Dr Mike Mullen, Dr Marcus Flather & Dr Shreenidhi Venuraju
Session chairs: Dr Mike Mullen & Dr Marcus Flather
Reporter: Sue Jackson RICR
Keywords
Cardiology, Endpoint selection, Migraine, Patent foramen ovale (PFO), Patient selection, Research management
PFO closure for migraine; Lessons from MIST
Dr Mike Mullen’s talk looked at Patent Foramen Ovale (PFO) and its relationship with migraine. We are all born with PFO; it is part of normal foetal physiology and without one you would have been unlikely to have been born. The PFO acts as a shunt, allowing the flow of oxygenated blood in the foetus. Once born this usually closes off over time; however, 25% of a normal population retain a PFO. If a large PFO remains and the shunting of blood from the right to left side of the heart continues, this can cause disease. We know PFO has a high association with stroke, transient ischemic attack (TIA), migraine and chronic obstructive pulmonary disease (COPD). Although the causal link is not known, we do know that something passes through this breach which causes problems, suggesting that the PFO should be closed.
Over the past 30 years, devices to close PFOs have been developed. The procedure is low risk, with 0% mortality and morbidity of less than10%; it is quick and easy and performed in a catheterisation laboratory. However, there was no clinical trial providing evidence of the benefit of this process, so the Migraine Intervention with STarflex (MIST) study was born.
Why was MIST needed?
MIST was the first randomized, controlled study looking at PFO closure. It was already know that PFO had a strong link to migraine and following closure the migraine had been shown to improve, so the trial was devised to show clinical evidence for this.
Evidence shows increases in prevalence in both migraine in patients with PFO and PFO in patients with migraine.
1 Additionally, this association appears strongest for migraine with aura, and both PFO and migraine are associated with cryptogenic stroke.
2 The estimated economic impact per year of migraine in 1996 was £650 million.
3 Patients who had had the PFO closed for other reasons had also reported benefit with regard to migraines, and PFO closures for migraine relief were beginning to be performed on the basis of this anecdotal evidence alone. The research was required to define clearly which patients would benefit and to determine the risk-benefit ratio before PFO closure should become accepted practice in treatment of severe migraine.
Trial design
The trial design for MIST brought its own ethical considerations, as it involved carrying out a sham procedure so that patients remained blinded as to which treatment they had received. It was therefore vital that good primary and secondary endpoints were also chosen to demonstrate clear efficacy. It was not really acceptable just to have a migraine reduction as the primary endpoint.
As such, secondary efficacy endpoints were defined:
- Incidence of migraine during the healing phase (the first 90 days post-procedure).
- Change in the severity of migraine attacks, based on migraine disability assessment (MIDAS) and headache impact test (HIT-6) assessment scales
- Change in the frequency of migraine attacks other than elimination of attacks
- Change in characteristics of migraine
- Change in severity, frequency and character of migraine relative to effective closure rate or presence of residual leak
- Change in quality of life based on the SF-36v2 general health outcome questionnaire
The study team included not only cardiologists but also a group of neurologists, along with a Data Safety Monitoring Board and medical monitor. Of the patients recruited to the study, 60% had a shunt of some type.
MIST results
The study failed to demonstrate any difference in the primary endpoint of migraine headache cessation or in any of the predefined secondary endpoints between the implant and sham groups.
Mike questions why the study showed these results: what went wrong? Why were the MIST results so different from those of the earlier observational studies? He explained that randomised, controlled studies are often less positive than observational studies. Some of the reasons for this are: prospective and contemporaneous measurement of outcomes, better recording and reporting of AEs and SAEs, and inclusion and exclusion criteria introducing bias to populations, making the study non-representative. On reflection, the MIST study measured the wrong patients. The group studied was a different patient population to the observational data patients:
- Patients with other indications for PFO closure (ie, who might have appeared in the observational studies) were excluded
- The inclusion criterion of severe migraine refractory to medical treatment was not the same as observational studies, where migraine was generally mild/moderate
- The International Heart School (IHS) guidelines lack precision and may lead to the inclusion of patients with chronic daily headache (CDH), depression etc.
- Although incidence of PFO was high, cardiologists reported that they were smaller and more difficult to find than in stroke patients
The endpoints studied were also wrong for a number of reasons. They were over-ambitious: complete cessation was unlikely in this population and reduction in migraine frequency and severity more clinically relevant. It was also important to establish baselines.
In conclusion, the results of the MIST study do not justify PFO closure for migraine, although they did demonstrate high incidence of PFO in the population of migraine sufferers. The study also established a paradigm for future studies, demonstrating the need for (and feasibility of) a sham arm and the importance of establishing clear baselines, and providing insights into more appropriate endpoints and patient inclusion/exclusion criteria.
Reliable evaluation of treatments
Dr Flather began his presentation by reminding us of some fundamental definitions:
- Research is the systematic collection, analysis and presentation of quality-assured information using accepted pre-specified methods with results subject to external review
- A clinical trial is the comparison of a new treatment with an existing treatment in human subjects, usually by random allocation, and using accepted methods for design, conduct and analysis
- Randomisation is a process that allocates patients to different treatment groups using the play of chance. This is distinct from ‘random selection’, where subjects are selected at random from a larger population. Randomisation is used so that, at the end of the study, you have two groups that are balanced for known and unknown variables.
We were then reminded and refreshed on the components of research, including development of protocols, review, reporting, publication, and innovation and changes in clinical practice.
The Clinical Trials and Evaluation Unit where Dr Flather is based has a large number of ongoing projects with a range of trials. He reviewed the seven most recent large phase III industry trials; three of these led to the product being licensed, but only one of these was actually a commercial success!
Clinical trials in the 3rd millennium
For the next segment of his presentation, Dr Flather reviewed the successes and pitfalls of clinical research as we enter the third millennium.
- He considers that there is no shortage of new treatments: pharmaceutical, cellular, genetic etc
- Increasing knowledge and capacity for clinical research is leading to more, better trained personnel
- Recognition is increasing that clinical research is essential for new knowledge, improved patient care, and both academic and commercial success
- He applauded better regulation, higher standards, and increased non-commercial and commercial funding
- He also stated that the UK ethics review process is efficient and of a high standard
- However, the hurdles for licensing and marketing are more stringent
- The administrative burden for regulators, industry and non-commercial sectors are increasing costs and delays
- There is an increasing disconnection between healthcare delivery and the design and conduct of clinical trials
- Trials are increasing in complexity (ie too much data, on-site monitoring and paperwork) with little evidence of additional benefit
- Dr Flather hit out against CROs that have little interest in efficiency and cost containment, and no clinical knowledge
- He also challenged the view that recruitment of patients into commercially-sponsored trials is a “service”
He also looked at areas that can be perceived as wasting time and resource:
- Too frequent and non-targeted on-site monitoring visits, which are costly with little hard evidence of benefit
- Mindless paperwork, eg, circulation of SAE reports to every investigator in a large phase 3 clinical trial leading to thousands of unread pages
- Over-reporting of SAEs for products on which prescribers have substantial clinical experience
- Clinical trial contracts with NHS Trusts, which in his experience take around 6 months from first contact to signing
- Over-collection of data that will never be used
- Wild over-estimates of enrolment targets
Finally, Dr Flather looked at how we could start to put things right. No individual group is to blame: collective responsibility of government, industry, regulators, non-commercial funders and academia should be acknowledged. There are currently very few cross-stakeholder conferences, groups or statutory bodies to discuss these issues, with each group working on its own (eg, development of GCP). He suggested that the future of clinical research depends on these groups working together to make clinical research simpler and cheaper while maintaining standards for safety, ethics, quality and speed. Dr Flather closed his presentation by proposing that ICR host a joint national conference to discuss “
efficiency and simplicity in clinical trials”.
References
- Schwedt TJ, Dodick DW, Trentman TL, Zimmerman RS. Occipital nerve stimulation for chronic cluster headache and hemicrania continua: pain relief and persistence of autonomic features. Cephalalgia 2006;26:1025-1027.
- Stang PE, Carson AP, Rose KM, et al. Headache, cerebrovascular symptoms, and stroke: the Atherosclerosis Risk in Communities Study. Neurology. 2005;64:1573-1577.
- CE Clarke, L MacMillan, S Sondhi, and NE Wells. Economic and social impact of migraine. QJM 89: 77-84.
Dr Mike Mullen is Consultant Cardiologist at the Royal Brompton Hospital, London.
Dr Marcus Flather is Deputy Director of Research and Head of Clinical Trials and Evaluation Unit at the Royal Brompton and Harefield Hospitals and Imperial College, London.
Sue Jackson RICR is Senior Manager Clinical Operations at PRA International.
