465
CLINICAL PRACTICE Comparison of continuous subcutaneous and intravenous hydromorphone infusions for management of cancer pain
To compare the safety and efficacy of subcutaneous and intravenous infusion of opioid analgesics, a randomised, double-blind, crossover trial was carried out in inpatients. 15 patients with severe cancer pain received two 48 h infusions of hydromorphone one subcutaneously and one intravenously in randomly allocated order. The study was made double-blind by the use of two infusion pumps throughout; during the active subcutaneous infusion the intravenous pump delivered saline and vice versa. Serial measurements of pain intensity, pain relief, mood, and sedation by means of visual analogue scales showed no clinically or statistically significant difference between the two infusion routes. Sideeffects were slight, and the mean number of morphine injections for breakthrough pain did not differ significantly between the routes (4·8 [SD 4·5] for intravenous vs 5·3 [5·6] for subcutaneous). Plasma hydromorphone concentrations measured at 24 h and 48 h of infusion showed stable steady-state pharmacokinetics; the mean bioavailability from subcutaneous infusion was 78% of that with intravenous infusion. Because of the simplicity, technical advantages, and cost-effectiveness of continuous subcutaneous opioid infusion into the chest wall or trunk, intravenous opioid infusion for the management of severe cancer pain should be abandoned. Introduction Continuous intravenous infusion of narcotic analgesics has been widely used in the management of patients with
pain caused by advanced cancer.l-4 It provides stable blood levels of the infused drug, which can be titrated to the needs of the individual patient so that peak level sedation and trough level breakthrough pain, sometimes associated with intermittent dosing regimens, can be avoided. The other major advantage of continuous parenteral infusion is that it bypasses the oral route-an obvious benefit for patients with intractable nausea and vomiting or non-functional gastrointestinal tracts. Continuous subcutaneous infusion of opioid drugs has been reported to be safe and efficacious5,6
and even the treatment of choice for parenteral management of cancer pain.7 This development was made possible by the availability of portable infusion pumps which allow continuous subcutaneous infusion into the chest wall or abdomen in an ambulatory patient. Continuous subcutaneous infusion therefore appears to carry all the advantages of continuous intravenous infusion with the added benefits of greater mobility, management on an outpatient basis, and avoidance of the need for intravenous access. However, the hypothesis that continuous subcutaneous infusion is as effective as continuous intravenous infusion in the management of advanced cancer pain needs to be examined critically, since it means that the drug has to pass through a subcutaneous compartment which might change the pharmacokinetics of the opioid infusion. By means of a randomised, double-blind, crossover design, we compared the pharmacodynamics and pharmacokinetics of continuous subcutaneous and intravenous hydromorphone infusions in patients with chronic cancer pain.
Patients and methods Between April, 1988, and April, 1990, 53 inpatients with poorly controlled cancer pain were started on continuous subcutaneous opioid infusions. Most of these patients were experiencing substantial side-effects during the regular administration of oral or rectal opioids and adjuvant analgesics. 10 patients were excluded from the trial because of clinical evidence of brain metastases or metabolic encephalopathy, and 23 refused to take part. The remaining 20 patients (10 male, 10 female) gave informed consent to take part in the study. Their mean age was 61 ’9 (11 -8) years and the mean and median weights were 61-0 (SD 18-3) and 53-6 (IQR 38-103) kg. 5 had breast cancer, 5 lung cancer, 4 cancer of the gastrointestinal tract, 2 cancer of the urogenital tract, 1 each melanoma and soft tissue sarcoma, and in 2 the primary tumour was unknown. The principal causes of pain were bone metastases in 8, soft tissue infiltration of chest wall or retroperitoneum in 8, brachial plexopathy in 2, and subacute bowel obstruction in 2. The Health Sciences Standing Committee on Human Research at the University of Western Ontario approved this double-blind, ADDRESS: Departments of Clinical Neurological Sciences and Oncology (D. E. Moulin, MD, N. Murray-Parsons, RN, A. I. Bouquillon, MSc), and Medicine (Clinical Pharmacology) (J. H. Kreeft, MD), University of Western Ontario, London, Ontario, Canada. Correspondence to Dr D. E. Moulin, Victoria Hospital, 375 South Street, London, Ontario, N6A 4G5, Canada.
466
study with hydromorphone hydrochloride (’Dilaudid’; Pharmaceuticals, Ontario). We used a portable programmable infusion pump that is accurate to within 10% of the set rate (Pharmacia CADD-PCA; Montreal). Patients were initially
crossover
Knoll
randomised
to
continuous
subcutaneous
or
intravenous
hydromorphone infusions. Before randomisation, dose titration to provide optimum analgesia with minimum side-effects was accomplished by means of subcutaneous infusions of morphine in 18 patients and hydromorphone in 2. Patients whose dose was established on morphine were switched to hydromorphone at one-fifth the morphine rate based on the five to six fold greater analgesic potency of hydromorphone." The double-blind nature of the study was made possible by use of the "double dummy" technique which uses simultaneous matching placebos. Each patient carried two infusion pumps (each weighing 475 g); during active subcutaneous infusion the intravenous pump delivered saline and vice versa. In randomised fashion each patient received
a
48 h infusion of either intravenous
or
subcutaneous
hydromorphone and then, after a 24 h wash-out with morphine, a 48 h infusion at the same dose by the other route. Fresh subcutaneous infusion sites were started on the anterior chest wall or abdomen at the beginning of each 48 h infusion. Morphine (subcutaneous injections) was used to treat breakthrough pain during hydromorphone infusions to avoid interference in the pharmacokinetic analysis of hydromorphone. Morphine was offered every 3 h at 50% of the equianalgesic dose8 of a 3 h infusion of hydromorphone. All co-analgesics and adjuvant analgesics were maintained unchanged throughout the study. While resting in bed, the patient was asked to estimate pain intensity, pain relief, general feeling (mood), and sedation by means of visual analogue scales. Each scale consisted of a 10 cm line on which the patient marked the point between extremes of effect. These assessments were made before each infusion (during treatment with oral or rectal opioids) and after 4 h, 8 h, 24 h, 32 h, and 48 h of each hydromorphone infusion. The visual analogue scales for pain intensity, pain relief, and mood have been validated in analgesic studies of cancer pain patients.9 Side-effects other than sedation were also monitored.1O In the arm opposite that bearing the intravenous infusion, another intravenous cannula connected to a heparin lock was inserted so that blood samples could be obtained at 24 h and 48 h during each hydromorphone infusion. By means of a technique modified from Svensson et al,l1 plasma hydromorphone concentration was determined by high-performance liquid chromatography with electrochemical detection. With 1 ml plasma volumes the lower limit of sensitivity of this assay was 2-5 ng/ml. Morphine or hydromorphone metabolites do not interfere with this assay. At presumed steady state, plasma concentration averaged for each hour between 24 h and 48 h (Css) was calculated from the area under the plasma concentration curve by means of trapezoidal summation. Because the infusion rates varied among patients, we had to normalise the data to drug concentrationfmg.h -1 for pharmacokinetic comparisons of the subcutaneous and intravenous infusion groups. This procedure makes an assumption of first order or linear kinetics for hydromorphone elimination, which we felt was valid for two reasons. Morphine is subject to first order elimination12 and morphine and hydromorphone are very similar opioid drugs (hydromorphone is a hydrogenated ketone of morphine) which are metabolised primarily by hepatic glucuronidation.13 Secondly, our findings of a significant linear correlation between intravenous hydromorphone infusion rates and average Css between 24 h and 48 h (r=0-85, p
CLINICAL PRACTICE Comparison of continuous subcutaneous and intravenous hydromorphone infusions for management of cancer pain
To compare the safety and efficacy of subcutaneous and intravenous infusion of opioid analgesics, a randomised, double-blind, crossover trial was carried out in inpatients. 15 patients with severe cancer pain received two 48 h infusions of hydromorphone one subcutaneously and one intravenously in randomly allocated order. The study was made double-blind by the use of two infusion pumps throughout; during the active subcutaneous infusion the intravenous pump delivered saline and vice versa. Serial measurements of pain intensity, pain relief, mood, and sedation by means of visual analogue scales showed no clinically or statistically significant difference between the two infusion routes. Sideeffects were slight, and the mean number of morphine injections for breakthrough pain did not differ significantly between the routes (4·8 [SD 4·5] for intravenous vs 5·3 [5·6] for subcutaneous). Plasma hydromorphone concentrations measured at 24 h and 48 h of infusion showed stable steady-state pharmacokinetics; the mean bioavailability from subcutaneous infusion was 78% of that with intravenous infusion. Because of the simplicity, technical advantages, and cost-effectiveness of continuous subcutaneous opioid infusion into the chest wall or trunk, intravenous opioid infusion for the management of severe cancer pain should be abandoned. Introduction Continuous intravenous infusion of narcotic analgesics has been widely used in the management of patients with
pain caused by advanced cancer.l-4 It provides stable blood levels of the infused drug, which can be titrated to the needs of the individual patient so that peak level sedation and trough level breakthrough pain, sometimes associated with intermittent dosing regimens, can be avoided. The other major advantage of continuous parenteral infusion is that it bypasses the oral route-an obvious benefit for patients with intractable nausea and vomiting or non-functional gastrointestinal tracts. Continuous subcutaneous infusion of opioid drugs has been reported to be safe and efficacious5,6
and even the treatment of choice for parenteral management of cancer pain.7 This development was made possible by the availability of portable infusion pumps which allow continuous subcutaneous infusion into the chest wall or abdomen in an ambulatory patient. Continuous subcutaneous infusion therefore appears to carry all the advantages of continuous intravenous infusion with the added benefits of greater mobility, management on an outpatient basis, and avoidance of the need for intravenous access. However, the hypothesis that continuous subcutaneous infusion is as effective as continuous intravenous infusion in the management of advanced cancer pain needs to be examined critically, since it means that the drug has to pass through a subcutaneous compartment which might change the pharmacokinetics of the opioid infusion. By means of a randomised, double-blind, crossover design, we compared the pharmacodynamics and pharmacokinetics of continuous subcutaneous and intravenous hydromorphone infusions in patients with chronic cancer pain.
Patients and methods Between April, 1988, and April, 1990, 53 inpatients with poorly controlled cancer pain were started on continuous subcutaneous opioid infusions. Most of these patients were experiencing substantial side-effects during the regular administration of oral or rectal opioids and adjuvant analgesics. 10 patients were excluded from the trial because of clinical evidence of brain metastases or metabolic encephalopathy, and 23 refused to take part. The remaining 20 patients (10 male, 10 female) gave informed consent to take part in the study. Their mean age was 61 ’9 (11 -8) years and the mean and median weights were 61-0 (SD 18-3) and 53-6 (IQR 38-103) kg. 5 had breast cancer, 5 lung cancer, 4 cancer of the gastrointestinal tract, 2 cancer of the urogenital tract, 1 each melanoma and soft tissue sarcoma, and in 2 the primary tumour was unknown. The principal causes of pain were bone metastases in 8, soft tissue infiltration of chest wall or retroperitoneum in 8, brachial plexopathy in 2, and subacute bowel obstruction in 2. The Health Sciences Standing Committee on Human Research at the University of Western Ontario approved this double-blind, ADDRESS: Departments of Clinical Neurological Sciences and Oncology (D. E. Moulin, MD, N. Murray-Parsons, RN, A. I. Bouquillon, MSc), and Medicine (Clinical Pharmacology) (J. H. Kreeft, MD), University of Western Ontario, London, Ontario, Canada. Correspondence to Dr D. E. Moulin, Victoria Hospital, 375 South Street, London, Ontario, N6A 4G5, Canada.
466
study with hydromorphone hydrochloride (’Dilaudid’; Pharmaceuticals, Ontario). We used a portable programmable infusion pump that is accurate to within 10% of the set rate (Pharmacia CADD-PCA; Montreal). Patients were initially
crossover
Knoll
randomised
to
continuous
subcutaneous
or
intravenous
hydromorphone infusions. Before randomisation, dose titration to provide optimum analgesia with minimum side-effects was accomplished by means of subcutaneous infusions of morphine in 18 patients and hydromorphone in 2. Patients whose dose was established on morphine were switched to hydromorphone at one-fifth the morphine rate based on the five to six fold greater analgesic potency of hydromorphone." The double-blind nature of the study was made possible by use of the "double dummy" technique which uses simultaneous matching placebos. Each patient carried two infusion pumps (each weighing 475 g); during active subcutaneous infusion the intravenous pump delivered saline and vice versa. In randomised fashion each patient received
a
48 h infusion of either intravenous
or
subcutaneous
hydromorphone and then, after a 24 h wash-out with morphine, a 48 h infusion at the same dose by the other route. Fresh subcutaneous infusion sites were started on the anterior chest wall or abdomen at the beginning of each 48 h infusion. Morphine (subcutaneous injections) was used to treat breakthrough pain during hydromorphone infusions to avoid interference in the pharmacokinetic analysis of hydromorphone. Morphine was offered every 3 h at 50% of the equianalgesic dose8 of a 3 h infusion of hydromorphone. All co-analgesics and adjuvant analgesics were maintained unchanged throughout the study. While resting in bed, the patient was asked to estimate pain intensity, pain relief, general feeling (mood), and sedation by means of visual analogue scales. Each scale consisted of a 10 cm line on which the patient marked the point between extremes of effect. These assessments were made before each infusion (during treatment with oral or rectal opioids) and after 4 h, 8 h, 24 h, 32 h, and 48 h of each hydromorphone infusion. The visual analogue scales for pain intensity, pain relief, and mood have been validated in analgesic studies of cancer pain patients.9 Side-effects other than sedation were also monitored.1O In the arm opposite that bearing the intravenous infusion, another intravenous cannula connected to a heparin lock was inserted so that blood samples could be obtained at 24 h and 48 h during each hydromorphone infusion. By means of a technique modified from Svensson et al,l1 plasma hydromorphone concentration was determined by high-performance liquid chromatography with electrochemical detection. With 1 ml plasma volumes the lower limit of sensitivity of this assay was 2-5 ng/ml. Morphine or hydromorphone metabolites do not interfere with this assay. At presumed steady state, plasma concentration averaged for each hour between 24 h and 48 h (Css) was calculated from the area under the plasma concentration curve by means of trapezoidal summation. Because the infusion rates varied among patients, we had to normalise the data to drug concentrationfmg.h -1 for pharmacokinetic comparisons of the subcutaneous and intravenous infusion groups. This procedure makes an assumption of first order or linear kinetics for hydromorphone elimination, which we felt was valid for two reasons. Morphine is subject to first order elimination12 and morphine and hydromorphone are very similar opioid drugs (hydromorphone is a hydrogenated ketone of morphine) which are metabolised primarily by hepatic glucuronidation.13 Secondly, our findings of a significant linear correlation between intravenous hydromorphone infusion rates and average Css between 24 h and 48 h (r=0-85, p
