In a much-quoted analysis, Lee found that most IDSA Guidelines were supported only by Level III quality data, the lowest category of evidence-based support[1]. The article concludes that more well-designed trials were needed, and clinical judgement should ultimately guide patient management.
This all sounds fine and good, and the new Candida Guideline by Pappas et. al. has a similar disclaimer[2]. However, the reality of medical practice is different. A guideline is not like a textbook chapter or article but has greater formative authority. That said, in a litigious society like ours, any deviation from chapter and verse of this ‘bible’ becomes a potential lawsuit, with the physician having to explain why (s)he deviated from expert recommendations by omission or commission. In other words, use your ‘judgement’ at your own risk.
It is a fallacy to blame the data situation for many poorly supported recommendations. We will never have “well-controlled” trials in some of the rare conditions described in the Guideline, hence, treatment decisions will have to be made based on extrapolation and inference from more common conditions. Over 50% of recommendations are, like in the Lee review, based on rather poor data, and the next version of the Guidelines will probably be no different again. Common sense justifies most of the recommendations, even if these were not supportable by high quality data, low quality data, or no data at all. Which is why, on the back of this document, we should see a banner stating the following: In Experts We Trust.
It should not come as a surprise that most of the “strong recommendations” supported by ”strong data” come from registration trials conducted by industry for new antifungals. These trials were conducted in mucosal candidiasis or candidemia for which patient recruitment into well-controlled trials was feasible.
It’s a pity that NIH and other similar sponsors have been much less important contributors to the clinical data situation. As industry runs trials with the same experts that later write our Guidelines, there is potential for conflict of interest, which we all need to be aware of.
FDA has a conflict of interest policy that pretty much prohibits these same experts from sitting on regulatory decision bodies, variously called AIDAC or AMDAC. That’s a good thing and should not be changed. Many of my colleagues argue that this ‘no conflict’ rule robs FDA of important expertise. Yes and no, there is a trade-off, of course. Democratic societies choose to keep the executive, legislative and judicial powers separate in government. Our panel experts already have strong influence in the clinical arena and are often expert witnesses in the judicial system; they should not be the regulatory decision makers too. The conflict of interest section at the end of the Guideline is amazingly long; it would certainly take up a full page if the type font were not reduced.
Thorough study of this new Guideline document easily takes a full day. IDSA should award at least 8 hours of CME credits for this 50 page monster of a document. Correction: the actual text is merely 38 pages long; the 560 references make up the rest. If you are willing to skip recommendations based on low-quality evidence, i.e., those more faith-based instead of evidence-based recommendations, you can pretty much eliminate 2/3 of the document making it more readable and better digestible.
By design, Guidelines are fairly dry and boring documents, so here some suggestions to spice it up a bit. Try figuring out why there are so many “strong recommendation, low quality data” recommendations. More importantly, the sections justifying recommendations are very insightful. There are some other ‘pearls’ too. The comment that isavuconazole may not have shown non-inferiority against an echinocandin is so far unpublished information. It explains why Cresemba is hardly mentioned in this Guidance document.
Spoiler alert: In the treatment category, it’s about azoles, echinocandins and AmB formulations. You already knew that? Well, you are ahead of the curve.
Another spoiler alert: AmB is still toxic and should be avoided unless you have a good lawyer.
Last spoiler alert: You can use fluconazole IF you can predict that it will work. Hmmh…
(Sorry folks: Garlic / allicin is not a treatment option for anything)
References:
[1] Lee, D. Analysis of Overall Level of Evidence Behind Infectious Diseases Society of America Practice Guidelines JAMA Internal Medicine 2011, Vol 171, No. 1
[2] Pappas, P. Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the Infectious Diseases Society of America. Clinical Infectious Diseases Advance Access. Published Dec 16, 2015
