffuses to the lumbar level, and the mean maximum concentration was 192 +/- 105 ng/ml at 4.5 +/- 1.3 hours. Ventricular and lumbar CSF morphine kinetics showed a similar decline during the elimination phase, with terminal half-lives of 3.8 +/- 0.6 hours and 4.2 +/- 1.6 hours, respectively. Pain relief was evaluated by a visual analog scale: the test showed a rapid onset of analgesia (less than 10 minutes). Analgesic effectiveness reached a maximum between 6 and 10 hours. The relationship between pharmacologic effect and morphine concentrations in ventricular CSF resulted in an anticlockwise hysteresis curve. The presence of morphine in lumbar CSF suggested an additive spinal action of morphine, which probably plays a role in the duration of analgesia.

Epidural and intrathecal opiates: cerebrospinal fluid and plasma profiles in patients with chronic cancer pain.

We studied the cerebrospinal fluid (CSF) and plasma concentration-time profiles of morphine, methadone, and beta-endorphin after lumbar epidural or intrathecal injection in 17 patients with cancer. After epidural injection, all three drugs reached peak levels in lumbar CSF within 34 minutes that were 50 to 1300 times higher than free drug concentrations in plasma. The rate of decline of CSF levels correlated with drug lipid solubility (methadone [t1/2 = 73 minutes] greater than morphine [126 minutes] greater than beta-endorphin [317 minutes]). Plasma levels were comparable with those after intragluteal injection of the same dose. In four patients given intrathecal morphine or methadone, CSF at the C1-2 level contained high levels of morphine as early as 1 hour after injection, but levels of methadone were lower or undetectable. Three of 17 patients reported improved analgesia initially, but none were improved at 2 weeks after chronic therapy. We conclude that analgesia induced by intrathecal or epidural morphine injections is caused by drug acting at both spinal and supraspinal sites. The use of spinal opiates such as morphine is of limited value in patients whose pain is not adequately managed by high systemic doses of morphine-like drugs.

Chronic morphine therapy for cancer pain: plasma and cerebrospinal fluid morphine and morphine-6-glucuronide concentrations.

Morphine-6-glucuronide (M-6-G) is an active metabolite that may contribute to the clinical effects produced by systemic administration of morphine. To help clarify the extent to which M-6-G may cross the blood-brain barrier and exert effects, we employed high-performance liquid chromatography with electrochemical detection to measure the concentrations of M-6-G and morphine in the plasma and either ventricular (three patients) or lumbar (eight patients) CSF of cancer patients receiving chronic morphine therapy. The mean ratio of morphine in ventricular CSF:morphine in plasma was 0.71; the same ratio for M-6-G was only 0.077. The average molar ratio of M-6-G: morphine in ventricular CSF was 0.207, and the average molar ratio in plasma was 1.89. Although sampling problems render the lumbar CSF results less reliable, they were very similar. Thus, plasma contained approximately twice as much M-6-G as morphine, whereas CSF contained only one-fifth to one-third as much. These data confirm that M-6-G in plasma is distributed into CSF, but to a far lesser extent than morphine. They help explain animal data demonstrating much higher potency of M-6-G on administration into CSF than systemic administration and indicate that the degree to which M-6-G contributes to morphine effects in humans remains an unresolved question.

Cerebrospinal fluid and plasma pharmacokinetics of morphine infusions in pediatric cancer patients and rhesus monkeys.

The cerebrospinal fluid (CSF) and plasma pharmacokinetics of morphine administered as a continuous infusion were studied in pediatric cancer patients and in monkeys with implanted Ommaya reservoirs. In monkeys administered a constant infusion of 0.15 mg morphine sulfate/kg/h, morphine steady-state plasma and CSF concentrations were 84.4 +/- 20.0 ng/ml and 25.3 +/- 4.9 ng/ml, respectively, for a CSF:plasma ratio of 0.30 +/- 0.05. For comparison, the monkeys also received morphine as an intravenous bolus at a dose of 0.45 mg morphine sulfate/kg. The CSF:plasma area under the concentration-time curve (AUC) ratio was 0.40 +/- 0.07, similar to that seen with continuous infusion. Morphine pharmacokinetics were also studied in cancer patients administered long-term infusions of morphine sulfate over a wide dosage range (0.04-31 mg/kg/h). The steady-state plasma concentration of morphine was linearly related to the infusion rate although variability was noted. The average clearance value was 23 ml/min/kg which is at the upper end of the estimates reported for morphine clearance using bolus administration. No evidence for morphine accumulation using long-term administration was observed. A limited number of CSF samples obtained by lumbar puncture showed comparable CSF and plasma concentrations of unbound morphine assuming morphine is approximately 30% bound in human plasma.

Morphine, morphine-6-glucuronide and morphine-3-glucuronide concentrations in plasma and cerebrospinal fluid during long-term high-dose intrathecal morphine administration.

The kinetics of morphine, morphine-6-glucuronide (M6G), and morphine-3-glucuronide (M3G) were studied in a 56-year-old female with lung cancer. Long-term treatment with morphine intrathecally 28 mg every 6 h was used for pain control. Cerebrospinal fluid (CSF) concentrations of morphine were high with an elimination half-life of 2.1 h. The plasma/CSF ratios for M6G and M3G were 1:0.8 and 4:1, respectively, suggesting that M6G penetrates the blood-brain barrier more easily than M3G. The low CSF concentrations indicate that M6G played hardly any major analgetic role in the patient studied.

CSF and blood pharmacokinetics of hydromorphone and morphine following lumbar epidural administration.

Sixteen consenting patients scheduled for elective thoracotomy were enrolled into a randomized trial of epidural morphine and hydromorphone. Each patient had a lumbar epidural catheter placed preoperatively for the purpose of post-thoracotomy analgesia. Shortly before the end of the operative procedure each patient received 5 mg of morphine and 0.75 mg of hydromorphone via the epidural catheter. Blood was sampled at regular intervals following the opiate administration and patients were randomized to 1 of 7 cervical CSF sampling times. Blood and CSF samples were assayed for morphine and hydromorphone concentration to determine blood and CSF pharmacokinetic profiles. A maximum blood morphine concentration of 60 +/- 25 ng/ml (mean +/- S.D.) was obtained at 11 +/- 6 min (mean +/- S.D.). The blood hydromorphone peak of 14 +/- 13 ng/ml (mean +/- S.D.) occurred 8 +/- 6 min (mean +/- S.D.). The mean peak CSF opioid concentrations of 1581 ng/ml for morphine and 309 ng/ml for hydromorphone occurred 60 min after epidural administration. The blood and CSF pharmacokinetic profiles for morphine and hydromorphone are presented. These profiles are similar for the two drugs after lumbar epidural administration.

CSF morphine levels after lumbar intrathecal administration of isobaric and hyperbaric solutions for cancer pain.

The objectives of this study were to compare the pharmacokinetic properties and the duration of analgesia following intrathecal administration (L5-S1) of 2 mg morphine in 2 forms: (1) an isobaric (NaCl 0.9%) and (2) a hyperbaric solution (7% dextrose). The study was carried out on 5 cancer patients with severe, intractable pain in the lower half of the body. Samples of CSF were collected at the level of the 10th thoracic vertebra at regular intervals for 15 h after administration. Morphine concentrations were determined by HPLC. The pharmacokinetic properties of the solutions (I and II) were quite different. Peak levels (I) were reached in 5-15 min (30 and 60 micrograms/ml); they then fell rapidly during the 1st hour (7 and 11 micrograms/ml) with an elimination half-life of 10 and 15 min, followed by a change in slope (elimination half-life of 108 and 140 min). Peak levels (II) were reached in 4-5 h (0.8-3.3 micrograms/ml); they then fell progressively according to a single exponential function (elimination half-life: 144-246 min). The duration of analgesia for a dose of 2 mg was 30 h for solution 2 and 24 h for solution 1. The hyperbaric solution, which produced the same degree of analgesia as the isobaric solution, limited the cephalad diffusion of morphine and reduced or abolished the central depressant effects of the drug.

CSF and plasma morphine concentrations in cancer patients during chronic epidural morphine therapy and its relation to pain relief.

Seventeen patients with advanced cancer pain, treated with chronic epidural morphine, were studied. Minimum plasma and CSF morphine concentrations (CSSmin) were determined at pharmacokinetic steady state. A linear relationship was found between epidural morphine dose and concentrations obtained in plasma (r = 0.92) as well as CSF (r = 0.90). The line for best fit was much steeper for CSF than for plasma. The CSF/plasma concentration gradient of morphine at CSSmin was 132 +/- 31 (mean +/- S.E.M.). Large interindividual variations of morphine concentrations in CSF were found. It is suggested that the variations are due to substantial differences in transdural morphine diffusion between individuals. No correlation was found between pain relief, evaluated with a visual analog scale, and CSF morphine concentrations at pharmacokinetic steady state, when calculated in 9 patients. Mean duration of treatment was 104 days (range 14-366) and the daily dose was increased from 18 +/- 2 to 87 +/- 31 mg/day (mean +/- S.E.M.). A total of 39 epidural catheters were inserted in 14 patients. The catheters were patent for 2-223 days with a mean of 38 days. When re-examined later during treatment, 2 out of 8 patients demonstrated decreased CSF morphine concentrations in spite of increased doses given. One patient with extremely high dose demand is reported on separately and data supporting the concept of a combined spinal and systemic brain morphine effect in such cases are presented. Side effects were not a major problem but the possibility of infectious complications should be considered during chronic epidural morphine therapy.

Pain characterization in cancer patients and the analgetic response to epidural morphine.

In 48 patients with pain related to malignancy, a pain characterization was performed during oral opioid therapy. After an optimal epidural morphine regimen had been established, the alteration in pain relief was evaluated by means of a visual analogue scale. The CSF and plasma morphine concentrations at minimum steady state were then analysed in 28 patients and related to the degree of pain relief. The efficacy of the spinal treatment ranked in the following order: somatic greater than visceral greater than radiating = 0, but the difference was only significant between the somatic and radiating pain groups. There was a tendency for continuous pain to be better relieved than intermittent pain. No correlations were found between the CSF or plasma morphine concentrations and the degree of pain relief, suggesting that not all pain impulses are modulated in a dose-dependent manner by morphine at the spinal level. Pain characterization may be instrumental in providing an optimal spinal opioid analgesia in malignancy. Moreover, there is a need for better defined diagnostic criteria for clinical pain characterization.

Neuropathologic lesions and CSF morphine concentrations during chronic continuous intraspinal morphine infusion. A clinical and post-mortem study.

Seven patients with chronic intractable pain due to cancer were given chronic intraspinal narcotic administration (CINA) and subsequently underwent post-mortem examination. All deaths were unrelated to CINA. Two of these patients were found to have clinically unsuspected posterior column degeneration. Both patients had had epidural catheters placed, and one had received prior radiotherapy to ports which included parts of the spinal cord. In retrospect, it is impossible to ascertain whether the degeneration occurred before or after infusion of morphine began. Review of the potential causes for posterior column degeneration suggests that neuropathy associated with malignant disease is more likely the cause of the degeneration rather than intraspinal infusion of morphine. However, continued vigilance at autopsy is recommended. In addition, utilizing a new method for measuring cerebrospinal fluid (CSF) concentrations of morphine via high-pressure liquid chromatography, CSF morphine levels at steady state were measured in 5 patients. These levels were much lower than peak levels previously reported following bolus intraspinal administration. The ability of these measurements to contribute to knowledge of efficacy, toxicity, lumbar-cisternal concentration gradients, and differentiation of tolerance from drug delivery problems is discussed.

Epidural mass associated with lack of efficacy of epidural morphine and undetectable CSF morphine concentrations.

A case is presented in which epidurally administered morphine failed to provide the expected extent and duration of pain relief. The patient had severe back pain in the upper lumbar and low thoracic regions and epigastric pain which was presumed to be caused by widespread metastatic deposits in the vertebral column from a prostatic carcinoma. The lack of clinical effect of epidural morphine was shown to be related to the unexpected presence of an epidural mass which was located below the dermatomal region for the patient's pain. Myelography and a CT scan indicated that the mass effectively compressed the subarachnoid space, thereby preventing the rostral spread of morphine in the cerebrospinal fluid (CSF) following lumbar epidural administration, resulting in inadequate pain relief.

Cephalad migration of morphine in CSF following lumbar epidural administration in patients with cancer pain.

This study examines the cephalad migration of morphine in CSF following lumbar epidural administration in cancer patients with pain. Fourteen cancer patients were administered 10 mg of morphine in 10 ml of normal saline via an epidural catheter inserted in the lumbar region (usually L2.3) and attached to a subcutaneously implanted portal for drug administration. There was a rapid vascular uptake of morphine from the epidural space with a mean (+/- S.D.) peak blood concentration of 110 +/- 32 ng/ml (range 76-182 ng/ml and the mean time associated with this peak blood concentration was 5.1 +/- 2.3 min (range 2-10 min). A cervical CSF sample was obtained from the C7-T1 interspace in each patient at one of the following times from the completion of the epidural morphine dose: 10, 30, 60, 120, 180, 240 or 360 min. There was a delay of at least 60 min before morphine was detected in significant concentrations (approximately 300 ng/ml) in the cervical CSF samples and peak CSF concentrations occurred after approximately 3 h. The results of this study are compatable with the hypothesis that the delayed onset of respiratory depression sometimes observed following epidural morphine in opioid naive patients results from significant amounts of morphine reaching the respiratory centre in the brain-stem as a consequence of passive CSF flow in a rostral direction from the lumbar region.