There is no real evidence on treating calcium channel blocker overdose

November 6, 2014, 8:06 pm

verapamil★★½☆☆

Treatment of calcium channel blocker poisoning: A systematic review. St-Onge M et al. Clin Toxicol 2014 Nov;52:926-944.

Abstract

This systematic review is massive, thorough, intimidating, and all but unreadable.

As a prelude to establishing a clinical guideline for treatment of calcium channel blocker (CCB) overdose, the 14 Canadian authors systematically searched and evaluated relevant literature. They initially identified 15,577(!) citations, from which two reviewers selected 216 for analysis. These included case reports, case series, animal trials, and observational studies. No eligible controlled trials were found.

Not surprisingly, the results of these disparate studies were all over the map, with the quality of evidence judged to be low to very low. Slogging through a discussion of all this that ultimately does not reach any clinically useful conclusion is very tedious. The authors conclude:

This systematic review found a low level of evidence supporting the use of high-dose insulin and extracorporeal life support, and a very low level of evidence supporting the use of calcium, dopamine, norepinephrine, and epinephrine for the treatment of CCB poisoning. This systematic review focused on important outcomes for decision-making in managing patients poisoned with CCB. Controlled clinical trials involving vasopressors, calcium, high-dose insulin, and extracorporeal life support should be performed.

Really? I think that such studies would be labor-intensive, expensive, and impossible to accomplish in any way that would yield convincing data. There are just too many variables and confounders. Therefore, we will never have real evidence-based guidelines on treating CCB overdose, and will have to muddle-through as best we can using available data. As Jerry Hoffman noted in a session at last week’s ACEP Scientific Assembly in Chicago, “Guidelines should regularly conclude that there is no evidence. Full stop.”

However, the paper is a valuable source of references (234 citations!) for anyone researching the topic.

 

12 Comments:

  1. Maude St-Onge Says:

    Thank you for commenting on our paper.

    Please note that an international working group is currently developing recommendations for the treatment of CCB poisoning, notably based on the evidence detailed in that systematic review. That will allow decision makers and guidelines users to clearly understand the nuances between the evidence in itself and the clinical guidance provided by a group of experts that integrates the evidence, the risks and benefits using a rigorous and transparent methodology.

  2. Leon Says:

    Dr. St-Onge:

    Thank you for the comment.

    I agree that in many cases where there is no clear and proven treatment protocol, the opinion of experts can provide guidance and aid decision making. But I’d emphasize two points:

    1) Such opinions should not be described as evidence-based is all the “evidence” is poor-quality or worse; and

    2) They should not be called “guidelines”, which implies somewhat more authority than justified by the data.

    I also agree that when the evidence is poor — as it is with many aspects of treating CCB poisoning — it is important to make clear distinctions between opinions and evidence supported fact.

  3. Guy Weinberg Says:

    I would like to add a few points looking particularly at the review’s perspective on lipid resuscitation (ILE). The paper points to animal studies that indicate efficacy in treating parenteral CCB, but cite two negative studies as providing evidence that ILE treatment of ORAL calcium channel blocker overdose is ineffective. However, the cited studies don’t negate the potential value of ILE in oral overdose but point instead to the need for further study.

    For instance, the citation of Chu et al is an abstract from the 2013 SAEM and hasn’t been published in a peer-reviewed journal. In this study, rats were infused with a total of ~ 18mL/kg 20% lipid in divided boluses over the first 5 minutes. However, no continuous infusion was given. This practice is based on the consistent laboratory observation that after an initial bolus, a continuous infusion is required to sustain hemodynamic benefit. This is particularly relevant to oral overdoses where toxicity is likely to persist much longer than in parenteral toxicities (e.g. bupivacaine). It is hardly surprising then that 115 minutes after the initial treatment there was no survival benefit in ILE v Saline controls.

    The cited study of Perichon et al also used a rat model of verapamil and treated with 20% lipid as a 4mL/kg load plus a following infusion of 4mL/kg/hour based, they say, on our original publication showing benefit of lipid infusion during resuscitation from i.v. bupivacaine overdose. However,we used 13.5mL/kg of 30% lipid. Moreover, they cite another study of ILE in verapamil (iv) toxicity but Perez et al found that the optimal dose for treating verapamil toxicity was 18mL/kg of 20% ILE. Clearly, the initial bolus dose in the study of Perichon et al was insufficient.

    It is unfortunate that based on such weak negative evidence the authors of the review decided to discount ILE. Clearly further studies are needed to identify a specific protocol for ILE in oral overdose. However, the negative studies don’t suggest this doesn’t exist; only that research is needed to optimize this approach. Meantime, the Conclusion references treatments for which only ‘low’ or ‘very low’ levels of evidence exist while ILE is overlooked. I don’t think the clinicians who’ve seen a strong, favorable response to ILE in CCB overdose would agree with this panel’s position.

  4. Leon Says:

    Dr. Weinberg:

    Thank you for your important and detailed analysis of the intralipid section of the St-Onge et al review. I too was surprised that the authors dismissed intralipid therapy based on a few animal models of doubtful validity. It seems to me that little useful information will ever come from rat studies, and that good convincing RCTs will never be done.

    Prompted by your remarks, I looked a little more closely at St-Onge et al’s discussion of intralipid. They reference a paper by Cohen et al to support the following statement: “One available human case series (five patients) demonstrated 60% mortality when using this antidote compared to a lower mortality reported in retrospective studies of CCB poisoning . . .” However, the Cohen paper they reference is a case report, not a case series, and the patient was not treated with intralipid. I don’t know if this is just a mistake and they intended to reference another paper, but the mishandling and mislabeling of evidence puts paid to any claims that the review is evidence-based.

    This reinforces my original comments on the St-Onge paper. It is unreadable. Careful analysis would require going back to many of the 234 references they site to evaluate how relevant they are and if they support the claims made by the authors. Also, I repeat that we likely will never have definitive “evidence” for the role of intralipid in treating overdose from CCBs and other cardiotoxins. This does not, or course, mean that at some point we won’t have a mass of convincing anecdotes.

  5. Maude St-Onge Says:

    Dear all,

    Concerning the word “guidelines”, the manuscript that will be submitted will mention “recommendations”, but will avoid the word “guidelines”.

    Concerning the use of lipid emulsion. It has been classifed as “low level of evidence”, but no “no evidence”. Therefore, it will be recommended as a rescue treatment or for patients in cardiac arrest in the manuscript that will be submitted (recommendations from the international working group).

    For the case series, we should have read ref 98 (as in the online supplemental material: Geib et al, 2009). That case series is only an abstract. The fact that the negative cases were only abstracts was considered when buidling the recommendations.

    Please avoid to take the systematic review results as being recommendations.

    Thank you again for sharing your comments,

    Maude St-Onge

  6. Maude St-Onge Says:

    Apologizes… I should mention “very low level of evidence” not “low level of evidence” as per the GRADE system.

    Maude St-Onge

  7. Steve Sheldon Says:

    Leon
    I have read the poison review for a long time and think your comments are a little off base regarding this paper.

    Everyone makes guidelines, they just don’t call them that! A book is a guideline, a blog can be a guideline, on-line sources like UpToDate are guidelines in that they all tell clinicians how to improve the care of the patients we are committed to. The difference is that most of these sources are really opinions of one or two authors (with or without a few editors) who cherry pick at the literature to support what they believe is a correct position.

    Sometimes we have great evidence (like aspirin for acute MI) and sometimes we do not have great evidence, but we always have to synthesize that evidence into a treatment recommendation. Many prestigious societies make guidelines based on very limited evidence, because without them, people not as smart and well read as yourself don’t have a clue as to how to practice medicine (or toxicology in this case).

    So, when a group of people get together and try honestly to collect all of the evidence and present it, where is the harm? The St. Onge paper was a challenging read for me, but contains everything I might want to read for myself in one place. It passed peer review of a journal that is fairly respected in Toxicology and your comments have additional implications on the peer review and editorial process. You’ve had a long and wonderful career commenting on other people’s works, but you do so without the benefit of real peer review. Might I respectfully suggest that your write a well crafted letter to the editor of the journal outlining your concerns.

    If a CCB guideline comes out of this group (and I suspect it will), they will make recommendations based on limited evidence, which is what I do every time I provide care for a patient with an unclear illness in the emergency department. If they are honest about the limitations and are willing to change their position as new evidence arises, then the guidelines will provide a good starting point for people who don’t know where to start.

    People always complain about guidelines inhibiting personalized care and free thinking, but the data suggest that guidelines improve care. They are not protocols (policies that cant be violated) they are “recommendations” from learned individuals, and like any recommendation from any consultant you accept it or you don’t.

  8. Sophie Gosselin Says:

    Could Dr Weinberg give any reference for his assertion or is it based on a convincing mass of unpublished anecdotes?

    “This practice is based on the consistent laboratory observation that after an initial bolus, a continuous infusion is required to sustain hemodynamic benefit”

  9. Leon Says:

    Dr. St-Onge:

    Thank you for clarifying the mis-identified reference in the paper.

    I am glad to hear that the forthcoming manuscript will not use the word “guidelines.” Perhaps I was misled by this passage from the paper:

    “in order to help clinicians to best treat CCB poisoning, the development of practice guidelines on the treatment of CCB poisoning is warranted.

    “Therefore, the goal of this systematic review was to document and characterize the available evidence to facilitate development of guidelines . . .”

    I may have also been misled by the fact that your “Conclusion” section omitted all mention of lipid infusion therapy, even in the list of those modalities supported by only a very low level of evidence. To my mind, this read as an outright rejection of the therapy.

    As for Geib et al, even though they called their abstract a “Case Series,” it did not describe well the study population or the exact treatment they received. The only thing their patients had in common was that they received lipid rescue therapy for cardiotoxicity. Five of the patients had taken a CCB; two had also ingested another cardiotoxic drug. To call this a case series is, to say the least, generous. This abstract was published in 2009, at time when — as now — lipid therapy was generally reserved for a last ditch effort in a crashing patient in whom all other modalities have failed. Therefore, it is not surprising that 3 of the 5 patients died. I would submit that this is not evidence of anything about lipid therapy, positive or negative, and this should have been clearly stated in your review. To say, as you do in your paper: “One available human case series (5 patients) demonstrated 60% mortality when using this antidote compared to a lower mortality reported in retrospective studies of CCB poisoning” — without discussing the severe limitations of the Geib abstract — certainly gives the impression that the abstract constituted negative evidence.

  10. Leon Says:

    Steve:

    Thank you for your comments. In general I agree with you, and where I disagree it may just be matter of terminology.

    The Institute of Medicine defines clinical practice guidelines thusly:

    “Clinical practice guidelines are statements that include recommendations intended to optimize patient care. They are informed by a systematic review of evidence and an assessment of the benefits and harms of alternative care options.”

    By this definition it is clear that in toxicology very few guidelines could really exist. Certainly, and review of evidence that is often of “very low quality” and at best “low quality” Since there have been, I submit, absolutely zero convincing and definitive RCTs in medical toxicology, the effort to develop evidence-based guidelines is doomed to failure.

    Given that, I think the opinions of respected experts can be extremely helpful in guiding decision making. (Or course, each person’s list of respected experts may differ.) That is why, unlike some of my colleages, I am very interested in the conclusions of the EXtracorporeal Treatments in Poisoning (EXTRIP) workgroup publications, which in some publications present their findings as “consensus-based recommendations.”

    As to your questions, “So, when a group of people get together and try honestly to collect all of the evidence and present it, where is the harm?” I think the harm lies in analyzing very poor evidence and then making guidelines based on that evidence. Much more honest, and helpful, to be upfront and admit that there is no real evidence, explain why the evidence is poor, and then give recommendations.

    As for writing a letter to the editor, I strongly believe in rapid post-publication review, a process at which FOAM excels, Traditional letters to the editor generally take a long time to appear in print and, it my opinion, have limited impact.

    I completely agree that expert recommendations are essential and often improve clinical care. I have certainly used them many times.

  11. Guy Weinberg Says:

    Dr. Gosselin makes a good point. It is more accurate to say the practice is based on my personal (and consistent) laboratory observation of a need for additional lipid dosing after the initial bolus. We saw this very early on, more than a decade ago, in a series of experiments using a dog model of bupivacaine toxicity. Initial hemodynamic recovery would fade after ~10 minutes unless a following infusion was used, in which case ROSC was robust and durable. This was reflected in early recommendations for clinical use of lipid in the setting of (parenteral) local anesthetic toxicity and in subsequent guidelines (lipidrescue.org, AAGBI, ASRA, etc).

    It is possible that repeated boluses would work just as well as an infusion but grant proposals to study this were never funded, so for the time being…it’s continuous infusion. This approach was based at least in part on the knowledge that a continuous infusion of any medicine generally leads to lower total doses than using repeated boluses which tends to repetitively ‘over’ and ‘under’ dose and thereby yields higher total doses over time.

    The need for the follow up delivery of lipid indicates that a simple lipid sink cannot account for the benefit of ILE. Our current view of the mechanism(s) (now under peer review) combines the inotropic effect of lipid with a dynamic partitioning effect where lipid droplets act as shuttle carrying drug away from target tissues to other organs. Both mechanisms require a consistent plasma triglyceride level to sustain the clinical benefit and given the half-life of triglyceride in plasma one can identify the need for a continuous infusion to sustain ROSC.

    Dr. Belinda Akpa, UIC faculty in the department of Chemical Engineering has established a computational model of lipid resuscitation that accurately predicts results in our animal models of bupivacaine overdose. She generated a simple graph of triglyceride plasma concentrations following a single lipid bolus with no following infusion showing why ROSC might not be durable in this context, since TG levels fall off rapidly without the follow up infusion. Nevertheless, as mentioned in the earlier note, the most effective regime for delivering lipid to treat enteral intoxication has yet to be identified.

    Plasma concentrations after single lipid bolus

  12. Guy Weinberg Says:

    Please see the following two papers for descriptions of the referenced in silico model:
    Kuo and Akpa, Anesthesiology 2013 Jun;118(6):1350-61.
    Fettiplace et al, Anesthesiology 2014 Apr;120(4):915-25

    See also the appended image, generated by Dr. Akpa, to predict plasma TG concentration as described in the previous post, a reply to Dr. Gosselin’s inquiry about evidence to support infusing lipid after a bolus.