Vitamin C and the ethics of Borrowing data.

I was recently amazed to be engaged in a Twitter kerfuffle which generated more than 10,000 Impressions within 24 hours. Passions were running high, libellous comments were being broadcast, and old friendships seemed to be at breaking point. The issue? The ethics of preserving endothelial (im)permeability. This Post reflects my current perspective on things; I think I am open-minded enough to reconsider my position, so please feel free to comment. I promise, no flaming if I disagree. Maybe just a little irony…

Paul Marik, of Norfolk VA, is the author of “the premier Evidence-Based Textbook in critical care medicine”, now in its 3rd Edition. I start with his CCM Editorial of June 2016; as first author of “Recommendations for the diagnosis and management of corticosteroid insufficiency in critically ill adult patients: Consensus statements from an international task force by the American College of Critical Care Medicine.” back in 2008, he is eminently qualified to comment on an interesting study of glucocorticoid receptor sensitivity in sepsis. But he goes much further, using the Editorial as a platform for his own hypothesis that the co-administration of vitamin C with hydrocortisone may have synergistic benefits through increasing adrenergic responsiveness, restoring the tight junctions between endothelial cells, and preserving the endothelial glycocalyx. He proposes that “patients with severe sepsis and septic shock may benefit from the early administration of IV glucocorticoids (vitamin S) and vitamin C. Clinical studies are required to test this highly cost effective and “radical” hypothesis.” He even Discloses “off-label product use (vitamin C: this compound is regarded as generally safe and is exempted from Food and Drug Administration approval).”

The Intercellular Junction.

Then came a peer-reviewed before/ after study published in the respected Journal Chest. 47 cases over a seven month period both Before and After introduction of vitamin C administration; The hospital mortality was 8.5% (4 of 47) in the treatment group compared to 40.4% (19 of 47) in the control group (p < 0.001).” “Additional studies are required to confirm these preliminary findings”, but noticably the Authors (five eminent professionals based in Norfolk, VA) do not specifically call for randomised controlled clinical trials (RCCT).

Wow! A Cure for Sepsis! Let’s call it the Norfolk Protocol for short. I expected Disbelief after so many years when sepsis has been the necropolis of so many hopeful therapeutic advances, a veritable Bermuda Triangle of Drug Research money. But I was taken aback by the bile and barely concealed hatred that anyone should dare claim that anything less than a standard ethically-approved RCCT be put forward as Evidence on which to Base life-saving Medical treatment. I suggested that others could attempt to replicate the Norfolk results, but was flamed for my impudence. Rather than risk allowing even more patients to die from sepsis for the sake of a pure and unadulterated RCCT, I suggested borrowing control data. There is a lot of it out there, we really don’t need another graveyard load more! If you won’t listen to me (and why should you?) consider the following paragraph extracted from an Open Access Paper ‘Use of historical control data for assessing treatment effects in clinical trials’ in the Journal Pharmaceutical Statistics;

Clinical trials rarely, if ever, occur in a vacuum. Generally, large amounts of clinical data are available prior to the start of a study, particularly on the current study’s control arm. There is obvious appeal in using (i.e., ‘borrowing’) this information. With historical data providing information on the control arm, more trial resources can be devoted to the novel treatment while retaining accurate estimates of the current control arm parameters. This can result in more accurate point estimates, increased power, and reduced type I error in clinical trials, provided the historical information is sufficiently similar to the current control data. If this assumption of similarity is not satisfied, however, one can acquire increased mean square error of point estimates due to bias and either reduced power or increased type I error depending on the direction of the bias. (Pharm Stat. 2014 Jan-Feb; 13(1): 41–54.)

And of course there is sequential analysis, in which a Treatment is used in every case until a subject dies, at which point the next sequence of subjects receive Control until one of them dies. Designed to minimise Control exposure, but not much used for some reason.

There was unconcealed anger that the Press should become involved.

Kurt Hofelich, Norfolk General’s president, said the protocol is being rolled out to other ICUs in the health system to validate the findings. We hypothesize that this new treatment will evolve into a national best practice and a new standard of care for patients with sepsis in an ICU level of care environment,” Hofelich said in a prepared statement.

Some of us clearly don’t want an informed Public to start demanding a Treatment we’d rather withhold. I am not one of those. I recall collecting a very sick toddler with meningococcal septicaemia from Basingstoke General to escort him to Southampton, and being asked by the parents if I was going to provide the new blood cleansing technique. I later found that that day’s Telegraph carried a report of a small case series of haemofiltration (from Yorkhill, I think?). I did line the child up as soon as I got to Southampton, filtered him aggressively, he recovered with all four limbs intact, and a Letter to the Lancet was born. But how firm can we be in denying a reasoned request for a nutritional supplement for a high RoM disease? Vitamin C is not even an investigative medicinal product.

So this is where I stand. When the signal for life-saving is loud and the harms are unknown but likely to be minor, the first concern of the statistician or research clinician must be to minimise or ideally prevent further untreated deaths in pursuit of probability estimates. The RCCT is the purest and most persuasive tool we have, but not at any cost. And even the most expensive CCM RCCTs (in terms of money spent and lives lost) have more often than not been found to be unreproducible. Let’s accept that it is OK for responsible ICUs to give vitamin C with patient or surrogate consent and to carefully assess the patients response and commit to sharing or publishing their data. If we ever get to the point of starting an ethically-approved RCCT (it will be several years) then the preferred pathway will be consented enrollment. But until then, don’t let your Puritanism stand between a patient and a possible Cure.

Post Scriptum. Marik Live in Belfast

Leave a Reply

Your email address will not be published. Required fields are marked *