1 Kaplan Comprehensive Cancer Center, New York University, New York, NY; 2 Department of Membrane Biochemistry, Walter Reed Army Institute of Research, Washington DC, USA
Received 14 March 2003; revised 2 May 2003; accepted 3 June 2002
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Pegylated liposomal doxorubicin (Doxil®) has been reported to cause immediate hypersensitivity reactions (HSRs) that cannot be explained as IgE-mediated (type I) allergy. Previous in vitro and animal studies indicated that activation of the complement (C) system might play a causal role in the process, a proposal that has not been tested in humans to date.
Patients and methods:
Patients with solid tumors (n = 29) treated for the first time with Doxil were evaluated for HSRs and concurrent C activation. HSRs were classified from mild to severe, while C activation was estimated by serial measurement of plasma C terminal complex (SC5b-9) levels. Increases in SC5b-9 were compared in patients with or without reactions, and were correlated with Doxil dose rate.
Results:
Moderate to severe HSRs occurred in 45% of patients. Plasma SC5b-9 at 10 min after infusion was significantly elevated in 92% of reactor patients versus 56% in the non-reactor group, and the rise was greater in reactors than in non-reactors. We found significant association between C activation and HSRs, both showing direct correlation with the initial Doxil dose rate.
Conclusions:
C activation may play a key role in HSRs to Doxil. However, low-level C activation does not necessarily entail clinical symptoms, highlighting the probable involvement of further, as yet unidentified, amplification factors.
Key words: allergy, anaphylatoxins, cancer chemotherapy, doxorubicin, liposomes, hypersensitivity reactions
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
One unsolved problem with Doxil is that the treatment is often associated with infusion or hypersensitivity reactions (HSRs) despite pretreatment of patients with corticosteroids and antihistamines. The reported frequency of HSRs to Doxil varies between 0% and 25%, with average and median values of 8% and 5%, respectively (Table 1). The symptoms include facial flushing, dyspnea, tachypnea, facial swelling, headache, chills, hypotension or hypertension, chest pain and back pain [1, 9, 17, 18]. Unlike IgE-mediated (type I) allergy, these reactions occur mostly at the first exposure to the drug without prior sensitization.
|
We have previously reported that Doxil can activate the complement (C) system in vitro in normal human sera, and that minute (milligram) amounts of Doxil can induce severe cardiopulmonary changes in pigs [20]. Because liposome-induced hemodynamic changes in pigs had been previously demonstrated to be due to activation of the C system [21, 22], we proposed that the underlying cause of HSRs to Doxil in humans might be the same process, i.e. C activation.
In order to test the hypothesis that C activation plays an important role in HSRs to Doxil, in the present study we measured C activation in patients following Doxil infusion and analyzed the relationship between C changes and clinical symptoms of hypersensitivity. We also evaluated the correlation of C activation and HSRs with Doxil dose rate.
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Treatment and assessment of HSRs
In accordance with the administration guidelines of Doxil and the underlying treatment protocol, the drug was dissolved in 250 ml 5% dextrose injection (USP) before administration, and was infused over 1 h at an initial rate of one-fifth of the final rate. Following the start of infusion, patients were observed for 30 min for the presence of any of the following symptoms: skin reactions (urticaria, erythema, facial edema, facial rash, pruritus, eruptions), hypotension or hypertension, respiratory problems (laryngospasm, laryngeal edema, bronchospasm, dyspnea), pain (joint pain, back pain, abdominal pain, chest pain) or other manifestations of hypersensitivity (fever, chills, rigors, diaphoresis, nausea, vomiting, neurological changes). The symptoms were graded as specified in Table 2. The clinical observations were blinded to the laboratory findings for the initial 19 patients.
|
Statistical analysis
Patients were divided into two groups according to the presence or absence of HSRs. Those who developed any type of reaction, regardless of severity, were referred to as clinical reactors, whereas those without symptoms were called clinical non-reactors. Comparisons of Doxil doses and initial dose rates in clinical reactors and non-reactors were carried out with the non-parametric MannWhitney test.
SC5b-9 values were expressed as means ± standard deviation (SD) of triplicate determinations, and the statistical significance of the differences between baseline and 10-, 30- and 120-min samples in each patients was established using analysis of variance followed by pairwise comparisons using the StudentNeumannKeuls test. Some patients had triplicate measurements only at baseline and 10 min, which were compared by Students unpaired t-test. Patients who displayed significant (P <0.05) increase of 10-min SC5b-9 values relative to baseline were called laboratory reactors.
Doxil-induced SC5b-9 formation in clinical reactor and non-reactor patients was compared by the MannWhitney test. The relationship between C activation and clinical reactions was analyzed by Fishers exact test and Cohens statistics [25]. Categorization of laboratory and clinical observations in a standardized 2 x 2 contingency table is described in the text (see Table 4).
|
|
|
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
Relationship between C activation and HSRs
We subjected the above data to statistical analysis addressing the following questions: (i) was C activation greater in clinical reactors than in non-reactors? (ii) was there association between C activation and HSR? (iii) was there agreement between two surveys identifying patients as reactors by clinical and laboratory criteria? and (iv) how reliable were serial SC5b-9 determinations in diagnosing HSRs?
For comparing C activation in clinical reactors and non-reactors we considered that the SC5b-9 values at 10 min were not normally distributed and that the sample sizes were small. Thus, we used the non-parametric MannWhitney test to compare the ranks of SC5b-9 values at 10 min in the reactor and non-reactor groups, as well as the ranks of another measure of C activation, the 10-min to baseline SC5b-9 ratios. While the former analysis resulted in borderline significance (P = 0.051), comparison of the ranks of SC5b-910min/SC5b-9baseline ratios indicated statistically significant (P <0.05) increase in C activation in the reactor group.
To assess the relationship between C activation and HSRs we used two methods of qualitative data analysis: Fishers exact test and Cohens statistics. They address different questions regarding the relationship between categorical variables; in our case Fishers method tests the degree of association between laboratory and clinical reactions, while
quantifies the agreement (or reproducibility) of two surveys identifying patients as reactors by clinical and laboratory criteria. Table 4 presents the 2 x 2 contingency table used for these analyses, which indicated a significant (P <0.05) relationship between C activation and HSR by both statistical methods. Hence, our questions regarding the association between C activation and HSRs, and also about the agreement between two surveys identifying patients as reactors by clinical and laboratory criteria, can be affirmatively answered.
As for the diagnostic value of the C test with regard to HSR, Table 5 specifies the sensitivity, specificity and positive and negative predictive values of 10-min SC5b-9 readings. In our case, sensitivity gives the proportion of clinical reactors correctly identified by increased C activation, specificity is the proportion of clinical non-reactors correctly identified by a lack of C activation, positive predictive value is the proportion of laboratory reactors correctly diagnosed as clinical reactors, while the negative predictive value is the proportion of laboratory non-reactors correctly diagnosed as clinical non-reactors. We computed these indices for all laboratory reactors (row 1), as well as for laboratory reactor subgroups differentiated by the extent of C activation (rows 24). It is seen in Table 5 that the C assay was highly sensitive in predicting HSRs, but the specificity and positive predictive value of the test was relatively low, particularly in patients in whom the rise of SC5b-9 at 10 min remained below 2x upper limit of normal SC5b-9 (row 2). However, when we restricted the criteria for laboratory reactivity to 10-min SC5b-9 values exceeding two-fold or four-fold the upper threshold of normal, the specificity and positive predictive value of the C assay remarkably increased with relatively less decrease in sensitivity. Thus, the extent of SC5b-9 elevation was proportional with the specificity and positive predictive value of the C assay with regards to HSRs. Taken together, our data and statistical analysis indicate that C activation plays a causal role in HSRs, although C activation per se does not guarantee HSR. Other pathogenic factor(s) may also be involved that become(s) rate limiting to HSRs in the case of low-level C activation.
Relationships among Doxil dose rate, C activation and HSRs
Figure 3 shows the 10-min SC5b-9 values of clinical reactors and non-reactors plotted against the initial rate of Doxil administration. Regression analysis revealed significant correlation between dose rate and SC5b-9 (P <0.01), indicating that C activation at 10 min was Doxil dose-dependent. Consistent with the correlation between HSRs and Doxil dose (Table 3) and the significant association between C activation and HSRs (Table 4), the upper right quadrant of the plot contained readings obtained exclusively from clinical reactors.
The data in Figure 3 also allowed quantification of the odds in favor of developing HSRs at different dose rates. Stepwise computation of odds ratios at increasing dose rate (legend to Figure 3) indicated significantly increased risk of HSR at and above 0.38 mg/min (dashed line), corresponding to 114 mg Doxil infused over 1 h at an initial rate one-fifth of the final rate.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The present report describes the first human study of the role of C activation in the development of HSRs to Doxil. In this study no premedications were routinely used; however, we did follow the recommendation in the Doxil package insert [19] with the warning to infuse the drug at an initial rate of <1 mg/min. The incidence of HSRs (45%) was much higher than reported earlier (Table 1), which could be due to a combination of such circumstances as the relatively high-dose Doxil regimens studied, the greater attention to HSRs and lack of premedication with antihistamines and steroids.
Measuring plasma SC5b-9 levels as an index of C activation [23, 24] we found significant rises in the majority (72%) of patients treated with Doxil for the first time. This reaction rate is very similar to our in vitro observations with normal human sera (70%) [20], indicating that in vitro C activation simulates the in vivo process. We also found that the frequency and the degree of C activation was greater in reactors than in non-reactors, and that there was significant association between C activation and HSRs. These facts, together with data from animal experiments [2022], strongly suggest that C activation plays a causal role in HSRs to Doxil. However, the fact that not all patients with C activation displayed HSR highlights the involvement of other essential factors in the pathogenesis that can limit the adverse consequences of anaphylatoxin release during C activation. Such mechanisms include the breakdown of anaphylatoxins by carboxypeptidase N, and the coupling of the anaphylatoxin signal to mast cell and/or leukocyte (mainly basophil) release of histamine and other inflammatory mediators. It is possible that patients who develop severe HSR to Doxil are prone to C activation, the anaphylatoxin clearance is slow, and at the same time their mast cells have increased susceptibility to release reactions. Consistent with this hypothesis, atopic constitution, which is characterized by mast-cell hypersensitivity to various stimuli, is a risk factor for liposome reactions as well.
Our data provide some potentially useful guidelines for the prevention of HSRs to Doxil and other liposomal drugs, inasmuch as they show an already increased risk of HSR at 0.38 mg/min initial infusion rate. This value is lower than the manufacturers recommended 1 mg/min, and is consistent with a study by Gabizon and Muggia [10] reporting HSR in only one of 25 patients when the initial infusion rate was 0.10.2 mg/min. The critical role of infusion rate in HSRs to Doxil is clearly in line with the prediction of the C hypothesis that if the rate of anaphylatoxin production exceeds the rate of clearance, the threshold for mast-cell and leukocyte activation may be reached sooner.
In conclusion, the present study confirms and extends former clinical evidence [29, 30] of C activation by parenteral liposomes. Our data indicate that C activation may play a key role in HSRs, although it may not be the only cause or rate-limiting factor, as the presence of other abnormalities also seems to be necessary. The high rate of reactions in this study strengthens the need for slow initial infusion, and is consistent with (although does not prove) the potential usefulness of premedication with steroids and antihistamines.
![]() |
Acknowledgements |
---|
![]() |
Footnotes |
---|
Present address: Division of Lymphoma/Myeloma and Bone Marrow Transplantation, Roswell Park Cancer Institute, Elm & Carlton Streets, Buffalo, NY 14263, USA
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2. Muggia FM. Liposomal encapsulated anthracyclines: new therapeutic horizons. Curr Oncol Rep 2001; 3: 156162.[Medline]
3. Lyass O, Uziely B, Ben-Yosef R et al. Correlation of toxicity with pharmacokinetics of pegylated liposomal doxorubicin (Doxil) in metastatic breast carcinoma. Cancer 2000; 89: 10371047.[CrossRef][ISI][Medline]
4. Gabizon AA. Pegylated liposomal doxorubicin: metamorphosis of an old drug into a new form of chemotherapy. Cancer Invest 2001; 19: 424436.[CrossRef][ISI][Medline]
5. Lyass O, Hubert A, Gabizon AA. Phase I study of Doxilcisplatin combination chemotherapy in patients with advanced malignancies. Clin Cancer Res 2001; 7: 30403046.
6. Muggia FM, Blessing JA, Sorosky et al. Phase II trial of the pegylated liposomal doxorubicin in previously treated metastatic endometrial cancer: a Gynecologic Oncology Group study. J Clin Oncol 2002; 20: 23602364.
7. Gabizon A, Catane R, Uziely B et al. Prolonged circulation time and enhanced accumulation in malignant exudates of doxorubicin encapsulated in polyethylene-glycol coated liposomes. Cancer Res 1994; 54: 987992.[Abstract]
8. Gabizon A, Isacson R, Libson E et al. Clinical studies of liposome-encapsulated doxorubicin. Acta Oncol 1994; 33: 779786.[ISI][Medline]
9. Uziely B, Jeffers S, Isacson R et al. Liposomal doxorubicin: antitumor activity and unique toxicities during two complementary phase I studies. J Clin Oncol 1995; 13: 17771785.[Abstract]
10. Gabizon AA, Muggia FM. Initial clinical evaluation of pegylated-liposomal doxorubicin in solid tumors. In Woodle MC, Storm G (eds): Long Circulating Liposomes: Old Drugs, New Therapeutics. Austin, TX: Springer-Verlag and Landes Bioscience 1998; 165174.
11. Northfelt DW, Dezube BJ, Thommes JA et al. Efficacy of pegylated-liposomal doxorubicin in the treatment of AIDS-related Kaposis sarcoma after failure of standard chemotherapy. J Clin Oncol 1997; 15: 653659.[Abstract]
12. Muggia FM, Hainsworth JD, Jeffers S et al. Phase II study of liposomal doxorubicin in refractory ovarian cancer: antitumor activity and toxicity modification by liposomal encapsulation. J Clin Oncol 1997; 15: 987993.[Abstract]
13. Koukouraski MI, Kourkouraki S, Giatromanolaki A et al. Liposomal doxorubicin and conventionally fractionated radiotherapy in the treatment of locally advanced non-small-cell lung cancer and head and neck cancer. J Clin Oncol 1999; 17: 35123521.
14. Gordon AN, Granai CO, Rose PG et al. Phase II study of liposomal doxorubicin in platinum- and paclitaxel-refractory epithelial ovarian cancer. J Clin Oncol 2000; 18: 30933100.
15. Hubert A, Lyass O, Pode D et al. Doxil (Caelyx): an exploratory study with pharmacokinetics in patients with hormone-refractory prostate cancer. Anticancer Drugs 2000; 11: 123127.[CrossRef][ISI][Medline]
16. Lyass O, Uziely B, Ben-Yosef R et al. Correlation of toxicity with pharmacokinetics of pegylated liposomal doxorubicin (Doxil) in metastatic breast carcinoma. Cancer 2000; 89: 10371047.[CrossRef][ISI][Medline]
17. de Marie S. Liposomal and lipid-based formulations of amphotericin B. Leukemia 1996; 10 (Suppl 2): S93S96.[ISI][Medline]
18. Skubitz KM, Skubitz AP. Mechanism of transient dyspnea induced by pegylated-liposomal doxorubicin (Doxil). Anticancer Drugs 1998; 9: 4550.[ISI][Medline]
19. ALZA Pharmaceuticals, I. Doxil Package Insert. Mountain View, CA, USA 2000.
20. Szebeni J, Baranyi B, Savay S et al. The role of complement activation in hypersensitivity to pegylated liposomal doxorubicin (Doxil®). J Liposome Res 2000; 10: 347361.
21. Szebeni J, Fontana JL, Wassef NM et al. Hemodynamic changes induced by liposomes and liposome-encapsulated hemoglobin in pigs: a model for pseudo-allergic cardiopulmonary reactions to liposomes. Role of complement and inhibition by soluble CR1 and anti-C5a antibody. Circulation 1999; 99: 23022309.
22. Szebeni J, Baranyi B, Savay S et al. Liposome-induced pulmonary hypertension: properties and mechanism of a complement-mediated pseudoallergic reaction. Am J Physiol 2000; 279: H1319H1328.[ISI]
23. Muller-Eberhard HJ. The membrane attack complex. Springer Semin Immunopathol 1984; 7: 93141.[ISI][Medline]
24. Buyon JP, Tamerius J, Belmont HM et al. Assessment of disease activity and impending flare in patients with systemic lupus erythematosus. Comparison of the use of complement split products and conventional measurements of complement. Arthritis Rheum 1992; 35: 10281037.[ISI][Medline]
25. Rosner B. Fundamentals of Biostatistics, 2nd edition. Boston, MA: Duxbury Press 1986; 424428.
26. Szebeni J. The interaction of liposomes with the complement system. Crit Rev Ther Drug Carrier Syst 1998; 15: 5788.[ISI][Medline]
27. Szebeni J, Muggia FM, Alving CR. Complement activation by Cremophor EL as a possible contributor to hypersensitivity to paclitaxel: an in vitro study. J Natl Cancer Inst 1998; 90: 300306.
28. Szebeni J, Alving CR, Savay S et al. Formation of complement-activating particles in aqueous solutions of Taxol: possible role in hypersensitivity reactions. Int Immunopharmacol 2001; 1: 721735.[CrossRef][ISI][Medline]
29. Sculier JP, Coune A, Brassinne C et al. Intravenous infusion of high doses of liposomes containing NSC 251635, a water-insoluble cytostatic agent. A pilot study with pharmacokinetic data. J Clin Oncol 1986; 4: 789797.[Abstract]
30. Brouwers AH, DeJong DJ, Dams ET et al. Tc-99m-PEG-liposomes for the evaluation of colitis in Crohns disease. J Drug Target 2000; 8: 225233.[ISI][Medline]