Learning from Northwick Park
pharmafile | February 19, 2007 | Feature | Research and Development |Â Â UK, biotech, clinical trials, drug safetyÂ
In December 2006, the official report on the Northwick Park incident was published. Although not the end of the affair, it is unlikely that much further hard information will now emerge, so we can safely sit back and look at the impact.
If anyone needs reminding, on 13 March 2006, eight healthy volunteers were dosed in a phase I first-into-man study of the agonistic anti-CD28 monoclonal antibody TGN 1412. As this was a single ascending dose study with a total planned population of 32 subjects, these eight men represented the lowest dose group, with two receiving placebo. As we now know, all six men who received TGN1412 rapidly developed severe reactions, now known to be a cytokine storm.
A number of aspects of this episode have interested me, particularly the way public statements by various bodies changed over time, and I have detected that the force of such statements is inversely proportional to the actual information underlying them.
The truth revealed?
Within days of the story breaking, the BBC aired a special early evening TV programme claiming to tell the truth about the disaster. The programme was filled with footage of the understandably tearful girlfriend of the most seriously affected subject, but hard facts were in short supply. It was opportunistic and exploitative broadcasting, which did no credit to the BBC.
About the same time, the lawyers were making all sorts of accusations before anyone had been able to analyse the problem. The MHRA acted very quickly by doing a for-cause inspection of the Parexel unit at Northwick Park, and also audited the lab work and the manufacture of the drug. Not surprisingly, the MHRA was roundly condemned by Ann Alexander, a well-known medical litigation solicitor acting for two of the victims, for producing a whitewash. The lawyers and media seized on the findings, which identified some compliance problems on the part of Parexel, but the MHRA had made it clear that these were unrelated to what had happened.
Pointing the finger of blame
Now, I have worked with a number of phase I units like Parexel and not one of them was completely free of compliance issues. There is not much that humans do that can be perfect – that's why we do inspections – but how we love to find someone to blame, as the lawyers and media tried very hard to do in this case. Significantly, nothing wrong was found with the manufacturing of the drug.
Meanwhile, reports came out periodically on the progress of the six men, one of whom suffered permanent physical disability. At three months, all had residual limitations affecting memory and concentration. Everyone was, quite rightly, very sympathetic and concerned that hitherto healthy people could be so reduced so quickly. This kept the story running for several months, while a specially appointed (by the Department of Health) Expert Scientific Group (ESG) carried out a wider investigation, chaired by the recently knighted Professor Gordon Duff.
I don't think it has been widely understood in lay circles the extent to which the ESG and MHRA reports differed in terms of remit. Essentially, the MHRA found that, a few administrative anomalies aside, the procedures surrounding the trial were followed correctly, but it was not concerned with unravelling the scientific conundrum which precipitated the disaster.
Conversely, the ESG was concerned with that, and a lot more besides. Interestingly, the ESG's report still did not point the finger of blame at any individuals or organisations, even though it did identify how the problem arose and how things should change for the better.
The lawyers might have branded this report as another whitewash, but in fact, they didn't. Indeed, Ann Alexander welcomed the report, and said: "We also welcome the finding that the pre-clinical development studies that were performed with TGN1412 did not predict a safe dose for use in humans, even though current regulatory requirements were met."
That's a far cry from the original sabre-rattling. Essentially, she is now focusing on what went wrong, rather than who was responsible. This dramatically changes the role of the lawyer, as the victims would be covered by a no-fault compensation scheme and the sponsor's insurance, instead of having the need to sue someone for negligence – something which seems unlikely to be proven.
As more information became available, the debate cooled off and became more rational; however, there was another subtle, but important, conceptual shift, which took place early on. Some commentators attributed the adverse reactions to an unpredictable (ie could not be predicted) action of the drug, whereas the MHRA and ESG reports referred to it as an unpredicted effect, meaning that it was not predicted – but the question remains, could it have been? It has become clear, despite continued noise from the media, that the pre-clinical studies met the current regulatory standards.
It was known that the TGN1412 binding site is not present in the usual species studied for toxicology, but only in primates (including humans). This is stated clearly in the study protocol. Therefore cynomolgus monkeys were used. A conventional No Adverse Effect Observed Level (NAEOL) of 50 mg/kg body weight was derived from these studies, and the starting dose for humans was set at 0.1 mg/kg, 500 times lower. This actually suggests a very cautious approach, but it was not good enough, as we now know, and to date, the exact reason for the extreme reaction in humans, in stark contrast to excellent safety in non-human primates, remains unclear.
Ticking all the right boxes
I don't propose to regurgitate the recommendations of the ESG report, but rather to put the episode into the context of the current drug development environment. For several years, I have been increasingly concerned at the emergence of a highly formulaic approach to the discipline.
There can be little doubt that ours is a very highly regulated industry – and so it should be – but now everyone expects there to be a procedure to follow; i.e. a recipe for everything, in which if you tick all the right boxes, you can't go wrong. In the Northwick Park case, all the boxes were ticked, but it still went wrong so, the answer must be that some of them were the wrong boxes, and/or some of the right boxes were missing.
The ESG report put a lot of emphasis on the uncharted waters we are now navigating with biological compounds. Not only are these humanised in an effort to reduce the potential for adverse reactions, but this process is increasing their species specificity. It seems likely, therefore, that predicting safety and tolerability in humans, on the basis of animal studies, will become more difficult.
This means that the box-ticking approach, which already has inherent insecurities, will become even more prone to risks. It means that the standard set of boxes won't work for everything. Above all else, it means that a more creative and critical approach is needed. The ESG report covered risk assessment in some detail, something which should be more deeply embedded in drug development.
An ethical matter
From the moment the Northwick Park story broke, I was surprised at the lack of involvement of the ethics committee. It's interesting to see that this was excluded from the ESG's remit, but they considered it so important that they covered it anyway. The ethics committee took responsibility for the approval of the study, and particularly for the protection of the subjects, yet there has been no statement from its chairman. The timing is especially relevant, with the advent of proper regulation of ethics committees via the EU Clinical Trials Directive.
I am not suggesting the ethics committee was at fault, but the ESG has called for changes in ethics reviewing. In particular, there needs to be more communication between the regulators and the ethics committees. Overall, open access to information was a major theme, with the aim of increasing alertness to potential problems, and underpinning the risk assessment process.
Of course, the key question is, why did these not predict the risk?
The use of a NAEOL model for calculating the human dose is a totally conventional approach and the sponsor was not criticised for following it. But its not the only approach, and the report included very useful information on alternative approaches. One such method is MABEL, or Minimum Anticipated Biological Effect Level. This would have given a much lower starting dose in humans, as under MABEL, the actual dose given would have been around the maximum not the minimum.
Clear objectives
When I train people in project management, I spend a great deal of time on defining the approaches to be used in a study. I like to see study objectives listed clearly, interpreted on a value basis (ie what will we do with the results if we meet this objective?). Then I get people to list all the alternative methods they could use to meet each objective. One key rule is that each method must be linked to an objective – it's common to see bits of methodology included without good reasons.
Another rule is that all possible approaches should be considered, on a brainstorming basis. There is a danger of the formulaic approach to study design. It's far too easy to use the standard approach that worked last time, instead of spending time on creative thinking. TeGenero used the standard approach, which the regulators accepted, but which, with hindsight, was not the right one. We don't know whether other approaches were considered, but it certainly behoves sponsors and regulators to make sure they are in the future.
The same principle applies to the in vitro studies included in the regulatory submission package. The ESG commissioned a lot of further studies, carried out by the National Institute for Biological Standards and Control (NIBSC), and it's very reassuring to learn that when the NIBSC repeated the studies done by TeGenero, it got the same results. Replication is a major part of the scientific method. Some key tests involved the possibility of cytokine release from human lymphocytes.
Without going into too much detail, it was found that free lymphocytes did not release cytokines when treated with TGN1412, but did do so when bound to a substrate. TeGenero did the free lymphocytes test but not the bound lymphocytes test, which is no reason for condemnation, but another example of the need to explore more possible approaches.
So why did TGN1412 cause such profound and long-term damage? First, it bypasses the normal control mechanism for cytokine release, which requires antigen stimulation to activate an immune response. Normally, only T-lymphocytes carrying the antigen receptor will proliferate. Instead, the drug expanded the population of T-lymphocytes irrespective of their antigen specificity – imagine an immune response to all possible infections, instead of just the one you have. Secondly, being a large molecule, it is not cleared by the kidneys, so has a half-life measured in weeks. Once in the system, it is not going to go away quickly.
Both these factors were well-known from preclinical studies, and ought to have been included in the risk management strategy, particularly as no antidote was known to exist.
Spreading the word
What impact has this episode had on clinical research in the UK, and on phase I trials in particular? Well, for a start, the general public now knows what phase I trials are.
I was quite surprised at how poorly understood they were, even to the extent that some people were horrified that drug companies contracted them out. I think the overall message came over that phase I studies have a very good safety record, as this incident was unique.
I have heard that enquiries from possible volunteers to some units have actually increased, although fewer volunteers agreed to participate after more careful reading of consent forms, and I don't have any concrete information suggesting that it's much more difficult to get people to participate. Certainly, this will have focused the minds of the ethics committees, as well as the regulators.
So far, phase I studies have been to some extent carried out almost on a production line basis. Until now, there has been a well-trodden path, starting with single ascending dose (SAD), and multiple ascending dose (MAD) studies, leading to a very standardised battery of studies which everyone accepted as the definitive package for a dossier.
The world is a bit different today. We haven't sailed off the edge of it, but we need not assume that we will always know exactly where we are.
One thing is certain, though; we are going to need a wider range of navigational aids as we venture further into the new territory of developing biological agents.
Les Rose is a freelance writer and independent clinical science consultant. Contact him at lesrose@ntlworld.com
Related Content
Cellbyte raises $2.75m to fund pharma drug launch platform
Cellbyte has announced that it has raised $2.75m in seed funding for the streamlining of …
Vesper Bio reports positive topline results for dementia candidate
Vesper Bio, a clinical-stage biotech developing novel oral therapies for neurodegenerative and neuropsychiatric disorders, has …
Von Willebrand disease – increasing awareness and access to vital care
Pharmafile talks to Anthea Cherednichenko, Vice President Franchise Head Haematology and Transplant at Takeda about …
