Screening for fall risk in patients with haemophilia.

Many risk factors for falls identified in the general population are found in patients with haemophilia. Furthermore, fall risk increases with age and patients with haemophilia are increasingly entering the over 65 age group. After a fall occurs, there are often behavioural changes that have significant health consequences and further increase fall risk. Fall risk can be quickly assessed in the clinical setting with specific questions in the medical history and by a variety of performance-based screening tools. Identification of fall risk enables early intervention, thereby preventing injury and fear of physical activity, both of which have been associated with falling and may carry an increased risk in patients with haemophilia. Review of the existing literature on assessment of fall risk reveals the importance of screening in the clinical setting, which is commonly done via a fall history and performance-based assessment tools. Selecting appropriate fall risk screening tools is an important step in identifying and providing optimal interventions for those at risk. Assessments of fall history, fear of falling, gait velocity, gait variability and vestibular dysfunction are suggested as screening tools for patients with haemophilia. Additional research is needed to determine the optimal screening, evaluation and treatment techniques for these patients. The longitudinal physical therapy care provided by Haemophilia Treatment Centres presents a unique opportunity for instituting measures that will reduce the incidence of falling in patients with haemophilia.

Msh2 deficiency prevents in vivo somatic instability of the CAG repeat in Huntington disease transgenic mice.

Huntington disease (HD), an autosomal dominant, progressive neurodegenerative disorder, is caused by an expanded CAG repeat sequence leading to an increase in the number of glutamine residues in the encoded protein. The normal CAG repeat range is 5-36, whereas 38 or more repeats are found in the diseased state; the severity of disease is roughly proportional to the number of CAG repeats. HD shows anticipation, in which subsequent generations display earlier disease onsets due to intergenerational repeat expansion. For longer repeat lengths, somatic instability of the repeat size has been observed both in human cases at autopsy and in transgenic mouse models containing either a genomic fragment of human HD exon 1 (ref. 9) or an expanded repeat inserted into the endogenous mouse gene Hdh (ref. 10). With increasing repeat number, the protein changes conformation and becomes increasingly prone to aggregation, suggesting important functional correlations between repeat length and pathology. Because dinucleotide repeat instability is known to increase when the mismatch repair enzyme MSH2 is missing, we examined instability of the HD CAG repeat by crossing transgenic mice carrying exon 1 of human HD (ref. 16) with Msh2-/- mice. Our results show that Msh2 is required for somatic instability of the CAG repeat.

Quantitative trait loci analysis affecting contextual conditioning in mice.

In an extensive backcross of mice between C3H/HeJ (C3H) and C57BL/6J (B6), we sought to map genes that influence learning and memory as measured by performance in a contextual fear-conditioning paradigm. Our results indicate that there are several genetic regions that have a strong influence on performance in this paradigm. The strongest influences map to the proximal and distal ends of chromosome 1 (lod scores of 5.14 and 4.76, respectively). Other chromosomal regions (chromosomes 3, 7, 8, 9 and 18) were also identified as candidates for regions containing genes influencing contextual fear conditioning, with lod scores ranging from 1.8 to 2.7.

Qat-4 and Qat-5, new murine T-cell antigens governed by the Tla region and identified by monoclonal antibodies.

Two new lymphocyte antigens, provisionally designated Qat-4 and Qat-5 have been identified with two different hybridoma-derived, monoclonal AKR antiC57BL/6 antibodies. These antigens are governed by genes located to the right (distal) end of the H-2 complex, within the Qa-2,3 region. Qat-4 and Qat-5 antigens which do not seem to be identical with Qa-2,3 or TL antigens are absent from Ig/ lymphocytes and thymocytes. They are only present on a fraction of peripheral T cells. Thus, Qat-4 is expressed on 70%, and Qat-5 on 30% of splenic and lymph node T cells, Qat-4 is also found on the majority of Ig- cells from athymic nude mice. These findings illustrate the complexity of the chromosome segment between the H-2D and Tla loci and they emphasize the role of major histocompatibility complex-associated genes for the differentiation of T cells into different subpopulations with possibly distinct immunologic functions.

H-Y antigen. Cell surface mapping and testosterone-induced supramolecular repatterning.

Previous work with the antibody-blocking technique showed that the map of surface components for thymocytes prefixed with paraformaldehyde is the same as the map for unfixed thymocytes, with the following exception: after exposure to anti-TL or anti-Db, TL and H-2Db occupy adjacent positions on unfixed cells but not on fixed cells. This was interpreted as an indication that activation of particular components of the surface phenotype initiates ordered changes in the display of cell-surface molecules, approximation of TL and Db in this instance. These studies have now been extended to the H-Y component on the surface of male cells. On fixed male mouse thymocytes, H-Y lies adjacent to TL and relatively distant from H-2Db, H-2Kb, H-2Lb, Lyt-1.2, and Lyt-2.2. However, on unfixed male mouse thymocytes, similarly exposed to H-Y antibody, H-Y and H-2Db are adjacent. Presumably, this engagement of H-Y sites by H-Y antibody brings H-Y and H-2Db together. Evidence that this change in pattern may be physiologically relevant comes from the finding that testosterone, but not estradiol, caused the same selective approximation of H-Y and H-2Db.

Serologic cross-reactivity between Class I MHC molecules and an H-2-linked differentiation antigen as detected by monoclonal antibodies.

Analysis of anti-Class I major histocompatibility complex (MHC) monoclonal antibodies by immunofluorescence and flow microfluorometry demonstrated an unexpected cross-reactivity. Two of fifteen antibodies examined (20-8-4, anti-Kb,Kd,r,s and 34-1-2, antiKd,Dd,Kb,r,s,q,p) were observed to detect an antigen determined by gene(s) mapping to the right of H-2D. Two-color immunofluorescence analysis demonstrated that this antigen, unlike classical H-2K and D antigens, was expressed in high amounts on peripheral T cells, but only weakly on Ia-positive cells and on small subpopulations of thymus and bone marrow cells. Mapping, absorption, blocking, and tissue distribution studies suggested that the cross-reactive antigen is Qa-like, but distinct from previously described Qa antigens. Thus, these data demonstrate serological cross-reactivity between a Class I MHC antigen and a differentiation antigen determined by genes linked to H-2. It seems likely that the gene responsible for this new antigen is one of the numerous Class I-like sequences detected by DNA hybridization analyses, but previously undefined in terms of tissue expression. These data suggest that many of these DNA sequences may be expressed in specific tissues and that cross-reactions of anti-Class I MAbs may provide useful probes for studying the products of such homologous genes.

The Qa2 subregion controls the expression of two antigens recognized by H-2-unrestricted cytotoxic T cells.

B6.KI mice were immunized with spleen cells from B6.K2, a Qa2-subregion congenic strain. Cytotoxic T cells were generated that recognize two target antigens controlled by this region. One of the target antigens is Qa-2. This was demonstrated by the findings that pretreatment of target cells with monoclonal anti-Qa-2 antibody blocked lysis of target cells, and Qa-2 target antigens and serological determinants had a concordant distribution on a panel of B10.W (wild) mice. The gene controlling the Qa-2 target antigen is not polymorphic because B6.K2 and three strains of Qa-2(+) B10.W mice express the same antigens, as determined by a CTL cold target competition assay. Anti-Qa-2 CTL were H-2 unrestricted because effector cells lysed Qa-2(+) targets irrespective of their H-2 haplotype, including five B 10.W strains, and lysis was not inhibited by pretreating target cells with anti-H-2 sera. The Qa2 subregion does not act as a restricting locus for anti-minor-H antigen CTL. A second target antigen was detected that was associated with the expression of the Qa-5 determinant. However, CTL activity could not be blocked by pretreating target cells with monoclonal anti-Qa-5. Therefore, the CTL target antigen may be expressed on a Qa-5(-) molecule. Although the Qa-5 associated CTL specificity is only detected on H-2D(b) strains, it is unlikely that CTL recognition is H-2 restricted because anti-H-2(b) sera has no effect in blocking this reactivity. Qa-2 and H-2 class I antigens share a similar structure and serve as target antigens for unrestricted CTL. However, unlike class I H-2 genes, Qa-2 neither restricts antigen-specific CTL nor is polymorphie. Therefore, it is likely that Qa-2 and H-2 are derived from a common ancestral gene and have evolved to serve different functions.

Qa-2, H-2K, and H-2D alloantigens evolved from a common ancestral gene.

The Tla region located on the murine 17th chromosome controls several serologically defined cell surface antigens. These antigens, referred to as Qa-1-5 and TL, are expressed on a variety of hematopoeitic cell populations. In the present studies we have immunoprecipitated isotopically labeled Qa-2 and H-2 molecules from mitogen-stimulated B6 spleen cells. Sequential immunoprecipitation experiments have shown that the determinants recognized by alpha Qa-2, alpha H-2Kb, and alpha H-2Db alloantisera reside on separate molecular species. Comparative mapping of the arginine-labeled tryptic peptides from Qa-2, H-2Kb, and H-2Db molecules indicate that Qa-2 is structurally distinct but that there is considerable structural homology; 21-43% of the Qa-2 peptides co-chromatograph with peptides derived from H-2Db and H-2Kb, respectively. Similar levels of homology are observed when Qa-2 is compared with H-2Kk or H-2Dd. The results show that the Qa-2 alloantigen is encoded by a locus separate from the loci encoding H-2K or H-2D alloantigens, but that the Qa-2, H-2K, and H-2D alloantigens are sufficiently related at the primary structural level to indicate that they evolved from a common primordial gene.

Molecular similarities between the Qa-2 alloantigen and other gene products of the 17th chromosome of the mouse.

The alloantigen Qa-2, whose gene is located on the 17th chromosome between H-2D and Tla, is identified as a molecule of 43,000 daltons which is associated with beta 2-microglobulin. Qa-2 comprises approximately 0.15% of the iodinateable cell surface protein of lymph node cells. Sequential precipitations demonstrated that Qa-2 is distinct from H-2D and H-2K molecules.

Nonconventional (TL-encoded) major histocompatibility complex molecules present processed viral antigen to cytotoxic T lymphocytes.

A large number of class I-like genes are located distal to the K and D regions of the murine major histocompatibility complex (MHC) within the Q and TL region. The function of the molecules encoded within this region is obscure since unlike conventional MHC gene products, these molecules have not been reported to present processed environmental antigens to T cells. In the present report, we demonstrate that a peptide corresponding to processed influenza virus hemagglutinin can be recognized by CD8+ T cell receptor alpha/beta-positive cytotoxic T lymphocytes (CTL) in association with a MHC class I-like product encoded within the TL region. Thus, nonconventional class I MHC molecules can bind and present processed environmental antigens, and TCR-alpha/beta CTL directed to such peptide MHC complexes are represented in the mature T cell pool. Our data imply that Q/TL region products may be charged by peptides generated through an antigen processing and presentation pathway distinct from the pathway used by conventional MHC molecules and not normally available to environmental antigens.

Mutation in a new H-2-associated histocompatibility gene closely linked to H-2D.

Sequential precipitations of soluble BALB/c antigen with antisera detecting private and public H-2 specificities indicated three distinct classes of molecules of 45,000 mol wt. However, only two of these classes of molecules were detectable in antigen from the loss mutant, BALB/c-H-2db. The class of molecules, detectable in the wild-type strain but missing in the mutant, does not bear private specificities but does react with an antiserum detecting H-2 public specificities. Absorption in mutant mice of the antiserum to public specificities, left antibodies specific for the antigen detectable in BALB/c but not BALB/c-H-2db. Genetic mapping studies using this specific antiserum indicated that the antigenic loss of this mutant is in a gene which maps in or close to the H-2D region, separable from the H-2K, S, G, Qa-2, and Tla regions.

Antigen-specific soluble helper activity for murine major histocompatibility complex-encoded molecules. I. Kinetics of factor production after skin transplantation and genetic mapping of the H-2 region specificity.

Antigen-specific soluble helper molecules are produced during major histocompatibility complex-disparate allograft priming. Genetic mapping studies with appropriate recombinant and mutant lines of mice have defined the antigen specificities of the soluble helper molecules described here as being directed against the H-2Dd molecules. The production of antigen-specific helper molecules is a relatively early event after H-2Dd-region allograft priming. A later phase of factor production near the time of graft rejection also contains nonspecific helper factors and IL-2.

Ir-genes in H-2 regulate generation of anti-viral cytotoxic T cells. Mapping to K or D and dominance of unresponsiveness.

H-2 dependent and virus-specific Ir genes regulate the generation of primary virus-specific K or D restricted cytotoxic T-cell responses in vivo. The following examples have been analyzed in some detail: first, Dk restricted responses to vaccinia in Sendai viruses are at least 30 times lower than the corresponding K-restricted responses irrespective of the H-2 haplotypes (k, b, d, dxs, dxq) of K and I regions; in contrast, LCMV infection generates high responses to Dk. These findings are consistent with but do not prove that this Ir gene maps to D. Second, Db restricted responses to vaccinia and Sendai viruses are high in strains possessing the Kq or KbIb, KbaIb haplotype, are very low in strains with Kk, and relatively low in mouse strains of the KdI-Ad haplotype; LCMV generates high Db restricted response in the presence of Kk. This Ir gene for the response to vaccinia and Sendai viruses maps to K since B10.BYR (KqIkdDb) is a responder and B10.A (2R) is a nonresponder (KkIkdDb). Third, virus and K or D allele specific nonresponsiveness is dominant with variable penetrance; in heterozygous mice the nonresponder Kk allele over-rides responsiveness normally found in KbDb or KqDb combinations. Fourth, when (responder X nonresponder)F1 lymphocytes are stimulated in an environment expressing vaccinia virus plus only a high responder Kb or Kq allelle and Db, response to vaccinia Db is high; in contrast when the same F1 cells are stimulated in an environment expressing the low responder allele Kk, response to vaccinia Db is low. Thus absence of Kk during immunization allows generation of high responsive Db restricted vaccinia specific cytotoxic T cells. The Dk dependent low response to vaccinia Dk can be explained by a preclusion rule or by failure of vaccinia to complex with Db; however the analysis of Kk dependent low response to vaccinia Db does not support these explanations or that self-tolerance is responsible for this Ir effect but is compatible with the interpretation that Kk vaccinia is immunodominant over Db vaccinia. These results are discussed with respect to (a) possible mechanisms of regulation by Ir genes and (b) H-2 polymorphism and HLA-disease association.

Localization of two new X-linked quantitative trait loci controlling corpus callosum size in the mouse.

Corpus callosum (CC) size is a complex trait, characterized by a gradation of values within a normal range, as well as abnormalities that include a small or totally absent CC. Among inbred mouse strains with defects of the CC, BTBR T(+)tf/J (BTBR) mice have the most extreme phenotype; all animals show total absence of the CC and severe reduction of the hippocampal commissure (HC). In contrast, the BALB/cByJ (BALB) strain has a low frequency of small CC and consistently normal HC. Reciprocal F(1) crosses between BTBR and BALB suggest the presence of X-linked quantitative trait loci (QTLs) affecting CC size. Through linkage analysis of backcross male progeny, we have localized two regions on the X chromosome, having peaks at 68.5 Mb (approximately 29.5 cM) and at 134.5 Mb (approximately 60.5 cM) that are largely responsible for the reciprocal differences, with the BTBR allele showing X-linked dominant inheritance associated with CC defects.

Phase II trial of paclitaxel and carboplatin in the treatment of advanced urothelial carcinoma.

PURPOSE: Both paclitaxel and carboplatin have single-agent activity against carcinoma of the urothelium. We evaluated the combination of paclitaxel and carboplatin in the treatment of advanced cancers of the urothelium. PATIENTS AND METHODS: Patients with cancers of the urothelium who had no prior chemotherapy (prior adjuvant chemotherapy > 6 months allowed) were eligible for treatment. Eligibility requirements were performance status of 2 or less, creatinine level less than 2.0 mg/dL, granulocyte count (AGC) 1,500/microL or greater, platelet count 100,000/microL or greater, and total bilirubin level less than 1.5 mg/dL. Paclitaxel 200 mg/m2 followed by carboplatin (area under the curve [AUC] 5, Calvert formula) were administered every 21 days. Patients were evaluated for toxicity weekly and assessed for response every 6 weeks. RESULTS: Thirty-six patients were entered onto the study and 35 patients were assessable for response. A total of 184 cycles were administered (median, six cycles per patient). Nine patients required one dose reduction, and seven patients required two dose reductions for a nadir AGC less than 500/microL, with only one episode of febrile neutropenia and sepsis. Myalgias and arthralgias of grades 1 to 2 occurred in 16 patients and usually lasted 2 to 3 days after treatment. There were no treatment delays because of toxicity. There were 18 responses; seven complete responses (CRs) and 11 partial responses (PRs) (response rate 51.5%; 95% confidence interval, 35 to 68). Median response durations for CR and PR were 6 and 4 months, respectively. Overall median survival was 9.5 months. CONCLUSION: The combination of paclitaxel and carboplatin is an active and well-tolerated regimen for the treatment of advanced urothelial carcinoma. Because of the modest toxicity of this combination, paclitaxel and carboplatin should be considered for addition to other agents with activity in urothelial carcinomas.

Phase II evaluation of oral estramustine and oral etoposide in hormone-refractory adenocarcinoma of the prostate.

PURPOSE: Estramustine and etoposide (VP-16) have been demonstrated to inhibit the growth of prostate cancer cells in experimental models. This led us to evaluate the effectiveness of this combination in the treatment of patients with metastatic prostate carcinoma refractory to hormone therapy. PATIENTS AND METHODS: Estramustine 15 mg/kg/d and VP-16 50 mg/m2/d, were administered orally in divided doses for 21 days. Patients were then taken off therapy for 7 days and the cycle then repeated. Therapy continued until evidence of disease progression. RESULTS: Forty-two patients have been enrolled onto this trial with a minimum of 40 weeks follow-up. Of 18 patients with measurable soft tissue disease, three demonstrated a complete response (CR) and six a partial response (PR) for longer than 2 months. Of these 18 patients, pretreatment prostate-specific antigen (PSA) levels decreased by at least 75% in five men (28%) and by at least 50% in nine (50%). The median survival duration has not been reached in those patients who demonstrated a response either by soft tissue or PSA criteria. Of 24 patients with disease limited to bone, six (25%) demonstrated improvement and nine (38%) demonstrated stability in their bone scans. Five men (21%) demonstrated a decrease of at least 75% in pretreatment PSA levels and 14 (58%) demonstrated at least a 50% decrease; the median survival duration has not been reached in these patients. Pretreatment performance status is an important predictor of survival. CONCLUSION: We conclude that the combination of estramustine and VP-16 is an active oral regimen in hormone-refractory prostate cancer.

Implications of microscopic satellites of the primary and extracapsular lymph node spread in patients with high-risk melanoma: pathologic corollary of Eastern Cooperative Oncology Group Trial E1690.

PURPOSE: To correlate the presence of extracapsular spread (ECS) of regional nodal metastases, and micrometastasis near the primary tumor, with disease outcome in the intergroup study E1690 in relation to the impact of recombinant interferon-alfa (rIFN alpha)-2b. PATIENTS AND METHODS: E1690 included 642 patients with American Joint Committee on Cancer stage IIB or III cutaneous melanoma. Patients were randomized into high- and low-dose rIFN alpha-2b treatment arms and an observation arm. Pathologic slides were reviewed for selected parameters from at least half of the subjects in all three arms. Evaluation of the primary tumor included notations regarding ulceration, mitotic activity, thickness, microscopic satellites (MS), and nodal ECS on a standardized pathology form. These data were collated in relation to relapse-free survival (RFS) and overall survival (OS) at 50 months' follow-up and studied using Cox regression analysis. RESULTS: Ulceration, mitotic activity, thickness, and size of tumor-bearing lymph nodes did not show a statistically significant correlation with either OS or RFS across all treatment arms. The presence of MS was correlated with RFS (P =.0008) and OS (P =.05). ECS correlated with RFS (hazard ratio = 1.44, P =.032) but not OS (P =.11). CONCLUSION: The presence of MS (in 6% [18 of 308 patients]) had a significant adverse impact on both RFS (P =.0008) and OS (P =.053). Ulceration, mitotic activity, thickness, and number of positive lymph nodes had no significant effect on OS in this subset study (univariate or multivariate Cox analysis). The presence of ECS in lymph nodes had a significant adverse effect on RFS (P =.032) but not on OS.

High- and low-dose interferon alfa-2b in high-risk melanoma: first analysis of intergroup trial E1690/S9111/C9190.

PURPOSE: Pivotal trial E1684 of adjuvant high-dose interferon alfa-2b (IFNalpha2b) therapy in high-risk melanoma patients demonstrated a significant relapse-free and overall survival (RFS and OS) benefit compared with observation (Obs). PATIENTS AND METHODS: A prospective, randomized, three-arm, intergroup trial evaluated the efficacy of high-dose IFNalpha2b (HDI) for 1 year and low-dose IFNalpha2b (LDI) for 2 years versus Obs in high-risk (stage IIB and III) melanoma with RFS and OS end points. RESULTS: A total of 642 patients were enrolled (608 patients eligible), of whom a majority (75%) had nodal metastasis (50% had nodal recurrence). Unlike E1684, E1690 allowed entry of patients with T4 (> 4 mm) deep primary tumors, regardless of nodal dissection, and 25% of the patients entered onto this trial had deep primary tumors (compared with 11% in E1684). At 52 months' median follow-up, HDI demonstrated an RFS benefit exceeding that of LDI compared with Obs. The 5-year estimated RFS rates for the HDI, LDI, and Obs arms were 44%, 40%, and 35%, respectively. The hazards ratio for the intent-to-treat analysis of HDI versus Obs was 1.28 (P(2) =.05); for LDI versus Obs, it was 1.19 (P(2) =.17). By Cox analysis, the impact of HDI on RFS achieved significance (P(2) =.03). The RFS benefit was equivalent for node-negative and node-positive patients. Neither HDI nor LDI has demonstrated an OS benefit compared with Obs at this time. A major improvement in the median OS of patients in the E1690 Obs arm was noted in comparison with E1684 (6 years v 2.8 years). An analysis of salvage therapy for patients who relapsed on E1690 demonstrated that a significantly larger proportion of patients in the Obs arm received IFNalpha-containing salvage therapy compared with the HDI arm; this therapy was unavailable to patients during E1684, and patients with undissected regional nodes were not included in E1684. This study did not specify therapy at recurrence. Analysis of treatments received at recurrence demonstrated significantly more frequent use of IFNalpha2b at relapse from Obs than from HDI, which may have confounded interpretation of the survival benefit of assigned treatments in E1690. CONCLUSION: The results of the intergroup E1690 trial demonstrate an RFS benefit of IFNalpha2b that is dose-dependent and significant for HDI by Cox multivariable analysis.

A phase I trial of recombinant interleukin-2 combined with recombinant interferon-gamma in patients with cancer.

Twenty-six patients with metastatic cancer were entered into a phase I trial of concurrent recombinant interleukin-2 (IL-2) and recombinant interferon-gamma (IFN-gamma). IL-2 was administered as a continuous intravenous infusion for 5 days. IFN-gamma was administered by a daily intramuscular (IM) injection during the 5 days of IL-2 administration. Treatment was repeated twice after 9-day rest periods. After a 2-week rest, patients without evidence of tumor progression were retreated. Natural killer (NK)- and lymphokine-activated killer (LAK)-cell activity were assayed in each patient before treatment, on day 1, and on day 5 of each cycle. Constitutional symptoms occurred in most patients but were not dose-limiting. Other toxicities included hypotension responsive to fluids, transient elevations in liver function tests, erythema/pruritus, eosinophilia, and transient leukopenia/thrombocytopenia. The maximum-tolerated dose (MTD) of the combination was 1 x 10(6) U/m2/d of IL-2 combined with 0.50 mg/m2/d of IFN-gamma. The dose-limiting toxicity was pulmonary manifesting as rales and shortness of breath. The dose of the combination that resulted in the optimal generation of in vivo LAK-cell activity was a dose of at least 0.25 mg/m2/d of IFN-gamma combined with 1 x 10(6) U/m2/d of IL-2. Objective clinical responses were seen in five of 26 patients. These included a partial response of 2 months duration in a patient with non-Hodgkin's lymphoma (NHL), mixed responses in a patient with NHL and two patients with renal cell carcinoma (RCC), and an ongoing assessable response in a patient with bone metastases from RCC. The recommended dose for phase II trials of this combination is 0.50 mg/m2 of IFN-gamma and 1 x 10(6) U of IL-2.

Outpatient biochemotherapy with interleukin-2 and interferon alfa-2b in patients with metastatic malignant melanoma: results of two phase II cytokine working group trials.

PURPOSE: The Cytokine Working Group performed a randomized phase II trial of two outpatient biochemotherapy regimens to identify an outpatient regimen with high antitumor activity and less toxicity than inpatient regimens which might be compared with chemotherapy or inpatient biochemotherapy regimens in future phase III trials. PATIENTS AND METHODS: Eighty-one patients with metastatic malignant melanoma received dacarbazine 250 mg/m(2)/d intravenously (IV) and cisplatin 25 mg/m(2)/d IV on days 1, 2, and 3, plus interferon (IFN) alfa-2b 5 mU/m(2) subcutaneously (SC) on days 6, 8, 10, 13, and 15, given every 28 days. Interleukin-2 (IL-2) was given daily on days 6 to 10 and 13 to 15. In group 1, IV IL-2 was given at 18.0 MU/m(2), and in group 2, SC IL-2 was given at 5.0 mU/m(2). RESULTS: In group 1 (IV IL-2), there were five complete responses (CRs) and 11 partial responses (PRs) among 44 patients (objective response rate [ORR], 36%; 95% confidence interval [CI], 22% to 51%). In group 2 (SC IL-2), there was one CR and five PRs among the 36 patients (ORR, 17%; 95% CI, 4% to 29%). The median survival was 10.7 months in group 1 and 7.3 months in group 2. Eleven patients in group 1 and four patients in group 2 remain alive as of the last follow-up. Toxicities in both groups were similar. No patient required hospitalization for neutropenic fever. CONCLUSION: Biochemotherapy has activity in these outpatient regimens with acceptable toxicity. The antitumor activity observed with the IV IL-2 regimen seems similar to that of inpatient biochemotherapy regimens. If inpatient biochemotherapy regimens develop an established role in the management of melanoma, future phase III trial comparisons with this outpatient IV IL-2 regimen would be appropriate.