Additional plerixafor to granulocyte colony‐stimulating factors for haematopoietic stem cell mobilisation for autologous transplantation in people with malignant lymphoma or multiple myeloma

Description of the condition

Malignant lymphomas are malignancies of the lymph nodes and lymphatic system with possible involvement of other organs, and can be differentiated into non‐Hodgkin lymphoma (NHL) and Hodgkin lymphoma (HL). Multiple myeloma (MM) is a malignancy of the bone marrow and is classified among the plasma cell neoplasms by the World Health Organization (WHO).

Non‐Hodgkin lymphoma derives from B and T lymphocytes or natural killer (NK) cells at varying stages of maturation. The most common classification of NHL is the WHO system, which was last updated in 2008. It classifies the NHL in B‐cell neoplasms and T/NK‐cell neoplasms, whereby about 85% of the NHL descends from mature and progenitor B lymphocytes (Herold 2013; Jaffe 2009). A clinical approach is the classification to indolent NHL (approximately 70% of all NHLs) and aggressive NHL (approximately 30% of all NHLs) (Herold 2013). The annual incidence of NHL varies between different regions in the world. In North America, the incidence is estimated to be 13.7 per 100,000 inhabitants, while in the European Union, it is 8.3 per 100,000 inhabitants, with the global incidence amounting to 5.1 per 100,000 inhabitants. The worldwide mortality is 2.5% and is quite similar in the various global regions (GLOBOCAN 2008). The median age at diagnosis is around 60 years, and varies only slightly between the diverse types of NHL (Herold 2013).

Hodgkin lymphoma arises from germinal centre or post‐germinal centre B cells, and is associated with unique neoplastic cells called Reed‐Sternberg cells, which morphologically differentiate HL from NHL. The WHO classifies HL into two types: the classic (approximately 93% of all HLs) and the nodular lymphocyte predominant (approximately 7% of all HLs) (Gobbi 2013; Herold 2013). HL accounts for about 10% of all lymphomas and has an annual incidence of 2 to 3 per 100,000 inhabitants in Western countries (Jemal 2009; Sant 2010; Smith 2011). Hence, HL is a comparatively rare disease, but it is one of the most common malignancies in young adults. In industrialised countries, the age distribution of HL shows a first peak in the third decade and a second peak around the age of 60 (Herold 2013).

Multiple myeloma, one of the mature B‐cell neoplasms, is characterised by the accumulation of malignant plasma cells in the bone marrow compartment, the increased production of a monoclonal immunoglobulin, and bone destruction. MM may be preceded by a pre‐malignant state called monoclonal gammopathy of undetermined significance. Overall, MM accounts for approximately 10% of all haematologic malignancies and 1% of all cancers (Siegel 2012). The annual incidence is approximately 4 to 5 per 100,000 inhabitants in Western countries (Phekoo 2004; Sant 2010; Smith 2011). MM is an adult‐onset disease, with a median age at diagnosis of 66 and only 2% of people younger than 40 years of age (Kyle 2003). To date, no curative approach exists for MM, and depending on the therapy, the overall survival varies between a few months and some years. The 10‐year survival rate of young patients with ideal treatment is around 50% (Herold 2013).

Description of the intervention

High‐dose chemotherapy in combination with autologous stem cell transplantation is a widely used effective treatment option for people with NHL, HL, and MM (Naumann‐Winter 2012; Rancea 2013; Schaaf 2012). Stem cell transplantation is directed at restoring a functioning bone marrow in patients after myeloablative chemotherapy. In contrast to allogeneic transplantation with donor cells, in the autologous setting stem cells are collected from the patient and re‐infused after the high‐dose chemotherapy (Jantunen 2012).

Commonly used agents for stem cell mobilisation are granulocyte colony‐stimulating factors (G‐CSF). The patient receives G‐CSF for up to eight days, and as soon as sufficient CD34‐positive cells are detectable in the patient's peripheral blood, the stem cells are collected by stem cell harvest (apheresis) and stored for transplantation after the myeloablative therapy (Jantunen 2012). The two types of traditional mobilisation methods, G‐CSF with chemotherapy (chemo‐mobilisation) or without chemotherapy ('steady‐state'), lead to a failure rate of 5% to 40% of the cases, depending on definition, disease, and mobilisation agents used (Jantunen 2010). In contrast to G‐CSF mobilisation alone, chemo‐mobilisation leads to higher stem cell yield, a higher probability of getting grafts with a large number of CD34+ cells, the need for fewer apheresis procedures, and even to an antitumour effect. But chemo‐mobilisation is associated with higher toxicity and increased probability of complications (Bensinger 2009; Gertz 2010). Major risk factors for mobilisation failure are advanced age, the diagnosis of NHL, progressive disease, previous radiotherapy, bone marrow involvement or prior treatment with lenalidomide or purine analogues, thrombocytopenia, and the failure of previous mobilisation attempts (Basak 2011; Malard 2011).

Plerixafor is a new agent that apparently improves mobilisation and stem cell harvest, thereby reducing the required number of apheresis procedures (Jantunen 2010). It has mainly been studied in combination with G‐CSF mobilisation. Before its approval in Europe, plerixafor was only available at transplant centres in the compassionate use program. Positive results of this programme evaluating plerixafor in combination with G‐CSF in patients who had previously failed a mobilisation attempt (60% to 75%) strongly indicate the efficacy of this agent (Basak 2011; Hubel 2011). Based on promising data from phase II studies, randomised controlled trials (RCTs) were initiated comparing G‐CSF in combination with plerixafor to G‐CSF and placebo (AMD3100‐3102 MM; AMD3100‐3101 NHL).

How the intervention might work

Plerixafor (formerly AMD3100) was originally developed for the treatment of HIV infection. It inhibits the CXCR4 chemokine receptor that blocks the receptor binding of the stromal cell–derived factor‐1α. This interaction leads to the mobilisation of haematopoietic stem cells from the bone marrow into the peripheral blood, in mice as well as in humans (Broxmeyer 2005; Hatse 2003; Maziarz 2013). Moreover, plerixafor and G‐CSF are synergistic. Results of clinical trials indicate that plerixafor plus G‐CSF leads to an increased augmentation of mobilisation and release of CD34+ cells into the peripheral blood, facilitating effective apheresis (Bensinger 2009).

Why it is important to do this review

Even though a Health Technology Assessment (HTA) report of the Ludwig Boltzmann Institute has already evaluated the role of plerixafor (Hintringer 2010), this systematic review will be of high relevance because a new systematic literature search will be performed and recent research results will be included. At the time of publication of the HTA report in March 2010, at least two clinical trials were still ongoing. No new systematic literature search was run for the HTA report, and it only evaluated the evidence used during the European approval process in the European Public Assessment Report of the European Medicines Agency, already published in 2009.

Based on the published trials, plerixafor might be an effective mobilisation agent in the treatment of people with NHL, HL, and MM. A systematic review and meta‐analysis will provide the conclusive evidence to clarify the role of this agent and will also examine differences of treatment effectiveness caused by population group or chemotherapy agents. Evidence for the best therapeutic options is needed for decision‐making at both the level of the individual patient and the healthcare system.

Criteria for considering studies for this review

Search methods for identification of studies

We used search strategies based on those described in Chapter 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Lefebvre 2011). We did not use any language constraints.

Data collection and analysis

Description of studies

Risk of bias in included studies

There was low risk of bias in AMD3100‐3101 NHL and AMD3100‐3102 MM; however, in NCT01301963 and NCT00665314 we considered the potential risk of bias as high since neither study has published outcomes nor reported reasons for stopping early. For detailed information, see the 'Risk of bias' table in Characteristics of included studies, Figure 2, and Figure 3.

Effects of interventions

See: Table 1

The trial NCT01301963 was prematurely closed and no analysis was being done on collected data. Nine participants were randomised in this trial. The trial NCT00665314, which had an enrolment of five participants, was completed in 2009 and also did not report any results.

Summary of main results

• As no hazard ratios for overall survival were provided, we analysed mortality at 12 months. No evidence for differences between plerixafor and control arm regarding mortality emerged in the meta‐analysis.

• The meta‐analysis of adverse events in study period one (stem cell mobilisation) did not show a statistically significant difference between the plerixafor and placebo group.

• Regarding the primary endpoint of successful stem cell collection, the meta‐analysis showed a significant advantage for those participants randomised to plerixafor, with important heterogeneity. Reasons for the heterogeneity could be different underlying condition or different targeted stem cell number. In AMD3100‐3102 MM, the primary endpoint was defined as the proportion of participants collecting more than or equal to 6 x 10⁶ CD34+ cells/kg in two or fewer apheresis days. In AMD3100‐3101 NHL, the primary endpoint was the percentage of participants who collected more than or equal to 5 x 10⁶ CD34+ cells/kg in four or fewer apheresis days. It is common practice to define a higher primary endpoint of collected stem cells for MM patients, as they often need a second transplantation. In both studies, the time to reach the primary endpoint was significantly shorter in the plerixafor group compared to the control group.

• In AMD3100‐3102 MM, 95.9% of participants in the plerixafor arm and 88.3% in the placebo arm underwent transplantation. By contrast, in AMD3100‐3101 NHL, 90% of participants in the plerixafor group but only 55.4% in the control group underwent transplantation.

• In both trials, there was no evidence for a difference between participants in the plerixafor and the control group in terms of time to neutrophil and platelet engraftment.

• None of the trials reported quality‐of‐life events or progression‐free survival.

Overall completeness and applicability of evidence

The following aspects should be considered when interpreting the results of this meta‐analysis.

• Trial NCT01301963, terminated in June 2013, did not analyse and publish collected data due to low enrolment (nine participants). The study investigator did not respond to email. Study NCT00665314, completed in June 2009 with randomisation of five participants, did not report any outcomes. Genzyme, the responsible party, was not able to provide the missing data.

• Trial NCT01767714 (classified as an ongoing study), completed in November 2014 with enrolment of 100 participants, has not to this date reported any outcomes, however the trial could publish results in the near future, which may affect the results of this review. The collaborator of the study, Genzyme, did not respond to email.

• The two analysed studies, AMD3100‐3101 NHL and AMD3100‐3102 MM, have been published several times.

• Besides mortality at 12 months, number of collected stem cells, and adverse events, the studies did not provide much clinical data (e.g. quality of life).

• Both trials were conducted by the same study director and were supported by the Genzyme Corporation, the manufacturer of plerixafor.

Quality of the evidence

The risk of bias in AMD3100‐3101 NHL and AMD3100‐3102 MM was low as they were randomised, double‐blind, placebo‐controlled studies. We rated the risk of performance, detection, attrition, and reporting bias as low for both trials. The quality of allocation concealment (selection bias) in AMD3100‐3101 NHL remained unclear as the publication provided no information. We judged the risk of selection bias in AMD3100‐3102 MM as low. Due to the sponsorship of the Genzyme Corporation, we rated the risk of other bias as unclear for both trials.

We could conduct a meta‐analysis for mortality at 12 months, successful stem cell collection, and adverse events. We rated the quality of the evidence for mortality after one year as moderate due to the wide confidence interval and imprecision. We judged the quality of evidence for successful stem cell collection and adverse events during study period one (stem cell mobilisation) as high.

There was a low risk of publication bias, but we did not downgrade the quality of evidence because although two trials, which were terminated early, did not report results, they were stopped early due to low recruitment, and included 14 participants only. We therefore judged potential risk of publication bias as low. The first of these two trials, NCT01301963, was conducted and sponsored by the Case Comprehensive Cancer Center in Ohio, USA and was terminated in 2013. On clinicaltrials.gov (https://clinicaltrials.gov/ct2/show/results/NCT01301963?sect=X01256#all), the investigators stated that they had recruited nine participants, but due to low enrolment (10% of anticipated participants), they were doing no analysis on data collected in this study. We contacted principal investigator Hien Duong, MD, to obtain the collected data, but he did not respond. The second trial NCT00665314 was terminated in 2009; study sponsor and collaborator Genzyme stated via email that the low enrolment (five participants) was probably due to the compassionate use programme for plerixafor that had started at the same time. Genzyme was not able to provide more information on the outcomes.

Potential biases in the review process

We tried to avoid bias by performing all relevant processes in duplicate. We are not aware of any obvious deficiencies in our review process. However, the small number of trials included in this analysis could lead to publication bias, as we could not generate a funnel plot. As stated above, there is a low risk of publication bias, because of three studies that were completed and have not yet published results.

Agreements and disagreements with other studies or reviews

Sheppard 2012 conducted a systematic review of RCTs of haematopoietic stem cell mobilisation strategies for autologous transplantation for haematologic malignancies. They analysed 28 articles including AMD3100‐3101 NHL and AMD3100‐3102 MM as well as studies with chemotherapy mobilisation strategies. Due to the different methodological quality of the trials, they were not able to compare the studies via meta‐analysis and draw definitive conclusions. Regarding AMD3100‐3101 NHL and AMD3100‐3102 MM, Sheppard 2012 stated that primary mobilisation failure in the control groups could be a result of a too‐low G‐CSF dose. Sheppard 2012 recommended a risk‐adapted, cost‐effective approach for the use of plerixafor. Duong 2011 and Abhyankar 2012 described a risk‐based algorithm for the use of plerixafor based on analysing the CD34+ cell count on the first day of collection. Duong 2011 identified participants with a day one CD34+ cell yield of <= 0.7 x 10⁶ CD34+ cells/kg as being at risk for mobilisation failure. Costa 2011 described a cost‐based algorithm to decide whether to add plerixafor on the basis of the CD34+ cell count on the fourth day of G‐CSF application.

Giralt 2014 discussed recommendations for optimal stem cell mobilisation strategies. With reference to AMD3100‐3101 NHL and AMD3100‐3102 MM, they agreed that plerixafor‐containing strategies reduce mobilisation failure rates. They recommend using plerixafor for remobilisation attempts or for predictable poor mobilisers.

Characteristics of included studies [ordered by study ID]

Characteristics of excluded studies [ordered by study ID]

Characteristics of ongoing studies [ordered by study ID]

Internal sources

• University Hospital of Cologne, Germany.

  Department I of Internal Medicine

External sources

• No sources of support supplied