Adrenaline Auto-injectors: A Review of Clinical and

Adrenaline Auto-injectors: A Review of Clinical and Quality Considerations

04 June 2014

Contents Abbreviations ....................................................................................................... 3 1

Lay Summary.................................................................................................. 4

2

Introduction .................................................................................................... 6 2.1

2.1.1

The issues.......................................................................................... 6

2.1.2

History of auto-injectors ...................................................................... 6

2.2

3

4

5

Background .............................................................................................. 6

Anaphylaxis.............................................................................................. 8

2.2.1

Incidence and treatment ...................................................................... 8

2.2.2

Pharmacokinetics of adrenaline ............................................................. 9

2.2.3

Doses needed to treat anaphylaxis ...................................................... 10

Quality Aspects ............................................................................................. 10 3.1

Drug Substance:- adrenaline .................................................................... 10

3.2

Design and Operating Principle of auto-injectors ......................................... 10

3.3

Finished product specification ................................................................... 12

Non-Clinical Evidence ..................................................................................... 14 4.1

Gelatine models ...................................................................................... 14

4.2

Pig models ............................................................................................. 16

4.3

Non-Clinical Conclusion ............................................................................ 16

Clinical Evidence ............................................................................................ 17 5.1

Intramuscular vs subcutaneous injection .................................................... 17

5.1.1 5.2

Site of injection ...................................................................................... 19

5.2.1 5.3

Conclusion ....................................................................................... 21

Appropriate needle length ........................................................................ 22

5.3.1 5.4

Intramuscular versus subcutaneous injection conclusions ...................... 19

Clinical Comment .............................................................................. 25

Post-marketing data ................................................................................ 26

5.4.1

Exposure data .................................................................................. 26

5.4.2

Clinical Comment:............................................................................. 28

6

Discussion and recommendations .................................................................... 28

7

Independent Advice Received .......................................................................... 30

References ......................................................................................................... 32


Abbreviations AAI

Adrenaline Auto-injector

ADR

Adverse Drug Reaction

BP

British Pharmacopoeia

BMI

Body mass index

Cmax

Maximum plasma concentration

CT

Computed tomography

DoH

Department of Health

EVDAS

EudraVigilance Data Analysis System

IM

Intramuscular

ISO

International Organization for Standardization

MAH

Marketing Authorisation Holder

NHS

National Health Service

Ph Eur

European Pharmacopoeia

PIL

Patient Information Leaflet

PK

Pharmacokinetic

RMS

Reference Member State

SC

Subcutaneous

STMD

Skin To Muscle Depth

Tmax

Time to maximum plasma concentration

3


1 Lay Summary Adrenaline auto-injectors (AAIs) are intended for self-administration of adrenaline solution as an emergency, on-the-spot treatment during the early onset of symptoms of an anaphylactic reaction. As the progression of anaphylactic shock can be rapid, individuals with known allergy profiles are prescribed AAIs to carry with them at all times and they should be familiar with the operation of their specific auto-injector. The Medicines and Healthcare products Regulatory Agency (MHRA) has undertaken a review of all AAIs licensed in the UK on the recommendation of a coroner’s report into a death of a patient who had used such a device to self-treat anaphylaxis. This paper mainly discusses two of the issues raised by the coroner: 1. The most effective site for injection and the clarity of instructions 2. The most appropriate auto-injector needle length for injections into the muscle (intramuscular or IM) rather than injections into the fatty layer under the skin (subcutaneous or SC) administration The review has also considered information supplied with these products and whether clearer instructions and advice to prescribers, patients and carers could be provided in order to improve outcome. Anaphylaxis is a severe type of allergic response and is a life-threatening condition that can escalate into something very serious extremely rapidly. It can be associated with marked swelling of the face and neck causing constriction of the throat and upper airway, tightness of the chest and difficulty in breathing, a raised skin rash and sometimes a marked decline in blood pressure causing collapse of the patient. Known factors affecting severity of an anaphylactic episode include the degree of exposure to the substance responsible for the allergic reaction (the “allergen”) and other factors such as associated poorly controlled asthma, recent illness or strenuous exercise after exposure to the allergen. Fortunately, fatalities occurring as a result of anaphylaxis are rare and even less common when AAIs have been used. It is vital however that these devices are used correctly and an important part of the MHRA review has been to clarify information provided with these products to ensure as far as possible their correct use. Anaphylaxis can be fatal and in these unfortunate cases, death usually occurs very soon after contact with the allergen. Some allergens act faster than others. Food allergens can cause breathing to stop (respiratory arrest) after approximately 30–35 minutes; insect stings can cause collapse from shock after 10–15 minutes; and allergic reactions to medicines given by injection can cause death within 5 minutes. Therefore the speed of treatment of an anaphylactic reaction is of great importance and can have a significant impact on the patient’s recovery. It is widely accepted that an injection into the muscle is the best way for treatment with adrenaline to be administered. Even if the injection does not reach into the muscle, it will still have some effect, but it may take longer to relieve the symptoms of anaphylaxis. The best place for the injection is considered to be the side of the thigh in the middle between the hip and the knee, as recommended in the Resuscitation Council Guidelines. This review considered the data regarding all possible injection sites and concluded that patients should continue to use the middle of the thigh, as this represented the best location and minimised the risk of the needle going too deep and hitting bone or accidentally injecting adrenaline into a blood vessel or tendon which could cause additional problems. As everyone has different body shapes, concerns were raised about the length of the needle within the actual auto-injector devices and whether or not these were long enough to inject adrenaline into the muscle of all patients needing treatment for anaphylaxis. It is difficult to study how deep the needle goes into the thighs of patients using these devices. Models using blocks of gelatine and pork tissue have been used to represent the thigh and measure how far the adrenaline travels after being propelled from an auto-injector device following injection. The pork tissue model is considered more like the human thigh than the gelatine model but both models provide some data

4


that shows that the spring-loaded auto-injectors can project the adrenaline beyond the end of the needle to as much as twice the depth that the needle penetrates. However, not all of the pork tissue studies confirm this. Furthermore, the models cannot fully mirror the real-life situation where other factors exist including local tissue blood flow and barriers such as fibrous tissue layers surrounding muscle, all of which may have an impact on how efficiently the adrenaline can penetrate into the muscle tissue. Two studies19,20 measured the skin to muscle distance in adults and in children and showed that the skin to muscle depth is greater than the length of the needle (15mm) in many people, particularly women due to a different distribution of fat from men. These studies also showed that Body Mass Index (BMI) and skin to muscle depth are not directly linked and people with low BMI may have still have thighs with a high skin to muscle depth. The AAI devices are spring loaded and the manufacturers claim that the adrenaline is injected forcibly into the muscle tissue. This is supported by non-human studies which provide some reassurance that the adrenaline does penetrate beyond the exposed needle length. However, as outlined above, there are additional factors that may influence how well the adrenaline penetrates. The MHRA’s report was presented to and evaluated by independent panels of experts (Commission on Human Medicines (CHM) and the Chemistry, Pharmacy and Standards Expert Advisory Group) in January 2014 and a number of recommendations were made. The experts advised on improvements to the information for healthcare professionals and patients on the management of an anaphylaxis episode, they proposed that manufacturers should conduct studies to evaluate injection delivery and should improve the quality standards for AAIs. The full list of recommendations made is provided in this report. The recommendations are currently being taken forward by the MHRA for consideration at a European level. This will enable the different AAIs authorised across Europe to benefit from this review.

5


2 Introduction 2.1 Background 2.1.1 The issues Adrenaline auto-injectors (AAIs) are intended for self-administration of adrenaline solution as an emergency, on-the-spot treatment during the early onset symptoms of anaphylaxis. As the progression of anaphylactic shock can be rapid, individuals with severe allergies are prescribed AAIs to carry with them at all times and they should be familiar with the operation of their specific brand of auto-injector. A coroner’s report raised four areas for consideration and investigation in relation to the death of a patient following use of an AAI for emergency treatment of an anaphylactic episode. The four areas were: 1. 2. 3. 4.

The need to contact emergency services after first use of auto-injector even if symptoms are abating The most effective site for injection and clarity of instructions The most appropriate auto-injector needle length for IM injection rather than SC administration The best position for transporting a patient following an anaphylactic event

The MHRA was asked to address the first three items. Item 1 was addressed by the MHRA during 2012. All marketing authorisation holders (MAHs) were required to clearly state in the Patient Information Leaflet (PIL) and/or labelling of all AAIs licensed in the UK that the patient should call 999 even if symptoms appeared to be abating. Although the review did not specifically address Item 4, it did consider whether improvements could be made to the information supplied by the manufacturers of these products, relating to instructions to be followed by the patient/carer and healthcare professionals at the scene of the emergency, as well as advice for follow-up. Therefore the scope of this paper is primarily to address items 2 and 3. In order to help with this review the MAHs for EpiPen (Meda Pharmaceuticals Ltd), Jext (Alk-Abello A/S) and Emerade (Namtall AB) were asked to provide: (a) (b) (c) (d) (e)

Evidence that a complete dose of adrenaline solution is delivered intramuscularly throughout the proposed shelf life of the product Evidence that the above can be delivered through clothing Any post-marketing clinical evidence that the product (adrenaline plus device) is effective in the treatment of acute anaphylaxis A summary of out of specification (OOS) results from stability studies conducted on all product strengths over the past three years Product complaints history (reported by either patients or healthcare professionals).

2.1.2 History of auto-injectors Auto-injectors were developed in the 1960s for military use following research between the American military and NASA. The original objective was to develop a self-injecting device that would inject atropine, the antidote for nerve agents in biological weapons. From this original design platform the AAI was developed and was introduced into the medical field approximately 25 years ago in the United States of America. The first marketing authorisation in Europe was for EpiPen® which was granted a Marketing Authorisation in Germany in 1989 and in the UK in March 1996. Subsequently other brands of AAIs were licensed: Anapen®, Jext® and most recently

6


Emerade®. Anapen® is no longer marketed in the UK but is still available in other EU countries. Scope of review Currently the following AAIs are licensed in the UK for use in adults and children and were included in the review: Table 1: Licensed AAIs Product name

Container closure detail

Product licence number/Type of licence

EpiPen® Adrenaline (Epinephrine) Auto-Injector 0.3mg

Pre-filled cartridge encased in an autoinjector

PL 15142/0245

Marketing Authorisation Holder’s name and address

MEDA Pharmaceuticals Limited, Skyway House Parsonage Road

EpiPen® Jr. Adrenaline (Epinephrine) Auto-Injector 0.15mg Jext 150 micrograms solution for injection in prefilled pen

PL 15142/0246

Pre-filled cartridge enclosed in an autoinjector

Jext 300 micrograms solution for injection in prefilled pen Emerade 150 micrograms, solution for injection in prefilled pen

United Kingdom

PL 10085/0052 ALK-Abelló A/S Bøge Allé 6-8 PL 10085/0053

Pre-filled syringe encased in an autoinjector

Takeley, Bishop’s Stortford, CM22 6PU

DK-2970 Hørsholm Sweden

PL 42457/0001

Emerade 300 micrograms, solution for injection in prefilled pen

PL 42457/0002

Emerade 500 micrograms, solution for injection in prefilled pen

PL 42457/0003

European

NAMTALL AB Rapsgatan 7, SE-754 50 Uppsala, Sweden

European

No new clinical studies were required to be submitted in support of the original applications.

7


2.2 Anaphylaxis 2.2.1 Incidence and treatment Anaphylaxis is a severe, life-threatening systemic reaction that can affect all ages. The clinical syndrome may involve multiple target organs, including skin, respiratory, gastrointestinal and cardiovascular systems. The essential underlying mechanism is the presence of biologically active chemical mediators such as histamine and tryptase released from mast cells or basophils. The complex signalling cascades that regulate mast cell activation have been extensively investigated and described in the literature 1. The true incidence of anaphylaxis is unknown. Epidemiological studies have shown differing results owing to differences in both definitions of anaphylaxis and the population groups studied; however the incidence is increasing in recent years. Prescribing of adrenaline increased by 97% between the years 2001 and 2005. It has been estimated that by the end of 2005 there were 37,800 people in England that had experienced anaphylaxis at some point in their lives2. There are very limited data on trends in anaphylaxis internationally, but data indicate a dramatic increase in the rate of hospital admissions for anaphylaxis in England, increasing from 0.5 to 3.6 admissions per 100,000 between 1990 and 2004: an increase of 700% (Figure 1)5.

Most of the data for the incidence of anaphylaxis have been derived from hospital databases, and it is widely believed that anaphylaxis is under-recognised and underreported3. Anaphylaxis can be triggered by any of a very broad range of allergens, but those most commonly identified include food, drugs and venom (including wasp and bee stings). The relative importance of these varies very considerably with age; with food being

8


particularly important in children and medicinal products being much more common triggers in older people. Anaphylaxis remains a significant cause of mortality. Of 164 fatal reactions identified between 1992 and 1998 in the United Kingdom, around half were caused by drugs. Of those not caused by drugs, half were related to venom and most of the remainder to food3. With the increase in food allergies, the Office of National Statistics (ONS) started recording deaths from anaphylaxis due to food allergies separately from anaphylaxis due to other causes in 2002. When anaphylaxis is fatal, death usually occurs very soon after contact with the trigger. From a case-series, fatal food reactions cause respiratory arrest typically after 30–35 minutes; insect stings cause collapse from shock after 10–15 minutes; and deaths caused by intravenous medication occur most commonly within five minutes. Death never occurred more than six hours after contact with the trigger Studies of fatal and near-fatal anaphylaxis in humans delineate risk factors for anaphylaxis such as pre-existing asthma, a current asthma attack, food allergies (particularly peanuts, tree nuts and shellfish), reaction to trace amounts of foods and use of non-selective β-blockers4. Other factors include recent infection, intense exercise after the exposure and concurrent exposure to other allergens such as pollen in pollen allergic individuals. Treatment Early intramuscular adrenaline is the optimal treatment for patients suffering anaphylaxis5. Most studies of fatal anaphylaxis show that a lack of, or delay in, administration of adrenaline is a frequent factor in death, whereas early administration of adrenaline even in severe attacks is associated with survival. The median time to respiratory or cardiac arrest is reported to be 30 minutes for food- and 15 minutes for venom-induced anaphylaxis, so adrenaline usually needs to be administered before medical help is available. However, self-injectable adrenaline is underused even when it is available4. The recommended dose for auto-injectors is 300-500 µg for adults and 150-300µg for children depending on body weight (10 µg/kg). One injection from an auto-injector should be given immediately when symptoms are recognised and a second injection can be given 5-15 minutes later if symptoms are not improving. Therefore patients known to be at risk of anaphylaxis should have access to at least two AAIs. The Resuscitation Council guidelines advise that patients should always be observed after treatment for anaphylaxis, for at least 6 hours and up to 24 hours in adults and for 12 to 24 hours in children, as symptoms can recur up to 24 hours after the initial reaction (this is called a biphasic reaction). The incidence of biphasic reactions is reported as 1-20% and unfortunately it is not possible to predict which patients will experience a biphasic reaction.

2.2.2 Pharmacokinetics of adrenaline Adrenaline is a naturally occurring substance produced by the adrenal gland in the body and secreted in response to exertion or stress. Endogenous plasma concentrations of adrenaline in normal subjects are in the range 30–160 ng/L. Adrenaline is rapidly destroyed in the gut if swallowed and therefore needs to be given by injection. The effects of adrenaline after subcutaneous (SC) injection (injection into the fatty tissue beneath the skin) are produced within 5 minutes but increase more slowly, taking 30 minutes to reach optimal levels compared with a more rapid peak after intramuscular (IM) injection (injection into the muscle)6.

9


The amount of adrenaline in the blood is halved in about 2.5 minutes. However, by subcutaneous or intramuscular routes, local constriction of the blood supply slows the absorption, so that the effects build up and last much longer than the half-life of 2.5 minutes would predict7. Adrenaline does have side effects, mainly on the heart (fast or irregular heartbeat, or angina).

2.2.3 Doses needed to treat anaphylaxis Even though adrenaline is considered to be the optimal drug for use in connection with anaphylactic or threatening anaphylactic reactions, very little is known about what doses or plasma concentrations are required in this context. The recommended dose of adrenaline is usually within the range 5-10 µg/kg bodyweight but higher doses may be necessary in some cases. When adrenaline is delivered by an auto-injector device the following are recommended doses: in children between 15 kg and 30 kg in weight the usual dose is 150 µg and in adolescents and adults the recommended dose is 300 to 500 µg. There is a risk of overdosing small children with a body weight of under 15kg with an auto-injector so these are not generally recommended for such small children. The following intramuscular doses are recommended in the Resuscitation Council Guidelines which are specified as being in the context of administration by a healthcare professional: > 12 years:

500 µg IM i.e. same as adult dose 300 µg if child is small or prepubertal

> 6 – 12 years:

300 µg IM

> 6 months – 6 years:

150 µg IM

< 6 months:

150 µg IM

Most patients only require one dose but the dose can be repeated after 5-15 minutes if symptoms do not improve or recur. The scientific basis for the recommended doses is weak. The recommended doses are based on what is considered to be safe and practical to draw up and inject in an emergency.

3 Quality Aspects 3.1 Drug Substance: adrenaline The European Pharmacopeia (Ph Eur) is a publication detailing the official European quality standards for ingredients of medicinal products. The quality of the drug substance adrenaline is controlled according to the Ph Eur. specification in all of the licenced AAIs.

3.2 Design and Operating Principle of auto-injectors All AAIs comprise a sterile adrenaline solution filled into a container consisting of either a glass cartridge (also known as a carpoule) or pre-filled glass syringe with a fixed needle. In all cases they are made from glass suitable for injections. There are two fundamental designs for AAIs licensed in the UK:

10


-

the cartridge type injector for example EpiPen® and Jext®

-

the syringe type injector for example Emerade®

Each auto-injector brand has a different delivery/administration system. Likewise the firing mechanism which provides the force behind the actual injection process which pierces the skin and enters the outer (antero-lateral) thigh is unique to each brand. The assembled auto-injectors are enclosed in a “carry case” to protect them from mechanical shock and damage. In the cartridge type injectors the volume of adrenaline solution that the auto-injector contains (the fill volume) is significantly larger than the actual volume of adrenaline solution intended to be injected (the delivered volume) so unused solution remains in the activated auto-injector after use. In the case of EpiPen the volume of adrenaline delivered is the same in both the adult and the paediatric injectors: the concentration of adrenaline in the solution is adjusted to give the different doses (150 µg for paediatric use and 300 µg for adult). Conversely the paediatric and adult versions of JEXT and Emerade auto-injectors contain the same concentration of adrenaline solution but the delivered volume is adjusted to achieve the correct paediatric dose. As delivered volume relates to the quantity of adrenaline actually injected it is a critical test for all AAIs to ensure delivery of the intended dose throughout the shelf life of the product. During a conventional manual injection i.e. one given by a healthcare professional to an individual in a medical setting, the force to move the solution in a pre-filled syringe is provided by the thumb pushing the plunger. An auto-injector is generally intended for self-administration by an individual or by a family member or friend. Prior to use the plunger and needle are concealed within a plastic shell. The injector is activated by pulling off a cap or pressing a button and either swinging the AAI towards the thigh or placing it against the thigh. A coiled spring is then released inside the auto-injector which pushes the plunger to inject the solution into the patient. The adrenaline solution is pressurised to varying degrees depending on the design of the AAI. When the autoinjector is used the needle is propelled forward to pierce the skin and deliver the solution. Discussion on the design of AAIs There has been considerable discussion in the medical community and in patient groups regarding the suitability of the needle length used with AAIs with respect to the ability of the injectors to deliver the adrenaline solution to the optimal body compartment i.e. into the thigh muscle tissue. All UK licensed products claim to deliver an intra-muscular injection of adrenaline. The way the AAI is used (its method of operation), the force behind the adrenaline solution and how these factors contribute to the dose delivery have also been debated. Evidence of how these factors influence the site of deposition in the tissue is based on limited studies using non-clinical models. These three issues (a) needle length (b) method of operation and (c) applied mechanical force are discussed in greater detail below. (a) Needle length As there are differences in AAI design and method of operation, the total needle length cannot be considered on its own, as a portion of the needle remains within the device once fired - unlike a manual injection. The extended needle length measurement

11


provides an indication of the effective needle length available to deliver the adrenaline solution into the body and should be controlled. Schwirtz and Seeger11 reported “the mean exposed needle length was 15.36 mm (standard deviation [SD] 0.22) for Jext and 15.02 mm (SD 0.25) for EpiPen”. Nonclinical evidence exists -using both ballistic gelatine and porcine models - to support that adrenaline solution from both EpiPen and JEXT auto-injectors penetrate some distance into body tissues beyond the needle tip (Refer to section 4. Non-Clinical Evidence, for study detail and discussion). This suggests that there are additional factors to take into consideration when determining where the adrenaline solution is actually deposited in the body:1. the method of operation of the respective auto-injectors, and 2. the force applied to the plunger by the firing mechanism. (b) Method of operation In all AAIs a safety cap is removed immediately prior to the actual injection sequence. The safety cap is at the opposite end of the device from the needle and once removed the device is considered to be “armed” for use. There are two principal methods used for the self-injection of adrenaline using auto-injectors. These are the “swing and jab” method or “place and press” method. EpiPen® utilises the swing and jab method of administration while JEXT® and Emerade® utilise the place and press method. The method employed is related to the activation force required for each delivery system. It is possible that there is some degree of tissue compression during both the “swing and jab” technique and the “place and press method”. This may result in a net decrease in the skin to muscle distance (STMD), enabling the solution to penetrate deeper into the tissues. (c) Force applied to the plunger by firing mechanism/power pack As liquids cannot be compressed, the adrenaline solution is pressurised to varying degrees depending on the individual device design and construction. This phenomenon theoretically causes the solution to be expelled beyond the needle tip to varying degrees and is device dependent. This is confirmed by studies where ballistic gelatine has been used as a substitute for human tissue (Refer to section 4. Non-Clinical Evidence). However, it is not known how this correlates to administration into live human tissue.

3.3 Finished product specification AAIs comprise a drug product i.e. the adrenaline solution, which is sealed in a glass container, with a device component (the injector) for delivering the solution. These elements form the finished product. The finished product specification is a set of characteristics and acceptance limits that each batch of finished product must comply with before it can be released for sale. As a part of this review the finished product specifications were examined for all licensed AAIs. The tests applied to auto-injectors can be sub divided into the following: a. Tests to meet the Ph Eur general requirements for injections b. Tests to monitor adrenaline content, degradation substances and other impurities and levels of important ingredients such as sodium metabisulphite (an antioxidant used to stabilise the adrenaline solution)

12


c. Functional tests to monitor the performance of the delivery mechanism d. The British Pharmacopeia (BP) monograph controls the quality of the adrenaline solution to a minimum standard in respect of the content of the active pharmaceutical ingredient, pH and degradation products. The above quality standards are reviewed for each individual product in an application for a marketing authorisation in the European Union. Functional tests related to the delivery mechanism Although these quality characteristics are controlled in the design of the products, they should be brought together in the finished product specification: 1. Delivered volume The delivered volume (the volume of adrenaline solution released when the autoinjector is deployed) requires tight control. 2. Delivery time The time taken to eject the adrenaline solution from the needle (the delivery time) is critical. As anaphylaxis progresses very rapidly, delivery time should be measured for all AAIs and should reflect a rapid delivery time in the order of seconds. 3. Exposed needle length The design of an auto-injector should ensure that consistent extended needle length occurs when the device is activated by a patient. 4. Activation Force All auto-injectors need to be activated by the patient before use. This is achieved by removing the safety cap and either swinging or pressing the needle end of the device to the thigh. These operations should be possible for both adults and children; however the safety cap should not come away too readily either, to prevent accidental removal. The force required to initiate the injection cycle should be consistent during storage to ensure that AAIs are usable throughout the shelf-life period. Discussion on functional testing The approach to functional testing varies between manufacturers. Our recommendation is that the acceptance criteria for functional tests should be based on a critical evaluation of historical long-term stability data with consideration of the impact on the delivered dose. Critical quality attributes which ensure the correct dose is delivered within defined time limits should be included in the release and shelf-life specification requirements. AAI product defect reporting and product recalls In the past two years quality defects have been reported regarding Anapen® and JEXT. A recall was issued by the MHRA Defective Medicines Report Centre (DMRC) for all strengths of Anapen in 2012, based on finished product testing failures to deliver the correct volume and/or delivery time failure. The JEXT quality defect was announced by the Reference Member State (RMS) Sweden and a Class II recall notification was issued by the DMRC in early December 2013. In January 2014 Sweden (RMS) issued a Class II recall notification for Emerade due to suspected technical defects, at this point the

13


product had not been launched within the UK. In addition there was a recall in October 2013 concerning another auto-injector product (which does not contain adrenaline) as a small number of syringes potentially had needles protruding through the needle shield. Assembly of “ready to use” injectable drug products is complex and these incidents suggest that additional controls might be necessary. AAIs are intended for single use and are classified as medicinal products with an integral delivery system (device). The device aspects of AAIs should be designed and qualified by the manufacturers to be fit for purpose with relevant supporting data on the development and manufacture of the device submitted in the marketing authorisation application reviewed by the Licensing Authority. Although the AAIs do not require a CE mark, they should be compliant with the relevant sections of standards published by the International Organization for Standardization (ISO), for example BS EN ISO 11608 Needle-based injection systems for medical use. They should meet the essential requirements of Annex 1 of the Medical Device Directive. The marketing authorisations for AAIs should be reviewed with respect to ensuring the finished product specifications and in-process controls (IPCs) for the device assembly process are adequately described and that a summary of the following is provided. 1. The design and qualification of the delivery system i.e. the device development history. 2. A summary of identified critical failure modes for the delivery system. 3. Updated finished product specifications including appropriate functional tests with sample size tested per batch and the acceptable quality level (AQL) for each test. 4. An overview of how the essential requirements of Annex 1 of the Medical Device Directive are met. Product Stability Update At the request of the MHRA, the manufacturers of the licensed AAIs provided updated stability data for their products to the MHRA for review.

4 Non-Clinical Evidence Two main non-clinical models have been cited in the MAH’s response to the MHRA’s request to provide further information on their products; gelatine and porcine tissue. Both have been used in the study of ballistics and weapons research. A brief discussion is presented below of both models in the context of their usefulness for assessing the performance of injector pens in delivering adrenaline to the muscle layer.

4.1 Gelatine models Ballistic gelatine is reported as being designed to simulate living soft tissue (Nicholas and Welsch, 2004)8. It is regarded by the US military as the standard for evaluating the effectiveness of firearms against humans because of its convenience and acceptability over animal or cadaver testing. Use by the military would appear to have resulted in the acceptance of the use of gelatine in ballistic and other research and it has been referred to in some publications (including those cited by the MAH) as a ‘validated tissue simulant’. However, its use appears to be based more on custom and practice than inherent suitability. It was first used in 1960 and various techniques were used to measure the kinetic energy of a projectile travelling through a block of gelatine. Early models were not compared to living tissue in a quantitative or reproducible way.

14


In the first of many papers in the mid to late 1980s, researchers at the Letterman Army Institute of Research (LAIR) used both live swine (50-70 kg) and gelatine blocks to test bullets and subsequently compared the results9. Although the paper did not include specific comparisons between gelatine and animal tissue, the LAIR team and many other researchers afterward cited this published paper as the foundation for using Fackler’s gelatine model as an approximate or equivalent substitute for animal tissue. A paper by Fackler and Malinowski (1985)10 states that the depth of penetration measured in living swine leg muscle was reproduced in the gelatine within 3%. These findings and the convenience of using non-animal or non-cadaveric tissue appear to have led to the use of gelatine on its own. While the conditions and preparation of the gelatine have been standardised to some extent, and can be used to compare the behaviour of projectiles within that limited context, the model cannot be regarded as fully representative of living tissue, primarily because it is homogeneous rather than heterogeneous. The different types and textures of animal tissue, particularly bones, cannot be regarded as being adequately simulated in a gelatine alone system. Also, differences in the gelatine such as method of preparation, concentration and temperature mean that consistency between laboratories cannot be guaranteed. The acceptance of the model for ballistics has led to its use in investigating the track of injections from auto-injectors. In this context, it could be suitable for investigating the depth to which a drug might be injected, as the only tissues to be penetrated are skin and fat, unlike in ballistics research, where the full range of tissues could be encountered. It might be acceptable, for example, to use gelatine models to rank devices against one another for depth of penetration, but it would not simulate clinical conditions as closely as live animal or human tissue. A study comparing three injector pens was reported by Schwirtz and Seeger (2012)11. Three AAIs (Jext, EpiPen® and Anapen) were tested for, amongst other features, the injection depth and estimated volume of black ink delivered into ballistic gelatine. The mean maximum injection depths in gelatine within 10 seconds were 28.87 mm (SD 0.73) for Jext, 29.68 mm (SD 2.08) for EpiPen® and 18.74 mm (SD 1.25) for Anapen (Figure 2). The length of the EpiPen® and Jext needles is 14.3 mm and the Anapen needle is 8.9 mm to 9.9 mm.

Figure 2: Photographs showing the total injection depth into gelatine 10 seconds after activation of Jext (A), EpiPen (B), and Anapen (C), measured as the vertical distance from the surface of the gelatine to the lowest part of the ink area using digital image processing.

(Photographs copyright of Schwirtz and Seeger, 201212)

A previous pilot study reported by the same authors (Schwirtz and Seeger, 2010) 12 included a simulation of firing two AAIs through clothes, EpiPen® Junior and Anapen® Junior. Each auto-injector was fired into ballistic gelatine in the presence or absence of a

15


piece of denim (a double seam of Levi’s blue jeans). The activation force was recorded, and the effective (exposed) needle length was measured by a calliper after the device was removed from the ballistic gelatine. The presence of denim did not alter the activation force or effective needle length of either of the AAIs.

4.2 Pig models Based on current knowledge, the pig as an animal model for human skin is generally accepted as being the most representative of human skin13 and it is commonly used in pharmaceutical development for local tolerance and skin penetration studies. Given the difficulty in generating clinical data on injector pens, the use of the pig for this purpose is considered appropriate and the most valid model currently available. The MAH for Epipen® has cited a study conducted by the US military on the depth of penetration into porcine thighs achieved by the EpiPen® to address the question of its performance in obese patients14. Adrenaline from 21 EpiPen® devices was mixed with methylene blue as a colour tracer and triggered into the lateral aspect of 21 cadaver pigthighs. The results show that with an exposed needle length of 14.3 mm, the mean ± SD delivery depth from the skin to the muscle was 26.9 ± 5.4 mm (p<0.0001), consistent with approximately twice the needle length. All injections delivered adrenaline beyond the needle length and into muscle. However in these pigs the depth of the fat layer was only 7mm according to a further communication from one of the authors (T Ted Song). In the publication by Schwirtz and Seeger (2012) cited above, a separate study to simulate the clinical conditions more closely was reported. A model was used in which the contents of AAIs were replaced with a contrast agent, the AAI was activated into fresh cadaver pork shoulder, and computed tomography (CT) scanning was used to examine the injection pathway. The maximum exposed needle lengths ranged from 14.5 mm to 15.2 mm for Jext and EpiPen® compared with 8.9mm to 9.9mm for Anapen. The maximum injection depths ranged from 22.45 mm to 35.05 mm for Jext and EpiPen® and from 10.00 mm to 14.80 mm for Anapen. The maximum injection depth for Jext and EpiPen® always reached the muscle, even when the skin-to-muscle distance was greater than the exposed needle length. The depth of fat ranged from 3.90 mm (‘lean’) to 19.40 mm (‘fat’). The delivery depths reported for EpiPen® are in broadly the same range as those published by Ferguson et al14 (mean of 26.9 mm), and confirm that the contents of the AAI can be delivered to a depth greater than the exposed length of the needle. However, although the adrenaline solution appears to reach the surface of the muscle even when the skin to muscle distance exceeds the length of the needle, it is not clear whether the adrenaline actually penetrates the body of the muscle tissue – which would be required for optimal clinical effect – or whether it merely reaches the surface of the muscle and surrounding connective tissue.

4.3 Non-Clinical Conclusion There are difficulties with attempting to generate clinical data on AAIs because of the nature of the clinical indication and the conditions surrounding the occurrence of anaphylaxis. A valid model would, therefore, be very useful. The similarity between porcine and human tissues indicates that the pig model is suitable for testing injector pens and the use of the model is acceptable as it is the best substitute for clinical data currently available. The MAH’s approach in the response in using porcine data is considered reasonable and the results are considered to be acceptably robust and biologically meaningful. The MAH’s conclusion that the EpiPen® can be reliable for use in patients with a fat thickness of up to twice the length of the needle is supported by the data from the papers by Ferguson and Schwirtz. However, it is unknown what the limitations of cadaveric material might be and how representative it is of live tissue with a functioning blood supply.

16


The data from ballistic gelatine are consistent with those obtained using porcine tissue. Consistent results were obtained in the study reported by Schwirtz and for the purposes of comparison of the three devices, the model is considered acceptable. While the results in a gelatine system show that the degree of penetration was not affected by the presence of denim, the results are not considered as robust as if porcine tissues covered with denim had been used. Nonetheless, the data are considered to be reasonably reliable and the conclusion that the performance of the AAI is not affected by the presence of clothing, including denim, is considered acceptable.

5 Clinical Evidence The following section contains evidence from published literature and evidence provided by the Marketing Authorisation Holders (MAH) at the request of the MHRA.

5.1 Intramuscular vs subcutaneous injection There is some debate over the most appropriate route of administration of adrenaline in the treatment of acute anaphylaxis. Many different authoritative recommendations have been made but these are largely based on descriptive studies, clinical experience and tradition rather than on prospective clinical studies, tailored for these products. Adrenaline is most effective when given immediately after the onset of anaphylaxis symptoms. The initial recommended adult dose is 300 - 500 µg, injected intramuscularly in the anterolateral aspect of the mid-thigh. When injected by other routes, adrenaline appears to have a less satisfactory therapeutic window; for example, onset of action is potentially delayed when it is injected subcutaneously, and the risk of adverse effects potentially increases when it is injected intravenously. An intravenous injection should only be given under medical supervision when continuous monitoring is available. Pharmacokinetics Simons et al15 conducted a prospective, randomized, blinded, placebo-controlled, 6-way crossover study of intramuscular versus subcutaneous injection of adrenaline in healthy allergic men aged 18-35 years. The objective of the study was to provide information regarding the optimal route and site of adrenaline injection in adults. During the course of the study, each participant received 4 injections of adrenaline 0.3 mg (0.3 mL) and 2 injections of saline solution (0.9% NaCl, 0.3 mL) through use of a variety of injection routes and sites. Adrenaline USP I: 1000, 0.3 mg (0.3 mL) was injected either IM into the thigh (vastus lateralis) muscle or the upper arm (deltoid) muscle or SC in the upper arm. To ensure blinding, all injections were given by a nurse not otherwise involved in the study, and at each visit both the thigh and upper arm sites were covered after the injection.

17


Figure 3: Mean plasma adrenaline concentrations versus time are shown after administration of an identical 0.3 mg (0.3mL) dose of adrenaline by IM or SC injection in 2 different sites. T; Thigh; A upper arm. Mean endogenous plasma adrenaline concentrations are shown after IM or SC injection of 0.9% saline solution (0.3 mL) in the upper arm. The plasma adrenaline concentrations shown were calculated by averaging (mean ±SEM) the adrenaline concentrations at each sampling time for each route and each site of injection.

The results showed a swift increase in plasma levels of adrenaline following IM injection into the thigh, which was greater than levels achieved from an IM or SC injection into the arm. Unfortunately the study did not investigate SC injection into the thigh. The time to maximum concentration in the blood (Tmax) for the IM injection was around 10 minutes Using the EpiPen a second peak in plasma concentration was seen at 40 minutes which the authors suggest may be due to further absorption of exogenous adrenaline at the injection site after a period of initial vasoconstriction at the site, or due to rebound endogenous adrenaline release. The latter seems unlikely as it is not seen with the other routes of administration. Another explanation could be that part of the dose from the EpiPen was delivered subcutaneously and was therefore absorbed more slowly giving a delayed onset of action. A further study by Simons et al16 in children measured the pharmacokinetics (PK) of adrenaline following subcutaneous injection (9 children) and intramuscular injection (8 children). The study was a prospective, randomised, blinded parallel group study in children with a history of anaphylaxis. The subcutaneous injection was administered via needle and syringe while the intramuscular injection was administered using an EpiPen Auto-injector. Results In the nine children who received a SC injection the mean maximum plasma concentration of adrenaline was 1802 ±214 pg/mL, achieved at a mean time of 34 ±14 minutes (range 5 to 120 minutes). Only two of the children achieved a maximum concentration of adrenaline by 5 minutes. In the eight children who received

18


intramuscular injection via EpiPen the mean maximum concentration of adrenaline was 2136 ±351 pg/mL achieved at a mean time of 8 ±2 minutes, which was significantly faster than the mean time at which maximum plasma concentrations of adrenaline were achieved using the SC route. Figure 4: Mean plasma adrenaline concentration versus time after injection of adrenaline subcutaneously or intramuscularly

The results of this study, despite its limitations, support the intramuscular route as the optimal route of injection of adrenaline in the treatment of anaphylaxis.

5.1.1 Intramuscular versus subcutaneous injection conclusions The data regarding subcutaneous versus intramuscular injections are sparse and the recommendation for intramuscular injection of adrenaline in the treatment of anaphylaxis appears to be mainly based on theoretical grounds. It is imperative that the adrenaline is absorbed quickly in order to minimise the risk of a fatal outcome in anaphylaxis and therefore the intramuscular route is the logical choice. The study by Simons et al in children with a history of anaphylaxis supports the assumption that the intramuscular route gives a faster time to maximum plasma concentration of adrenaline, although the data are limited by the small number of children included in the study. It also lends some support to the supposition that EpiPen delivers its dose intramuscularly at least in the children studied; but it should be borne in mind that children, in general, have less subcutaneous fat than adults. Owing to the nature of anaphylaxis no clinical studies to compare the relative effectiveness of the two routes during an actual anaphylactic reaction have been conducted, nor would they be ethical. It may be that the auto-injectors actually deliver some of the dose intramuscularly and some subcutaneously. As the subcutaneous portion would be absorbed more slowly this may be beneficial in some cases where the anaphylactic reaction is prolonged, but it is imperative that the adrenaline is delivered quickly to halt the allergic cascade and therefore the major part of the dose should be delivered intramuscularly.

5.2 Site of injection The Resuscitation Council Guidelines state that the best site for IM injection is the anterolateral aspect of the middle third of the thigh and that the subcutaneous or inhaled routes for adrenaline are not recommended for the treatment of an anaphylactic reaction because they are less effective. Injection in the anterolateral aspect of the middle third

19


of the thigh is emphasised in the auto-injector Patient Information Leaflet (PIL) for EpiPen and in the DVD given to patients. A prospective study by Bewick et al17 recruited 93 children (age range, 1-16 years) with food allergies who attended the authors’ regional paediatric allergy outpatient clinics over a 6-month period in mid-2012. Using a MicroMaxx portable ultrasound machine with a linear HFL38/13.6 MHz probe the authors measured the distance from the skin surface to the vastus lateralis muscle interface at 3 distances along the outer thigh (onefourth [proximal thigh], one-half [mid-thigh], and three-fourths [distal thigh] the distance from the greater trochanter to the lateral epicondyle of the femur) as determined with a tape measure. Weight, height and waist circumference were also measured, and BMI as well as age- and sex-appropriate BMI centiles were calculated (Table 2). Table 2: Anthropometric measures of 93 children referred to the local paediatric allergy service Parameter* No. (%) Age (y), median (IQR) Boys, no. (%) Weight (kg), median (IQR)

Children <30 kg weight 62 (67) 4 (2-6) 35 (57) 16.6 (12.2 – 20.8) 102 (88-114) 16.1 (15.5-17.1) 52 (49-56)

Children >30 kg weight 31 (33) 12 (8-14) 19 (61) 43.8 (38.4-53.3)

Height (cm), median (IQR) 150 (140-159) BMI (kg/m2), median (IQR) 19.9 (18.2-22.4) Waist circumference (cm), 75 (68-80) median (IQR) Skin surface to muscle depth (mm), median (IQR) Proximal thigh 10.0 (8.3-13.2) 19.2 (12.8-25.7) Mid-thigh 8.4 (7.0-10.2) 12.2 (7.8-16.5) Distal thigh 6.8 (5.8-8.5) 9.7 (7.2-12.2) Mid-calf 7.0 (6.2-7.2) 9.5 (8.6-10.5) Skin surface to muscle depth >12.7 >15.9 greater than needle length (mm) Proximal thigh, no. (%) 17 (27) 19 (61) Mid-thigh, no. (%) 10 (16) 9 (29) Distal thigh, no. (%) 1 (2) 4 (13) Mid-calf, no. (%) 0 (0) 0 (0) IQR, Interquartile range *The median (IQR) is based on triplicate measurements

All children 93 (100) 6 (3-10) 54 (58) 20.8 (14.5-38.6) 114 (96-141) 16.8 (15.7-19.1) 56 (51-68)

12.0 (8.6-16.9) 8.8 (7.0-12.9) 7.9 (5.9-9.6) 8.5 (7.0-9.8)

36 (39) 19 (0) 5 (5) 0 (0)

20


Figure 5 Percentage of children whose skin surface-to-muscle depth was greater than the Epipen (children > 30kg) or Epipen Junior (children <30kg) auto-injector exposed needle length, based on level of obesity defined by age-corrected BMI. Healthy weight (BMI <85th percentile) n = 67 (72%), overweight (BMI 85th-94th percentile) n = 9 (10%), obese (BMI ≥95th percentile) n = 17 (18%). Differences within all groups were statistically significant, with a P <.001 (2 test).

A possible concern of injections into the distal thigh is the risk of the needle going into the bone. Skin surface-to-bone depth, therefore, was measured at the distal thigh in 11 children ages 1 to 15 years, with BMI units (kg/m2) that ranged from 14 to 27 (median, 17). The median depth was 29.5 mm (interquartile range, 21-36 mm). The thinnest skin-to-bone depth was 16.2 mm in a 5-year-old child with a BMI of only 14 and a weight of 15 kg.

5.2.1 Conclusion The site of injection as the anterolateral aspect of the mid-thigh has been accepted since the AAIs were developed and provides the best balance between safety and efficacy. It is also the area most readily accessible to the patient when self-administering the adrenaline. The study reported by Bewick et al in children demonstrated that over 50%, even in the obese category, would have a skin to muscle depth within the exposed needle length of the respective Epipen injectors (junior versus adult, prescribed according to body weight) at the mid-thigh. Furthermore, this study has not made any allowance for compression during the activation of the device or for the adrenaline being expelled beyond the needle length as demonstrated in the ballistic gel and porcine models, which would also increase the possibility of penetration into muscle. The proviso is that the data are circumstantial and do not directly demonstrate evidence for penetration of adrenaline into the body of the muscle tissue.

21


5.3 Appropriate needle length There are concerns that, owing to the increasing obesity (BMI ≥30) of the population in the UK, the needle lengths in the currently licensed AAIs are not adequate to deliver the dose of adrenaline to the muscle tissue of the thigh. A survey published in 2012 found that just over a quarter of all adults (26%) in England are obese. The report compiled by the Health and Social Care Information Centre, relates to information gathered during 2011. There has been a marked increase in obesity rates over the past eighteen years – in 1993 13% of men and 16% of women were obese; in 2011 this rose to 24% for men and 26% for women. For children attending reception class (aged 4-5 years) during 2011-12, 9.5% were obese18. A study by Song et al19 investigated whether EpiPen auto-injector, with a needle length of 14.3 mm, is sufficient for intramuscular delivery of adrenaline in men and women. The distance from skin to muscle in the anterolateral aspect of the thigh was measured in 50 men and 50 women who had undergone computed tomography (CT) of the thighs for other medical reasons. For each individual, body mass index (BMI; a measure of weight in kilograms divided by the square of height in meters) was also calculated, and the individuals were classified as underweight (BMI, 18.5), normal (BMI, 18.5–24.9), overweight (BMI, 25.0 –29.9), and obese (BMI, 30.0) using standard definition. The CTs were analysed for measurement of the distance from the skin surface to the muscle. This is the path the needle traverses before reaching the fascia of the vastus lateralis muscle.

Results The 50 men included 39 white individuals (78%), 4 African American individuals (8%), 1 Asian individual (2%), and 6 individuals of other races (12%). The 50 women included 35 white individuals (70%), 12 African American individuals (24%), 2 Asian individuals (4%), and 1 individual of another race (2%). In the study participants the mean ±SD distance from skin to muscle was 6.6 ±4.7 mm for men and 14.8 ±7.2 mm for women (P <.001). One man (obese at a BMI of 42.2) and

22


21 women (11 obese with a mean BMI of 35.2, 6 overweight with a mean BMI of 30.1, and 4 normal with a mean BMI of 24.5) had a greater distance from skin to muscle than the EpiPen extended needle length of 14.3 mm.

As a certain pressure is required to activate the EpiPen device, in order to investigate the role of any subsequent compression, the distance to muscle was measured with an ultrasound machine in 1 representative man and 1 representative woman with and without 8 lb. of weight applied. The 8 lb. of weight decreased the distance to muscle by 25% in the woman and 19% in the man. Assuming a liberal estimate of 25% compression of distance to muscle in both sexes, the authors recalculated the distance to muscle for all study participants. The single man with a distance to muscle of 34.7 mm would not be affected, whereas the number of women with a distance to muscle greater than 14.3 mm was calculated to still be 14 (28%). These results demonstrate that the EpiPen needle length is adequate to reach the muscle and therefore deliver adrenaline intramuscularly in most men but not in a number of women. Even when allowance was made for BMI the gender difference remained as seen in the figure above. Applying the pressure needed to trigger an EpiPen device decreased the skin to muscle distance in a representative man and woman but not sufficiently to ensure that an intramuscular injection of the dose would be delivered in all women or in very obese men. From this study it would seem that even the longer needle length of the Emerade auto-injector would not be adequate for all subjects. Another study conducted by Stecher et al20 in children demonstrated that the needle on AAIs is not long enough to ensure delivery of the medication intramuscularly in a significant number of children. Patients between the ages of 1 and 12 years who presented to a children’s hospital were enrolled in the study. Ultrasound was used to determine the depth from the skin to the vastus lateralis muscle. The patient’s body mass index was also recorded. The data were analysed using simple descriptive statistics, and logistic regression was used to identify variables that might predict whether or not the needle length was exceeded. In addition, the data were analysed using an estimate of 25% for displacement of tissue with applied pressure from the adult study cited above.

23


Results A total of 256 children were enrolled. Of these, 158 children weighed less than 30 kg and would be prescribed the 0.15 mg AAI (extended needle length of 10.16 to 15.24 mm). Nineteen of these children (12%) had a skin to muscle surface distance of >12.5 mm and would not receive adrenaline intramuscularly from current auto-injectors. There were 98 children weighing ≥30 kg who would receive the 0.3 mg AAI. Of these 98 children, a total of 29 (30%) had a skin to muscle surface distance of >16 mm and would not receive adrenaline intramuscularly. Figure 6: Scatter plot of depth to muscle from skin surface vs BMI (<30 kg group). The vertical line represents the length of the needle (12.7 mm).

Figure 7: Scatter plot of depth to muscle from skin surface vs BMI (>30 kg group). The vertical line represents the length of the needle (15.8 mm).

24


From these data there is no clear correlation between the muscle depth and the BMI in this population of children. Also, not surprisingly, unlike the adult population, there is no marked difference between the genders. A further study was conducted by Bhalla et al21 in order to measure muscle depth and evaluate predictors of auto-injector needle length inadequacy. This was a prospective cross-sectional study of a sample of low acuity emergency department patients aged 18 to 55 years. Demographic data and thigh circumference were recorded and body mass index (BMI) was calculated. Depth-to-muscle measurements of the vastus lateralus in a standing position, with and without gentle pressure to simulate muscle compression that occurs with correct auto-injector use were made using ultrasound. Results One hundred and twenty (120) subjects were enrolled with a mean BMI of 29.2 kg/m2. Thirty-one percent (31%) of the sample were found to be failure risks (36/116; confidence interval, 22.6%-39.5%) because these ED patients had compressed muscle depths exceeding 15.9 mm. Women were 6.4 times more likely than men to be a failure risk (54.4% vs 5% for men failure rate; P <.001). Failures were more likely to be shorter, have a higher BMI, and have larger thigh circumference (P <.001). Unlike the study conducted by Song et al 19, Bhalla et al found significant associations between compressed muscle depth and BMI (r = 0.48; P b .001) as well as between compressed muscle depth and thigh circumference (r =0.62; P b .001; Figs. 2 and 3 below).

5.3.1 Clinical Comment These three studies demonstrate that in adults and in many children a needle length of 14.3 mm and even the longer needle at 23 mm is not adequate to consistently ensure delivery of an intramuscular injection. Only the study reported by Bhalla et al demonstrated a correlation between BMI and skin to muscle distance. However, it may be difficult to recommend the appropriate prescription for the individual patient should devices with varying needle lengths be available. Although the conclusion from these studies is that the currently available needle lengths in AAIs are not adequate to ensure intramuscular injections in most patients this does

25


not take into account any projection of the adrenaline solution beyond the end of the needle by the firing mechanism of the device.

5.4 Post-marketing data The following relates to post-marketing data for the EpiPen® and Jext® AAIs. There are no post-marketing data for Emerade AAI as it is only recently marketed.

5.4.1 Exposure data Since AAIs are used in case of emergency only, it is difficult to estimate the number of doses actually administered as an unknown number of packages expire without having been used. Of over 2.3 million devices sold each year it is estimated by one of the manufacturers that only approximately 2% are actually used. Currently, an EpiPen Patient Register Survey is on-going in Belgium initiated by the manufacturer Meda Pharmaceuticals Ltd. One of the objectives is to evaluate how EpiPen® is used by the patients in terms of purchase, therapeutic administration and disposal. The survey started in March 2012 and about 400 patients are to be included. End of the survey was planned for November 2013. Final results are expected for end of the first quarter of 2014. Auto-injector not working in a critical situation Based on information provided in post marketing data, it has been estimated that ~40 units/million sold fail to activate. However it is often not possible to distinguish between device failures and handling error. The auto-injectors are used by any age group and by patients, caregivers or health care professionals in life threatening and high-stress emergency situations. Accidental injection During a two-year period 128 cases (medically confirmed and consumer cases) of accidental injections were reported. Of these 128 cases, 5 cases were additionally classified as “auto-injector not working in a critical situation”. This could have occurred due to people putting pressure on the wrong end of the device leading in many cases to accidental injection into a finger. Designs and labelling of devices have now improved to minimise this risk. When the injection is not administered in the correct way, this misadministration puts the patient suffering from anaphylaxis at significant risk. However backup measures – a second auto-injector or emergency treatment in an ambulance or in the hospital – are available and the overall number of these incidents is low. Recently the design of the EpiPen auto-injector has been improved to make handling easier and safer. The product information has also been updated to provide clearer advice and there are additional training materials available from the MAH. No or reduced effect from injection Over a two-year period, for one AAI nineteen (19) cases (medically confirmed and consumer cases) of “no or reduced effect from injection” were reported. Of these 19 cases, 5 cases were additionally classified as “auto-injector not working in a critical situation”. Of the 14 that were classified as “no or reduced effect from injection” only, 12 were considered serious and 5 were associated with a fatal outcome. The MAH for another AAI received 3 reports of lack of efficacy

26


Lack of efficacy reported to the MHRA Figure 8: Reports from the UK Yellow Card system

Figure 9: The EudraVigilance Data Analysis System (EVDAS) data are shown below. This database contains reports of adverse drug reactions (ADRs) from round the world.

27


The rise in the number of cases in 2013 in the EVDAS data can be attributed to new legislation being actioned which now requires companies to submit details of any reports of lack of efficacy in addition to any reports of suspected side effects. There is an apparent peak of reports around 2007 for which there is no explanation.

5.4.2 Clinical Comment: Lack of efficacy of the treatment of anaphylaxis with adrenaline may be due to several factors such as delayed administration at the point when circulatory collapse is already severe, the patient’s overall health, co-existing poorly controlled asthma, the amount of allergen exposure, poor or slow adrenaline absorption (perhaps due to a subcutaneous injection), and potentially, adrenaline resistance (for example due to concomitant betablocker medication). Lack of an initial response may require a second injection administered 5 to 15 minutes after the first. From the narratives of the cases reported to the MHRA the majority report difficulty in activating the device or failure of the device to activate. Many of the reports in the EVDAS database also relate to ‘device failure’. However the failure may not be a quality issue but more one of lack of training of the patient and/or carer who needs to be able to administer the adrenaline in an emergency situation. Although the percentage of devices reported to the MAH as possibly failing to deliver an effective dose is small, in a life threatening situation such as acute anaphylaxis even one failure may have fatal consequences. Of the 22 cases of ‘no or reduced effect’ from the auto-injector that were reported, 5 were associated with a fatal outcome. The possibility of providing two auto-injectors in a single pack has been proposed in order to facilitate the recommendation that the patient should always carry two auto-injectors with them. It is acknowledged that in some cases, a single injection is not sufficient to achieve a response for a number of reasons, including severity of attack as well as the possibility that a dose has not been effectively administered; a second injection may therefore be needed. If the patient routinely carries two auto-injectors this will provide a safeguard.

6 Discussion and recommendations The EpiPen® AAI was first licensed in the UK in 1996 for use in the treatment of acute anaphylaxis by non-medically trained staff or patients. The accepted optimal route of administration is by intramuscular injection into the anterolateral aspect of the mid-thigh and this is supported by limited pharmacokinetic data demonstrating that adrenaline by the intramuscular route is absorbed more quickly (shorter Tmax) and gives a higher maximum plasma concentration (Cmax) than by the subcutaneous route. This is important for a life-threatening condition such as anaphylaxis where swift treatment decreases the risk of a fatal outcome. However the data are sparse and mainly in healthy volunteers; no clinical trials have been conducted in anaphylaxis because of practical and ethical considerations. The recommendation that adrenaline is ideally administered by the intramuscular route in the community setting is accepted. The site of injection is accepted as the anterolateral aspect of the mid-thigh. The study by Stecher et al20 in children suggested that the distal thigh would provide a shorter distance to the muscle tissue but the risk of the needle hitting bone in this area is greater. This could lead to the needle not deploying correctly or breaking in the tissue or, theoretically, osteitis. It could be suggested that the distal thigh be recommended for more overweight and obese patients but recommending different sites in different situations could cause confusion. On balance it is accepted that the anterolateral aspect of the mid-thigh continues to be the recommended site of injection.

28


The clinical data describing the skin to muscle distances in adult and child populations support the concern that the needle length on AAIs is too short to be sure that the dose of adrenaline is consistently delivered intramuscularly in all patients. This is a particular problem in females in whom the skin to muscle distance is more frequently longer than the exposed length of the needles of the currently available AAIs. Even the most recently licensed auto-injector (Emerade) with an exposed needle length of 23 mm cannot ensure an intramuscular injection in all patients. The AAIs were originally developed for the military where the population would have been mainly male and with less variability in weight and distribution of adipose (fatty) tissue than in the current civilian population. Unfortunately there appears to be no consistent correlation between BMI and skin to muscle distance in males or females, adults or children so if a range of needle lengths were available it would be difficult to provide any practical recommendations for the prescription of the different devices. It is not known how closely the studies measuring skin to muscle depth mimic the real life use of an auto-injector. These devices are spring loaded and the manufacturers of EpiPen and JEXT auto-injectors claim that the adrenaline is injected forcibly into the muscle tissue. The only evidence to support this theory is non-clinical with the use of porcine or ballistic gel models and these studies do provide some reassurance that the adrenaline does penetrate beyond the exposed needle length of 14.3 mm and can penetrate tissue up to depths of 15 mm to 26 mm. This however would still not be sufficient to reach the muscle tissue in a substantial number of adults and children according to the studies reported by Song and Stecher19,20. A pressurised liquid will follow the path of least resistance. The resistance offered by a uniform block of ballistic gelatine is useful to provide a qualitative comparison tool to estimate the performance characteristics of different auto-injector products. However mammalian limb tissue is a composite of different strata from the skin surface to sub cutaneous adipose tissue to the muscle layer and other less permeable structures are present within and between each discrete layer, such as blood vessels and connective tissue forming a sheath-like covering over the muscle surface. The relative resistance of each physiological layer to needle penetration and solution penetration/diffusion is not known but is likely determined by multiple tissue architectural properties and vascularity. The Emerade auto-injector range has only recently been marketed (December 2013) in some EU markets but there are no comparative studies in the public domain involving this product versus JEXT and EpiPen auto-injectors. The selection of a 25 mm needle length (23 mm exposed needle length) in this product was based on recommendations (UK Resuscitation Guidelines 2008) for manual intramuscular injections. However even this length of needle will not deliver an intramuscular injection to all patients and it is not known if a 25 mm needle length is the optimum for an auto-injector. From the post-marketing data it is demonstrated that in a small number of instances the current AAIs are either failing to activate or failing to deliver an effective dose. The reasons for this are uncertain. In a 2-year period between 2011 and 2013, five (5) such cases were associated with a fatal outcome. It is also possible that in some cases even when delivered appropriately that the severity of the anaphylactic reaction is such that adrenaline will not be effective. Research is required to understand the factors that affect severity but is beyond the remit of this review. In order to ensure that all patients receive an intramuscular dose of adrenaline in the case of an acute anaphylactic reaction it is recommended that the needle length of all AAIs should be reviewed and the licence holders (MAHs) requested to provide data to demonstrate that an intramuscular injection is delivered consistently.

29


7 Independent Advice Received This review was presented to the Chemistry, Pharmacy and Standards Expert Advisory Group and the Commission on Human Medicines; panels of independent experts who advise the Licensing Authority. They advised on improvements to the information for healthcare professionals and for patients on the management of an anaphylaxis episode, they proposed that manufacturers should conduct studies to evaluate injection delivery and should improve the quality standards for AAIs. The full advice is summarised below: 1. In view of information that indicates that AAIs will not deliver an intramuscular injection in all patients the needle length for all AAIs should be reviewed by the manufacturers and increased, if necessary, to ensure that an intramuscular injection is delivered to a greater proportion of patients. This is of particular importance in patients with a high BMI or those with increased skin to muscle distance. 2. MAHs should include penetration and absorption studies using stable isotopes or other traceable markers and suitable imaging techniques and a range of patients to demonstrate the characteristics of patients that limit the likelihood of intramuscular injection. Currently approved AAI MAHs should submit such studies to the relevant competent authority. The design of such studies should reflect the technique for administration with an exact replica of the commercial product including any features deployed automatically after solution injection in order to prevent needle stick injuries. Pharmacokinetic measurements should be included if feasible. The resulting data should be reflected in the SmPC. 3. Although intramuscular injection cannot be guaranteed for all patients, the recommended site of injection should remain unchanged as the anterolateral aspect of the mid-thigh until evidence is presented from penetration and absorption studies that an alternative injection route is superior and does not present additional risks. 4. In the absence of evidence that the devices can deliver adrenaline intramuscularly to all patients, that the SmPC for all AAIs should be amended to state that “Successful intramuscular (IM) administration is dependent on the skin to muscle distance (SMTD) and in certain cases the administration may be subcutaneous. Care should be taken to ensure IM administration as far as possible”. 5. The outcome of this review on AAIs should be communicated to practitioners via a DHCP letter and an article in Drug Safety Update (DSU). Consideration should be given to the possibility of including on the MHRA website links to other websites that give helpful information on the use of AAIs (e.g videos). Consideration should also be given to informing the Ambulance Service of the outcome of this review. 6. Advice to the patient and carers and healthcare professionals in the PIL should be amended to: a. Ensure the instructions for use are clear and clearly illustrated. Reference to websites with video instruction on device use should be included. b. Reinforce the instruction to dial 999 immediately after use of an AAI c. Advise that the patient should not be left unattended and if alone should seek assistance immediately after using the AAI.

30


d. Advise that the optimum position for the patient while waiting for medical assistance is lying in a head down position to maintain effective circulation 7. Those food allergy sufferers with co-existent allergic asthma should be informed that exposure to other allergens could result in increased susceptibility to an anaphylactic reaction. This should be reflected in the SmPC. 8. Patient education with respect to their condition and training in the use of their prescribed devices should be reviewed to ensure that MAHs provide such information to all users of AAIs. Other points for consideration 9. The finished product specifications for all AAIs should be reviewed with a view to including or updating tests and their respective limits for functional performance. 10. The marketing authorisations for auto-injectors should be critically reviewed with a view to updating the in-process controls (IPCs) in place to ensure they are suitable for the delivery device assembly phase. 11. MAH’s should submit a summary of: (a) The design and qualification of the delivery system i.e. device development history (b) A summary of identified critical failure modes for the delivery system (c) Updated finished product specifications including appropriate functional tests with sample size tested per batch and the acceptable quality level AQL for each test (d) An account of how the essential requirements of Annex 1 of the Directive 93/42/EC on Medical Devices are met. 12. An appropriate ISO standard should be developed specifically for AAIs. 13. The excipients guideline regarding drug product labelling and patient leaflet information should be reviewed and amended to include related allergens to peanuts/tree nuts. Patients may be unaware of the importance of potential crossreactivity to related legumes for example fenugreek, chick pea and lupin flour. 14. MAHs should be encouraged to develop a 500 µg strength AAI.

AAI products are available in the UK and throughout Europe and have been authorised by either an independent national procedure or a European regulatory procedure. Consequently, the recommendations arising from this review are currently being taken forward for consideration at a European level by consulting the appropriate European scientific committees. This will enable the different AAIs authorised throughout the European Community to benefit from this review.

31


References 1

Metcalfe DD. Peavy RD. Gilfillan AM. Mechanisms of mast cell signalling in anaphylaxis. Journal of Allergy and Clinical Immunology, 2009. 124(4):639-46.

2

Sheikh A. Hippisley-Cox J. Newton J. Fenty J. Trends in national incidence, lifetime prevalence and adrenaline prescribing for anaphylaxis in England. Journal of the Royal Society of Medicine, 2008. 101(3):139-43.

3

McLean-Tooke AP. Bethune CA. Fay AC. Spickett GP. Adrenaline in the treatment of anaphylaxis: What is the evidence? British Medical Journal, 2003. 327(7427):133235.

4

Levin ME. Adrenaline for anaphylaxis - what is the evidence? Current Allergy and Clinical Immunology, 2012. 25(3):117-22.

5

Resuscitation Council (UK): Emergency treatment of anaphylactic reactions. January 2008. Annotated with links to NICE guidance July 2012.

6

Moffat AC. Osselton DM. Widdop B. Watts J. Clarkes Analysis of Drugs and Poisons, 4th Edition, 2011. Pharmaceutical Press, London, UK.

7

Emerade pre-filled pen; Summary of Product Characteristics.

8

Nicholas NC. Welsch JR. Ballistic Gelatin, Institute for Non-Lethal Defence Technologies Report, Pennsylvania State University, February 2004.

9

Fackler ML. Surinchak JS. Malinowski JA. Bowen RE. Bullet fragmentation: a major cause of tissue disruption. Journal of Trauma-Injury Infection & Critical Care, 1984. 24(1):35-9.

10 Fackler ML. Malinowski JA. The wound profile: a visual method for quantifying gunshot wound components. Journal of Trauma-Injury Infection & Critical Care, 1985. 25(6):522-9. 11 Schwirtz A. Seeger H. Comparison of the robustness and functionality of three adrenaline auto-injectors. Journal of Asthma and Allergy, 2012. 5:39-49. 12 Schwirtz A. Seeger H. Are adrenaline autoinjectors fit for purpose? A pilot study of the mechanical and injection performance characteristics of a cartridge versus a syringe-based autoinjector. Journal of Asthma and Allergy, 2010. 3:159-67. 13 Sullivan TP. Eaglstein WH. Davis SC. Mertz P. The pig as a model for human wound healing. Wound Repair and Regeneration, 2001. 9(2):66-76. 14 Ferguson JW. Merrill N. Song TT. Delivery Depth of Epinephrine by Auto-injector into Subcutaneous Tissue of Swine. Journal of Allergy and Clinical Immunology, 2008. 121(2)Abstract 98. 15 Simons FE. Gu X. Simons KJ. Epinephrine absorption in adults: Intramuscular versus subcutaneous injection. Journal of Allergy and Clinical Immunology, 2001. 108(5):871-73

32


16 Simons FE. Roberts JR. Gu X. Simons KJ. Epinephrine absorption in children with a history of anaphylaxis. Journal of Allergy and Clinical Immunology, 1998. 101(1 Pt 1):33-37. 17 Bewick DC. Wright NB. Pumphrey RS. Arkwright PD. Anatomic and anthropometric determinants of intramuscular versus subcutaneous administration in children with epinephrine autoinjectors. Journal of Allergy and Clinical Immunology in Practice, 2013. 1(6)692-94. 18 NHS Choices: Latest obesity stats for England are alarming. http://www.nhs.uk/news/2013/02February/Pages/Latest-obesity-stats-for-Englandare-alarming-reading.aspx (accessed 19/11/2013). 19 Song TT. Nelson MR. Chang JH. Engler RJ. Chowdhury BA. Adequacy of the epinephrine autoinjector needle length in delivering epinephrine to the intramuscular tissues. Annals of Allergy, Asthma and Immunology, 2005. 94(5):539–54. 20 Stecher D. Bulloch B. Sales J. Schaefer C. Keahey L. Epinephrine Auto-injectors: Is Needle Length Adequate for Delivery of Epinephrine Intramuscularly? Pediatrics, 2009. 124(1):65-70. 21 Bhalla MC. Gable BD. Frey JA. Reichenbach MR. Wilber ST. Predictors of epinephrine autoinjector needle length inadequacy. American Journal of Emergency Medicine, 2013. 31(12):1671-76.

33

Adrenaline Auto-injectors: A Review of Clinical and

Recommend Documents