An unsound AAN Practice Advisory on poststroke etanercept

KEYWORDS:

• TNF
• etanercept
• AAN clinical advisory
• poststroke
• treatment

The American Academy of Neurology (AAN) has promulgated guidelines [1] for its Practice Advisory recommendation system (see http://www.neurology.org/content/86/23/2208.full and its condensed forms, the downloadable summaries for clinicians and patients on the same page of the AAN webpage). Here, the case is presented that the lowest Level U Advisory on employing etanercept, a widely used specific anti-tumour necrosis factor (TNF) biological agent, for poststroke disability has been generated either disingenuously or incompetently.

This editorial makes three broad observations:

1. Clinical observations, by their nature, need to be seen before being judged.

2. Authors unaware of literature on rapid, long-lasting response to treatment.

3. The possibility of poststroke etanercept toxicity is profoundly overstated.

1. Clinical observations, by their nature, need to be seen before being judged

When clinical observations are argued to have created insights that may have uncovered an important new use for an established drug, in this case perispinal etanercept for poststroke disability [2–4], prudence and objectivity require such claims to be first observed in the clinic by any organization planning to make pronouncements. Although the clinical observations doubted in the Advisory – time of onset of improvement and its duration after transiently acting medication – are very simple to assess, it does not appear as if they have yet been observed by AAN members, despite many years of invitation to do so. So far as the author can ascertain, the authors of this Advisory [5] do not appear to have even talked to the increasing numbers of people who continue to witness this treatment and its consequences. This may have been a useful starting point for formulating advice, as expert opinion along defined lines [1], to clinicians and to the public.

The AAN would be aware of an attempt in 2014 by a Fellow of their Academy to stop this treatment through the Court system. In essence, the trial judge, with access to a wide range of information, gave more credence to witnesses who had observed patient outcomes than those who had not, including AAN members (California Case No. 04–2012-222007, 22 May 2015). The attempt to stop this off-label treatment failed, with the court declaring poststroke etanercept to be within the standard of care. In addition, the court stated that there were sufficient data and research about the drug’s safety and potential effectiveness to support this treatment. The decision was adopted in its entirety by the Medical Board of California, through whom the case was brought. Nevertheless, and still without having observed patient outcomes after years of opportunity, the AAN, in this author’s opinion, now disregards this court decision, and attempts to over-ride it by issuing this Advisory. As was the court case, it is clearly directed toward Tobinick, the only person publishing outcomes of giving this treatment, which he pioneered.

2. Authors unaware of literature on rapid, long-lasting response to treatment

As noted earlier, the AAN Advisory expresses very low confidence in the accuracy of two reported observations, claims of rapid onset of benefit after treatment, and long-term effects of administration of a transiently active medication. Regarding the former, introducing 25 mg of etanercept into the cerebrospinal venous system, followed by Trendelenburg positioning for 5 min, has been reported in peer-reviewed literature to produce rapid clinical responses, initially in 2003 [6] and 2008 [7], and regularly confirmed since [2–4,8,9]. Observers have included non-neurological medical practitioners, nurses, speech pathologists and neuroscientists, and indeed the author of this editorial. A parallel outcome has apparently been reported independently (https://www.youtube.com/watch?v=VGuBMroX7c8) by a Stanford group [10] by injecting stem cells known to generate a cytokine with anti-TNF activity into the brain [11,12]. These as yet uncontrolled observations reinforce each other, and are both highly plausible from the TNF literature. For example, TNF has many rapid experimental responses (within minutes), including regulation of G-proteins that would be expected to influence multiple neurotransmitter systems [13]. Likewise, as reviewed [14], turnover of cerebral extracellular glutamate – central to synaptic function but in excess central to the harmful effects of TNF in the brain, and which etanercept inhibits [15] – is very rapid [16]. These data predict equally rapid clinical responses to an anti-TNF agent, such as etanercept, provided it can quickly get to where it is needed.

The second of these doubted observations is whether a transient treatment with etanercept could plausibly have a long-term effect. This displays, in brief, an unawareness by the AAN of the literature that cerebral TNF generated by a single event has a propensity to maintain its own production by keeping the main cells that produce it, microglia, in the activated state that continues its production [17]. Hence removing this positive feedback with etanercept can be predicted to prevent microglia generating and releasing excess TNF.

3. The possibility of poststroke etancercept toxicity is profoundly overstated

The complete AAN Advisory, as well as the downloadable summary directed at clinicians and the public, lists essentially the full range of adverse events that have been reported among the millions regularly treated with etanercept. Although the list is formidable, we must recall that these risks are so low in practice that they have not prevented etanercept from becoming one of the world’s biggest selling biological drugs. Since the risk to any individual patient depends on the dose administered, it is startling to see toxicity given by the AAN as a major reason for warning clinicians against a single 25 mg dose when the norm is very much higher than this, and it is administered over an indefinite period. This is discussed in the next paragraph.

These potential adverse effects are referenced to a particular publication [18] that, while noting this list in the abstract, actually deals with safety and efficacy of etanercept in the treatment of ankylosing spondylitis (AS). These authors summarize a series of AS trials in which some 40 times (typically 50 mg weekly for 24 weeks) the total dose of etanercept administered to the poststroke patients (25 mg, once) for which AAN professes toxicity concerns. This 40-fold higher dose recorded <5% of patients affected, and this ‘often unrelated to treatment.’ Likewise, these authors refer to a long-term trial [19] in which 500 times the total dose of etanercept given to poststroke patients, 50 mg administered weekly for 264 consecutive weeks, generated 0.03 serious events per subject year. Likewise, the recent subcutaneous etanercept trial for Alzheimer’s disease reported no adverse events [20] during 50 mg weekly for 24 weeks, i.e. again a 40-fold higher total dose. Moreover, the authors of the AAN Advisory seem unaware of the 2011 JAMA editorial that argued, from the experience with rheumatoid arthritis (RA), that the toxicity of specific anti-TNF biological agents is proving to be much less than predicted [21]. Since the mid-1990s RA has been by far the largest approved indication for these agents, and employing etanercept entails it being administered weekly for an indefinite period at double the single 25 mg, one-time, dose given to the poststroke patients the AAN Advisory claims to be concerned about. It is important to note that these doses of etanercept, very much higher than those given by Tobinick, acquired US FDA approval many years ago. For these many reasons, it is most difficult to see objectivity in the AAN Advisory specifically warning a single 25 mg injection is so dangerous as to warrant including the two-line summary justifying a Level U rating, their lowest level.

We also note that specifically drawing attention to reactivation of tuberculosis in any list of adverse events after treatment with anti-TNF agents, as does this Advisory, has largely become redundant since wide publicity of the need to test for the presence of the causal organism before treatment with any anti-TNF biologicals. In any event it would apply much more to the considerably larger, repeated doses given with FDA approval in AS and RA.

In summary, the AAN may wish to create a version of this Advisory in which the plausibility of the described outcomes is considered in the light of experiencing, for itself, the clinical observations it currently disputes. In the author’s view, not doing so is particularly unconscionable when, as here, it involves any long-term life-changing condition for which the relevant specialist group can at present offer no treatment. The AAN could also link its arguments, including those on adverse events, to scientific literature such as that provided earlier. For the sake of responsibility to patients, withdrawal of this Advisory in its present state is indicated.

Declaration of interest

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Additional information

Funding

This paper was not funded.