Representing the first gene therapy to be approved in the Western world, alipogene tiparvovec (Glybera; Uniqure) has recently been said to have had a “substantial impact from a regulatory perspective” (Nature Rev. Drug Discov. 11, 664; 2012)1. The therapy was granted marketing authorization in the European Union for the treatment of lipoprotein lipase deficiency, which results in a clinically heterogeneous condition with a risk of potentially life-threatening pancreatitis2, at the end of 2012. The decision followed a positive opinion by the European Medicines Agency (EMA)'s Committee for Medicinal Products for Human Use (CHMP)3, and a previous recommendation of the EMA's Committee for Advanced Therapies (CAT)4,5.

The approval process for Glybera was extensively discussed in the scientific community, sometimes critically6,7,8,9,10,11,12. During the process, the opinions of the CHMP and the CAT differed: although the opinion of both committees was originally negative, in a “re-examination procedure” the opinion of the CAT became positive5, whereas the CHMP maintained its negative opinion13. However, both committees finally recommended approval of the medicine. As regulators who have been involved in this approval process, we would like to provide insight into why the Glybera procedure was challenging, and give assurance to the scientific community regarding confidence in both orphan drug and gene therapy regulation in Europe.

The approval process for Glybera

In Europe, gene therapies undergo a centralized approval procedure via the EMA. For advanced therapy medicinal products (ATMPs), which include gene therapies, the CAT as an expert committee first performs a scientific assessment of the application dossier and prepares a draft opinion on its approvability for a final decision by the CHMP, a committee with considerable long-term expertise, which also ensures consistency in the opinions.

The assessment process for Glybera was long and complex owing to multiple reasons. These included the complexity of the product class for which there was little previous regulatory experience (this was the first procedure for a gene therapy to correct a genetic deficiency); the long product development time, during which science evolved and specific regulatory requirements were about to be established; the complex disease scenario (a very rare disease) with a fluctuating clinical outcome (pancreatitis); and the fact that the company was small with academic origins (as usually also seen for other ATMPs)14.

The applicant's total clinical programme included 27 patients with lipoprotein lipase deficiency on a low-fat diet2. At first, the main measure of efficacy was based on a reduction in blood triglyceride levels. However, this was later changed to postprandial chylomicrons, as this biomarker was thought to more specifically address the pharmacodynamic effect of Glybera, whereas the effect on triglyceride levels was only short-lived. This raised additional issues during the scientific review: were the data robust enough for a previously non-validated biomarker (postprandial chylomicrons) in the presence of more inconclusive clinical evidence (pancreatitis)? Should data from that biomarker be accepted as pivotal evidence for activity, as it was scientifically better fitted to measure the treatment effect (as recommended by an ad hoc scientific advisory group to the CAT and CHMP) than the one originally defined in the protocol (triglyceride levels)? Or should one rather focus on that (formally failed) primary end point? There were only a handful of patients from the study population for whom such data on the newly proposed biomarker were available, and even fewer for a sustained period.

Although different opinions were issued by the two committees during the procedure, both committees were consistent in their scientific judgement when the details were considered. The CHMP acknowledged the promising nature of the data by an almost unprecedented “positive tone” when drafting its initial negative opinion, so as to demonstrate that it would be prepared to reconsider the case once more specific and supportive data had been collected (and here the vast majority of the CHMP agreed). The CHMP was clearly of the view that development should continue. We consider that both committees had already, at that time, taken major steps towards finding a way for ATMPs to be used for the treatment of very rare diseases, by considering all data rather than single outcome measures. Both committees considered that evolution of scientific knowledge can make the appreciation of an emerging, biologically more plausible biomarker necessary, if well justified and supported by data. Both committees were well aware that the acceptance of a limited data set was a double-edged sword, as it could be perceived as lowering the standards. However, this was not the case: both committees agreed that the limited clinical efficacy and safety data set needed to be supplemented by additional data.

The opinion of the two committees differed on the stage at which such additional data should be submitted: before approval (as initially preferred by the CHMP) or after approval (as recommended by the CAT). It is important to note that the final positive outcome was, to a major extent, also driven by an appreciation of the specific clinical scenario of lipoprotein lipase deficiency. Assessing orphan drugs in 'ultra-rare' conditions (here defined as a prevalence of less than 1 in 100,000) was not new to the CHMP; several drugs are already licensed for conditions with such low prevalence (Fig. 1). However, all of these conditions follow either a continuously progressive course of deterioration or a very active disease course if left untreated (Table 1). The effect of a therapeutic intervention for such conditions is therefore easier to measure within a relatively short timeframe.

Figure 1: Authorized orphan drugs and ultra-orphan conditions.
figure 1

Orphan conditions with authorized treatments (at least one per condition) are ordered by prevalence (number of patients per 10,000 in the European Union). Analysis from data on file at the European Medicines Agency (EMA; cut-off: November 2012); in situations where a condition was listed more than once (and with varying prevalence), the lowest prevalence was listed. Ultra-rare conditions (the boxed indications in the figure) were defined as having a prevalence of equal to or less than 0.1 per 10,000 persons in the European Union. Lipoprotein lipase deficiency, targeted by Glybera (alipogene tiparvovec), represents the third rarest disease for which a medical treatment has so far been authorized. The pivotal data submitted for these ultra-rare conditions are listed for comparison in Table 1. ALL, acute lymphoblastic leukaemia; CEL, chronic eosinophilic leukaemia; CFTR, cystic fibrosis transmembrane conductance regulator; HPC, haematopoietic progenitor cell; Ph+, Philadelphia chromosome positive; SGCA, subependymal giant cell astrocytoma. *Treatment in patients who require intrathecal analgesia. ‡With platelet-derived growth factor receptor (PDGR) gene rearrangements. §C1 esterase inhibitor deficiency.

PowerPoint slide

Full size image
Table 1 Pivotal data for authorized treatments in ultra-rare conditions*
Full size table

Furthermore, surrogate markers for clinical outcome existed for these conditions (Table 1). By contrast, lipoprotein lipase deficiency results in pancreatitis in otherwise often phenotypically healthy individuals — a complication that is not only influenced by various factors including diet, but that also exhibits a much more unpredictable and fluctuating disease course with intermittent healthy periods. Therefore, it would — in theory — be necessary to conduct either a large clinical study enabling for a sufficiently high event rate (which is not feasible for an ultra-rare disease) or a long follow-up in order to detect a sufficient number of events in a smaller patient population (how long would be feasible in a pre-approval situation?). So, to counterbalance the inherent limitations of the clinical database, an approval “under exceptional circumstances”15 with post-approval supplementation of the database was the most appropriate route.

The way forward

In our view, the message to the scientific community is positive: both committees had complementary approaches. They delivered within the spirit of the legislation and took a scientific approach recognizing the rarity of the disease and the difficulties in obtaining comprehensive data. They acknowledged the evolution of science and considered the data as a whole. It is common in science that in borderline situations two experts can reach different conclusions when balancing certainties and uncertainties. However, this balanced and weighted scientific view of both committees, as briefly highlighted here, resulted in a seemingly 'black and white' outcome, as for a regulatory opinion there are only two possible decisions (positive or negative), and only this was visible to the scientific community.

During the approval procedure the two committees closely collaborated, and they will continue to strengthen their collaboration based on increasing experience. This will be facilitated by the recent inauguration of the EMA Scientific Coordination Board16, which involves all chairpersons of the EMA committees. We do not agree that this is “the perfect bureaucratic response”17, as it is not another committee that complicates procedures but a body that takes a more strategic and coordinative role under the umbrella of the EMA's missions and visions. Its creation is not a consequence of the Glybera procedure, but a logical step towards the fulfilment of the EMA's goal to ensure efficient operation of its core business — principles that were already published in early 2011 (Ref. 18).

The role of the CAT was recently challenged6,7,8,9,10,11,12, given that the CHMP is required by legislation to make the final opinion. This may be perceived as though the CAT is not heard, or that it lacks the power to foster innovation and create a fruitful environment for its stakeholders to develop and gain the approval of ATMPs. However, this assumption is not only incorrect but also oversimplified. Besides the fact that the CHMP fully supports innovation, the CAT has considerable influence at the early stage of ATMP development, for example, by systematically reviewing and discussing the CHMP's scientific advice procedures for ATMPs, by participating in the innovation task force meetings concerning ATMPs at the EMA, or when issuing certificates on quality and/or non-clinical data. The approach that the legislator had taken when drafting the ATMP regulation was not intended to separate the licensing procedure for ATMPs from that of other human medicines; it would be difficult to explain to patients why an ATMP should have a different standard compared to a conventional medicine, and it would not be helpful to exclude the CHMP's extensive experience of reviewing dossiers including those for drugs targeting rare diseases.

In conclusion, we submit that both committees' deliberations and conclusions illustrate their openness towards innovative approaches to drug development and their willingness to exercise flexibility where scientifically appropriate. The history summarized above also provides a reminder of the regulators' challenge to balance early availability of promising treatments with a need to uphold appropriate standards of evidence, and in borderline situations a different opinion of independent experts can never be fully excluded. We contend that a 'black and white' view of the regulatory procedure for Glybera (or any other similar scenario), as has been expressed in the literature, is not correct; rather, the opinions of the two committees illustrate the shades of grey involved in such procedures and also highlight the shared dedication of the committees to patients and innovation.

Disclaimer

This paper reflects the view of the authors, based on their observations of the discussions held at the committees. It may not reflect the view of all members of the Committee for Advanced Therapies (CAT) or the Committee for Medicinal Products for Human Use (CHMP). The views expressed in this article are the personal views of the authors and may not be understood or quoted as being made on behalf of, or reflecting, the position of the agencies and organizations the authors work for.