Drugs

, Volume 65, Issue 16, pp 2405–2427

Gemtuzumab Ozogamicin

A Review of its Use in Acute Myeloid Leukaemia

Authors

    • Adis International Limited
  • Caroline M. Perry
    • Adis International Limited
Adis Drug Evaluation

DOI: 10.2165/00003495-200565160-00014

Cite this article as:
Fenton, C. & Perry, C.M. Drugs (2005) 65: 2405. doi:10.2165/00003495-200565160-00014

Summary

Abstract

Gemtuzumab ozogamicin (Mylotarg®) is a conjugate of a monoclonal antibody and calicheamicin, which targets the membrane antigen CD33 in CD33-positive acute myeloid leukaemia (AML) and, after cell internalisation, releases a derivative of the cytotoxic calicheamicin component. In the US, it is approved as monotherapy in patients aged ≥60 years with a first relapse of AML who are ineligible for other cytotoxic therapy.

Monotherapy with gemtuzumab ozogamicin results in complete remission (CR) or CR with incomplete platelet recovery (CRp) in ≈25% of adults (including those aged ≥60 years) with CD33-positive AML in first relapse. Preliminary data indicate a potential role for gemtuzumab ozogamicin as a component of induction or consolidation regimens in adults and, based on an early study, in the treatment of children with AML, although randomised, controlled studies are needed. Serious adverse events, notably hepatotoxicity, characterise its tolerability profile, but gemtuzumab ozogamicin is comparatively well tolerated by most patients. Gemtuzumab ozogamicin is a valuable new treatment option for patients aged ≥60 years with CD33-positive AML in first relapse for whom other cytotoxic chemotherapy is not considered appropriate; patients with a first CR (CR1) of >12 months are likely to have the best outcome.

Pharmacological Properties

Gemtuzumab ozogamicin is a humanised monoclonal antibody conjugated to a cytotoxic calicheamicin derivative, which targets the CD33 antigen expressed by leukaemic blasts in most patients with AML. After internalisation by the leukaemic cell, the linker between the antibody and calicheamicin is hydrolysed, calicheamicin dimethyl hydrazide is released and reduced; the reduced species binds to DNA in the minor groove, causing site-specific double-stranded breaks and cell death.

In vitro, gemtuzumab ozogamicin displays good activity against certain CD33+ AML cell lines. At low concentrations (0.01–0.025 ng/mL), in vitro sensitivity of AML cells to gemtuzumab ozogamicin correlates with CD33 expression, but at high concentrations (1–10 μg/mL), CD33-independent uptake may occur. Various gemtuzumab ozogamicin resistance mechanisms have been suggested, including cell escape because of surface antigen expression or cell cycle phase, expression of proteins causing drug efflux, altered signalling pathways, antiapoptotic expression and patient antigen load. Sensitivity to gemtuzumab ozogamicin can be enhanced in vitro by ciclosporin, an inhibitor of p-glycoprotein-mediated drug efflux. In patients, a clinical response is inversely correlated with peripheral blood antigen load and drug efflux ratios, and is not predicted by pharmacokinetic parameters.

The maximum plasma concentrations of the CD33 antibody (hP67.6) and calicheamicin occur shortly after the end of the intravenous infusion of gemtuzumab ozogamicin. Distribution is mainly within plasma and antibody distribution in bone marrow, spleen and liver occurred in radiolabelling studies. The pharmacokinetic parameters of gemtuzumab ozogamicin differ after the first and second doses and vary widely between patients, particularly after the first dose; however, they do not depend on age or gender. The calicheamicin derivative remains conjugated in plasma and hP67.6 is believed to be eliminated from plasma mainly by binding to CD33 expressed on peripheral blast cells.

Therapeutic Efficacy

The clinical efficacy of gemtuzumab ozogamicin has been evaluated in noncomparative trials in adults. In a pooled analysis of three trials, one to three doses of gemtuzumab ozogamicin 9 mg/m2 monotherapy 14–28 days apart, each administered as a 2-hour intravenous infusion, resulted in CR in 13% of 277 patients with a first relapse of primary AML and CRp in another 13%, giving an overall remission (OR) rate of 26%. CR and CRp rates in patients aged ≥60 years, representing 57% of the trial population, were 12% and 12%. The OR rate was 34% in patients with a CR1 of ≥1 year and 11 % in those with a CR1 of ≤6 m onths. Median relapse-free survival for all those in OR was 5.2 months. Median overall survival was almost 5 months for the total patient population, and was >1 year in those patients achieving CR or CRp; overall survival was longer in patients in CR or CRp who proceeded to haematopoietic stem cell transplantation (HSCT) than in those receiving no further treatment.

In a dose-finding noncomparative trial in 29 children (median age 12 years) with relapsed or refractory AML, 14% achieved CR and 14% CRp after one or two gemtuzumab ozogamicin doses of 6–9 mg/m2, but dose-limiting toxicity occurred.

Early-phase trials of gemtuzumab ozogamicin in combination therapy regimens, usually with cytarabine and anthracyclines, resulted in CR rates of 35–83% in adults with primary and secondary AML, including those aged >60 years and, in a consolidation regimen, maintenance of remission in 32% of patients after 1 year.

Tolerability

One to three doses of gemtuzumab ozogamicin 9 mg/m2 were generally fairly well tolerated in adults with relapsed primary AML, although serious adverse events were reported. Patients receiving gemtuzumab ozogamicin monotherapy plus prior or subsequent HSCT had a 17% incidence of hepatic veno-occlusive disease (VOD) and a 10% VOD-associated fatality rate. VOD also occurred in patients who did not have HSCT; other severe hepatotoxicity has affected recipients of gemtuzumab ozogamicin.

Infusion-related events were common but generally transient and reversible. Gemtuzumab ozogamicin has, however, been associated with severe hypersensitivity reactions, including anaphylaxis, infusion reactions and pulmonary events. Almost all patients in phase II trials experienced severe neutropenia and thrombocytopenia, the latter associated with serious bleeding in 13% of patients. Other severe adverse effects had a relatively low incidence; grade 3 or 4 sepsis, pneumonia or nausea or vomiting affected 8–17% of patients. The overall incidence of early treatment-related mortality was 16%. Paediatric patients in a small dose-finding trial experienced similar adverse events, although there was less myelosuppression; VOD occurred at dosages of 6–9 mg/m2 in children and, as in adults, was more common in HSCT recipients, affecting 40%.

1. Introduction

Acute myeloid leukaemia (AML) is an uncommon heterogeneous blood disease arising from the abnormal proliferation and differentiation of a clone of haematopoietic progenitor cells (blasts), which replace normal bone marrow.[1] AML affects people of all ages, but the median age at diagnosis is 68 years and the annual incidence in the US is 11–24 cases per 100 000 in patients aged 65–84 years (vs 3.5 per 100 000 in the general population).[2]

Treatment in AML generally aims to eliminate the leukaemic cell clone and achieve and maintain remission with cytotoxic chemotherapy and/or haematopoietic stem cell transplantation (HSCT).[1] About 45–50% of treated patients aged >60 years achieve a first complete remission (CR1);[35] most (>80%)[3] subsequently relapse, after which 20–40% attain a second remission.[6] Chemotherapy alone is rarely curative in relapsed AML.[7] Haematological relapse, discussed in this review, is defined as the reappearance of >5% blasts in bone marrow; molecular relapse, which is particularly relevant to acute promyelocytic leukaemia (APL), and extramedullary relapse may also occur.[8]

The 5-year relative survival rates of patients with AML aged <65 and ≥65 years are 33.1% and 3.7%.[2] The unfavourable prognosis in elderly patients is attributable to disease-related factors and a ≈25% mortality rate associated with standard treatment.[3,5,9,10] Elderly patients have reduced bone marrow reserves[9] increasing their susceptibility to bleeding and systemic infections (often a result of gut mucosal damage coincident with the neutropenic nadir).[11] Cardiotoxicity, particularly with anthracyclines and mitoxantrone,[11] and neurotoxicity after high-dose cytarabine (cytosine arabinoside, ara-C) also occur more frequently in elderly patients.[12]

AML karyotype predicts the response to current treatment and is accordingly classified broadly as favourable, intermediate and poor in about 23%, 68% and 9% of patients aged <55 years[13] and 4%, 63% and 33% of patients aged ≥60 years.[5] Age has an independent impact; approximate 5-year survival with a favourable karyotype is 72% in patients aged <55 years[13] and 30% in patients aged ≥60 years,[5] and 17%[13] and 1%[5] with unfavourable karyotypes. Patients with a multidrug resistant (MDR) phenotype associated with drug efflux (section 2), which is more likely in the elderly, have a CR rate of 35% (vs 58% in those without MDR phenotype).[5] In relapsed AML, age and karyotype still affect outcomes, but the most significant prognostic factor, especially in elderly patients,[10] is CR1 duration.[8]

Increasing knowledge of the molecular basis of AML has led to the investigation and development of targeted therapies, which may improve results in specific patient groups.[14] Gemtuzumab ozogamicin (Mylotarg®)1 comprises the humanised anti-CD33 antibody hP67.6, which is used as a carrier and facilitates uptake of the linked cytotoxic component, a calicheamicin derivative.[15] The membrane antigen CD33 is expressed on the surface of leukaemic blasts in ≈75–90% of adults with AML.[16] CD33 is also expressed on the surface of early haematopoietic progenitors and myelomonocytic precursors, but not normal pluripotent haematopoietic stem cells.[15] The presence of CD33 on progenitor cells may not always correlate with blast cell CD33 expression; therefore, definitions of CD33-positive AML based on the percentage of CD33+ blast cells identified by conventional immunophenotyping techniques may not predict the effectiveness of gemtuzumab ozogamicin and are not used in all trials.[17]

This article reviews the pharmacology and clinical profile of gemtuzumab ozogamicin, administered as a 2-hour intravenous infusion, in patients with a first relapse of CD33-positive AML. Gemtuzumab ozogamicin is approved in the US for the treatment of such patients who are aged ≥60 years and are not candidates for other chemotherapy. Also reviewed briefly are phase I and II trials assessing gemtuzumab ozogamicin in younger patients and in combination chemotherapy regimens; these are not approved indications. Gemtuzumab ozogamicin has been reviewed previously in Drugs.[18]

2. Pharmacodynamic Properties

The pharmacodynamic properties of gemtuzumab ozogamicin monotherapy have been reviewed in detail[15,18] and are summarised here. Data are available mainly from in vitro[1926] studies; some are sourced from published patient studies[27,28] and prescribing information.[29]

2.1 Mechanism of Action

Gemtuzumab ozogamicin consists of a recombinant humanised IgG4 anti-CD33 monoclonal antibody (clone P67.6), joined by a bifunctional linker to a cytotoxic calicheamicin antibiotic derivative (N-acetyl γ-calicheamicin dimethyl hydrazide [DMH]);[22,30] the linker is stable at physiological pH in the extracellular compartment.[15] The average overall ratio of calicheamicin to antibody is ≈2–3;[22] approximately 50% of the antibody is unconjugated. After binding of the anti-CD33 antibody to the CD33 antigen, the resulting complex is rapidly internalised[29] and, at the acidic pH inside myeloid lysosomes, the linker is hydrolysed.[15] Calicheamicin DMH is released from the antibody in the leukaemic cell and reduced; the reduced species binds to and breaks DNA strands.[15,29]

Calicheamicin is a potent enediyne antibiotic, ≈4000 times more active than doxorubicin against murine tumours in vitro.[31] Gemtuzumab ozogamicin is more potent in vitro against CD33+ HL-60 human promyelocytic leukaemia cells than earlier tested carbohydrate conjugates and the parent drug;[22] potency is important as levels of CD33 expression on AML cells limit the amount of conjugated agent that can be delivered.[22] Calicheamicin binds to DNA in the minor groove and, by forming a p-benzene diradical, causes site-specific double-stranded breaks and cell death.[30]

Susceptible cells responded to calicheamicin by cell cycle arrest, apparently via the ataxia-telangiectasia (AT)-mutated/AT-related Ch1/Ch2 kinase pathway.[32] Apoptosis of myeloid leukaemia HL-60 and NB-4 cells, but not THP-1 cells, then occurred via caspase 3 activation; the THP-1 cell line was resistant to the proapoptotic effects of gemtuzumab ozogamicin, responding by G2 arrest without apoptosis.[32]

2.2 In Vitro Activity

Gemtuzumab ozogamicin shows good in vitro activity against certain CD33+ AML cell lines. Three of four human myeloid leukaemia cell lines (HL-60, THP-1 and NB-4) tested were very sensitive in vitro to gemtuzumab ozogamicin, with concentrations required to inhibit 50% growth (IC50 values) of 2–6 ng/mL.[32] This inhibition was specific for the conjugated antibody. However, the KG-1 line was resistant at these concentrations, with a much higher IC50 of >1000 ng/mL.[32]

The growth of NB-4 APL cell lines and patient blast cells resistant to tretinoin (all-trans tretinoin, retinoic acid) and/or arsenic trioxide was dose-dependently suppressed by gemtuzumab ozogamicin.[26] However, MDR cell lines NOMO-1/ADR and NB-4/MDR were resistant to gemtuzumab ozogamicin at a concentration of 10 000 ng/mL, while dose-dependent inhibition of the growth of NOMO-1 and NKM-1 cell lines occurred at concentrations of 5–100 ng/mL.[18]

CD33-negative cell lines are, however, generally unresponsive to gemtuzumab ozogamicin. Significantly different levels of endogenous CD33 expression have been observed in different AML cell lines,[21] as has a direct correlation between levels of cell surface expression of CD33 and the extent of CD33 internalisation after binding of gemtuzumab ozogamicin.[24] Increased sensitivity to gemtuzumab ozogamicin, particularly at the lowest concentrations tested, was dependent on multiplicity of infection.[24] Cells expressing low levels of CD33 (parenteral OCI-AML3 cells) are sensitive to gemtuzumab ozogamicin at dose-equivalent concentrations of about 0.01–0.025 ng/mL.[24] Internalisation of anti-CD33 antibody was controlled by CD33 immunoreceptor tyrosine-based inhibitory motifs (ITIMs); mutations in these amino acid sequences, particularly in the proximal ITIMs, reduced both internalisation of and susceptibility to gemtuzumab ozogamicin.[24]

By contrast, CD33-independent uptake of gemtuzumab ozogamicin has been observed and provides a rationale for the efficacy of gemtuzumab ozogamicin in patients with CD33-negative haematological malignancies with endocytic capacity, such as CD33-negative acute lymphocytic leukaemia (ALL).[19] At concentrations of 1–10 μg/mL, the percentage of lysis of CD33+ cells was 12–32% at 24 hours and 25–68% at 40 hours.[19] However, CD33-negative ALL cells exposed to gemtuzumab ozogamicin 1–10 μg/mL also underwent 20–55% lysis at 24 hours.[19] Susceptibility to gemtuzumab ozogamicin was greatest in the activated phases (G1, S and G2/M) of the cell cycle, whereas cells in the resting G0 phase appeared to be protected against the cytotoxic effects of gemtuzumab ozogamicin.[19] Non-CD33-mediated uptake of gemtuzumab ozogamicin may occur via a receptor-independent mechanism, i.e. endocytosis.[19] Background levels of CD33 or some CD33+ blasts, even in patients with CD33-negative disease, may also explain responsiveness to gemtuzumab ozogamicin.[15]

2.3 Resistance

A range of potential gemtuzumab ozogamicin resistance mechanisms have been proposed, related to patient antigen load, cell escape because of surface antigen expression or cell cycle phase, expression of proteins causing drug efflux, altered signalling pathways and antiapoptotic expression.[15] In NB-4 APL cells, the mechanism of resistance to gemtuzumab ozogamicin differs from that of resistance to tretinoin or arsenic trioxide.[26]

Several studies have assessed the effect of cell surface expression on gemtuzumab ozogamicin activity.[19,24] At concentrations of ≤10 ng/mL, which are cytotoxic to parental KG-1a CD33+ cell lines, CD33-negative AML cell lines were resistant to gemtuzumab ozogamicin, although susceptibility could be induced by lentivirus-mediated CD33 expression.[24] Additionally, an in vitro study showed increased resistance to gemtuzumab ozogamicin in the presence of CD34-positive cells.[23]

Resistance to gemtuzumab ozogamicin is also associated with, but not fully explained by, drug efflux[20,27] and transporter proteins including p-glycoprotein (Pgp) and multidrug resistance protein 1 (MRP1).[15,20,23,25,32] The MDR phenotype is frequently associated with Pgp expression[21] and Pgp may export calicheamicin from drug-resistant CD33+ cell lines.[20,23]

Indeed, Pgp was implicated in the clinical resistance to gemtuzumab ozogamicin observed in studies of blast cells from patients with relapsed AML who were eligible to receive the drug in clinical trials.[20,27] Blast cell Pgp function (as evidenced by diethyloxacarbocyanine iodide [DiOC2] dye efflux assays) was associated with a less favourable clinical response to therapy with gemtuzumab ozogamicin, suggesting that the MDR phenotype may be an independent prognostic variable.[20]

A low blast cell dye efflux, indicating low Pgp activity, correlated with the achievement of CR or CR with incomplete platelet recovery (CRp) in gemtuzumab ozogamicin recipients.[20,27] The median baseline dye efflux ratio was 1.24 in samples from 38 adults who achieved CR or CRp after gemtuzumab ozogamicin, versus 1.58 for 88 patients who did not (p = 0.002).[20] Samples from paediatric patients with advanced CD33-positive AML who did or did not respond to one or two doses of gemtuzumab ozogamicin 6–9 mg/m2 also had significantly different mean dye efflux ratios (p = 0.005), and five of eight patients with a ratio of <1.2, versus 0 of 14 with a ratio ≥1.2 (p = 0.002), achieved CR or CRp.[27]

Variation with respect to Pgp was shown in this and two other in vitro studies,[20,25,32] which also demonstrated the ability of several agents to enhance the susceptibility of gemtuzumab ozogamicin-resistant AML cells taken from patients in a first relapse of CD33-positive AML.[20,25] While there was no correlation between either cell surface or intracellular Pgp expression and resistance to gemtuzumab ozogamicin in AML cell lines in the first study,[32] ciclosporin, a Pgp inhibitor, decreased efflux and increased gemtuzumab ozogamicin sensitivity in at least some patient blast samples in another.[20] However, Pgp function was strongly correlated with surface Pgp expression only in ciclosporin-sensitive samples (r = 0.796, p < 0.0001 vs r = 0.025, p > 0.05 in insensitive samples).[20] Similarly, MK-571 (an MRP inhibitor) further enhanced sensitivity in 3 of 23 patient samples but did not affect the sensitivity of six Pgp-negative samples; overall, ≈30% (5 of 17) of CD33+ and Pgp-positive blasts were resistant to gemtuzumab ozogamicin despite the presence of both ciclosporin and MK-571.[25]

In vitro, the peripheral benzodiazepine receptor ligand RP52028 (PK11195) overcomes multiple resistance mechanisms, thereby increasing the sensitivity of AML cells to gemtuzumab ozogamicin.[21] RP52028, a proapoptotic agent and an inhibitor of Pgp-mediated efflux, was often more effective than ciclosporin or MK-571 and further investigation was suggested.[21]

The variable response to gemtuzumab ozogamicin observed in patients with AML was also attributed partly to different molecular pathways induced by the drug in vitro.[32] The different responses of four AML cell lines (HL-60, THP-1, NB-4 and KG-1) to gemtuzumab ozogamicin did not correlate with MDR proteins or with levels of CD33 expression, although drug efflux appeared to be the mechanism for resistance in the KG-1 cell line. Response patterns observed in CD33-positive AML cells from ten patients were similar to those seen in the cell lines.[32]

2.4 Effects in Patients with Acute Myeloid Leukaemia (AML)

The likelihood of remission after gemtuzumab ozogamicin treatment in patients with CD33-positive AML (section 4) is inversely correlated with peripheral blood CD33 antigen loads, probably because peripheral consumption depletes gemtuzumab ozogamicin concentrations and limits bone marrow saturation.[28] Pre-dose CD33 antigen levels in patients achieving or not achieving CR after gemtuzumab ozogamicin were 3.67 and 4.58 log10 units (p < 0.001 and p ≤ 0.014 with univariate or multivariate analysis [n = 92]). In AML blast cells taken 24 hours after a single 9 mg/m2 dose of intravenous gemtuzumab ozogamicin (n = 6), mean CD33 saturation was nearly complete in peripheral blood (92%), but not in bone marrow (74%, p < 0.05).[28] High bone marrow CD33 expression was correlated with lower CD33 saturation (r = 0.867, p < 0.05); a gemtuzumab ozogamicin concentration of ≥0.2 μg/mL was needed for ≥90% saturation and ≥5% lysis of AML-193 cells.[28]

3. Pharmacokinetic Properties

The pharmacokinetic properties of gemtuzumab ozogamicin, administered as two separate 2-hour intravenous infusions, have been evaluated in 59 adult patients with CD33-positive AML in first relapse (table I),[30] and in a dose-escalation study in 29 paediatric patients (table II).[27,33] A separate analysis found no significant differences in pharmacokinetic parameters between adult patients aged <60 years (n = 34) and ≥60 years (n = 24) or between men (n = 29) and women (n = 29).[34] Pharmacokinetic properties are also unaffected by body surface area and weight.[29] No formal drug interaction studies with gemtuzumab ozogamicin have been undertaken,[29] although there is potential for interaction between gemtuzumab ozogamicin and other drugs.
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Tab1.jpg
Table I

Pharmacokinetic parameters of gemtuzumab ozogamicin (GO) in adult patients (pts) with untreated first relapse of CD33-positive acute myeloid leukaemia. Mean values for anti-CD33 antibody (hP67.6) and total and unconjugated calicheamicin after one or two doses (14–28 days apart) of GO 9 mg/m2 administered as a 2-hour intravenous infusion, 1 hour after paracetamol (acetaminophen) and diphenhydramine.[30] Pts had a mean age of 52.7 years, mean weight of 81.1kg and mean body surface area of 1.9m2

https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Tab2.jpg
Table II

Pharmacokinetic parameters of gemtuzumab ozogamicin (GO) in paediatric patients (pts) aged 1–16 years with first relapse of or refractory CD33-positive primary acute myeloid leukaemia. Pts aged 3–16 years received one or two doses, a median 14 days apart, of GO 6 (n = 11), 7.5 (n = 2 [not shown]) or 9 (n = 12) mg/m2 as a 2-hour intravenous infusion.[33] Corresponding doses in infants aged 0–2 years were 0.2 (n = 3), 0.25 (n = 0) or 0.3 (n = 2) mg/kg. One child received a second GO cycle 9 months later; data are shown as an additional pt. Mean values for anti-CD33 antibody (hP67.6) and total and unconjugated calicheamicin after GO 6 and 9 mg/m2 or equivalent infant doses

After intravenous administration, the gemtuzumab ozogamicin complex is rapidly internalised by AML cells and split into the two components via hydrolysis of the bifunctional linker (section 2.1). Measurements of the humanised antibody hP67.6, which represents 97% of gemtuzumab ozogamicin and is the targeting component of the drug as well as being the basis for dose administration units, are regarded as the primary surrogate for the pharmacokinetic profile of gemtuzumab ozogamicin.[30] However, the pharmacokinetic parameters of both hP67.6 and unconjugated calicheamicin as well as total calicheamicin have been reported (table I).[30]

As expected, the maximum plasma concentrations of hP67.6 and calicheamicin occur soon (specific times not stated) after the infusion ends.[30] Plasma concentration-time curves for total calicheamicin and hP67.6 parallelled each other in most patients, suggesting that the calicheamicin DMH derivative remains conjugated in plasma.[30] There were significant differences in pharmacokinetic parameters between the first and second gemtuzumab ozogamicin doses (table I) and between patients.[30]

Compared with values after the first dose, the values for the area under the plasma concentration-time curve (AUC) after the second dose were almost twice as high for hP67.6 and about 35% higher for unconjugated calicheamicin (table I). These increases were attributed to changes in patient antigen levels, although these are difficult to determine, and suggest that binding to CD33 expressed by peripheral blast cells is the principal mechanism for the elimination of hP67.6 from plasma. However, pharmacokinetic parameters did not predict patient remission.[30] The pharmacokinetic parameters for calicheamicin (total and unconjugated) varied less between patients with the second than the first dose.[30] Gemtuzumab ozogamicin appears to be distributed mainly within the plasma compartment; uptake in bone marrow and antibody distribution in spleen and liver were reported in radiolabelling studies.[30]

The pharmacokinetic profile of gemtuzumab ozogamicin in 29 children (table II) was broadly similar to that in adults (table I); individual results for paediatric patients also varied widely.[33] Pharmacokinetic parameters were similar across different paediatric age groups (0–2, 3–11 and 12–16 years), despite numerically (81% and 78%) lower clearance in infants and children than in adolescents.[33] Similar changes to those in adults were observed between the first and second doses. After dose-limiting toxicity (DLT) in children at the 9 mg/m2 dose (section 5),[27] a dose of 7.5 mg/m2 was administered to two children before trial enrolment stopped for extraneous reasons; results for these two children[33] are not shown.

4. Therapeutic Efficacy

The therapeutic efficacy of gemtuzumab ozogamicin monotherapy in patients with CD33-positive, primary AML has been assessed in noncomparative trials.[27,3551] Adults (n = 277) experiencing an untreated first relapse participated in three multicentre, phase II trials analysed and reported together[35] (results for the first 142 patients[52] and for 101 of the 157 patients aged ≥60 years[53] were published earlier). Four other small (n = 12–43) monotherapy trials in adults[3639] are summarised briefly in section 4.1.1. A phase I dose-escalation trial in 29 children aged 1–16 (median 12) years with refractory AML or an untreated first relapse has also been completed (section 4.1.2);[27] two smaller paediatric compassionate-use case series (n = 15[54] and 12[55]) had shorter follow-up periods and results are discussed briefly (section 4.1.2).

The inclusion of gemtuzumab ozogamicin in induction,[4045] post-relapse[43,45,46] or consolidation[42,47] combination therapy regimens in small (n = 44–74), open-label trials in patients aged ≥17 years with untreated, refractory or relapsed CD33-positive and other AML or myelodysplastic syndrome (MDS) is discussed briefly in section 4.2. Preliminary noncomparative trials (n = 14–19) of other potentially feasible combination regimens are also summarised;[4851] one was a pilot study[48] for randomised trials and one a dose-escalation/phase II trial.[43] Phase III trials of gemtuzumab ozogamicin are ongoing.[17,56,57]

4.1 Monotherapy

Eligibility criteria in adult and paediatric monotherapy trials in AML (sections 4.1.1 and 4.1.2),[27,35] included a white blood cell (WBC) count of <30 × 109/L and adequate renal and hepatic function. Hydroxycarbamide (hydroxyurea) was permitted to reduce the peripheral WBC count to <30 × 109/L prior to gemtuzumab ozogamicin treatment[27,35] and was administered to 68 adult patients (25%).[35]

Scheduled treatment was two doses of gemtuzumab ozogamicin monotherapy, 14–28 days apart in adults[35] and 14 days apart in children,[27] each administered as a 2-hour intravenous infusion.[27,35] Paracetamol (acetaminophen) and antihistamines[35] (diphenhydramine[27]) were administered before gemtuzumab ozogamicin and paracetamol allowed afterwards; no corticosteroids were used.[27,35]

Patients who recovered from reversible, non-haematological adverse events and had no disease progression, uncontrolled infection or gemtuzumab ozogamicin-related antibodies received a second gemtuzumab ozogamicin dose.[27,35,53] In adults who did not achieve remission, a third dose was permitted, subject to the same criteria and a bone marrow biopsy indicating a partial response (>50% decrease in bone marrow blasts and >15% bone marrow cellularity).[53] One course of gemtuzumab ozogamicin was defined as two to three doses in adults.[35] Subsequent treatment was determined by the patient’s treating physician.[27,35]

The primary efficacy endpoint in both trials was the CR rate.[27,35] CR was evaluated ≈28 days after the last gemtuzumab ozogamicin infusion in adults[35] and monthly in children,[27] and defined as:

  • no leukaemic blasts in peripheral blood;

  • ≤5% leukaemic blasts in bone marrow (aspirate or biopsy sample);

  • peripheral blood haemoglobin ≥90 g/L, absolute neutrophil count ≥1.5 × 109/L and platelets ≥100 × 109/L; and

  • independence from transfusions of red blood cells for ≥2 weeks and platelets for ≥1 week.[27,35]

Trials also assessed a secondary endpoint of CRp, which required all CR criteria except platelet count to be met, and reported some results for CR plus CRp patients as an overall remission (OR) group.[27,35] Patients who did not achieve CR or CRp were classified as ‘no remission’.[27,35] Relapse-free and overall survival rates were also assessed in adults.[35]

4.1.1 Adult Patients

Study designs in the three phase II trials in adults with an untreated first relapse of CD33-positive primary AML were generally similar and efficacy data are pooled; specified differences in inclusion criteria are shown in table III.[35] Eligible patients had an Eastern Cooperative Oncology Group performance status of 0–2. Patients were aged a median 61 (range 20–87) years, with 57% aged ≥60 years. Almost all (95%) were White and 55% were male. Few (2%) had prognostically favourable cytogenetics at first relapse, 40% had normal or intermediate and 23% had poor-risk karyotypes; in 35% of patients, karyotyping was not performed. Karyotype distribution was similar in patients aged ≥60 years.[35]
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Tab3.jpg
Table III

Gemtuzumab ozogamicin trial inclusion criteria. Adults (n = 277) with an untreated first relapse of CD33-positive primary acute myeloid leukaemia participated in noncomparative, multicentre phase II studies[35]

The median CR1 duration at enrolment was 10.6 months with 87% of patients having a CR1 of over 6 months and 42% having a CR1 lasting >1 year (table IV).[35] Intensive consolidation treatment had been received during CR1 by most patients (93%) and 27 (10%) had undergone allogeneic (n = 2) or autologous (n = 25) HSCT (table III) a median 11.7 months before gemtuzumab ozogamicin treatment.[35]
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Tab4.jpg
Table IV

Efficacy of gemtuzumab ozogamicin (GO) monotherapy in adult patients (pts) with untreated first relapse of CD33-positive primary acute myeloid leukaemia (AML). Pts achieving complete remission (CR),a CR other than adequate platelet count (CRp) or no remission of AML in a combined analysis of three noncomparative, multicentre phase II trials. Pts received one to three doses of intravenous GO 9 mg/m2; the second dose (in 203 pts [73%]) was administered 14–28 days after the first (subject to no disease progression, GO-related antibodies, infection or ongoing non-haematological adverse event) and a third dose (in seven pts [3%]) was permitted if second dose criteria were met and a partial response was evident. Response was evaluated ≈28 days after the last GO dose. CR + CRp constitutes overall remission[35]

Monotherapy with gemtuzumab ozogamicin showed efficacy in adults with an untreated first relapse of CD33-positive primary AML. The primary endpoint (CR) was achieved by 35 patients (13%) and CRp by 36 (13%), giving an OR rate of 26% (24% and 28% in patients aged ≥60 and <60 years [table IV]) in a median time of ≈66 days.[35] OR was most likely after a CR1 of ≥12 months and least likely after a CR1 of ≤6 months or previous HSCT (table IV). Age and cytogenetics did not significantly affect remission rates, but only 2% of patients had favourable cytogenetics at baseline. CR and CRp were each achieved by 12% of patients aged ≥60 years, versus 13% and 14% for patients aged <60 years (table IV).[35]

An early response to treatment (≤5% leukaemic blasts in bone marrow on days 5–11) occurred in 111 of 277 patients (40%).[35] The proportion of early responders was significantly greater (p < 0.0001) among patients who went on to OR (56 of 66 patients [85%] tested) than in those who did not (55 of 169 [33%] tested).[35]

Median overall survival was 4.9 months in all patients and 12.6 months in patients achieving CR or CRp; patients aged <60 or ≥60 years survived a median 5.3 and 4.5 months overall (figure 1 a).[35] In patients who achieved a second remission, median overall survival times in those who received allogeneic HSCT, autologous HSCT, other chemotherapy or no further treatment (n = 14, 11, 11 and 35) were >18.3, 16.5, 12.2 and 11.2 months (log rank p-value = 0.007). Median overall survival in the 24 of 27 previous HSCT recipients who did not respond to gemtuzumab ozogamicin was 3.3 months.[35]
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Fig1.jpg
Fig. 1

Overall survival after gemtuzumab ozogamicin (GO) monotherapy in adult patients (pts) with untreated first relapse of CD33-positive primary acute myeloid leukaemia. Pts (n = 277) received one to three doses of GO 9 mg/m2, as specified in table IV. Trial outcomes were complete remission (CR), CR without adequate platelet count recovery (CRp), or no remission. (a) Shows median survival duration for all GO recipients; numbers above bars show the no. of pts in group. (b) Shows, for the 46 of 277 pts who underwent haematopoietic stem cell transplantation (HSCT) after GO therapy, the no. who survived or had died by 100 days after HSCT and by the end of the studies. Those HSCT pts alive at the end of the studies survived 2.0–4.8 years from the first dose of GO.[35,58] med. = median.

Figure 1b shows the number of patients surviving (16 of 46 [≈35%] at last observation) after post-gemtuzumab ozogamicin HSCT, including 6 of 21 who did not achieve CR or CRp.[58] In patients aged <60 years, median post-HSCT survival in those who did not achieve remission (n = 19) was 3.3 months, versus 21.2 and >15.6 months for those in CR (n = 7) and CRp (n = 11) [only 9 recipients of post-gemtuzumab ozogamicin HSCT were aged ≥60 years].[35]

Median relapse-free survival was 6.4 months for CR patients (n = 35), 4.5 months for CRp patients (n = 36) and 5.2 months for the OR group.[35] However, among those who received no post-remission treatment (about half of the OR patients), median relapse-free survival was significantly longer after CR than CRp (3.8 vs 2.4 months; p = 0.03). Relapse-free survival, also possibly reflecting different post-remission treatment, was also significantly different (quantitative data not reported; p = 0.008) in patients aged ≥60 and <60 years. Patients who achieved CR or CRp required fewer days’ hospitalisation (median 18 vs 27 days overall and 30 days for those not in CR or CRp; p-values not stated); 42 patients (15%) were hospitalised for ≤7 days, including 2% who were not hospitalised at all.[35]

Three other trials reported remission rates of 14%,[36] 17%[38] and 25%[39] after intravenous gemtuzumab ozogamicin 9 mg/m2 monotherapy in populations with poor-prognosis AML, including remission in patients aged ≥65 years;[36,38,39] however, only 5% of patients aged >75 years achieved remission.[38] Molecular remission was also attained after three gemtuzumab ozogamicin doses in 13 of 16 patients aged 17–77 years with relapsed APL (molecular relapse in APL, in patients who may have morphologically normal bone marrow and blood, is detected with a reverse transcriptase-polymerase chain reaction assay).[37]

4.1.2 Paediatric Patients

The use of gemtuzumab ozogamicin in paediatric patients is not an approved indication, and the small sample size (n = 29) in the fully published study with longest follow-up[27] means results should be interpreted with caution. Children (median age 12 years) had refractory (n = 10) or relapsed (n = 19) CD33-positive AML (>80% leukaemic blasts with increased CD33-immunoflourescence staining) with a median CR1 of 144 days (range 1.1–27.8 months).[27] Patients with CNS or testicular involvement or who had undergone HSCT were ineligible. Children were required to have a Lansky play index or Karnofsky performance status of ≥60%.[27]

A primary study objective was to determine the maximum tolerated paediatric dose of gemtuzumab ozogamicin.[27] The planned gemtuzumab ozogamicin doses were 6 mg/m2, increasing, in the absence of DLT (section 5), to 9 then 12 mg/m2, all administered twice, 14 days apart; equivalent doses for patients aged <3 years[27] (n = 5)[33] were 0.2, 0.3 and 0.4 mg/kg. During the trial, the design changed slightly (section 5) and children received doses of 6, 7.5 and 9 mg/m2 or 0.2, 0.25 or 0.3 mg/kg (table V).[27]
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Tab5.jpg
Table V

Efficacy of intravenous gemtuzumab ozogamicin (GO) 6–9 mg/m2 monotherapy in 29 paediatric patients (pts) with acute myeloid leukaemia (AML). Pts with refractory (n = 10) or a first relapse of (n = 19) CD33-positive, primary AML without prior haematopoietic stem cell transplantation (HSCT) received GO in a noncomparative dose-escalation trial with a primary objective of determining the maximum tolerated paediatric dose; doses 1 and 2 were administered a median 14 days apart.[27] Some children subsequently received HSCT. Outcomes included complete remission (CR)a [primary endpoint], CR except for adequate platelet count (CRp) or no remission

Four patients achieved CR (primary outcome) and four CRp, giving an OR rate of 28% (table V); these eight children included five previously in first relapse and three with previously refractory AML.[27] Five children who achieved CR or CRp received gemtuzumab ozogamicin 9 mg/m2, but DLT (section 5) at this dose resulted initially in dose de-escalation to 6 mg/m2 (n = 7), then an increase to 7.5 mg/m2 (n = 2). Trial enrolment then stopped for extraneous reasons. Most children received two gemtuzumab ozogamicin doses (table V).[27]

Fifteen children, including three who received only one dose of gemtuzumab ozogamicin, received HSCT (table V); these children lived longer than those who did not receive HSCT (quantitative data not reported). One HSCT recipient and one non-recipient were alive at last follow-up (table V).[27]

In 15[54] and 12[55] children aged 0–17 years who received gemtuzumab ozogamicin 1.8–9 mg/m2[55] (one or two doses) or 4–9 mg/m2 (1–3 doses)[54] on compassionate-use bases, CRp was achieved in five patients (33%)[54] and blast cell reduction to ≤5% in 5 of 12 (41%)[55] and 3 of 15 (20%)[54] patients; no children achieved CR.[54,55] Two of the children who achieved CRp[54] and one who achieved blast cell reduction to ≤5%[55] were in CR 6–9 months after post-gemtuzumab ozogamicin HSCT.

4.2 Combination Regimens

4.2.1 Induction and Post-Relapse Therapy

Potentially feasible AML combination induction therapy regimens (also not an approved indication), including gemtuzumab ozogamicin, cytarabine and (apart from one group[41]) an anthracycline or mitoxantrone[4042] in previously untreated patients are shown in table VI;[4042] one study was reported as an abstract.[41] One study,[42] which was a prelude to a large, randomised study (AML-15),[17] also evaluated gemtuzumab ozogamicin in consolidation regimens (section 4.2.2).
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Tab6.jpg
Table VI

Efficacy of gemtuzumab ozogamicin (GO) as part of a combination induction therapy regimen in acute myeloid leukaemia (AML). Patients (pts) with untreated AML received GO as a 2-hour intravenous (IV) infusion, combined with cytarabine (cytosine arabinoside, ara-C [A]), and (except in one group[41]) an anthracycline (daunorubicin [D], idarubicin [Id] or mitoxantrone, an anthracenedione [M]), etoposide (E), fludarabine (FL) or subcutaneous granulocyte colony-stimulating factor (G-CSF). Phase II[40,41] or pilot,[42] sequential cohort[42] or noncomparative[40,41] trials (one reported as an abstract[41]) assessed the rate of complete remission (CR),[40,41] vs CR other than adequate platelet count (CRp) or treatment failure/death (F) [includes partial response[40]]. Except for G-CSF, drugs were administered IV. Pt numbers are evaluable pts

Remission (CR) rates in the larger studies[40,41] (primary endpoint in one[40]) were 35–83% (table VI) and combination induction therapy appeared to offer benefits over gemtuzumab ozogamicin monotherapy.[40] For example, CR and CRp rates of 23% and 12% after the gemtuzumab ozogamicin phase of induction treatment,[40] which (irrespective of the response to gemtuzumab ozogamicin) was followed by mitoxantrone, cytarabine and etoposide (MICE) chemotherapy, increased to 35% and 19% after MICE (table VI). In total, another 11 of 57 participants achieved CR or CRp; disease-free survival after CR or CRp was a median 190 days.[40] Cytogenetics influenced remission rates after the complete induction sequence, but neither cytogenetics nor CD33 status affected the response to gemtuzumab ozogamicin.[40]

Induction or post-relapse combination therapy, including gemtuzumab ozogamicin 4.5–9 mg/m2 and cytarabine, but excluding anthracyclines, achieved CR rates of 10–46%[4346] (plus CRp of 2–15%) in trials in elderly[46] or poor-prognosis patients (n = 44–59) with refractory,[43,45] relapsed,[43,45,46] or untreated[44] CD33-positive[43,46] or other[44,45] AML (and MDS[44]) [two studies reported in abstracts[43,46]]. Regimens included one dose of gemtuzumab ozogamicin 4.5[45] or 6[44] mg/m2 combined with cytarabine (0.5 g/m2 twice daily on days 2–6),[44,45] ciclosporin (6 mg/kg loading dose before gemtuzumab ozogamicin, then 16 mg/kg on days 1 and 2)[44,45] and fludarabine (15 mg/m2 twice daily on days 2–6) [MFAC regimen];[44,45] two doses of gemtuzumab ozogamicin 9 mg/m2 (days 4 and 19) plus oblimersen 7 mg/kg/day (days 1–6 and 15–21);[46] and cytarabine (3 g/m2/day for 5 days) [HiDAC regimen] followed by one dose of gemtuzumab ozogamicin 9 mg/m2 on day 7.[43]

The percentage of patients achieving remission in some smaller trials, reported fully[48,49] or in abstracts,[50,51] supports further investigation of the following induction chemotherapy combinations with gemtuzumab ozogamicin: idarubicin and cytarabine (MIA) [CR or CRp in 6 of 14 patients with refractory or relapsed AML];[48] mitoxantrone and intermediate-dose cytarabine (MIDAM) [CR or CRp in 12 of 17 patients with refractory or relapsed AML];[50] mitoxantrone and high-dose cytarabine followed by amifostine (5 of 7 patients aged ≤55 years leukaemia-free at median 227 days, but only 1 of 12 aged >55 years);[51] and tretinoin plus, in 3 patients, idarubicin (CR in 16 of 19 patients with APL).[49]

4.2.2 Consolidation Therapy

In trials of consolidation regimens that combined other intensive chemotherapy with single doses of gemtuzumab ozogamicin at 3[42] and 4.5[47] mg/m2, including a trial in patients in CR1 who had received gemtuzumab ozogamicin during induction therapy[42] (section 4.2.1), CR1 was maintained in 32% of patients at 1 year[47] and 29 of 31 patients survived with satisfactory haematological recovery.[42]

5. Tolerability

Tolerability data for gemtuzumab ozogamicin are sourced from the clinical trials in section 4.1, including the earlier analysis of patients aged ≥60 years,[53] other trials,[27,3842,59] analyses,[6064] and drug approval summary[6] and prescribing information.[29]

5.1 Adult Patients

Intravenous gemtuzumab ozogamicin is usually fairly well tolerated in adults with AML, although serious adverse events that are either infusion reactions or otherwise related to drug toxicity are common.[29,35] Severe (grade 3 or 4) haematological adverse events affect almost all adult monotherapy recipients and severe hepatotoxicity (section 5.3), including hepatic veno-occlusive disease (VOD), particularly in recipients of HSCT, has occurred.[35,59,64] The prescribing information includes a boxed warning of severe hypersensitivity reactions, including infusion reactions, anaphylaxis and pulmonary events,[29] based on post-marketing experience of severe drug reactions, some fatal.[6,29]

The tolerability of 1–3 doses of gemtuzumab ozogamicin 9 mg/m2 monotherapy was generally fairly good in adult patients (57% aged ≥60 years) in the three phase II trials;[35] the data for 101 of the 157 patients aged ≥60 years indicates similar tolerability in these patients.[53] Grade 3 or 4 treatment-emergent adverse events affecting ≥5% of the 277 total patients, which were not infusion-related, hepatotoxic (section 5.3) or myelosuppressive, are shown in figure 2.[35]
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Fig2.jpg
Fig. 2

Tolerability of intravenous gemtuzumab ozogamicin (GO) monotherapy in adults with untreated first relapse of CD33-positive primary acute myeloid leukaemia. Grade 3 or 4 treatment-emergent adverse events affecting ≥5% of patients (pts) receiving one to three doses of GO 9 mg/m2, excluding infusion-related, hepatotoxic and myelosuppressive events. Results from a combined analysis of three noncomparative phase II trials (n = 277). LDH = lactate dehydrogenase.

Adults (including those aged ≥60 years[53]) were significantly more likely to experience grade 3 or 4 infusion-related events with the first than second gemtuzumab ozogamicin dose (30% vs 10% of patients,[35,53] p < 0.0001[35]). Such events (occurring on the day of the gemtuzumab ozogamicin infusion and similar to those observed after other monoclonal antibody infusions[61]) included chills and fever (in 8% and 6% of patients), hypotension that was generally reversible with fluids (4%), nausea (3%) and hypertension (2%).[35] Between 66% and 82% of patients experienced grade 1–4 infusion-related fever, nausea or chills; vomiting also affected 58%.[29]

Grade 3 or 4 sepsis, pneumonia and nausea or vomiting affected a relatively low 17%, 8% and 10% of patients, respectively. Oral mucositis or stomatitis affected 69 patients (25%), but was of grade 3 or 4 severity in only 9 patients (3%) after the first gemtuzumab ozogamicin dose; 2 of these had received prior hydroxycarbamide (hydroxyurea) treatment.[29] There were no detectable antibody responses to gemtuzumab ozogamicin.[35] Grade 3 or 4 cardiotoxic events occurred in 3 of 12 monotherapy recipients in a pilot study.[39]

Early treatment mortality was similar in patients aged <60 and ≥60 years (14% and 17%), with 44 patients (16%) dying within 28 days of their last gemtuzumab ozogamicin dose.[35] Most early deaths resulted from disease progression or infection (each affecting 13 patients); other known causes were intracranial haemorrhage, multiorgan or respiratory failure, amphotericin anaphylaxis and VOD[35] (section 5.3).

Haematological events were grade 3 or 4 neutropenia and thrombocytopenia in almost all patients (98% and 99%, and 99% for both in the analysis of patients aged ≥60 years[53]);[35] grade 3 or 4 anaemia affected 52%.[29] The median time to recovery of neutrophil counts to 0.5 × 109/L was unaffected by age and was similar in patients in CR or CRp (40 and 43 days), but was longer (51 days, log rank p < 0.001) for those not in remission.[35]

Severe bleeding (grade 3 or 4) affected 36 patients (13%), including one who achieved CR and four who achieved CRp,[35] but it was more common in patients who did not achieve remission (15% vs 7% in OR).[29] Most frequent were epistaxis (3%) and intracranial haemorrhage; the latter affected nine patients (3%) and caused eight early treatment deaths.[35] Patients in CR achieved platelet counts of 25 × 109/L in a median 36 days (35 and 38 days in those aged <60 and ≥60 years), and 100 × 109/L in a median 50 days. Those in CRp, however, required a median 51 days (39 and 72 days in those aged <60 and ≥60 years) to recover platelet counts to 25 × 109/L (p < 0.001 vs CR). Platelet counts also recovered in 61 patients not achieving CR or CRp in a median 32 days.[35]

A similar range and incidence of adverse events to those occurring in the overall final study population was experienced by patients aged ≥60 years,[29,53] 77 of whom (76%) received two gemtuzumab ozogamicin doses.[35] Eight of the 101 elderly patients in the separate analysis received only one dose of gemtuzumab ozogamicin because of adverse events, including infection (in three patients), bleeding (two patients) and bradycardia/hyperkalaemia, diffuse fluid retention and tonic-clonic seizure, each affecting one patient.[53] Severe infections occurred in 27% of patients (mostly sepsis [15%] and pneumonia [6%]) and severe bleeding in 11%, including cerebral haemorrhage in 3% (two of whom died) and epistaxis in 3%. The incidence of severe mucositis, which probably influenced the incidence of sepsis, was low (4%); two affected patients had received hydroxycarbamide treatment.[53]

Limited data indicate that adverse events in patients receiving gemtuzumab ozogamicin plus other chemotherapy are generally similar to those in recipients of gemtuzumab ozogamicin monotherapy, with hepatotoxicity (section 5.3), myelosuppression and infusion-related events all common.[40,42,44,45] However, two doses of gemtuzumab ozogamicin 9 mg/m2 14 days apart as induction therapy appeared toxic in patients aged >75 years, with 7 of 22 (32%) dying during induction in one trial.[38]

5.2 Paediatric Patients

Paediatric patients in the dose-escalation trial generally tolerated gemtuzumab ozogamicin 6 mg/m2, but hepatotoxicity (section 5.3) was common.[27] DLT, first observed at the 9 mg/m2 dose, resulted in a change in trial design. DLT was defined as: three or more patients experiencing a grade 3 adverse event that was not infusion-related, haematological or a transient elevation in bilirubin, ALT or AST; or any patient experiencing a grade 4 event or VOD[27] (section 5.3).

Infusion-related events in children receiving gemtuzumab ozogamicin 6–9 mg/m2 monotherapy (or infant equivalent) were similar to those in adults, with nausea, fever, vomiting or chills affecting 28–55% of patients and headache, body or neck pain, hypertension, hypotension, sweating or tachycardia affecting 7–14%.[27] Grade 3 or 4 infusion-related events occurred at all dose levels, affecting eight patients (28%). Common non-hepatic, non-infusion-related grade 3 or 4 adverse events included leucopenia and thrombocytopenia in 48% and 35% of children, and sepsis, fever, hypokalaemia, hypochromic anaemia, pleural effusions and pneumonia each affecting 17–24%. Three children (10%) died of progressive disease within 28 days of their last dose.[27]

5.3 Hepatotoxicity

Hepatotoxicity, including VOD, was common in adult and paediatric recipients of gemtuzumab ozogamicin in the trials discussed in section 4.1; additional analyses of potential risk factors for post-gemtuzumab ozogamicin VOD in adults have also been published[59,6264] (two as abstracts).[62,63]

5.3.1 Adult Patients

Grade 3 or 4 liver function abnormalities were common in adults, including those aged ≥60 years, in the phase II trials after gemtuzumab ozogamicin monotherapy, although most were transient and reversible.[35] They included elevated bilirubin, AST and ALT levels in 29%, 18% and 9% of patients and moderate ascites in 3% of patients. One adult with ascites died after multiorgan failure 35 days after receiving gemtuzumab ozogamicin.[35] Grade 3 or 4 AST or ALT elevations (to more than five times the upper limit of normal) affected 15% of patients aged ≥60 years in the interim analysis, mostly after the first gemtuzumab ozogamicin dose.[53]

There were 16 cases, 9 fatal, of hepatic VOD (table VII);[35] one patient had two episodes, the second fatal. VOD mostly affected recipients of HSCT, who had an incidence of 17% and a VOD-related fatality rate of 10% (table VII). VOD was common irrespective of whether HSCT occurred before or after gemtuzumab ozogamicin therapy,[35] but did not occur in patients receiving HSCT >3.8 months after gemtuzumab ozogamicin.[62] Similarly, VOD occurred in nine of ten patients receiving myeloablative allogeneic HSCT within 3.5 months of gemtuzumab ozogamicin in a separate study, but in none of four patients with a >3.5 month interval before HSCT (p = 0.007).[64] Initial results (n = 10) of a dose-finding trial suggest the risk of post-gemtuzumab ozogamicin VOD may also be lower after autologous than allogeneic HSCT.[59]
https://static-content.springer.com/image/art%3A10.2165%2F00003495-200565160-00014/MediaObjects/40265_2012_65162405_Tab7.jpg
Table VII

Incidence of hepatic veno-occlusive disease (VOD) after gemtuzumab ozogamicin (GO) in patients (pts) with acute myeloid leukaemia (AML). Pts received intravenous GO 9 mg/m2 monotherapy for a first relapse (FR) of CD33-positive primary AML (n = 277).[35,63] a Some pts received haematopoietic stem cell transplantation (HSCT) before FR or after GO treatment, two received it at both times[29]

No specific baseline demographic or disease-related risk factors for the development of post-gemtuzumab ozogamicin VOD or liver function abnormalities have been identified;[63,64] some liver biopsies showed damaged hepatic endothelial cells with increased collagen deposition.[60]

Hepatotoxicity, including VOD, in adults with AML receiving gemtuzumab ozogamicin in combination chemotherapy regimens varied.[4042,44,47] In those (including elderly patients)[40] receiving induction regimens incorporating one or two gemtuzumab ozogamicin doses of 3–9 mg/m2,[40,42] increased liver toxicity and associated fatalities occurred only in patients also receiving thioguanine, which is associated with liver toxicity. Gemtuzumab ozogamicin did not increase the hepatotoxicity of subsequent treatment.[42] Similarly, while patients with newly diagnosed[44] or resistant or relapsed[47] AML receiving chemotherapy (section 4.2.1) had a 7%[44] and 9% (all fatal in this trial)[47] incidence of VOD, no VOD occurred with the same combination alternating with idarubicin plus cytarabine (IA) chemotherapy in post-remission patients.[47]

5.3.2 Paediatric Patients

Paediatric recipients of gemtuzumab ozogamicin also commonly experienced hepatotoxicity, with grade 3 or 4 elevated bilirubin, AST and ALT levels in 7%, 10% and 10% of children; as in adults, these events were mostly transient and reversible.[27] The incidence of VOD was 24% (7 of 29 patients) overall; however, 40% (6 of 15 patients) of those receiving HSCT developed moderate-to-severe VOD 1–22 days after HSCT. Five of these patients had received the 6 mg/m2 dose and one 9 mg/m2; four died from disease progression in 1–12 months and two from sepsis within 1.5 months.[27]

The child who developed VOD 3 days after one gemtuzumab ozogamicin dose without HSCT recovered in 13 days then underwent HSCT without another VOD episode.[27] A temporal relationship has been suggested, as VOD occurred in children receiving HSCT 20–106 days after their last dose of gemtuzumab ozogamicin, but not in the two children receiving HSCT after ≥120 days. However, five of the eight children who received HSCT within 55 days of gemtuzumab ozogamicin did not develop VOD.[27]

6. Dosage and Administration

Gemtuzumab ozogamicin is approved in the US for the treatment of patients aged ≥60 years with CD33-positive AML in first relapse who are not candidates for other cytotoxic chemotherapy.[29] The recommended dose is 9 mg/m2, administered over 2 hours as an intravenous infusion. A treatment course of two doses 14 days apart, with or without full haematological recovery between doses, is recommended. Leukoreduction should be considered prior to gemtuzumab ozogamicin administration in patients with a WBC count of >30 × 109/L; all patients should receive oral diphenhydramine 50mg and paracetamol 650–1000mg prior to gemtuzumab ozogamicin, with two further doses of paracetamol, as required, after gemtuzumab ozogamicin.[29]

Currently, gemtuzumab ozogamicin is approved for use only as single-agent chemotherapy and not as a component of combination therapy.[29] Caution is advised if gemtuzumab ozogamicin is administered to patients with hepatic impairment (bilirubin >2 mg/dL), in whom no studies have been conducted.[29] Local prescribing information should be consulted for other details including warnings and contraindications.[29]

7. Place of Gemtuzumab Ozogamicin in the Management of AML

The treatment of AML, particularly relapsed AML in elderly patients, is difficult and the prognosis poor. Induction, post-remission and post-relapse treatment options consist of full-dose chemotherapy, autologous or allogeneic HSCT after myeloablative or nonmyeloablative chemotherapy, investigational treatment and, for patients not eligible for trials or well enough to undergo intensive treatment, low-intensity chemotherapy and palliative care.[3,9]

The potential efficacy of more intensive treatments for AML is balanced against toxicity and treatment-related mortality. In an elderly, relapsed patient, the balance may (depending on individual assessment[3]) result in a different decision to that in a younger, untreated patient, but in any group, there is a place for any new treatment that is more effective or better tolerated than established treatments. In the elderly, the largest group of patients with AML, treatments that provided a long-term survival rate of ≥5% and resulted in a mortality rate of <25% in patients considered well enough to receive them would be an improvement (section 1).

In the US, practice guidelines recommend clinical trial participation for all patients with AML.[65] Other treatment is recommended only in the absence of this option.[65] In untreated patients, standard induction chemotherapy is one or two courses of cytarabine then an anthracycline or mitoxantrone.[7,65] Tretinoin is recommended in APL[9,65] and arsenic trioxide in relapsed APL.[7,65]

Without post-remission therapy, most patients with AML relapse within 9 months.[65] Patients’ treatment options are theoretically the same irrespective of age.[3,9] Recommended consolidation treatment in the US is cytarabine with or without an anthracycline, with HSCT an option for patients with refractory AML or without favourable cytogenetics.[65] In the absence of a clinical trial, recommended post-relapse therapy in the US in patients aged ≥60 years is gemtuzumab ozogamicin or supportive care in patients with a CR1 <6 months; gemtuzumab ozogamicin or chemotherapy is recommended for those with a CR1 >6 months.[65]

Cytarabine plus an anthracycline is also recommended as induction therapy in Europe.[66] No specific post-remission strategy is favoured;[5,66] active treatment options are chemotherapy (usually intermediate or high-dose cytarabine and/or investigational agents) or upfront allogeneic HSCT, which can be recommended only in patients with a relatively low marrow blast percentage.[8] Similarly, no specific post-relapse therapy other than allogeneic HSCT for second or later remissions is recommended in Europe.[66]

Allogeneic HSCT in early first relapse or second CR is potentially curative in AML,[7] but requires high-dose chemotherapy and/or radiotherapy conditioning treatment, with attendant high mortality risks, especially in elderly patients.[3] Non-myeloablative therapy then allogeneic HSCT, particularly if tolerability can be improved, may be an option in elderly patients,[1,3] as may autologous HSCT. In one small study, patients aged >60 years with mostly primary AML and a WHO performance status of 0–1 receiving autologous HSCT survived a median 19 months.[67] Adverse events were similar to those in younger patients.[67]

Although patients aged ≥60 years may be ideal candidates for investigative therapies,[8] fewer enter trials than patients aged <60 years and <60% receive treatment after CR.[68] This may reflect age, comorbidity, secondary disease, CR1 duration or personal choice,[1,5] but selection bias probably means clinical trial results in elderly patients are optimistic.[1]

The few available published studies on outcomes specifically in elderly patients with relapsed AML receiving chemotherapy[10,69,70] (including an analysis of therapeutic practice[10]) illustrate the difficulty of meaningfully improving survival in this group. Other possible approaches are to restrict intensive chemotherapy to patients who may be eligible for allogeneic HSCT[5] or to treat elderly patients primarily on the basis of karyotype, with chemotherapy plus investigational agents in those with adverse karyotypes, and experimental therapy in those with serious comorbidity or early relapse unlikely to respond to intensive chemotherapy.[1]

Several investigational AML treatments[5,7,8,14] have shown promise and may be available in clinical trials, some as adjuvant therapy. They include: monoclonal antibodies using antibody-radionucleotide conjugates that aim to improve post-HSCT results (e.g. [88Re] anti-CD66); inhibitors of mutated tyrosine kinase receptors, farnesyltransferase inhibitors (e.g. tipifarnib); histone deacetylase inhibitors (e.g. phenylbutyrate); apoptosis inducers targeting antiapoptotic proteins such as bcl-2, (e.g. oblimersen, a proteasome inhibitor); less toxic MDR modifiers that may reduce Pgp expression (section 2) [e.g. zosuquidar]; nucleoside analogues (e.g. clofarabine); and hypomethylating agents that irreversibly inhibit DNA methyltransferase (e.g. decitabine). Trials include patients with relapsed or poor-prognosis AML and elderly patients.[5,7,8,14]

Gemtuzumab ozogamicin is the first therapy approved in the US that specifically targets a known AML antigen (CD33), using a humanised monoclonal antibody (hP67.6) conjugated to a cytotoxic derivative of the potent antibiotic calicheamicin (section 2). In vitro activity against AML cell lines and patient blast samples is good, although MDR and other resistance mechanisms are evident (section 2). CR or CRp are not predicted by pharmacokinetic parameters (section 3), although increased plasma exposure to both hP67.6 and calicheamicin after the second versus the first dose was attributed to changes in patient antigen levels (section 3).

Patient efficacy and tolerability data are preliminary, pending randomised, head-to-head comparative trials, but in noncomparative trials, gemtuzumab ozogamicin resulted in a CR rate of 13% and a CRp rate of 13%, giving an OR rate of 26% in adult patients (57% of whom were aged ≥60 years) with a first relapse of CD33-positive primary AML and a median CR1 of 10.6 months (section 4.1.1). OR was associated with a median overall survival of >1 year. CR or CRp was achieved by 24% (12% and 12%) of patients aged ≥60 years. By comparison, CR rates in other studies in elderly patients receiving other chemotherapy for relapsed[10,69,70] or refractory[69,70] AML were 26–48%, with early[10,69] or chemotherapy-related[70] deaths in 12–36%; these trials did not use the CRp criteria. In an early trial, gemtuzumab ozogamicin monotherapy also resulted in molecular remission in 13 of 16 adults with relapsed APL.

Preliminary results in trials of gemtuzumab ozogamicin in non-approved indications include an OR rate of 28% (CR and CRp each 14%) in 29 children with a first relapse of or previously refractory AML participating in a dose-finding gemtuzumab ozogamicin monotherapy trial (just over half of the patients went on to receive HSCT) [section 4.1.2], and CR rates of 35–83% in adults receiving gemtuzumab ozogamicin plus adjunctive induction chemotherapy (section 4.2.1). Combination induction therapy regimens in adults, including gemtuzumab ozogamicin, cytarabine and an anthracycline or mitoxantrone appear to offer benefits over gemtuzumab ozogamicin monotherapy (table VI, section 4.2.1) and consolidation regimens of chemotherapy plus gemtuzumab ozogamicin maintained remission in 32% of patients after 1 year (section 4.2.2). These early results indicate a role for gemtuzumab ozogamicin in combination chemotherapy regimens in patients with newly diagnosed AML; several ongoing studies are assessing these outcomes.[17,56,57]

The tolerability of gemtuzumab ozogamicin appears better than that of conventional treatments,[67,69,70] although there are no comparative trials. Severe infusion-related hypersensitivity reactions, such as anaphylaxis, have occurred with gemtuzumab ozogamicin, but in clinical trials, grade 3 or 4 severe adverse events were mainly haematological or hepatotoxic (section 5). The favourable gastrointestinal profile, particularly the low level of mucositis, may reduce the incidence of infections and mortality. Despite an ≈100% incidence of myelosuppression, severe infection and gastrointestinal adverse events affected 8–17% of gemtuzumab ozogamicin monotherapy recipients, and early treatment mortality was 16% overall and 17% for those aged ≥60 years (section 5). By comparison, although data in elderly, relapsed patients are limited, severe infection rates after chemotherapy were >50%[69,70] and mucositis affected 32% of those receiving autologous HSCT.[67]

Median hospitalisation times for all treated patients also appear slightly shorter with gemtuzumab ozogamicin, at 27 days (and ≤7 days in 15% of patients) versus 32 days with chemotherapy[10] and 31 days with autologous HSCT[67] in similar patient groups; 2% of gemtuzumab ozogamicin monotherapy recipients were treated as outpatients (section 4). However, there are no published pharmacoeconomic data comparing the treatment costs of gemtuzumab ozogamicin, based on the final full trial results, with outcomes of conventional therapy in similar patients.

VOD is a significant adverse effect of gemtuzumab ozogamicin monotherapy and HSCT (17%, mostly fatal, in adults and 40% in the paediatric patients [section 5]) and treatment with gemtuzumab ozogamicin is a VOD risk factor in allogeneic HSCT recipients.[64] VOD also occurred in non-recipients of HSCT. Grade 3 or 4 hepatotoxicity after gemtuzumab ozogamicin appears more common than with chemotherapy, which resulted in an incidence of ≤13% in elderly, relapsed patients;[69] VOD was not reported.[69,70] VOD itself is not well understood, although hepatic sinusoidal epithelium appears to be the specific injury site.[71] Theories of pathogenesis include damage from circulating unconjugated calicheamicin and uptake of the antibody-antigen complex (section 2), affecting CD33+ liver Kuppfer cells, sinusoidal endothelial cells or stellate cells.[71]

Delaying HSCT until >3.5 months after gemtuzumab ozogamicin may reduce the risk of VOD (section 5), but leaves patients, especially those in CRp, vulnerable to a relapse (section 4.1.1). Without HSCT, treatment of relapsed AML is unlikely to be curative, and even with VOD, adults in remission after gemtuzumab ozogamicin had a significantly better outcome if they progressed to HSCT (section 4.1.1). A trial is investigating autologous HSCT or a lower dose of gemtuzumab ozogamicin after allogeneic HSCT as better tolerated options;[59] a greater understanding of VOD risk factors might also contribute to a reduced incidence in gemtuzumab ozogamicin recipients.

More specific characterisation of AML and affected patients will help target treatment,[1,5] including identifying patients most likely to benefit from and tolerate gemtuzumab ozogamicin. Currently, a second remission after the recommended dosage of gemtuzumab ozogamicin monotherapy appears more likely in patients with a CR1 of >12 months, a lower CD33 antigen load and a DiOC2 dye efflux ratio of <1.2 (section 2); other possible predictors of remission are CD34-negative AML (section 2) and an early response to the first drug dose (section 4.1.1). A phase II trial assessing gemtuzumab ozogamicin plus ciclosporin in patients aged >60 years with relapsed AML may clarify any clinical benefit accruing from Pgp inhibition (section 2)[57] and other trials may identify more effective dosage regimens.

Following early-phase trial results (section 4), later phase trials of gemtuzumab ozogamicin as monotherapy and adjunctive therapy in adults, including elderly patients, and children, with newly diagnosed, refractory and relapsed AML are ongoing.[17,57,72] Given the short survival times in AML, comparative studies assessing quality of life outcomes, as well as remission and survival rates, in elderly patients with AML would help both physicians and patients by providing accurate information.

In conclusion, monotherapy with gemtuzumab ozogamicin results in CR or CRp in ≈25% of adults (including those aged ≥60 years) with CD33-positive AML in first relapse. Preliminary data indicate a potential role for gemtuzumab ozogamicin as a component of induction or consolidation regimens in adults and, based on an early study, in the treatment of children with AML, although randomised, controlled studies are needed. While serious adverse events, notably hepatotoxicity, characterise its tolerability profile, gemtuzumab ozogamicin is comparatively well tolerated by most patients. Gemtuzumab ozogamicin is a valuable new treatment option for patients aged ≥60 years with CD33-positive AML in first relapse for whom other cytotoxic chemotherapy is not considered appropriate; patients with a CR1 of >12 months are likely to have the best outcome.

Footnotes
1

The use of trade names is for product identification purposes only and does not imply endorsement.

 

Acknowledgements

At the request of the journal, Wyeth Pharmaceuticals provided a non-binding review of this article.

Copyright information

© Adis Data Information BV 2005