Adherence and persistence with teriparatide among patients with commercial, Medicare, and Medicaid insurance.

UNLABELLED: Adherence to, and persistence with, treatments for osteoporosis are low. Adherence with teriparatide decreases over time. Higher copayments in the commercial/Medicare population were associated with worse persistence. Understanding factors such as prior screening, prior treatment history, and out of pocket costs that influence persistence with teriparatide may help clinicians make informed decisions. INTRODUCTION: The purpose of this study was to evaluate adherence and persistence with teriparatide. METHODS: Beneficiaries with at least one claim for teriparatide in 2003 or 2004 and continuous enrollment in the previous 12 months and subsequent 6 months were identified in a national commercial/Medicare and Medicaid administrative claims database (MarketScan®). Adherence was assessed through calculation of the medication possession ratio (MPR). Persistence was measured by time until discontinuation and time until first 60-day gap in treatment. Factors associated with persistence were assessed using Cox proportional hazards models. RESULTS: The average MPR at 6 months was 0.74 (N=2,218) and at 12 months, was 0.66 (N=1,303). At 6 months, 64.6% of patients remained on therapy and at 12 months, 56.7% remained. Bone mineral density screening and use of antiresorptive therapy within the 12 months pre-period, and lower patient copayments were associated with increased persistence. CONCLUSION: Patients appear to have good adherence with teriparatide over the first 6 months which declines over time. Prior screening and treatment of osteoporosis and out of pocket costs appear to impact persistence. To optimize patient outcomes, clinicians should consider clinical factors that impact persistence, while healthcare decision makers should consider the negative effect of higher patient copayments on persistence.

Relationship between duration of teriparatide therapy and clinical outcomes in postmenopausal women with osteoporosis.

SUMMARY: The extent to which fracture protection and safety varies with increasing time on teriparatide [rhPTH(1-34)] therapy is a clinically relevant unanswered question. In postmenopausal women with osteoporosis, increased duration of teriparatide versus placebo treatment was associated with a progressive decrease in the rates of nonvertebral fragility fractures and back pain. INTRODUCTION: The impact of duration of teriparatide [rhPTH(1-34)] therapy on patient outcomes is a relevant unanswered question. METHODS: Postmenopausal women with osteoporosis were randomized to once-daily subcutaneous injection with placebo (N = 544), teriparatide 20 microg (TPTD20; N = 541), or teriparatide 40 microg (TPTD40; N = 552) plus calcium and vitamin D supplementation. The time to first nonvertebral fragility fracture and new or worsening back pain following treatment initiation was analyzed using Cox partial likelihood regression treating time on therapy as a linear, time-dependent covariate. RESULTS: Compared with placebo, the relative hazard for nonvertebral fragility fractures decreased by 7.3% for each additional month of TPTD20 [hazard ratio = 0.927, 95% CI (0.876 to 0.982), p = 0.009] and by 7.6% for each additional month of TPTD40 [hazard ratio = 0.924, 95% CI (0.871 to 0.981), p = 0.009]. Clinical vertebral fractures appeared to increase over time in the placebo group and occurred primarily in the first time interval in the teriparatide treatment groups. Compared with placebo, the relative hazard of back pain was decreased by 8.3% for each additional month of TPTD20 [hazard ratio = 0.920, 95% CI (0.902 to 0.939), p < 0.001] and 8.7% for each additional month of TPTD40 [hazard ratio = 0.917, 95% CI (0.898 to 0.935), p < 0.001]. CONCLUSIONS: These findings suggest increased nonvertebral fracture protection, reduced back pain, and reduced occurrence of side effects with longer duration of teriparatide therapy.

Plasminogen activator inhibitor type-1 interacts exclusively with the proteinase domain of tissue plasminogen activator.

Two different techniques have been used to study the complex formation of recombinant human plasminogen activator inhibitor type-1, PAI-1, with either recombinant human two-chain tissue plasminogen activator, tc tPA (EC 3.4.21.68), or the tPA deletion variants tc K2P, containing the kringle 2 domain and the proteinase domain, and P, containing only the proteinase domain. The same value for Kon, 2.10(7) M-1s-1 for binding of PAI-1 was found for the three tPA forms by direct detection of the complex formation in real time by surface plasmon resonance, BIAcore, or indirectly by monitoring the time course of the inhibition of tPA using the chromogenic substrate N-methylsulfonyl-D-Phe-Gly-Arg-4-pNA-acetate. Apparently, no conformational change is involved in the rate-limiting step, since the kon value was found to be independent of the temperature from 20 to 35 degrees C. By the BIAcore technique, it was found that the complex between PAI-1 and tPA covalently coupled to the surface, was stable at 25 degrees C, since no dissociation was seen in buffer. However, extended treatment with 1 M NH4OH destroyed the complex with t 1/2 = 5 h. The same kon values and complex composition were found by measuring either the binding of tPA to PAI-1 captured on the monoclonal antibody MAI-11 or the binding of PAI-1 to tPA captured on the monoclonal antibody 2:2 B10. Quantification of the complex composition between PAI-1 captured on the monoclonal antibody MAI-11 with either tPA, K2P or P gave a one-to-one ratio with the fraction of active PAI-1, consistent with the results from SDS-PAGE and the specific activity of PAI-1. The complexes of the three tPA forms with PAI-1 captured on a large surface of MAI-11 dissociated more rapidly from MAI-11, with the same apparent koff, kdis, = 2.10(-3) s-1, compared with 0.7-10(-3) s-1 for the dissociation of PAI-1 alone. In consistance, the Kd, calculated from the direct determination of the kon and koff for the association of different form of PAI-1 to a small surface of MAI-11, was found to be higher for PAI-1 in complex with tPA than for free active PAI-1. Apparently, upon complex formation, a change is induced in PAI-1 at the binding epitope for MAI-11.(ABSTRACT TRUNCATED AT 400 WORDS)

Sustained reflow in dogs with coronary thrombosis with K2P, a novel mutant of tissue-plasminogen activator.

Coronary artery reocclusion after thrombolysis with human recombinant tissue-type plasminogen activator (rt-PA) is related to the short half-life of this agent in plasma. K2P, a mutant of rt-PA lacking the fibronectin fingerlike, epidermal growth factor-like and first kringle domains (amino acids 6 to 173) and having the glycosylation site Asn184 mutagenized to Gln, has been produced in Chinese hamster ovary cells. In this study we compared the thrombolytic effect of K2P and rt-PA in dogs with electrically induced coronary artery thrombosis. Both agents were given intravenously in equimolar amounts over 20 min after the occlusive thrombus was stable for 30 min; dogs were monitored for 1 h after reperfusion if flow occurred. Coronary blood flow was restored by rt-PA in 6 (60%) of 10 dogs. The restored flow lasted for 49 +/- 12 min and mean flow at 60 min from the start of reperfusion was 7 +/- 3 ml/min. The reocclusion rate was 50% (three of six dogs). Flow was restored in five (100%) of five dogs by K2P. The restored blood flow lasted during the entire 1-h observation period in all but one dog and mean flow at 60 min was 49 +/- 16 ml/min (p less than 0.02 vs. flow in rt-PA-treated dogs). Restored coronary blood flow showed marked cyclic flow variations in rt-PA-treated but not in K2P-treated dogs.(ABSTRACT TRUNCATED AT 250 WORDS)

Drug resistance factors in acute myeloid leukemia: a comparative analysis.

To compare the clinical relevance of drug resistance factors in de novo acute myeloid leukemia (AML), we determined their relationship to both response to induction chemotherapy and survival of the patients in univariate as well as multivariate analyses. The drug resistance factors immunocytochemically studied in 111 patients at the time of diagnosis included the lung resistance protein (LRP), P-glycoprotein (P-gp), multidrug resistance protein (MRP1) and bcl-2. In the univariate analyses, age (P = 0.005), karyotype (P = 0.03), LRP (P = 0.003), P-gp (P = 0.02) and bcl-2 (P = 0.03) predicted for response to induction chemotherapy, whereas MRP1 had no predictive value. Age (P = 0.05), karyotype (P = 0.05) and LRP (P = 0.03) retained their predictive value in the multivariate logistic regression analyses. With regard to overall survival, age (P = 0. 008), karyotype (P = 0.006), LRP (P = 0.001) and P-gp (P = 0.01) were of prognostic value in the univariate Cox regression analyses but only age (P = 0.01), karyotype (P = 0.02) and LRP (P = 0.01) retained their prognostic significance in the multivariate analyses. A risk score based on the number of independent prognostic factors allowed division of patients into four groups with different outcome. In these groups, the complete remission rates were 93%, 75%, 47% and 33%, respectively, and median overall survival was 2.4, 1.2, 0.6 and 0.2 years, respectively. Thus, several drug resistance factors did predict outcome in the univariate analyses but LRP was the only drug resistance factor with independent predictive and prognostic significance. The proposed risk score might be useful for risk-adapted treatment in the future. Leukemia (2000) 14, 68-76.

Expression of the lung resistance protein predicts poor outcome in patients with multiple myeloma.

Expression of the lung resistance protein (LRP) is associated with resistance to various anticancer drugs including melphalan and, therefore, may affect the clinical outcome in multiple myeloma (MM). To determine the clinical significance of LRP, we have compared LRP expression in bone marrow plasma cells with clinical parameters including response to chemotherapy and survival of previously untreated patients with MM (n = 72). LRP expression immunocytochemically assessed by means of the LRP-56 monoclonal antibody was positive (> or =10% staining plasma cells) in 44 (61%) samples. There was no correlation between LRP expression and age, sex, type of the paraprotein, serum creatinine, stage, beta2-microglobulin, serum lactate dehydrogenase, or C-reactive protein. However, LRP expression was more frequently observed in patients with a p53 deletion than in those without such a deletion (P = 0.01). The overall response rate for all of the patients evaluable for response to induction chemotherapy (n = 58) was 67%. The response rate was 87% for patients without LRP expression but only 54% for patients with LRP expression (P = 0.01). Kaplan-Meier analysis revealed that patients with LRP expression had a shorter overall survival (median, 33 months) than those without LRP expression (median not reached; P = 0.04). These data show that LRP expression is an important marker for clinical drug resistance and predicts a poor outcome in MM.

Human tissue-type plasminogen activator synthesized by using a baculovirus vector in insect cells compared with human plasminogen activator produced in mouse cells.

A cDNA fragment encoding the human tissue-type plasminogen activator was inserted into the baculovirus Autographa californica nuclear polyhedrosis virus downstream from the polyhedrin promoter. The induction kinetics of t-PA was followed, after infection of Spodoptera frugiperda cells, at both mRNA and protein levels. Fibrinolytically active plasminogen activator accumulated in the culture medium and reached 2.5 micrograms/ml after 120 h. The protein was compared with recombinant plasminogen activator produced in mouse cells and was found to be slightly smaller. This difference in size was found to be caused by N-linked oligosaccharides which are shorter in the recombinant activator obtained from insect cells. The molecules produced in such cells contain at least two different types of N-linked glycans, since only one out of three oligosaccharides is sensitive to endoglycosidase H. However, all glycan structures bind strongly to concanavalin A-Sepharose.

Synthesis and secretion of a fibrinolytically active tissue-type plasminogen activator variant in Escherichia coli.

A gene encoding a variant (lacking amino acids 6-173) of human tissue-type plasminogen activator (t-PA), consisting only of the second kringle domain (K2) and the serine protease domain (P), was fused to a DNA segment coding for the signal peptide of staphylococcal protein A and a synthetic gene coding for a protein with ability to bind immunoglobulin G (IgG). The fusion protein which was synthesized in Escherichia coli and secreted into the growth medium, was found to be fibrinolytically active. Purification of the fusion protein was performed in a single step by affinity chromatography with immobilized IgG. Enzymatically active K2P was liberated from the fusion protein by cleavage at a unique Asn-Gly dipeptide sequence using hydroxylamine. These results demonstrate that a variant of human t-PA can be synthesized and secreted by E. coli as a fibrinolytically active fusion protein, which upon specific cleavage yields an active variant t-PA of the expected size.

Porcine tissue plasminogen activator. Immunoaffinity purification, structural properties and glycosylation pattern.

Tissue plasminogen activator was purified in high yield from pig heart by immunoaffinity chromatography and characterized by analysis of the glycosylation pattern and the N-terminal amino acid sequence. Comparisons with the human enzyme reveals residue exchanges in the A-chain at positions 3 (porcine Arg/human Gln) and 5 (Thr/Ile), and in the B-chain at positions 6 (Tyr/Phe), 10 (Thr/Ala) and 20 (Val/Ala). The glycosylation pattern for the porcine activator was determined by endoglycosidase treatment followed by gel electrophoresis. The A-chain contains a single high-mannose type of N-linked glycan structure and the B-chain contains a complex type of oligosaccharide. A similar but not identical pattern has been observed for the human activator, purified from melanoma cells.

Differential proteolysis and evidence for a residue exchange in tissue plasminogen activator suggest possible association between two types of protein microheterogeneity.

The N-terminal part of native one-chain tissue plasminogen activator from melanoma cells is not homogeneous. The protein chain starts at two different positions, in all probability representing a processing difference in the N-terminus. Both 'long' L-chains and 3-residue shorter S-chains are present in the preparations. In addition, results compatible with a positional Ser/Gly microheterogeneity were obtained at a single position (position L-4 which is equal to S-1). The N-terminal tripeptide difference seems to be coupled to the possible microheterogeneity: L-chains contain Ser in this position, while S-chains appear to contain predominantly Gly.

Differences between uterine and melanoma forms of tissue plasminogen activator.

Tissue plasminogen activator purified from human uterine tissue exhibits differences in N-terminal starting positions in relation to the melanoma cell plasminogen activator usually studied. A new starting position is compatible with an additional N-terminal processing apart from those already known. Like the melanoma activator, the uterine activator was found to yield protein chains starting at either of two positions. One of these was identical between uterine and melanoma activators, whereas the other was unique in each case. The most abundant starting position for the uterine preparation was at a valine residue, apparently from cleavage of a Gln-Val bond, and corresponding to Val-7 of the longest form of the melanoma activator chain.

Synergism between tissue-type plasminogen activator and a genetically engineered variant lacking the finger domain, the growth factor domain and the first kringle domain.

A modified variant of human tissue-type plasminogen activator (t-PA) lacking the finger domain (F), the growth factor domain (G) and the first kringle domain (K1), has an extended plasma half-life in vivo, compared to that of t-PA. When the variant (denoted K2P) was tested in vitro for its ability to lyse human plasma clots we found that the activity was characterized by a time lag phase and a sigmoidal dose-response curve. However, an attenuation of the lag phase in vitro was observed both when K2P was mixed with t-PA in a w/w ratio of 4:1 and when K2P was allowed to lyse a clot that had been pre-exposed to t-PA i.e. submitted to a limited plasmic digestion. Dosis that in vitro caused 50% lysis within 6 h were calculated from individual dose-response curves and were for K2P, t-PA and K2P/t-PA (4:1 w/w) 540 ng/ml, 360 ng/ml and 310 ng/ml, respectively. These results indicated a synergistic effect between K2P and t-PA. However, the data from individual dose-response curves showed that the effect of the K2P/t-PA mixture never was better than that of t-PA alone, and the synergistic effect in vitro is therefore considered to be of limited use. The thrombolytic activity in vivo was evaluated in a rabbit jugular vein thrombus model. Despite the lag phase observed in vitro, K2P was approximately 3 times as effective as t-PA in vivo (bolus injection). The thrombolytic effect of K2P was further potentiated when it was administered together with a small amount of t-PA (4:1 w/w).(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of ipratropium bromide treatment on oxygen saturation and sleep quality in COPD.

STUDY OBJECTIVES: Patients with COPD are at risk of experiencing a deterioration in arterial oxygen saturation (SaO2) during sleep, which is generally most pronounced during rapid eye movement (REM) sleep. Increased cholinergic tone has been suggested as a contributing factor to this decrease in SaO2. Therefore, we investigated whether 4-week treatment with ipratropium bromide inhalation solution 0.02% (qid) could improve sleep characteristics in COPD. DESIGN: Randomized, placebo-controlled, double-blind, two-arm parallel study of 4 weeks of treatment with ipratropium bromide solution or placebo. SETTING: Multicenter investigation. PATIENTS: Thirty-six patients with moderate-to-severe COPD (FEV1 < 65% of predicted). MEASUREMENTS AND RESULTS: Evaluation included polysomnographic, pulmonary function, and subjective quality of sleep (visual analog scale [VAS]) assessments. It was found that 4 week of treatment with ipratropium bromide solution in patients with COPD led to the following: (1) a significant (p = 0.05) improvement in mean nocturnal SaO2 with the more severe the nocturnal desaturation, the greater the improvement in SaO2; (2) significant (p = 0.03) improvement in perceived sleep quality (VAS: 5.5 +/- 0.5 after placebo; 7.2 +/- 0.5 after ipratropium); (3) a significant (p = 0.05) increase in REM sleep time (48.6 +/- 6.3 min after placebo; 66.5 +/- 6.4 min after ipratropium) with no effect on other sleep stages or total sleep time; and (4) a significant (p = 0.01) increase in pre-sleep FVC and flow rate at 50% of the vital capacity. CONCLUSIONS: These findings demonstrate that ipratropium bromide therapy can improve sleep SaO2 as well as sleep quality in patients with moderate-to-severe COPD.

Production of a biologically active variant form of recombinant human secretin.

A biologically active variant form of recombinant human secretin was produced using a gene fusion system designed to facilitate the purification of the protein. The fusion protein was recovered from the culture medium of Escherichia coli by IgG affinity chromatography, and recombinant secretin was released by cyanogen bromide treatment. A novel approach involving addition of a C-terminal Gly-Lys-Arg extension, was used to overcome the lack of amidation of recombinant proteins in Escherichia coli. The biological activity of the recombinant variant of secretin was at least 80% of the porcine secretin standard.

Production of human tissue-type plasminogen activator in different mammalian cell lines using an Epstein-Barr virus vector.

Human tissue-type plasminogen activator (t-PA) cDNA inserted into an Epstein-Barr virus (EBV) derived expression vector was transfected into human HeLa, 293, K-562 and hamster CHO-K1 cells and the expression of t-PA was studied. The best t-PA producing cell clones were found among CHO-K1 cells (up to 11 micrograms d-1 per 10(6) cells). However, HeLa and 293 cells were most efficiently transfected, e.g. about 70% of the selected cell clones were t-PA positive. The vector DNA copy numbers correlated with the mRNA levels and the protein levels for all cell lines analysed, with the exception for the K-562 cell line, where the production of t-PA was very low. The results obtained indicated that the highest expression levels were achieved in low density cultures.

A bleeding model in rabbits demonstrate fresh clot selectivity for a genetically engineered variant of tissue-type plasminogen activator and for streptokinase.

The way in which three thrombolytic agents, tissue-type plasminogen activator (t-PA), streptokinase (SK) and a genetically engineered variant of t-PA composed of the second kringle and the protease domain (K2P), cause the dissolution of haemostatic plugs of differing ages was investigated in a novel rabbit model. Standardized incisions were made on the rabbit ear and the wounds were left to heal for 0.5 h or 24 h, before the thrombolytic agents were infused. In the absence of heparin, t-PA showed little discrimination between clots of different ages (36% and 28% lysis of the 0.5 h and 24 h wounds, respectively). In contrast, K2P and SK showed a pronounced fresh clot selectivity since they were significantly more effective in lysing fresh clots than old ones (68% and 4% lysis for K2P and 72% and 36% for SK, respectively). In the presence of heparin the potency of t-PA on fresh clots was considerably increased whilst the effect on old clots was not affected, a fresh clot selectivity for t-PA (64% lysis of fresh clots, 24% lysis of old clots) was thus effected. Heparin did not significantly affect the fresh clot selectivity of K2P or SK, although lysis of old clots was increased (from 4% to 36%) when it was given together with K2P. Furthermore, heparin did not affect the time to onset of bleeding nor was the bleeding time prolonged by its addition. The bleeding time observed for t-PA (20-25 min) was markedly shorter than that found for K2P or SK (40-50 min).(ABSTRACT TRUNCATED AT 250 WORDS)

Binding of tissue plasminogen activator to endothelial cells. The effect on functional properties. Localization of a ligand in the B-chain of tPA.

The binding of 125I-labelled tissue plasminogen activator (tPA), the tPA A- or B-chain to endothelial cells (EC) were studied in suspensions of cultured human umbilical vein EC (HUVEC) or immortalized microvascular EC (HMEC). By determinations of the concentration-dependent binding it was shown that both the A-chain and the B-chain, which were isolated after partial reduction of two-chain tPA, contain ligands for binding to EC. The affinity for the B-chain was much higher than for the A-chain according to Scatchard analysis (Kd 24 and 515 nM, respectively), whereas the number of binding sites was higher for the A-chain than for the B-chain (Bmax 8 x 10(5) and 1.2 x 10(5), respectively). There were no cross interactions between the A- and B-chains and their binding sites. The binding of tPA to EC induced an almost 100-fold increase of the activation rate when compared to the same amount of enzyme in free solution, which in contrast to the fibrin-induced stimulation was not inhibited by antibodies against fibrin. The enzymatic activity of the B-chain was much less affected by the association to the cells. Both tPA and the tPA B-chain were largely protected against inhibition by an excess plasminogen activator type-1 (PAI-1) when bound to EC, whereas the same amount of free tPA was totally inactivated. The competition studies strongly indicated that an N-terminal segment in the B-chain, AKHRRSPGER, may be the ligand part of the B-chain. It is interesting to note that this polypeptide segment also participates in a binding site for PAI-1, necessary for effective inhibition. This implies a possible competition between PAI-1 and a tPA-receptor for binding of tPA. High molecular weight urokinase had no quenching effect on the binding of the B-chain to EC.

Expression of the lung resistance protein in primary colorectal carcinomas.

To assess whether the lung resistance protein (LRP) is of clinical significance in colorectal carcinomas, we immunohistochemically determined LRP expression of colorectal carcinoma specimens (n = 68) by means of the monoclonal antibody LRP-56 and compared this expression with clinical parameters. LRP expression was negative in 7 (10%), low in 36 (52%) and high in 25 (38%) carcinomas. LRP expression was independent of histological grade, tumor size, lymph node involvement and distant metastasis. Survival of the patients with LRP-positive tumors was similar to the survival of patients with LRP-negative tumors. However, patients with high LRP expression in their carcinomas had a prolonged survival. Thus LRP is frequently expressed in colorectal carcinomas and high expression might indicate improved survival.

Expression of the lung resistance protein (LRP) in primary breast cancer.

BACKGROUND: To determine the clinical significance of the lung resistance protein (LRP) in breast cancer, we have studied its expression in primary breast carcinomas (n = 99) and assessed the association of this expression with clinical parameters of the patients. MATERIALS AND METHODS: LRP expression was immunohistochemically determined by means of the monoclonal antibody LRP-56 on frozen tumor sections. RESULTS: LRP expression was negative in 12%, low in 20%, intermediate in 47% and high in 21% of the carcinomas. LRP expression was independent of age of the patients, histology, tumor grade, estrogen receptor as well as progesterone receptor status, tumor size and lymph node involvement. Kaplan-Meier analyses revealed that both overall survival and disease-free survival were independent of the degree of LRP expression. CONCLUSIONS: LRP is frequently expressed in breast carcinomas, but is neither associated with known prognostic factors, nor a prognostic factor by itself.

Expression of the multidrug resistance protein (MRP1) in breast cancer.

BACKGROUND: The multidrug resistance protein (MRP1) is expressed in human breast carcinomas but its clinical significance remains unclear. The aim of the present study was to determine the clinical significance of MRP1 in breast cancer patients. MATERIALS AND METHODS: MRP1 expression of primary carcinomas from 100 breast cancer patients was immunohistochemically determined by means of the monoclonal antibodies QCRL-1/QCRL-3. RESULTS: MRP1 was negative in 20 (20%) and positive in 80 (80%) breast carcinomas. MRP1 expression was more frequent in both estrogen receptor-negative carcinomas and progesterone receptor-negative carcinomas (p = 0.1 in both cases), but was independent of tumor size and lymph node involvement. Patients with MRP1-negative carcinomas had prolongations of overall survival (p = 0.01 for death due to any cause, p = 0.04 for breast cancer-related death) and disease-free survival (p = 0.07) as compared to those with MRP1-positive carcinomas. Also in subsets of patients (negative lymph nodes; positive lymph nodes; positive estrogen receptor; T1/T2 tumors), overall survival was longer for patients with MRP1-negative carcinomas. In univariate Cox regression analyses, MRP1 positivity was associated with relative risks of 4.9 (95% CI 1.2-20.6; p = 0.03) for death due to any cause, 6.4 (95% CI 0.9-48.0; p = 0.07) for breast cancer-related death and 3.5 (95% CI 0.8-14.9; p = 0.09) for relapse. In multivariate Cox regression analyses, MRP1 positivity had relative risks of 5.1 (95% CI 1.2-21.7; p = 0.03) for death due to any cause, 6.5 (95% CI 0.8-50.1; p = 0.07) for breast cancer-related death and 3.4 (95% CI 0.8-15.1; p = 0.1) for relapse. CONCLUSIONS: Our results suggest that MRP1 might be an important factor in breast cancer indicating excellent prognosis for patients with MRP1-negative carcinomas.