Antisense oligodeoxynucleotide effects on the hypothalamic-neurohypophysial system and the hypothalamic-pituitary-adrenal axis

The possibility of sequence-dependent, transient, and local inhibition of neuropeptide or neuropeptide receptor expression within the brain makes antisense targeting an attractive approach for those interested in the involvement of brain neuropeptide systems in behavioral and neuroendocrine regulation. Here, I describe our attempts to manipulate the synthetic activity of peptidergic systems of the hypothalamic-neurohypophysial system, i.e. , oxytocin and vasopressin, and the hypothalamic-pituitary-adrenal (HPA) axis by antisense oligodeoxynucleotides. Detailed experimental protocols including different approaches for intracerebral antisense application in anesthetized or conscious rats are provided. As a consequence of local oxytocin or vasopressin antisense treatment within the hypothalamic supraoptic nucleus, various aspects of the neuronal activity are already altered after a few hours. Thus, we monitored electrophysiological parameters of oxytocinergic and vasopressinergic neurons, stimulus-induced expression of the Fos protein in oxytocin neurons, and stimulated release of oxytocin or vasopressin into blood as well as within the hypothalamus by dendrites and cell bodies as measured by simultaneous microdialysis in blood and brain, shortly after a single acute antisense infusion. We also employed chronic antisense infusion via osmotic minipumps or by repeated local infusion into the targeted brain region; for example, septal vasopressin receptor downregulation impairs the ability of male rats to discriminate between juvenile rats. Further, reduction of the amount of available CRH, vasopressin, and oxytocin within the hypothalamic paraventricular nuclei alters the neuroendocrine stress response of the HPA axis.     

A randomised dose ranging study of recombinant tissue plasminogen activator in acute myocardial infarction

To assess the thrombolytic efficacy and the effect on the systemic fibrinolytic system of recombinant tissue plasminogen activator doses of 20 mg, 50 mg, and 100 mg were compared in a randomised study. Tissue plasminogen activator was infused intravenously over 90 minutes in 50 consecutive patients with acute myocardial infarction of four hours'' duration or less; on average the infusion was started 135 minutes (range 20 to 240) after the onset of pain. The affected artery was patent at the end of the 90 minute infusion in 14/17 (82%) of those who received 100 mg, 12/17 (71%) of those who received 50 mg, and 8/16 (50%) of those who received 20 mg. Regardless of dose, reperfusion rates were significantly better for patients treated within two hours of the onset of symptoms (81%) than for those treated in the third and fourth hours (54%). At the end of the infusion serum fibrinogen concentrations fell to 86% of the preinfusion value after 20 mg, 75% after 50 mg, and 63% after 100 mg, and similar dose dependent changes occurred in plasminogen, α₂ anti-plasmin, and fibrinogen and fibrin degradation products. The mean infarct related regional third ejection fraction was 46% for patients with grade 2 or 3 reperfusion and 35% for those with grade 0 or 1. Ventricular fibrillation occurred in six (12%) patients during the infusion of tissue plasminogen activator, but no late ventricular fibrillation occurred. Bleeding was minimal, reocclusion occurred in three patients, and four patients died from cardiac causes.Recombinant tissue plasminogen activator is an effective thrombolytic agent which produces better reperfusion rates after a 50 or 100 mg dose than after a 20 mg dose. The effect on the systemic fibrinolytic system is dose dependent. Successful reperfusion results in improvement of left ventricular function.     

Induction of midtrimester abortion by serial intravaginal administration of 15(S)-15-methyl-prostaglandin F2α (THAM) suppositories

Midtrimester abortion was successfully induced in 13 of 22 patients by serial intravaginal administration of 15(S)-15-methyl-prostaglandin F_2α (THAM) suppositories. Nine patients, 4 nulliparas and 5 multiparas, failed to abort after 24 hours of prostaglandin administration and a concomitant infusion of oxytocin was initiated. Seven of the nine patients aborted within 7 hours of the combined therapy and one patient on methadone maintainence aborted after 17.5 hours of combined therapy, 41.5 hours after the first dose of prostaglandin. A single patient failed to abort, despite the concomitant prostaglandin-oxytocin administration and underwent surgical evacuation. The mean abortion time for the 21 successful abortions was 22.56 hours. Nulliparous patients aborted somewhat faster, mean 21.79 hours, than multiparous patients, mean 23.80 hours, but this difference was not statistically significant. In this study, one patient aborted in less than 12 hours, and 62% of the successful cases aborted within 24 hours. The plasma levels of 15-ME-PGF_2α were analyzed by radioimmunoassay in 10 patients. Plasma prostaglandin levels rose significantly 30 minutes after the insertion of the first suppository, but there was a wide variation in levels from patient to patient. It was observed that the 2 patients with the highest levels had the fastest abortion times and episodes of gastro-intestinal side effects appeared related to a rise in prostaglandin levels. Sixty-four percent of the patients in this study had no gastro-intestinal side effect related to prostaglandin administration.     

Exposure of the Isolated Frog Skin to High Potassium Concentrations at the Internal Surface : II. Changes in epithelial cell volume,resistance and response to antidiuretic hormone

Isolated frog skin epithelia undergo marked, but reversible swelling when the external skin surface is bathed with conventional NaCl Ringer''s and the internal surface with KCl Ringer''s solutions. In 2 hours, epithelial thickness increased by over twofold. When NaCl Ringer''s was replaced on both sides of the skin, volume returned to control levels in less than 1 hour. When sulfate, rather than chloride, was the predominant anion, exposure of the internal surface to high potassium concentrations did not evoke changes in epithelial cell volume. With both KCl and K₂SO₄ Ringer''s, an immediate drop in DC resistance across the skin occurred. This was followed by partial recovery. Both the immediate drop and partial recovery were unrelated to changes in volume. A slow, sustained secondary drop in resistance was observed with KCl but not K₂SO₄ Ringer''s. This slower drop was associated temporally with swelling. When epithelial cell swelling occurred (i.e. with KCl Ringer''s), the characteristic response of the skin to vasopressin was abolished. However, with sulfate as anion, vasopressin elicited an increase in short-circuit current and/or in cell volume despite high internal potassium concentrations. It is concluded that epithelial swelling increased the permeability of the sodium-selective barrier at the external surface of the cells; and the possibility exists that stretching of cell membranes altered dimensions of pathways through which Na and water move, thereby mimicking the effects of vasopressin.     

Pregnancy complicated by psittacosis acquired from sheep.

Two cases of chlamydial infection in pregnant women are described, the first serologically proved and the second suspected. In both cases the infection was probably contracted from sheep suffering with enzootic abortion. Both patients were farmers'' wives who had helped their husbands and lambing and developed a non-specific febrile illness in late pregnancy. In the first case as there was no clinical improvement after 26 hours the patient was delivered by caesarean section of a live infant in good condition; the patient recovered fully. The second patient had presented a year earlier, the fetus had died in the uterus, and the patient himself died after spontaneous labour and forceps delivery 14 hours after admission. Both patients developed disseminated intravascular coagulation. As the casual agent in enzootic abortion in ewes has a predilection for the placenta, early delivery may be the management of choice in late pregnancy if infection with this organism if suspected.     

Induction of midtrimester abortion by serial intravaginal administration of 15(S)-15-methyl-prostaglandin F2alpha (THAM) suppositories.

Midtrimester abortion was successfully induced in 13 of 22 patients by serial intravaginal administration of 15(S)-15-methyl-prostaglandin F2alpha (THAM) suppositories. Nine patients, 4 nulliparas and 5 multiparas, failed to abort after 24 hours of prostaglandin administration and a concomitant infusion of oxytocin was initiated. Seven of the nine patients aborted within 7 hours of the combined therapy and one patient on methadone maintainence aborted after 17.5 hours of combined therapy, 41.5 hours after the first dose of prostaglandin. A single patient failed to abort, despite the concomitant prostaglandin-oxytocin administration and underwent surgical evacuation. The mean abortion time for the 21 successful abortions was 22.56 hours. Nulliparous patients aborted somewhat faster, mean 21.79 hours, than multiparous patients, mean 23.80 hours, but this difference was not statistically significant. In this study, one patient aborted in less than 12 hours, and 62% of the successful cases aborted within 24 hours. The plasma levels of 15-ME-PGF2alpha were analyzed by radioimmunoassay in 10 patients. Plasma prostaglandin levels rose significantly 30 minutes after the insertion of the first suppository, but there was a wide variation in levels from patient to patient. It was observed that the 2 patients with the highest levels had the fastest abortion times and episodes of gastro-intestinal side effects appeared related to a rise in prostaglandin levels. Sixty-four percent of the patients in this study had no gastro-intestinal side effect related to prostaglandin administration.     

Treatment with tumour infiltrating lymphocytes and interleukin-2 in patients with metastatic melanoma: A pilot study

Tumour infiltrating lymphocytes (TIL) were isolated and expanded from biopsy samples of 4 patients with metastatic melanoma. The patients were treated with autologous expanded TIL and continuous or bolus infusion of Interleukin 2 (IL-2) at a dose of 18 × 10⁶ International Units/m²/day for 5 days starting 36–48 hours after administration of cyclophosphamide at a dose of 1 g/m². The number of TIL infused ranged from 10¹⁰ to 5,56 × 10¹⁰ cells. Two patients had stable disease (SD) lasting for 2 1/2 and 4 months respectively and they died 24 and 13 months after therapy. One patient died during therapy due to a pseudomonas septicaemia and another patient developed progressive disease (PD). He died 3 months after the start of therapy. The side effects were substantial but most of them were reversible upon cessation of the treatment.The majority of the expanded TIL of all patients were of the CD8+ phenotype. Cutaneous metastases from two patients, removed after treatment with IL-2 and TIL, showed moderate lymphocytic infiltration also mainly of CD8+ T cells.The treatment with IL-2 and TIL is feasible, but further investigations should continue in an attempt to improve the efficacy of the therapy, to reduce toxicity and to diminish the costs and labour of the culture methods.     

Impaired Lignocaine metabolism in patients with myocardial infarction and cardiac failure.

Plasma concentrations of lignocaine were measured during and after infusion of lignocaine at 1.4 mg/min for 36-46 hours in 12 patients with myocardial infarction and one patient with cardiac failure due to uncontrolled ventricular tachycardia. In six patients without cardiac failure the plasma concentrations of lignocaine rose progressively during the infusion and the mean lignocaine half life was 4.3 hours compared with 1.4 hours in healthy subjects. Mean plasma lignocaine concentrations were significantly higher in seven patients with cardiac failure, and concentrations also rose during the infusion and the half life was considerably prolonged to 10.2 hours. Lignocaine concentrations rose rapidly to toxic levels when cardiogenic shock developed in one patient and did not fall when the infusion was stopped. The mean plasma antipyrine half life was moderately prolonged (19.4 hours) in a larger group of patients with myocardial infarction and cardiac failure but returned to normal during convalescence (13.2 hours). The metabolism of lignocaine is grossly abnormal in patients with cardiac failure and cardiogenic shock after myocardial infarction.     

Role of sodium and water excretion in the antihypertensive effect of vasopressin in the spontaneously hypertensive rat.

Mean arterial pressure (mmHg (1 mmHg = 133.322 Pa)), sodium excretion rate (mumol.kg-1.min-1), and urine flow (microL.kg-1.min-1) were measured in conscious unrestrained spontaneously hypertensive rats (SHR) and normotensive Wistar-Kyoto rats (WKY) before, during, and after a 3-h intravenous infusion of arginine vasopressin (20 ng.kg-1.min-1), an equipressor dose of phenylephrine, or an infusion of the vehicle. Cessation of the phenylephrine infusion was associated with a return of arterial pressure to preinfusion control values in both SHR and WKY. Cessation of the vasopressin infusion was also associated with a return of arterial pressure to preinfusion values in WKY. In contrast, in the SHR, arterial pressure fell from a preinfusion control level of 164 +/- 6.2 to 137 +/- 4 mmHg within 1 h of stopping the vasopressin infusion. Five hours after stopping the infusion, pressure was 134 +/- 3 mmHg (29 +/- 5 mmHg below preinfusion levels). Similar to the WKY, cessation of a vasopressin infusion was associated with a return of arterial pressure to preinfusion values in Sprague-Dawley rats. Thus, the failure to observe a hypotensive response in normotensive rats was not a peculiarity of the WKY strain. Sodium excretion rates increased during the infusions of vasopressin to a greater extent in SHR than in WKY. However, the natriuresis induced by phenylephrine was not significantly different from that generated by vasopressin in SHR, and in WKY, the natriuresis was greater for phenylephrine than for vasopressin. Urine output increased to a greater extent during the infusions of phenylephrine in both SHR and WKY than during vasopressin infusion.(ABSTRACT TRUNCATED AT 250 WORDS)     

Defective blood glucose counter-regulation in diabetics is a selective form of autonomic neuropathy.

Administration of a low-dose insulin infusion to normal subjects results in a mild drop in blood glucose concentration (1.1 mmol/1 (20 mg/100 ml)) and the resetting of the basal glucose at the lower concentration. Clinical hypoglycaemia does not develop, and there is a significant release of glucagon, growth hormone, and cortisol. A similar infusion in insulin-requiring diabetics results in hypoglycaemia accompanied by a release of growth hormone and cortisol but no significant release of glucagon. Subsequently giving arginine to these patients results in a significant release of glucagon, indicating that the alpha cell is intact and can respond to local, direct stimulation. In one patient the defect in glucagon response to impending hypoglycaemia developed after two years'' insulin treatment. This type of dissociated response'' of the alpha cell has been reported in animals after denervation of the pancreas, and insulin-requiring diabetics may develop a selective form of autonomic neuropathy affecting the vagal control of glucagon release.     

Comparison of spontaneous ascites filtration and reinfusion with total paracentesis with intravenous albumin infusion in cirrhotic patients with tense ascites.

OBJECTIVE--To compare the effectiveness and safety of spontaneous ascites filtration and reinfusion and total paracentesis plus intravenous albumin infusion in cirrhotic patients with tense ascites. DESIGN--Randomised trial of the two treatments. SETTING--Teaching hospital and district general hospital in Milan. PATIENTS--45 consecutive cirrhotic patients with recurrent tense ascites and urinary sodium excretion rate less than 20 mmol/day. 35 fulfilled admission criteria and completed the study. 17 received spontaneous ascites filtration and 18 paracentesis plus albumin infusion. MAIN OUTCOME MEASURES--Body weight; urinary volume; serum and urinary electrolyte, serum fibrinogen, and plasma aldosterone concentrations; and plasma renin activity before the procedure and 24 hours and eight days afterwards. RESULTS--Both procedures were effective in all patients. Weight decreased in both groups and showed no substantial increase after eight days. In patients receiving ascites filtration, values decreased significantly (p less than 0.01) after 24 hours for platelet count (mean relative change 0.92; 99% confidence interval 0.86 to 0.98) and serum fibrinogen concentration (0.92; 0.88 to 0.98) but returned to pretreatment values after eight days; no laboratory and clinical signs of disseminated intravascular coagulation were noted. Three patients in this group had fever, which receded spontaneously. One patient in each group had dilutional hyponatraemia. CONCLUSIONS--Spontaneous ascites filtration and reinfusion is an effective treatment for tense ascites. Reinfusion of the patient''s concentrated proteins provides savings without compromising safety.     

Minimal Erythema Dose (MED) Testing

Ultraviolet radiation (UV) therapy is sometimes used as a treatment for various common skin conditions, including psoriasis, acne, and eczema. The dosage of UV light is prescribed according to an individual''s skin sensitivity. Thus, to establish the proper dosage of UV light to administer to a patient, the patient is sometimes screened to determine a minimal erythema dose (MED), which is the amount of UV radiation that will produce minimal erythema (sunburn or redness caused by engorgement of capillaries) of an individual''s skin within a few hours following exposure. This article describes how to conduct minimal erythema dose (MED) testing. There is currently no easy way to determine an appropriate UV dose for clinical or research purposes without conducting formal MED testing, requiring observation hours after testing, or informal trial and error testing with the risks of under- or over-dosing. However, some alternative methods are discussed.     

Treatment of Diabetic Coma with Continuous Low-dose Infusion of Insulin

Thirty-eight patients in diabetic coma from four different centres were treated with a continuous low-dose intravenous infusion of insulin at an average dose of 7·2 IU/hr. All patients recovered rapidly except for one profoundly shocked patient who died. The mean fall in plasma glucose was 58% four hours after the start of insulin. Blood ketone bodies and plasma free fatty acids showed a similar response. There was no significant difference in plasma glucose response according to severity of acidosis or previous treatment with insulin. Hypokalaemia was uncommon. In the treatment of diabetic coma this technique has proved simple, safe, and effective.     

Differences in the effects of vasopressin and oxytocin on rabbit myometrial activity and a possible mediation of prostaglandins

Uterine responses to vasopressin and oxytocin were monitored in non-pregnant and 3- or 6-8-day-pregnant rabbits by recording the intrauterine pressure. Oxytocin stimulated uterine activity in all groups, but the effect of vasopressin was stimulatory in non-pregnant animals, inhibitory in those 3 days post coitum and weakly stimulatory in those later in pregnancy. Inhibition of prostaglandin (PG) synthesis, by the administration of indomethacin, reduced the spontaneous uterine activity as well as the responses to oxytocin and vasopressin in the non-pregnant rabbits, but had little effect in the pregnant animals. During infusion of PGF-2alpha, PGE-1 or PGE-2 in 6-8-day-pregnant rabbits, the stimulatory response to vasopressin, although slight before the infusion, was inhibited whereas the stimulatory response to oxytocin remained virtually unchanged. The results suggest that vasopressin and oxytocin under certain hormonal conditions, are able to activated the uterine contractions by mechanisms in which the involvement of PG is not obligatory.     

A phase I trial of recombinant tumor necrosis factor (rTNF) administered by continuous intravenous infusion in patients with disseminated malignancy

rTNF was administered to 28 patients with advanced metastatic cancers by continuous intra venous infusion for 5 consecutive days every 2 weeks. The dose levels were 30, 40, 70, 110, 180 and 290 µg/M²/day. Groups of 3 patients were started at each successive dose level and then on subsequent courses treated with the next dose level through 4 escalations as tolerated. Tumor types were: colon cancer 14; adenocarcinoma of unknown primary, 2; renal cancer, 2; leiomyosarcoma, 2; lung cancer, 1; prostate cancer, 1; thymona, 1; bladder cancer; 1; parotid, 1; Kaposi's sarcoma 2; ovarian 1. Toxicities included fever and chills (usually within the first 8 hours of infusion), fatigue, headache, decreased performance status, hypotension and CNS. All patients experienced leukopenia and thrombocytopenia within 24 hours or less after start of infusion with return of baseline by 72 hours after rTNF was stopped. The fall in these counts averaged 50% and was not dose related. No major changes in liver or renal function, coagulation or blood lipids were seen. Major dose limiting toxicities were fatigue, confusion, thrombocytopenia, seizures, hypotension and decreased performance status. NK cell activity measured against K562 target cells was augmented from about 30% target cell lysis to about 70% target cell lysis over the first 7 days of treatment. Two patients, both with metastatic colon cancer showed transient, objective tumor regression which did not qualify as a partial response. One patient with ovarian cancer had a stable partial response but progressed after 13 courses of treatment. Continuous infusion of TNF can be safely administered to patients with a maximum tolerated dose of only between 30 and 40 µg/M²/day. In addition, the MTD with continuous infusion seems to be highly variable and unpredictable from patient to patient. These data suggest that continuous infusion will not be an optimal way to administer TNF.     

Gorlin's syndrome,or nevoid basal cell carcinoma syndrome.

Gorlin''s syndrome is a condition inherited in an autosomal dominant fashion. It involves many organs, but principally affects the skin, skeleton, and endocrine and nervous systems. The most common features are multiple nervi and basal cell carcinomas of the skin, benign jaw cysts, dyskeratotic pits in the palms and soles, rib and vertebral abnormalities, brachymetacarpalism, and calcification of the falx cerebri. In 14 patients, 4 of whom belonged to one family, the age at the time of diagnosis ranged from 11 to 63 years. Ten patients are alive, but five are severely disfigured by carcinomas. Two patients died of complications resulting from uncontrolled tumours, and two died of other cancers. New skin tumours constantly develop; small ones can be excised, but large ones require extensive surgery with or without radiotherapy.     

Fatal ischaemic brain oedema after early thrombolysis with tissue plasminogen activator in acute stroke.

Two patients with acute major, disabling cerebral infarction with presumed middle cerebral artery occlusion were treated with the clot specific thrombolytic agent tissue plasminogen activator roughly three and a half hours after the onset of symptoms. Both patients had a normal computed tomography (CT) scan before treatment. No appreciable systemic bleeding complications occurred, apart from bruising. One patient had bleeding into the subarachnoid space from a microscopic angioma, which was found at necropsy. Haematological monitoring of the two patients showed pronounced fibrinogenolysis and alpha 2 antiplasmin consumption in one. One patient showed transient improvement during the infusion. In both cases extensive infarction, partly haemorrhagic in one, with massive concomitant oedema was found on repeated CT. Both patients deteriorated and eventually died as a consequence of transtentorial herniation. In the one patient who came to necropsy a moderate, probably pre-existing smooth stenosis of the ipsilateral carotid artery was found, all cerebral vessels being patent. It is concluded that thrombolytic treatment with a clot specific agent such as tissue plasminogen activator started three to four hours after a major ischaemic stroke may be hazardous, not because of haemorrhagic transformation of the original ischaemia but because early reperfusion may promote massive, potentially fatal cerebral oedema.     

Patient and general practitioner delays in acute myocardial infarction

The longest component of the total delay in coming under coronary care is patient delay, and it has been suggested that public education might be used to make it shorter. The patterns of patient delay were studied in 450 patients with acute myocardial infarction uncomplicated by cardiac arrest out of hospital, of whom 243 had a previous history of ischaemic heart disease. Patient delays had a skewed distribution with a modal delay of up to one hour, a median delay of two hours, and a mean delay of 10 hours. Two thirds of patients had sought help from their general practitioners within four hours of the onset of symptoms. During the first four hours the longer that patients delayed the lower was the subsequent mortality (27%, 18%, and 9% for delays of one hour or less, up to two hours, and up to four hours, respectively), but patients who delayed four to eight hours had the highest mortality of all (38%). Neither the median value nor the pattern of patient delays was altered by a previous history of ischaemic heart disease.There were pronounced differences in doctor delays, depending on the patient''s age, delay time, and ultimate place of treatment, showing that the doctors'' behaviour was influenced before they had seen their patients. Nevertheless, the median total delay for patients aged up to 70 was one hour 35 minutes, and a higher proportion of patients were seen early after infarction than in recent hospital trials of thrombolytic treatment.These findings suggest that the patients'' call for help and the doctors'' response may be at an instinctive level according to the patients'' distress; these patterns of behaviour may be difficult to modify by public education.     

Toxic effects of catecholamines on skin

The purpose of this study was to examine the effects of catecholamines on skin necrosis independent of their vasoactive effects. Rat abdominal or human breast skin was excised, pinned flat, and incubated at 37 degrees C for 6 hours in a buffered salt solution containing catecholamine. At 0.1 and 6 hours the lactate dehydrogenase (LDH) released from the skin and appearing in the buffer was determined spectrophotometrically. All groups showed similar LDH levels at 0.1 hour. Rat skin treated with greater than or equal to 10(-7) M epinephrine (33 times less than the 1:200,000 used clinically) or greater than or equal to 10(-5) M norepinephrine showed a significant increase in the LDH released at 6 hours versus controls (18.75 +/- 1.25 versus 13.75 +/- 1.25 and 29.25 +/- 2.96 versus 22.00 +/- 1.96 IV, respectively). Total tissue LDH levels were not significantly different at 0.1 or 6 hours. The toxic effect of epinephrine was eliminated by the addition of propranolol or selective beta 2 blockade, but not by alpha or beta 1 blockade. Therefore, this effect appears to be mediated largely by beta 2 receptors. Similar toxic effects were seen in human breast skin treated with 1:200,000 epinephrine and were blocked with propranolol. Phenylephrine at 1:20,000 demonstrated toxicity, but angiotensin II and vasopressin did not. These studies indicate that addition of catecholamine to ischemic rat or human skin accelerates skin death within 6 hours, but that the toxicity can be reversed with beta blockade.     

Subcutaneous narcotic infusions for cancer pain: treatment outcome and guidelines for use.

OBJECTIVE: To provide guidelines for the institution and maintenance of a continuous subcutaneous narcotic infusion program for cancer patients with chronic pain through an analysis of the narcotic requirements and treatment outcomes of patients who underwent such therapy and a comparison of the costs of two commonly used infusion systems. DESIGN: Retrospective study. SETTING: Tertiary care facilities and patients'' homes. PATIENTS: Of 481 patients seen in consultation for cancer pain between July 1987 and April 1990, 60 (12%) met the eligibility criteria (i.e., standard medical management had failed, and they had adequate supervision at home). INTERVENTION: Continuous subcutaneous infusion with hydromorphone hydrochloride or morphine started on an inpatient basis and continued at home whenever possible. OUTCOME MEASURES: Patient selectivity, narcotic dosing requirements, discharge rate, patient preference for analgesic regimen, side effects, complications and cost-effectiveness. RESULTS: The mean initial maintenance infusion dose after dose titration was almost three times higher than the dose required before infusion (hydromorphone or equivalent 6.2 v. 2.1 mg/h). Eighteen patients died, and the remaining 42 were discharged home for a mean of 94.4 (standard deviation 128.3) days (extremes 12 and 741 days). The mean maximum infusion rate was 24.1 mg/h (extremes 0.5 and 180 mg/h). All but one of the patients preferred the infusion system to their previous oral analgesic regimen. Despite major dose escalations nausea and vomiting were well controlled in all cases. Twelve patients (20%) experienced serious systemic toxic effects or complications; six became encephalopathic, which necessitated dose reduction, five had a subcutaneous infection necessitating antibiotic treatment, and one had respiratory depression. The programmable computerized infusion pump was found to be more cost-effective than the disposable infusion device after a break-even point of 8 months. CONCLUSIONS: Continuous subcutaneous infusion of opioid drugs with the use of a portable programmable pump is safe and effective in selected patients who have failed to respond to standard medical treatment of their cancer pain. Dose titration may require rapid dose escalation, but this is usually well tolerated. For most communities embarking on such a program a programmable infusion system will be more cost-effective than a disposable system.