In May 2010, the EMA held a conference to evaluate 10 years of Orphan Regulation in Europe. Over the past decade, this legislation has resulted in the approval of 62 OMP's out of > 700 OMP designations 3 . Of these 62 orphan drugs, the majority were for cancer treatments (35.5%), with the next highest being for metabolic diseases and miscellaneous disorders (both 24.2%), followed by cardiovascular and respiratory disorders (8.1%), immunological (4.8%) and musculoskeletal and nervous system disease (3.2%). Post-authorization requirements differed: 54% were authorized without specific requirements, while 41% of OMP's were authorized "under exceptional circumstances" and 5% received "conditional approval", the latter meaning that results of additional studies will be needed for full authorization. Within the relatively large group of metabolic diseases, the majority are LSD's (8 products for 7 diseases, including one product for two diseases and in two cases two enzymes for one disease). These orphan products are mainly enzymes produced in genetically engineered cell lines, which need to be administered intravenously. Oral substrate inhibitors have been developed as well, although miglustat (Zavesca) is the only one that has received approval so far. An authorization "under exceptional circumstances" and a requirement to install post-marketing drug registries are common for OMP's that are indicated for LSD's. With the exception of Myozyme for Pompe's disease, all OMP's for LSD's have been authorized under exceptional circumstances. This is most likely explained by the extreme rarity of LSD's: the prevalence of the diseases for which an OMP has been approved is higher than 1 in 10.000 in more than half of the cases, whereas the prevalence of LSD's vary between 1 in 600.000 and 1 in 37.000 4 5 6 7 , see table 1. Since the LSD's for which OMP's are currently available, usually have a slow progressive course and require lifelong treatment, it takes a long time before additional effectiveness and safety data becomes available. The European authorities therefore depend heavily on the outcomes of these registries for final authorization. The need for a post-marketing surveillance system to improve our understanding of long term effectiveness and safety of treatments for LSD's is not debated and LSD registries have proven to add meaningful data 8 . However, the EMA requires separate post-marketing registries for each drug, even when more OMP's are available for one rare disease. In addition, the current drug registries do not sufficiently fulfill the requirements for a regulatory registry. The following examples will illustrate the limitations and challenges of the current drug registries.

Fabry disease is an X-linked lysosomal storage disorder resulting from deficient activity of the enzyme alfa-galactosidase A 9 . Storage of globotriaosylceramide in the vascular wall underlies the pathophysiology 10 . Male patients typically develop painful acroparesthesia followed by proteinuria, renal failure, cardiac hypertrophy and cerebral white matter lesions. Females are carriers and may present a wider range of symptoms. For Fabry disease, two enzymes were granted marketing approval under exceptional circumstances by the EMA in 2002. Both enzymes were tested in small placebo-controlled studies of short duration, that heavily lent on surrogate end points and limited clinical effectiveness parameters 11 12 . Following approval, both pharmaceutical companies, Shire Human Genetic Therapies for Replagal and Genzyme Corporation for Fabrazyme, were asked to provide data for annual reassessment of the benefit/risk profile of their product. Both companies set up a drug registry, (called the Fabry Outcome Survey (FOS) and Fabry Registry respectively). Although each has a board of independent advisors, the pharmaceutical company manages each database. The situation whereby two competitive enzymes were approved at the same time each being evaluated separately resulted in treatment centers being generally involved with only one of the products. In essence, identical datasets were being collected and separate working groups established to address, for example, data on females or children, or effects of treatment on the kidney. It needs to be acknowledged that since these registries did not only address long term effectiveness and safety outcomes, but also focused on natural history data and diversity of disease manifestations, they have added important information. Increased awareness and a better understanding of the natural course of the disease, specific involvement such as the heart, and symptoms in females and pediatric patients, has been achieved 13 14 15 16 17 18 . The existence of two registries, however, has also resulted in publications that were quite similar 13 14 . Importantly, however, reports from the registries on the longer term outcomes of ERT, (the primary motive behind the requirements for these registries) have been scarce and have only been published using the FOS registry 19 20 . From these publications it is clear that datasets that are complete and accurate enough for analysis are usually extremely small compared to the total number of patients in the registry 20 . This may also be caused by the fact that the quality of the data is variable due to the lack of standardization of assessments. For example, renal function can be measured in different ways with different levels of accuracy 21 . Whether the selection of such a small sample results in a bias is difficult to ascertain, but is certainly a risk. In addition, comparator data from untreated patients is usually lacking. The paucity of high quality long term studies makes it difficult to understand the true health benefit of enzyme therapy in Fabry disease 22 . This is a particular challenge since the disease is highly heterogeneous. For example, it only gradually became clear that in patients with advanced disease the process of deterioration could not be stopped 23 24 25 . Although it is believed that early therapy could be beneficial, no studies have convincingly shown that occurrence of organ damage can be prevented. How the enzymes compare to each other is another unresolved issue 24 . So far, physicians and researchers have not sufficiently joined forces to initiate multi-center studies independent from industry to address these important longer term outcomes. The lack of collaboration and the separation of data in the two drug registries have undoubtedly contributed to the fact that after almost 10 years experience with Fabrazyme and Replagal, the most important questions such as the benefit of early therapy, the comparison of drugs and the influence of antibodies, remain largely unanswered resulting in having to conduct further clinical studies.

Gaucher disease is the first disorder for which an enzyme therapy was developed. This disorder results from deficient activity of the lysosomal enzyme glucocerebrosidase and is clinically characterized by hepatosplenomegaly, cytopenia with fatigue and bleeding, and devastating bone complications 26 . Cerezyme (imiglucerase) was licensed in Europe in 1994, before the European Orphan Drug legislation was enacted. This treatment has proven to be extremely beneficial in reversing the major manifestations of Gaucher disease 26 . Following this tremendous success, many questions remained unanswered. Since the disorder is extremely heterogeneous, ranging from asymptomatic cases to severe forms in childhood, debates followed on optimal dosing, costs, timing of initiation and prevention of long term complications 27 . The Gaucher Registry was launched in 1991 by Genzyme to collect data on longer term outcomes as well as on the natural history of the disease. Following a Health Technology Assessment conference at the National Institutes of Health (NIH) in 1995, it was suggested that NIH initiate and foster the establishment of a cooperative group of investigators in order to create a patient registry for the analysis of data on natural history and response to therapy 28 . This initiative was not followed through and the Gaucher Registry was further expanded and now contains information on over 5,500 patients. Some interesting publications have arisen from its data, including a number of studies on the variability and rate of improvement of different disease parameters 29 30 31 . However, variability of assessments and lack of completeness limits its use for robust analyses. For example, in a dose-efficacy analysis, only a subgroup of 366 patients from the registry were analyzed and although the study showed that responses were dose dependent, important outcomes on rates of complication or quality of life could not be extracted 32 . In a study on cost effectiveness of ERT, Connock and coworkers concluded that "Gaucher Registry data, which potentially represented the richest source of observational data for this purpose, were inadequate for the task in hand" 33 . They also expressed concern that the decisions about analyses of Gaucher Registry data largely depend on people who have an interest in the product and recommended that analysis of registry data should be undertaken in such a way that analytical, reporting and publication biases are minimised 33 .

Although the Gaucher Registry allows the addition of data on untreated patients as well as on patients treated with competitive products, fragmentation of data with new OMP's entering the market remains a real concern. A second treatment for Gaucher disease with the oral substrate inhibitor Zavesca was licensed in 2002 as an OMP 34 . As part of post-marketing commitments to the EMA, a safety database was set up but this was not designed to evaluate efficacy. Apart from a switch study, no comparative data is available and the registries operate separately. In the near future, data on effectiveness of Gaucher treatment will become increasingly difficult to assess: two new enzymes have recently been developed: velaglucerase (h-GCB, Shire Human genetic Therapies, MA, USA), and taliglucerase (pr-GCD, Protalix Biotherapeutics, Carmiel, Israel). Velaglucerase has recently received marketing authorization in Europe and the EMA has again required, post marketing safety data which will be collected by the company in a drug registry. If this will also be required for taliglucerase and also for the new substrate inhibitor eliglustat (Genzyme Corp) in the course of development, five different drug registries for a single rare disorder will exist. The rapidly increasing number of OMP's for LSD's illustrates that this will only lead to further fragmentation of data (table 1).