Long-term survival results of a randomized trial comparing gemcitabine/cisplatin and methotrexate/vinblastine/doxorubicin/cisplatin in patients with locally advanced and metastatic bladder cancer.

PURPOSE: To compare long-term survival in patients with locally advanced and metastatic transitional cell carcinoma (TCC) of the urothelium treated with gemcitabine plus cisplatin (GC) or methotrexate/vinblastine/doxorubicin/cisplatin (MVAC). PATIENTS AND METHODS: Efficacy data from a large randomized phase III study of GC versus MVAC were updated. Time-to-event analyses were performed on the observed distributions of overall survival time and progression-free survival. RESULTS: Four hundred and five patients were randomized, 203 to the GC arm and 202 to the MVAC arm. At the time of this analysis, 347 patients have died (GC 176, MVAC 171). Overall survival was similar in both arms (HR 1.09; 95% confidence interval [CI] 0.88-1.34, P = 0.66) with a median survival of 14.0 months (95% CI 12.3-15.5 months) in the GC, and 15.2 months (95% CI 13.2-17.3 months) in the MVAC arm. The median progression-free survival was 7.7 months with GC (95% CI 6.8-8.8) and 8.3 months with MVAC (95% CI 7.3-9.7) with a HR of 1.09 (95% CI 0.89-1.34). Significant prognostic factors favoring overall survival included performance status (>70), TNM staging (M0 vs. M1), low/normal alkaline phosphatase expression, number of sites of disease <3, and the absence of visceral metastasis. By adjusting for these prognostic factors, the HR was 0.99 for overall survival and 1.01 for progression-free survival. CONCLUSIONS: Long-term overall and progression-free survival following treatment with GC or MVAC are similar. These results strengthen the role of GC as a standard of care in patients with locally advanced and metastatic transitional-cell carcinoma (TCC).

Prophylactic strategies to meet infectious complications in fludarabine-treated CLL.

Fludarabine has emerged as salvage therapy in chlorambucil-resistant CLL. However, encouraging response rates have been compromised by a high incidence of serious infectious complications. Prophylactic measures to reduce the frequency of infections are needed, but up to now, there are no established standards for supportive therapy in fludarabine-treated CLL. Clinicians have observed an increasing frequency of life-threatening opportunistic infections but only some of these may be explained by fludarabine-induced impairment of cell-mediated immunity. Neutrocytopenia commonly found during initial fludarabine treatment may not have been addressed sufficiently as risk factor for infections. Thus, G-CSF supplementation may improve the rate of infectious complications by reducing the duration of fludarabine-induced neutrocytopenia. The changing spectrum of infectious complications should stimulate additional trials on the value of IVIG replacement in fludarabine-treated CLL patients and on the role of low-dose co-trimoxazole in patients at high risk of Pneumocystis carinii infections.

Increases of sICAM-1 during neutropenic pneumonia in leukemic patients.

Aggressive chemotherapy of leukemia increases the risk of severe infections during treatment-induced myelosuppression. However, the assessment of an infectious origin of neutropenic fever is often difficult. Leukocyte adhesion molecules such as E-selectin, intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) are involved in early inflammatory response. We studied plasma concentrations of their soluble isoforms during 48 treatment courses with myeloablative chemotherapy in 32 leukemic patients. There were 35 febrile episodes during neutropenia. Pneumonia was clinically and microbiologically documented in 15 cases, six had proven infections but normal chest radiograph, and 14 were classified as fever of unknown origin. Longitudinal studies revealed a sustained increase of sICAM-1 plasma levels associated with pneumonia. Increase of sICAM-1 plasma levels distinguished patients with pneumonia from those with fever not related to pneumonia (positive predictive value 0.87, negative predictive value 0.94). Plasma levels of sICAM-1 were elevated in both, fungal and non-fungal pneumonia. Increases of sICAM-1 paralleled first radiographic evidence of pulmonary infiltrations in most cases. In contrast, no elevation of sVCAM-1 or sE-selectin was documented during febrile events prior to recovery of leukocyte counts.

Serum erythropoietin concentrations in patients with myelodysplastic syndromes.

Medullary dyserythropoiesis with reduced production of erythrocytes is an early and consistent feature of myelodysplastic syndromes (MDS). The mechanism underlying the disturbed red cell proliferation and maturation is presently unknown. In order to study the role of erythropoietic growth factors, we determined by radioimmunoassay the serum concentrations of immunoreactive erythropoietin (Epo) in 42 non-transfused patients with primary and secondary MDS. Their median hemoglobin concentration at the time of Epo measurement was 9.1 g/dl (range, 5.7-14.6). Compared with the control group, 83% of the MDS patients had increased serum Epo levels, ranging from 26-4530 mU/ml. Although in the entire patient population an inverse relationship between serum Epo and hemoglobin concentrations was noted (r = -0.35; p = 0.02), Epo titers differed markedly between patients at comparable degrees of anemia. In 7 patients presenting with a hemoglobin concentration between 5.9 and 11.9 g/dl, excessive elevations of Epo levels (greater than 500 mU/ml) were found. In contrast to previous observations, serum Epo concentrations were not shown to correlate with the percentage of erythroblasts in the bone marrow. There was, however, a significant relationship between the Epo activity and the degree of medullary dyserythropoiesis, as assessed by morphological criteria (p less than 0.01). From these data we conclude that the anemia in MDS is not due to an endogenous Epo deficiency. The marked variability of Epo production in these disorders is not fully explained by the degree of anemia, but may also reflect inherent abnormalities of the myelodysplastic erythropoiesis.

Pharmacokinetics of high-dose cytarabine and its deamination product--a reappraisal.

Cytarabine is intracellularly activated and correlations have been established between the pharmacokinetic behaviour of active metabolites and their antileukemic effect. Recently, a good response to high-dose treatment of leukemias has additionally been attributed to a so-called low deamination phenotype of cytarabine inactivation. Consequently, these findings would support plasma level monitoring of cytarabine and its metabolite uracil arabinoside in high-dose cytarabine regimens. This pharmacokinetic study presents data attempting to reevaluate these observations. Thirty-seven patients were treated by 3-h high-dose cytarabine infusions (9 patients 1000 mg/m2, 28 patients 3000 mg/m2) as part of their treatment for acute leukemia. Serial blood samples during and post infusion were analysed for cytarabine (araC) and its deamination product uracil arabinoside (araU) using HPLC with UV-detection. Considerable interindividual variation was observed in end-infusion plasma concentrations of araC (1000 mg/m2: 2.1-fold, 3000 mg/m2: 5.5-fold) and araU (1000 mg/m2: 2.7-fold, 3000 mg/m2: 2.9-fold). The median ratio of end infusion concentrations araU/araC (on a molar basis) was 5.6 (S.D. 3.0), extreme ratio values were 2 and 14. No differences of the araU/araC ratio were found between the two dosages used. Minimum plasma araC concentrations at the end of infusion were 10.5 micromol/l and 22.0 micromol/l at a dose of 1000 and 3000 mg/m2, respectively. In our European study population a "fast" deamination phenotype of cytarabine (araU/araC ratio > 14) was not be observed.

Monitoring of multiple myeloma patients by simultaneously measuring marker substances of bone resorption and formation.

Fifteen patients (13 males and two females; mean age, 63 years; age range, 46-84 years) with multiple myeloma were studied prospectively (range of follow-up period, 2-6 months) to elucidate the diagnostic validity of biochemical markers of bone formation (bone alkaline phosphatase and the C-terminal propeptide of type I procollagen) and bone resorption (urinary excretion of pyridinium cross-links) for monitoring these patients. Eleven of 15 patients received melphalan i.v. and prednisone p.o. every 4 weeks. All patients were given pamidronate i.v. for inhibition of bone resorption. The mean values of the urinary excretion of pyridinium cross-links were significantly higher in the patients fulfilling the criteria of 'progression' or 'relapse' than in those showing 'response' and those in the 'plateau phase' (P < 0.05). In contrast, neither bone alkaline phosphatase nor C-terminal propeptide serum values differed significantly between these two groups (P > 0.05). The concentrations of both bone formation markers were significantly lower in the patients than in the samples obtained from apparently healthy persons (P < 0.001). There was a significant inverse correlation between the number of pamidronate courses and the serum concentrations of bone alkaline phosphatase (P < 0.05). A lack of correlation was observed between the urinary excretion of pyridinium cross-links and all other laboratory parameters measured (serum concentrations of total protein, calcium, creatinine and (beta 2-microglobulin). In conclusion, the urinary excretion of pyridinium cross-links might be a useful parameter for monitoring multiple myeloma patients. Decreased values of bone formation markers may be due to a suppressive effect of the bisphosphonate agents administered or reflect the severity of osteolytic lesions which have been described as being associated with unbalanced bone remodelling.

The influence of gemcitabine on the CD4/CD8 ratio in patients with solid tumours.

Gemcitabine (dFdC) is a novel pyrimidine antimetabolite with documented antineoplastic activity against metastatic non-small cell lung cancer (NSCL), pancreatic carcinoma, ovarian and breast cancer. The side effects of gemcitabine are generally mild; severe infections are reported in less than Ilo of patients. In contrast, other new nucleoside analogues such as the purine antimetabolite fludarabine lead to a significant alteration of the CD4/CD8 lymphocyte ratio associated with an increased risk for opportunistic infections. This study investigates the effect of gemcitabine on different lymphocyte subsets during consecutive applications. 16 patients with solid rumours (3 non-small cell lung cancer, 3 pancreas, 3 testicular, 2 breast, ovarian germ-cell, 1 ovarian, 1 small cell lung, 1 gastric cancer, 1 carcinoma of unknown primary); 15 patients were previously treated, received at least 3 applications of gemcitabine (1,000 mg/m(2) as a 30 min infusion, at days 1, 8, 15; q 4 weeks). Lymphocytes surface antigens were analysed by standard technique flow cytometry prior to every infusion. The median number of leukocytes before therapy was 7823/mu l, with lymphocytes 875/mu l, including 68% T-cells (CD3(+)), 9% B-cells (CD19(+); CD20(+)) and 15% NK-cells (CD56(+); CD16(+); CD3(-)), the CD4/CD8 ratio was 1.7. After gemcitabine therapy the median number of leukocytes was 5136/mu l, with lymphocytes 1012/mu l, including 77% T-cells, 8% B-cells and 10% NK-cells and a CD4/CD8 ratio of 2.2. Severe complications or opportunistic infections were not seen in these 16 patients. No significant change of CD4/CD8 ratios and NK-ccll numbers was seen in our patients with solid tumours during weekly treatment with gemcitabine. A severely increased risk for opportunistic infections following treatment with the new antimetabolite gemcitabine appears unlikely.

A phase III trial of pemetrexed plus gemcitabine versus gemcitabine in patients with unresectable or metastatic pancreatic cancer.

BACKGROUND: This randomized phase III study compared the overall survival (OS) of pemetrexed plus gemcitabine (PG) versus standard gemcitabine (G) in patients with advanced pancreatic cancer. PATIENTS AND METHODS: Patients with unresectable locally advanced or metastatic pancreatic cancer and no prior systemic therapy (including 5-fluorouracil as a radiosensitizer) were randomized to receive either 1,250 mg/m(2) gemcitabine on days 1 and 8 plus pemetrexed 500 mg/m(2) after gemcitabine on day 8 (PG arm) of each 21-day cycle, or gemcitabine 1,000 mg/m(2) on days 1, 8 and 15 of each 28-day cycle (G arm). RESULTS: Five hundred and sixty-five patients with well-balanced baseline characteristics were randomly assigned (283 PG, 282 G). OS was not improved on the PG arm (6.2 months) compared with the G arm (6.3 months) (P=0.8477). Progression-free survival (3.9 versus 3.3 months; P=0.1109) and time to treatment failure (3 versus 2.2 months; P=0.2680) results were similar. Tumor response rate (14.8% versus 7.1%; P=0.004) was significantly better on the PG arm. Grade 3 or 4 neutropenia (45.1% versus 12.8%), thrombocytopenia (17.9% versus 6.2%), anemia (13.9% versus 2.9%), febrile neutropenia (9.9% versus 0.4%; all P <0.001) and fatigue (15% versus 6.6%; P=0.002) were significantly more common on the PG arm. Four treatment-related deaths occurred on the PG arm and none in the G arm. CONCLUSIONS: Pemetrexed plus gemcitabine therapy did not improve OS. Single-agent gemcitabine remains the standard of care for advanced pancreatic cancer.

[Prophylaxis against mycoses in neutropenic patients]. / Mykose-Prophylaxe bei neutrozytopenischen Patienten.

During the last years, the proportion of cancer patients who develop systemic fungal infections has increased steadily. These infections are characterised by high mortality, especially in patients with persistent granulocytopenia and in those receiving allogeneic bone marrow transplants. The most important pathogens in neutropenic patients are Candida and Aspergillus spp. Usually, Candida infections arise from overgrowth in the gastrointestinal tract, while Aspergillus infections are acquired by inhalation of spores. Prophylaxis of systemic fungal infections seems mandatory since optimal strategies for diagnosis and treatment of these infections are lacking. Treatment with the non-absorbable polyenes nystatin and amphotericin B is useful for prophylaxis of superficial fungal infections, provided that compliance of the patients is optimal. The imidazoles ketoconazole and miconazole can reduce the incidence of superficial fungal infections, but there are conflicting data regarding their value for prevention of systemic mycoses. There are several studies indicating that prophylactic use of fluconazole reduces the incidence of mucosal and systemic fungal infections, especially in patients receiving allogeneic bone marrow transplants. Fluconazole shows reduced activity against several Non-albicans spp. and is not active against Aspergillus spp. Itraconazole has in vitro and in vivo activity against several Aspergillus spp. but high serum and tissue levels are necessary. However, bioavailability of itraconazole is reduced in patients with raised gastric pH and no i.v. formulation is available. Although there is some evidence for its prophylactic activity against Aspergillus infections in neutropenic patients, more studies are necessary to confirm these findings. Intravenous amphotericin B cannot be recommended for routine prophylactic use because of its toxicity.(ABSTRACT TRUNCATED AT 250 WORDS)

Prevention of amphotericin-B-induced nephrotoxicity by loading with sodium chloride: a report of 1291 days of treatment with amphotericin B without renal failure.

An open, prospective study was performed to evaluate the clinical usefulness of sodium chloride loading for prevention of amphotericin-B-induced nephrotoxicity in 37 patients requiring 44 courses of amphotericin B treatment. The median duration of the treatment course was 22 days (range, 9-136 days), and mean cumulative dose per patient was 1117 mg (range, 231-7831 mg). During amphotericin B treatment, all patients received 50 to 100 ml of 10% sodium chloride (85 to 171 mmol NaCl) via an intravenous line for prevention of amphotericin-B-induced nephrotoxicity evaluated by serum creatinine levels. Using this regimen, none of the patients developed significant nephrotoxicity (increase in serum creatinine of more than twice baseline level, or serum creatinine level greater than or equal to 2.0 mg/dl, respectively) despite the co-administration of other potentially nephrotoxic drugs. It was not necessary to discontinue treatment with amphotericin B in any of the patients. There were no side effects due to sodium chloride loading. Our results demonstrate that sodium chloride loading is useful for the prevention of amphotericin-B-induced nephrotoxicity.

[The influence of the duration of infusion on the acute toxicity of amphotericin B. Comparison of two hour versus four hour infusions]. / Einfluss infusionsdauer auf die Akuttoxizität von Amphotericin B. Vergleich von zwei- versus vierstündigen Infusionen.

A prospective, randomized, double-blind crossover study was performed to compare the incidence and severity of amphotericin-B-induced acute toxicity (chills, fever, nausea and vomiting) after two- and four-hour infusions in 33 leukemic patients with suspected or microbiologically proven systemic fungal infections. Each patient was treated in an alternating fashion of two- and four-hour infusions every other day. Toxicity was graded according to modified WHO-criteria. Evaluation of 264 infusions revealed no difference between the two schedules neither in incidence nor severity of acute toxic reactions. These data indicate that amphotericin B given over 2 hours is equally well--or poorly (!)--tolerated than the four-hour regimen.

[Atypical joint symptoms as initial symptoms of varicella infection in chronic myeloid leukemia]. / Atypische Gelenksymptomatik als Erstsymptom einer Varizelleninfektion bei chronischer myeloischer Leukämie.

A 25-year-old man, known to have chronic myeloid leukaemia for four years, acutely developed a fever of 39.5 degrees and severe pain in the shoulder and hip joints. There was no evidence of joint disease. Treatment with indomethacin briefly improved the symptoms, but within 72 hours the patient developed a fulminant illness with high fever and clinical as well as biochemical signs of a severe consumption coagulopathy. Herpetiform efflorescences appeared over the head and trunk shortly before death. At autopsy there was histological and immunohistochemical evidence of a varicella infection with hepatitis, oesophageal involvement and severe internal organ bleedings. In immune-compromised patients with atypical prodromal symptoms a varicella infection must be considered so that causal treatment with acyclovir and hyperimmune-globulin can be begun in time.

Neutrophilic dermal infiltrates in granulocytopenic patients with acute leukemia.

Acute febrile neutrophilic dermatosis (AFND, Sweet's syndrome) is clinically characterized by fever, neutrophilic leukocytosis, and tender dermal plaques. Histological examination typically reveals infiltration of the dermis by neutrophils. In three patients (2 female, 1 male, 54-59 years) with acute leukemia (2 myelogenous, 1 lymphoblastic) dermal plaques developed during febrile episodes in chemotherapy-induced pancytopenia. The clinical appearance was compatible with AFND. The diagnosis was substantiated by skin biopsies which showed dense neutrophilic dermal infiltrates without leukemic cells. Leukocytoclastic vasculitis was considered as differential diagnosis. Plasma levels of soluble adhesion molecules ICAM-1, VCAM-1, and E-selectin regulating leukocyte transendothelial migration were in the normal range. Systemic glucocorticoids were avoided because of the high risk of infection during prolonged bone marrow aplasia. The lesions were treated with topical steroids and resolved without scarring within 1-5 weeks. AFND has been reported in association with acute leukemia at normal or elevated white blood cell counts. Although implausible from a pathophysiological point of view, similar neutrophilic dermal infiltrates were found in three patients during chemotherapy-induced pancytopenia with white blood cell counts distinctly below 1 x 10(9)/l.

[A lethal course in pseudomembranous enterocolitis during the parenteral administration of vancomycin and imipenem]. / Letaler Verlauf einer pseudomembranösen Enterokolitis unter parenteraler Gabe von Vancomycin und Imipenem.

A 48-year-old woman required mechanical ventilation after aortic valve replacement for decompensated aortic valve stenosis when bleeding complications developed and rethoracotomy had to be performed. Acute renal failure necessitated haemodialysis. Septic fever of unknown aetiology failed to respond to oxacillin, cefotaxim and tobramycin. The endotracheal cannula and central venous catheter were changed on the 24th postoperative day and the antibiotic treatment altered to 250 mg imipenem and 125 mg vancomycin three times daily intravenously. The fever soon subsided, but recurred on the 32nd postoperative day, accompanied by increasing leucocytosis. The patient was obstipated but had no intraabdominal signs. Four days later ultrasonography demonstrated thickening of the intestinal wall and coloscopy showed typical pseudomembranous colitis. Intestinal contents were positive for Clostridium difficile toxin. Despite immediate rectal and intragastric administration of 250 mg vancomycin four times daily the patient died of pseudomembranous colitis, confirmed at autopsy. The case demonstrates that vancomycin cannot always prevent the development of pseudomembranous colitis.

Infection prophylaxis in neutropenic patients with acute leukaemia--a randomized, comparative study with ofloxacin, ciprofloxacin and co-trimoxazole/colistin.

Preliminary results are presented of an ongoing, prospective, randomized, study comparing ofloxacin, ciprofloxacin and co-trimoxazole/colistin for the prevention of infection in patients with acute leukaemia. The results for 59 patients (median age 47 years, range 21-72) included 88 episodes of neutropenia, each associated with a course of cytotoxic therapy. The main factor measured was the time elapsed from the beginning of neutropenia (neutrophils less than 500/microliter) until the first infectious febrile episode. The median time for the period was 12 days (range 1-56) for the cotrimoxazole/colistin group, 15 days (range 1-38) for the ofloxacin group and 20 days (range 1-36) for the ciprofloxacin group (differences not significant). Microbiologically proven major infections occurred in 10/27 treatment courses with co-trimoxazole/colistin 7/31 courses with ofloxacin and 7/30 courses with ciprofloxacin (P not significant). These were mostly due to Gram-positive cocci. There were no Gram-negative infections in the quinolone groups compared with one major Pseudomonas aeruginosa infection in the co-trimoxazole/colistin group. No Pneumocystis carinii infections were encountered. Adverse reactions associated with co-trimoxazole/colistin required discontinuation of medication in 11/27 treatment courses because of compliance problems, skin reactions or gastrointestinal intolerance. There were significantly fewer discontinuations in the ofloxacin (n = 2) and in the ciprofloxacin groups (n = 3). Major side effects of the quinolones included persistent icterus in one patient receiving ofloxacin and psychiatric symptoms in one patient receiving ciprofloxacin. It is concluded from these data that there were no statistically significant differences between the three treatment groups in respect of the prevention of infection.(ABSTRACT TRUNCATED AT 250 WORDS)

Septicemia due to Streptococcus mitis in neutropenic patients with acute leukemia.

Eight neutropenic patients with acute lymphocytic or nonlymphocytic leukemia had septicemia due to different strains of Streptococcus mitis (St. mitis), a microorganism not commonly recognized as a special pathogen in leukemic patients. Four of the patients had been treated with high-dose cytosine arabinoside as part of the cytostatic regimen, six had a central venous line and four patients had oral lesions prior to the infection. Selective gut decontamination consisted of co-trimoxazole/colistin in five patients and quinolones in three patients. The first three patients died, either due to interstitial pneumonia with the adult respiratory distress syndrome (ARDS), or due to infection-triggered disseminated intravascular coagulation despite prompt empiric antibiotic therapy including vancomycin. The other patients improved after empiric supplementation of penicillin G (30 Mega/day) to the antibiotic regimen. Beginning ARDS in two of these patients dramatically responded to high-dose steroids. We conclude that St. mitis is a major pathogen in neutropenic leukemic patients. Infection appears to occur independently of acute leukemic cell type, regimen of selective gut decontamination, venous access, visible oral lesions or treatment with high-dose cytosine arabinoside. The clinical course of our patients raises questions about the value of commonly recommended empiric antibiotic regimens, which were clearly ineffective to control infections with St. mitis in this patient group. Our data indicate that immediate antibiotic therapy with penicillin G is indicated and may be life-saving for suspected St. mitis infections in neutropenic leukemic patients.

Granulocyte colony-stimulating factor (G-CSF) treatment in a neutropenic leukemia patient with diffuse interstitial pulmonary infiltrates.

Adult respiratory distress syndrome (ARDS) in patients suffering from acute leukemia usually occurs during chemotherapy-induced neutropenia. In addition, intensified chemotherapy with high-dose cytosine arabinoside and mediastinal irradiation may contribute to the development of ARDS. This complication is usually refractory to conservative treatment with antibiotics, steroids, and mechanical ventilation. In this report, we describe a 25-year-old patient with acute lymphoblastic leukemia who developed ARDS during the phase of chemotherapy-induced neutropenia. Subcutaneous administration of granulocyte colony-stimulating factor (G-CSF) at doses of 300-600 micrograms/day led to a prompt increase of peripheral granulocyte counts. With resolution of neutropenia, respiratory function gradually improved, and mechanical ventilatory support was stopped after 2 weeks. From this observation we surmise that the application of G-CSF may be an effective therapeutic approach for preventing the fatal outcome of ARDS in leukemia patients with bone marrow aplasia.

Pulmonary toxicity during infusion of liposomal amphotericin B in two patients with acute leukemia.

Pulmonary toxicity with acute dyspnea occurred during infusion of a liposomal amphotericin B preparation (AmBisome) in two adult leukemic patients. The preparation was administered as a one hour infusion at a dose of 3 mg/kg body weight. Within 15 min after starting the infusion, both patients experienced sudden onset of dyspnea and chest tightness. Physical examination showed the patients to be anxious and restless with tachycardia and orthopnea but without other cardiopulmonary findings. No elevation of body temperature, rigors or chills were recorded. Symptoms disappeared within minutes after discontinuing the infusion. At present, the pathophysiologic mechanisms underlying these side effects are unknown.

Infusion-related toxicity of three different amphotericin B formulations and its relation to cytokine plasma levels.

A prospective study was performed to compare the infusion-associated toxicity of three different amphotericin B preparations and to correlate acute side-effects with plasma levels of tumour necrosis factor-alpha (TNF-alpha), interleukin-6 (IL-6) and interleukin-1 receptor antagonist (IL-1-RA) during and after the infusions. Six adult neutropenic patients with acute leukaemia suffering from suspected or documented systemic fungal infections were treated on three consecutive days with conventional amphotericin B (AmB), liposomal AmB (AmBisome) and AmB mixed in lipid emulsion (AmB/lipid). Drugs were given over 1-2 h. Drug-induced toxicity was monitored every 30 min for 4 h. Plasma levels of the three cytokines were determined using commercially available enzyme-linked immunosorbent assay (ELISA) techniques. Four of six patients showed toxicity after AmB and AmB/lipid infusions; only one patient reacted to liposomal AmB. Clinical toxicity was associated with increases in TNF-alpha plasma levels during two of four infusions of AmB and three of four infusions of AmB/lipid. Major increases in IL-6 occurred during three of four infusions of AmB and during all four AmB/lipid infusions associated with clinical toxicity. Three of four AmB infusions and all four AmB/lipid infusions accompanied by clinical toxicity were associated with major increases in IL-1-RA plasma concentrations. Liposomal AmB was better tolerated than AmB and AmB/lipid. This formulation also caused the lowest liberation of all three cytokines tested. The severity of clinical symptoms did not correlate closely with absolute cytokine plasma levels. The findings provide further evidence that expression of TNF-alpha, IL-6 and IL-1-RA plays an important role in mediating AmB-related acute toxicity in vivo.