A comment regarding the paper “Adverse Event Reporting in Randomized Clinical Trials for Multiple Myeloma”
This is a very interesting article that should kindle in-depth conversations about the descriptions we use to summarize clinical trials in myeloma. The overarching message of this paper is that it is not uncommon to minimize toxicities, even when some clearly have a significant impact on patients. Some things have always bothered me, such as how language can minimize toxicity or seem outright erroneous, and this paper is important documentation. I have some points of agreement with the interpretation but also some of disagreement. I realize the authors have much more to say as the paper is a medical publication focused on the data presented, but this paper stirred many thoughts. I will consider it a conversation starter. Let me explain.
The term “manageable”
Overall, I react negatively to hearing the term “manageable” as I consider it vacuous. Arguably, everything is “manageable,” but is this management effective?
The Oxford Dictionary states that manageable is an adjective as defined by something that is:
“able to be managed, controlled, or accomplished without great difficulty.”
Prescribing opioids or medications like gabapentin for painful neuropathy is simple – and can be done without much difficulty. But the results are often suboptimal. The definition lacks clarity in the outcome of the management. The portrayal of something being manageable is further problematic for enduring toxicities like peripheral neuropathy (more on this later). The infusional toxicities of daratumumab have been greatly ameliorated by the use of premedication and the subcutaneous route. Furthermore, they rarely occur after the first episode of administration. To some degree, they can be “minimized,” and they could clearly be said to be “manageable.” So manageable is incomplete and needs to be clarified further.
Toxicity attribution
I would also highlight (not the subject of this paper) that the attribution of toxicities for experimental drugs needs to improve. As Dr. Tom Martin (UCSF) rightfully pointed out at the IMS meeting in 2022, infections occurring in the setting of bispecifics antibodies should never be unliked to the study drug. At a minimum, they should be considered possible (e.g., bacterial pneumonia) or probable (e.g., PJP, adenoviral infections). Anything less is minimizing.
Context and nuance
The key concern by the authors is that using these terms inappropriately allays concerns and diminishes precautions by doctors and patients. While patients regularly read the medical literature, and I’m glad for that, the intended audience for medical publications is professionals. Accordingly, how messages are portrayed has to be evaluated in the context of the intended recipient.
I would argue that the word “acceptable” can be appropriate when describing medical treatments. Let me use an extreme example: we can describe the mortality rate of a medical intervention as “acceptable.” This never means that physicians or healthcare tolerates death, but merely that the risk-benefit for said intervention may be better than perhaps prior ones. We obviously want to have treatments that have no mortality. Still, lifesaving treatments, like allogeneic stem cell transplants or CARTs have a mortality rate that is deemed acceptable by the medical community. A new acne treatment should never be associated with an acceptable mortality. But is it inappropriate or misleading to say that a high-complexity procedure has an acceptable mortality compared to predecessors?
Additionally, I don’t think terms such as “well-tolerated” are necessarily inappropriate or minimizing, not in the medical literature or the context of patient care. For example, it will be hard to argue against the term describing the long-term administration of single-agent daratumumab. “Well tolerated” is a rhetorical description that can help patients when considering therapy. The same is true for tolerable.
Because of the fiduciary responsibility of physicians, patients often place utmost trust in the recommendations being made by physicians. I believe patients ultimately make decisions more based on the trust they have in their physician’s wisdom (e.g., What would you do in my case?”) as opposed to our technical knowledge (e.g., rates of the various toxicities). Of course, this trust must be built by our objective interpretation of the best available data. As much as physicians need to know the numerical descriptors of efficacy and toxicity, they must provide general descriptors of proposed treatment regimens. I feel that it is ethical, appropriate, and accurate to describe so much of our treatments in that way. One strategy I use is to preface these conversations with the words “in general.” I often start by saying: “In general, this treatment is well tolerated, but we must be aware of X, Y, and Z.”
It goes both ways!
Now, medical publications could exclude any editorial comments in the discussion and just present data, but this would greatly reduce messaging for learners at all levels.
As physicians, we need to be as accurate as possible as we describe these treatments' potential benefits and risks to patients. As this article shows, industry will naturally be more inclined to portray a treatment more favorably. In such cases, cross-referencing adjectives and adverbs to the data presented will be paramount, and the medical community must remain vigilant.
But I would suggest adjectives and minimization can also be misleading when used in the opposite direction. Let me provide some personal subjective observations. I have observed the use of exaggerated negative adjectives by those more reserved about using new medications because of their associated cost.
Let’s compare a decision between using VRD versus daratumumab-Rd as induction therapy for a transplant-ineligible patient. I would argue that peripheral neuropathy associated with bortezomib, a form of mostly irreversible damage, is minimized by many in the myeloma community. I suspect this is born out of the aspiration that the now generic medicines can be more widely used in LIMC. If I were ineligible and newly diagnosed with myeloma, I would rather be treated with daratumumab over bortezomib. I believe the correct interpretation of current data would lead to most (if not all) patients with a new diagnosis of MM should be treated with the MAIA regimen as opposed to VRD. Even if this is an unpleasant message for those with more limited access. Likewise, the “cardiorenal” toxicity of carfilzomib tends to be exaggerated.
Other serious toxicities have also been exaggerated for medications that have clear anti-myeloma effects, even if not at par with the best available drugs. In the past, patients of mine refused to be treated with belantamab because of the risk of them “going blind.” And the ultimate decision is always that of patients, but hopefully not because they felt blindness was an inevitable consequence. Adherence to selinexor was quite problematic when used in the doses and schedule as originally approved, but much improved recently. I found it unusual to see myeloma doctors say – “let’s balance this conversation.” Even if selinexor ultimately has limited usage in myeloma, several other indications are being pursued, some with potential promise. If future patients with other such disease states that may ultimately prove to derive benefit from its use read the hyperbolic rhetoric, they might opt to forgo treatment.
What patients want
Indeed, having clinical trials routinely consider PROs is a laudable idea, and we should all support this idea. It will probably become more affordable and common with the advent of digital tools. However, data summaries and general statements about patients’ preferences have to be confronted with the nuances of individual choice at the bedside. Everyone is an N of 1!
Solutions
The authors conclude, “Instead of using these terms, trial investigators should highlight event rates and PROs to allow clinicians and patients to better evaluate the true tolerability of AEs.” Perhaps these would be papers with only methods and results – some would prefer that. The current world is rife with examples of individuals or groups wanting to regulate language for multiple reasons. An inadvertent consequence of this regulation is a lack of candor and transparency. I believe that eliminating certain terminology will not truly advance medical research or provide better care. The data presented in the paper is invaluable and provides clear examples of potential misuse of adjectives. But I don’t think we should eschew the editorial perspectives that medical articles contain in the discussion - it would be paternalistic. And many important possibilities and additional hypotheses would be lost.
The authors also state that “Cancer treatment can result in burdensome toxic effects that profoundly affect patient quality of life.” Correct. But there is nothing more toxic or detrimental to the QOL than progression or death.
My disclosures are below.
Consulting: AbbVie, Adaptive Biotechnologies, AMGEN, AZeneca, Bayer, Binding Site, BMS (Celgene), Millenium Takeda, Jansen, Juno, Kite, Merck, Pfizer, Pharmacyclics, Regeneron, Sanofi.
Scientific Advisory Boards: Adaptive Biotechnologies, Caris Life Sciences, Oncotracker.
Board of Directors: Antegene (for profit), AZBio (not for Profit)
Patent for FISH in MM - ~$2000/year
[1] https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2811645