Investigational New Drugs

, Volume 31, Issue 2, pp 473–478

Differences in drug approval processes of 3 regulatory agencies: a case study of gemtuzumab ozogamicin


    • Division of Social Communication System for Advanced Clinical Research, Institute of Medical ScienceThe University of Tokyo
  • Masaharu Tsubokura
    • Division of Social Communication System for Advanced Clinical Research, Institute of Medical ScienceThe University of Tokyo
  • Jinichi Mori
    • Tokyo Metropolitan Cancer and Infectious CentreKomagome Hospital
  • Monika Pietrek
    • Pietrek Associates GmbH
  • Shunsuke Ono
    • Laboratory of Pharmaceutical Regulatory Science, Graduate School of Pharmaceutical SciencesThe University of Tokyo
  • Masahiro Kami
    • Division of Social Communication System for Advanced Clinical Research, Institute of Medical ScienceThe University of Tokyo

DOI: 10.1007/s10637-012-9877-8

Cite this article as:
Tanimoto, T., Tsubokura, M., Mori, J. et al. Invest New Drugs (2013) 31: 473. doi:10.1007/s10637-012-9877-8


Major discrepancies concerning risk-benefit assessments and regulatory actions are frequent among regulatory agencies. We explored the differences by scrutinizing a case of gemtuzumab ozogamicin (GO) in patients with acute myeloid leukaemia (AML). Assessment reports of GO were retrieved form the websites of the US Food and Drug Administration (FDA), the European Medicines Agency (EMA) and Japanese regulatory agency, and we also reviewed published clinical trials. While GO was approved by the US FDA under the accelerated approval program in 2000, it was withdrawn from the market in 2010, based on the required post-marketing commitment failure. The EMA refused granting marketing authorization for GO in 2008 on the grounds that there were no randomised controlled trials (RCTs). GO was approved as an orphan drug in Japan in 2005, and the Japanese regulatory authority decided to continue with the approval in 2010 on the condition that post-marketing surveillance is strengthened. Under these situations, promising new results of RCTs appeared in 2011, and the role of GO in AML treatment was refocused worldwide. The stringent regulation may not be suitable in case of an orphan drug of targeted therapy, and more room should be kept to facilitate effective developments of new anti-neoplastic agents.


Gemtuzumab ozogamicinPharmaceutical and Medical Devices Agency and Ministry of HealthLabour and Welfare of JapanRegulatory AgencyThe European Medicines AgencyThe US Food and Drug Administration


With increasing knowledge on the molecular mechanisms underlying cancer biology, the number of anticancer drugs, especially those for targeted therapies, submitted for marketing approvals has been increasing. In order to guarantee better access to novel anticancer drugs to patients with life-threatening diseases, the drug development strategies have been under intensive discussions [1, 2]. Regulatory agencies worldwide have tried to improve the process for earlier access to novel anticancer drugs. The US Food and Drug Administration (FDA) introduced an accelerated approval process in 1992, according to which a new investigational drug was approved based on the results of early phase trials with surrogate endpoints and the validation of its safety and efficacy was obtained by post-approval studies [3]. Similarly, the European Medicines Agency (EMA) introduced two instruments that regulate early market access: conditional approvals and approvals under exceptional circumstances [4]. The conditional marketing authorization adopted in 2006 applies to drugs that show benefits for life-threatening diseases, with a yearly renewal of approval. Furthermore, the companies must annually reassure the EMA that the confirmatory studies remain feasible and are under way. The Japanese regulatory authorities, i.e. the Pharmaceutical and Medical Devices Agency (PMDA) and the Ministry of Health, Labour and Welfare (MHLW), may approve a new drug without randomised controlled trials (RCTs) if it is defined as an orphan drug. In these situations, under the Japanese health care system with universal coverage [5], post-marketing surveillance (PMS) for all prescribed patients throughout Japan are required. In the Japanese system, post-approval clinical trials of the drugs are not mandatory.

Although cooperation among regulatory authorities worldwide has been strengthened over the decades, each has their individual evaluations of risk-benefit assessments, often resulting in different regulatory actions in different countries.

This lack of coordination among the agencies causes confusion among healthcare providers and patients worldwide, and has a big effect on daily clinical practice and public health. One striking example is the case of gemtuzumab ozogamicin (GO), an orphan drug for acute myeloid leukaemia (AML). GO was obtained by conjugating calicheamicin, an antitumor antibiotic, to the CD33 humanised monoclonal antibody in the late 1990s [6]. The risk-benefit assessments, requirements for approval, and phase 4 commitments for GO approval were completely different among the FDA, the EMA, and PMDA/MHLW. In this study, we explored these differences in GO approval among the regulatory agencies in order to understand the current obstacles for the development of targeted drugs in the era of personalised medicine.

Design and methods

We used the assessment reports for GO by the US FDA [7], EMA [8], and PMDA [9], and analysed the results of the studies submitted to the authorities for GO approval, such as time of approval and interpretation of the submitted studies by each authority. In addition, phase 4 commitments, i.e. the required post-approval study or PMS of the US FDA, and the PMDA/MHLW were also examined by using the reports available on their websites [10, 11]. A literature search of Pubmed databases and proceedings of the American Society of Hematology annual meetings was also conducted to identify important trials relevant to regulatory actions or clinical practice changes, including RCTs on AML patients treated with GO. We included articles published up to April 2012 and indexed under the subject headings leukaemia, clinical trial, and GO.


Figure 1 shows the differences, in chronological order, of the approaches to GO approval among the regulatory authorities including the US FDA, EMA, and PMDA/MHLW, and the results of published important trials including RCTs. Table 1 shows the reported published clinical trials relevant to regulatory actions and pivotal RCTs.
Fig. 1

The differences of the approaches to GO approval among the regulatory authorities including the US FDA, EMA, and PMDA/MHLW in chronological order NDA new drug application; FDA the Food and Drug Administration; EMA the European Medicines Agency; MRC Medical Research Council; AML acute myeloid leukaemia; ALFA the Acute Leukemia French Association; NCRI the National Cancer Research Institute; PMDA/MHLW the Pharmaceutical and Medical Devices Agency/the Ministry of Health, labour and Welfare

Table 1

Selected trials of GO including RCTs




Age, Median (range)

No. of Pts

Dosage and administration of GO

Chemotherapeutic regimens

Interpretation of RCT results



CD33-positive AML in first relapse

61 (22–84)


9 mg/m2, day 1 and 15

Single Agent GO




Previously untreated AML

46.5 (18–59)


3 or 6 mg/m2, day1 of each chemotherapy course

GO + chemotherapy as induction and consolidation




Previously untreated AML

68 (61–73)


9 mg/m2, day 1 and 15

GO followed by chemotherapy




CD33-positive AML in first recurrence

61 (20–87)


9 mg/m2, day 1 and day15

Single Agent GO




Previously untreated AML

76 (61–89)


9 mg/m2, day 1 and 15

Single Agent GO




Relapsed or refractory AML

55.5 (16–71)


9 mg/m2, day 4

GO + chemotherapy




Relapsed or refractory AML

58 (28–68)


6–9 mg/m2 in phase I, 9 mg/m2 in phase II, day1 and 14

Single Agent GO




AML or RAEB in first CR

67 (60–78)

113 in GO arm

3 cycles of GO at 6 mg/m2 every 4 weeks

Single agent GO as consolidation therapy

Postremission treatments of GO does not provide benefits.



Previously untreated AML


277 in GO arm at induction

6 mg/m2, day 4 during induction, and 3 doses of 5 mg/m2 every 28 days after consolidation

Chemotherapy ± GO

GO did not show any outcome improvement but rather an excess of induction deaths.



Previously untreated AML

49 (0–69) at induction

556 in GO arm at induction, 473 in GO arm at consolidation

3 mg/m2, day 1 at induction and/or consolidation

Chemotherapy ± GO

There were no overall difference in response or survival. However, a predefined analysis showed significant survival. benefit for patients with favorable cytogenetics.



Favorable risk AML

48 (18–60)

132 in GO arm

6 mg/m2 before transplantation

± GO followed by autologous transplantation

The addition of a single dose of GO did not improve outcome.



De novo AML, Secondary AML, MDS

67 (51–84)

1,155 in total

3 mg/m2, day 1 at induction

Chemotherapy ± GO as induction

Significantly reduced rate of relapse and improved overall survival.


II, randomised

Previously untreated AML

69 (60–83)

57 in GO arm

6 mg/m2, day1 and 4 mg, day 8

Chemotherapy ± GO as induction

GO did not show significant superiority over standard therapy



Previously untreated de novo AML

62.2 (IQR 58.5–66.3)

140 in GO arm

3 mg/m2, day 1, 4, 7 during induction and day 1 of each consolidation

Chemotherapy ± GO

Improved event-free and overall survival.

GO gemtuzumab ozogamicin; RCT randomised controlled trial; AML acute myeloid leukaemia; NA not applicable; IQR interquartile range

GO was approved in May 2000 by the US FDA under the accelerated approval program for the treatment of patients with CD33-positive AML in first relapse who were ≥60 years of age and who were not eligible for other cytotoxic chemotherapies. The data submitted for approval included an interim analysis of 3 single-arm open-label trials that enrolled 142 patients with an overall response rate of 29.6 % [12]. The recommended dose by the US FDA 9 mg/m2 as a single agent, and the recommended treatment course was 2 doses with a 14-day interval between the doses.

However, the required post-approval study (SWOG S0106) combining chemotherapy with GO failed to demonstrate improved survival compared with chemotherapy alone in patients with previously untreated AML [13]. In this study, 627 patients with untreated AML aged 18 to 60 years were randomly assigned to receive induction chemotherapy with or without GO, at a dose of 6 mg/m2 for induction, followed by a maximum of 3 doses of 5 mg/m2 given at 28-day intervals. This study was designed to confirm the efficacy of GO in AML, but an interim analysis showed similar complete response rate and disease-free survival between the 2 arms, and the rate of fatal adverse events was higher in the group receiving GO compared with the control arm. Based on these results, the FDA recommended GO withdrawal from the market. Furthermore, similar results were reported by the MRC15 trial, a large study of more than 1,100 young, newly diagnosed AML patients who were randomly assigned to receive or not receive a single dose of GO (3 mg/m2) on day 1 of the induction course 1 with other induction chemotherapeutic regimens [14]. The 948 patients in remission were randomly assigned to GO in course 3 in combination with consolidation chemotherapeutic regimens. The MRC15 showed that the addition of GO did not affect either the overall response or survival.

In June 2010, based on the results of SWOG S0106, Pfizer Inc. decided to discontinue the sales of GO in the US, and to withdraw any new drug application for GO [10]. GO is still available for selected investigator-initiated clinical trials, but not for general healthcare providers in daily clinical practice.

In Japan, GO was approved in June 2005 as an orphan drug based on the results of the same 3 phase II studies submitted to the US FDA [12] and an additional phase I study and a phase I/II study conducted in Japan [15]. The indication for GO use is recurrent or refractory CD33-positive AML, and it is administered with the same dosage and administration approved by the FDA. The PMDA/MHLW demanded to conduct PMS for all treated Japanese patients receiving GO to collect data on the safety and efficacy of the drug. While the PMDA/MHLW re-evaluated the safety and efficacy of GO based on its withdrawal in several countries including the US, the Japanese authority decided to continue the approval process of GO in November 2010 [11]. At the time of re-evaluation, the PMS in Japan included 852 registered cases. The PMDA/MHLW confirmed the use of GO only at the approved dosage and conditions, and requested further continuation of PMS for all treated patients. Furthermore, the Japanese Society of Hematology and the Japanese Society of Clinical Oncology launched the academic registry system in August 2011 [16].

On January 2008, the EMA refused to grant market authorization for GO due to insufficient data on its clinical efficacy compared to other treatment options and the severe adverse events reported in the available single-arm clinical trials [8]. In Europe, the drug has been available on a compassionate basis for relapsed AML and investigator-initiated clinical trials.

As shown in Table 1, 2 European clinical trials with positive results have been published in 2011. In the Acute Leukemia French Association trial, the addition of GO with a new dosing schedule of 3 mg/m2 on days 1, 4, and 7 to standard chemotherapy significantly improved both event and overall survival in patients with newly diagnosed AML who were aged 50 to 70 years [17, 18]. In another trial, the National Cancer Research Institute AML 16 (intensive), researchers showed a significant overall survival benefit for newly diagnosed AML patients with a median age of 67 years at a dose of 3 mg/m2 given on day 1 of induction chemotherapy [19]. It should be noted that the disease status, dosage, and administration in these positive RCTs were different from the approved labels in the US and Japan.


Cancer therapies are moving rapidly and irreversibly in the direction of personalised medicine with precise characterization of disease subsets and development of targeted drugs. Historically, the most novel anticancer agents have been approved after demonstration of clinical benefits in RCTs that compare standard of care with the new treatment. However, the time and resources necessary to conduct well-designed RCTs are enormous, and it is often difficult to develop novel therapeutics for life-threatening diseases with limited numbers of patients such as leukaemia. To improve access to new drugs for such patients, the US FDA has established accelerated approval regulations, in which sponsors are allowed to begin marketing novel drugs on the basis of trials that identify improvements in surrogate endpoints that are likely to predict clinical benefit. In Japan, although the PMDA/MHLW has no formal regulations similar to the accelerated approval program of the US FDA, novel agents designed for orphan diseases could obtain approvals without confirmatory trials in patients with life-threatening diseases. In the EU, the EMA has also implemented regulations of conditional approvals and approvals under exceptional circumstances, but their application for anticancer drugs are limited. Thus, each regulatory agency has its own rules and requirements for the approval of new drugs, and this often results in different risk assessments, especially when there are no confirmatory RCTs. The current approval system in each regulatory agency has its drawbacks and advantages, and the different requirements and actions of the regulatory agencies could cause confusions in pharmaceutical industries and among global medical professionals as well as patients. In addition, the most appropriate indication as well as dosage and administration of a new drug may not be properly determined in accelerated approval processes because of the limited number of patients and trials before approval.

In this case study, we found substantial differences in the regulatory actions among regulatory agencies concerning GO approval. As shown in Fig. 1, after the first approval by the US FDA in 2000, GO was voluntarily withdrawn from the market in the US and other countries in 2010 and the EMA refused to grant marketing authorization in 2008. In Japan, the PMDA/MHLW approved GO in 2005, and after their reassessment confirmed the market approval in 2010. Thus, Japan has become the only country that still sustains the market authorization of GO. It should be noted that 2 RCTs with negative results (SWOG S0106 and MRC15) were performed with a modified dosing schedule and mainly in younger patients, i.e. in conditions which were different from the approved indication and dosages in the US and in Japan. The previous single-arm studies suggested that GO is effective for patients with acute promyelocytic leukaemia [20, 21] and refractory AML [22, 23]. We speculate that if RCTs had been performed in the relapsed or refractory settings or more specified disease categories, the US FDA might have reached another conclusion.

No regulatory system is flawless, especially for an orphan drug for a life-threatening disease. A large time-consuming pivotal trial for approval usually examines only 1 treatment at a time, and it is often difficult to determine the most appropriate doses and candidates based on its results. It is important to ensure that in these studies the efficacy of promising drugs is not overlooked. In the case of GO, the subgroup analysis of cytogenetic-defined risk groups showed a significant survival benefit for patients with favourable risk disease and a trend for those with intermediate risk in the MRC15 RCT [14]. The subgroup analysis in the SWOG S0106 RCT also showed a significant benefit for the favourable risk patients treated with GO [13]. Therefore, when subsets of AML patients that benefit from the addition of GO to their initial and salvage therapy are not common, the current regulatory system could fail to approve the drug for these less common groups of patients. Furthermore, investigator-initiated randomised trials reported in 2011 showed promising results with dose modification of GO in patients with untreated AML [1719]. Although the Japanese authority has sustained its approval for GO use, they restricted the indication to refractory disease and the original high dosage. Because the standard chemotherapy for patients with newly diagnosed AML remained unchanged for decades [24], GO could be a potential breakthrough treatment if the results were confirmed, and it should be reflected into clinical practice in a timely manner without lengthy regulatory processes. The stringent regulations may not be suitable in case of an orphan drug, and more flexibility should be allowed to healthcare providers and patients with poor prognosis to facilitate effective developments of new anti-neoplastic agents. The straightforward refusal of GO approval may deprive patients with AML of potential therapeutic opportunities. Moreover, patients may seek to be treated with GO outside clinical trials [25]. In the era of personalised medicine, the collaboration between industry and academia for the drug development process should become more important than ever, and regulatory agencies should reconsider traditional regulatory processes, which depend largely on conventional clinical trial designs. The regulatory policy in each region is influenced by its own health care system as well as cultural and historical context with jurisdiction.[33] However, differences in access to novel therapeutics should be minimized in such a globalized society by making further efforts on harmonizing decision making among regulatory systems. Promotion of regulatory science would have an important role if more concrete researches were accumulated from all over the world.

In summary, this study highlighted the differences in the approach to GO approval among the 3 regulatory authorities. More studies are needed to establish the optimal evaluation methodology for anti-cancer drug approval in order to demonstrate safety, and to ensure the access to treatment for every patient who is expected to receive benefits from new drugs.


Not applicable.


The authors declare no conflict of interest.


Not applicable.

Copyright information

© Springer Science+Business Media, LLC 2012