by Beat U. Steffen, Founder & CEO of confinis ag, on MDR Article 117

After the summer break I’m eager to continue! Hope everybody had a great time!

Many thanks to all of you who provided their feedback to Part 1, 2, 3, 4 and 5 of this series of articles – highly appreciated and overall a lot of great information for all stakeholders!

Flashback: In May and June 2018, TOPRA ( published two highly interesting articles written by Mark Chipperfield and Tim Chesworth that list a number of concerns and open questions that require clarification. As in part 1 and 2, please find below the respective text form the TOPRA articles (italic) and my interpretation / suggestion:

Applicability of normative standards. Normative standards and pharmacopoeia can have slightly different definitions and testing methods and monographs may be similar, but different. Considering the product primarily as a medicinal product or medical device can influence the testing performed. In certain cases, this leads to a medicinal product device satisfying an applicable International Organisation for Standardisation (ISO) requirement while the medicinal product or primary packaging constituent does not satisfy the ‘parallel’ pharmacopeia requirement. Additionally, the expectation is that normative standards are adapted to recognise specific risk. In an integrated product, the primary risks are often arising from the medicinal product constituent. This risk must be translated into the relevant level of technical requirements for the medicinal product device. An example could be the therapeutic window. If an accurate injectable dose is required, the ISO11608-1 requirements may not be sufficient. Based on the potential for harm, acceptance criteria and sample size should be adjusted for increased statistical confidence. Reviewers must recognise the most applicable standards and allow these to be justified and utilised.

Annex I of the MDR (and this is comparable to the requirements of Annex I of the MDD) requires a device to be safe and effective considering the generally acknowledged state of the art which is often defined in specific standards. If such a standard defines tolerances which are not good enough to lead to a safe product (e.g. because of the therapeutic window), the device would not meet Annex I and could therefore not be placed on the market.

Clinical data. The terms commonly used in the context of clinical assessment of products likely require further clarification within guidance. Clinical evaluation, clinical investigation of device and medicinal product, respectively, are often misunderstood. Several drug-device scenarios exist, such as:

  • Investigational medicinal product clinical trial, using an investigational medicinal product device (medicinal product endpoints, ‘unassessed’ medicinal product device to facilitate, possibly secondary device endpoints)
  • Investigational medicinal product clinical trial using an approved CE-medical device (medicinal product endpoints, CE-medical device to facilitate)
  • Investigational medicinal product device clinical trial using an investigational drug (device endpoints delivering an unapproved drug, possibly secondary medicinal product endpoints)
  • Investigational medicinal product device clinical trial using an approved drug (device endpoints, while delivering an approved drug)
  • Clinical bridging study (pharmacokinetic endpoints, comparing proposed device with a prior device).

The medicinal product clinical strategy may be key to NB review of the medicinal product device. For each of the above scenarios, clarity is needed around what data (if any) the NB would need to review to form its opinion. There is concern that any data submitted may not be reviewed from the appropriate perspective. The CA remains accountable for the overall approval decision. However, it must be clear what elements of any clinical investigation data need to be, or have been, reviewed by the NB. In some cases, the clinical intended use may be different to the final (commercial) intended use. For example, during a clinical study the medicine may have been delivered by a healthcare professional in a controlled study environment. Ultimately, however, the product may be self-administered in the home. Given that certain safety and performance requirements are closely associated to use, use error and user safety, it is suggested that the NB opinion should be formed regarding the commercial product use only, and so allowance for any differences in intended use must be made when reviewing any clinical material.

The NBs task is to assess if the device meets the GSPR. As part of this, the intended use is of course key which means that data acquired in a medicinal product clinical trial can only be used if it is relevant. Human factors are a classical example which may need to be assessed separately in a non-interventional summative human factors study. There may also be other relevant data, e.g. injection depth or other that may be relevant. This is very much dependent on the final embodiment of the device but guidance may indeed be helpful.

We consulted many companies over the last years regarding device aspects in medicinal product clinical trials to set it up in a way it also generates useful data for meeting Annex I. In several cases the Competent Authorities actually required to adding elements of EN ISO 14155 to the study protocol.

Equivalence and representativeness. Clearly, the earlier in a program that NB opinion is required or sought, the more likely it is that the device is not “final”. Additionally, any samples or material provided to the NB may not be fully equivalent or representative of the final product. This can be for several reasons, including ongoing manufacturing industrialisation, limited material availability, incomplete labelling negotiations with authorities, and the need for final adjustments to instructions and training arising from human factors design validation studies. It is not clear how the NB should manage evolving design or labelling finalisation relative to an opinion review. NBs should be cautious with issuing a positive opinion on a non-final product, especially where there is potential liability. Some allowance might be made for a provisional opinion, a clear statement of the opinion limitations, or even a staged process with supplementary review once final. The NB will need to understand and potentially review changes and ensure it remains comfortable that all relevant GSPRs will continue to be met by the final commercial product. However, the requirement for an applicant to notify and the change review process are not currently defined.

As far as I’m concerned, an opinion to be used in a market authorization dossier can only be provided on a final device. Notified Bodies should mention in the opinion which device configuration and documents (including version / revision) were assessed to prevent the use of an “invalidated” device configuration which does no longer represent the final device.

The design of a device will always change over time. Reasons may be that some raw material is no longer available, the design needs to be modified to eliminate flaws which led to complaints, manual assembly will be replaced by semi-automated or fully-automated assembly or because new features are added which is particularly likely (and easy to integrate) with devices containing software. As the authors point out, there is a loophole because nothing in article 117 requires sponsors to approach a Notified Body for an update of the opinion unless there is a new marketing authorization dossier that will be filed. I pretty much see two ways on how to approach this:

  • Either by a guidance that requires sponsors to seek an updated opinion in case of substantial changes and have it “on file” in case a medicinal product Competent Authority requires to see it, or
  • An update of the Medicinal Product Directive defines instances where an updated / new opinion needs to be obtained and submitted to the medicinal product Competent Authority.

Imagining how long a revision of the MPD may take, the only real option is probably the first one.

Regarding provisional opinions: since a Notified Body is a private organization, I’m pretty sure they will provide them and (hopefully) include a disclaimer that the opinion was provided on a non-final device and/or mention the specific configuration of the device.

Labelling. Additional device symbols are not currently added to medicinal products. In the case of a medicinal product device the CE-mark would not be applied, but neither should any other device symbol as dictated by, for example, ISO 15223-1. Another aspect of labelling is the need for the device manufacturer (eg, producer, chief marketing officer) to be mentioned on the labelling of the product – whether that be the (legal) manufacturer or the producer. Clarification would be highly beneficial.

Annex I of both the MDD and the MDR define requirements regarding labeling. Therefore, Annex I can’t be met if the labeling requirements are not fulfilled. For drug device combination products regulated as medicinal products, the labeling requirements of the medicinal product directive take precedence in case of conflicting requirements.

However, I’ve seen and been involved in many DDC products out there that have symbols as part of their labeling and to be frankly honest, this makes sense. It should not be seen as an additional regulatory burden but as a necessity to inform users about risks or precautions to be taken. There are fairly complex devices on the market and in development that contain electronics, near-field communication, software and the like. Symbols are important for providing critical information to the users and therefore minimize risk.

Expiry date, product stability and shelf-life. The effect of age and storage conditions on device components, device sub-assemblies, drug product, and container components need to be characterised, and their impact upon each other and the finished product understood in the context of the processing and storage conditions during final supply and process chains.

Device manufacturers have built up many years of experience regarding the expected changes that may occur to moulded and fabricated components. These data can be generated both in real-time and under specific engineering-science driven accelerated aging strategies. Testing of the finished combination product is normally only performed in a real-time aging model, since accelerated aging techniques adopted for drugs do not age integrated medicinal product devices at the same rate and vice-versa. The influence of one set of conditions on the other constituent is normally detrimental and significantly reduces the validity or usefulness of certain collected data. Any party involved in reviewing expiry date, stability or shelf-life data for a combination product must understand what the data truly represents in the context of the final product, before concluding suitability. There is a need for clarity on how the responsibility for review of such data will be split between a CA and NB.

This is indeed a complex area which requires a lot and diverse expertise. Many sponsors we consult have extensive programs in place to ensure (as far as possible) that the design is robust with regard to ageing under the expected storage and handling conditions (e.g. challenge testing that goes beyond any standard). From a Notified Body perspective, I would expect a solid design verification strategy explaining the purpose of each test and a rational why this is considered to be “good enough”. Given the divided responsibility regarding the assessment of a drug device combination product, the NB should focus on the device components including any negative impact from the container closure (e.g. break-loose force in the case of prefilled syringes or cartridges) and the drug product but not the drug product itself. The CA on the other hand should focus on the drug product, the container closure and any influences from the device part (e.g. fluid path).

This whole discussion reminds me of several discussions during scientific advice meetings with the FDA that expects that certain shelf-life tests are conducted in the “final (to be marketed) configuration” (see In the case of prefilled syringes provided with a needle-safety device and also in an autoinjector we got very inconsistent feedback whether such tests are acceptable if only performed with the prefilled syringe or whether it also needs to be performed in the autoinjector configuration.

Notified body expertise. The selection of a suitable NB will require evaluation of competency. Currently, the November 2017 NBOG Best Practice Guide: ‘Applied-for scope of designation and notification of a Conformity Assessment Body’ (NBOG F 2017 3 MDR) does not contain a designation code specifically for this type of product. Several codes may infer expertise:

  • MDA 0102 Active implantable devices for delivering drugs or other substances
  • MDA 0307 Active non-implantable respiratory devices
  • MDA 0309 Active non-implantable ophthalmologic devices
  • MDA 0318 Other active non-implantable devices
  • MDN 1202 Non-active non-implantable devices for administration, channeling, and removal of substances, including devices for dialysis
  • MDN 1206 Non-active non-implantable ophthalmologic devices
  • MDN 1214 General non-active non-implantable devices used in healthcare and other non-active non-implantable devices
  • MDS 1001 Devices incorporating medicinal substances (this code typically used for ancillary drug product)
  • MDS 1010 Devices with a measuring function
  • MDT 2007 Devices which require knowledge regarding the manufacturing of pharmaceuticals
  • MDT 2011 Devices requiring packaging, including labelling.

The codes MDN 0318, MDT 2007 and MDT 2011 could be relevant to the example of an integrated single-use spring-powered auto-injector, as an active non-implantable device. However, the same can be said of the example of a standalone reloadable injection pen (CE-medical devices). NB competence in the latter may not automatically qualify that NB to review the former, given all the points raised elsewhere in this article.

It is strongly recommended that the NBOG considers the need for specific designation code(s) for integrated drug-device products.

Working for a Notified Body I know how difficult it sometimes is to find a suitable code for a specific medical device and this is nothing different for drug-device combination products. The problem is really the vast number of different devices and embodiments that need to be covered by a finite number of codes. This approach will (in most cases) end up in a “puzzle” of different codes that need to be covered by the NB. I don’t really see another option because also integrated drug-device products may look very different, e.g. a bolus injector vs. a transdermal patch.

Regarding the example made above (single-use spring-powered auto-injector vs. standalone reloadable injection pen) I actually believe this is pretty close. If I compare this with CE marked medical devices, this is much closer to each other than some of the medical devices belonging to the same code.

Documentation to be issued by NB. Documentation appropriate to the liability of the NB in the provision of an official opinion must be considered. It is recommended that the NB issues a single-sheet certificate to the applicant and there should be harmonisation of certificate format across NBs. An initial proposal is outlined:

  • Unique document number (to enable referencing)
  • Name of NB
  • Clear title, ie, NB opinion certificate, Art.117 MDR (EU) 2017/745
  • Name of applicant
  • Product name
  • Design version (to support later change and notification management)
  • Scope of assessment
  • Criteria against which the assessment has been made (also see opinion criteria, below) – typically those relevant safety and performance requirements of Medical Device Regulation (EU) 2017/745, Annex 1
  • Clear statement of opinion outcome – eg, “Positive”, “Granted”, or similar
  • Date of issue
  • Expiry date (if ultimately deemed applicable)
  • Name of individuals performing the opinion assessment
  • Signature of responsible NB representative
  • No CE-mark.

It may be beneficial to delineate clearly between the look of an NB’s CE-marking certification and any Art.117 opinion certificate.

In addition to a certificate the NB should be required to issue a formal assessment report. It is assumed that such a report would be issued for all assessments performed, whether successful or not. It is vital that industry, NBs and CAs collaborate to define expectations in this area.

I definitely see a need in harmonizing the format of the “opinion”, particularly to make the life of the CA easier. In this way they know what they can expect. We know of Notified Bodies that will simply take a checklist according to Annex I, list the evidence they assessed with regard to all applicable sections and conclude that Annex I is met / not met. We know of other Notified Bodies that will write a formal report as for a medical device to be CE marked according to the NBOG document.

More to follow in Part 7, the last part of this blog series… Keep your eyes open!

Again, I strongly encourage other professionals involved in this field to provide their opinion and start a lively conversation on this platform – the whole industry can only benefit from this!

#MDR #Article117 #CombinationProducts #TOPRA #confinis