Lipid therapy in oral poisoning: a not-so-systematic review

December 27, 2016, 4:12 pm

intralipid-161x300★★☆☆☆

No support for lipid rescue in oral poisoning: A systematic review and analysis of 160 published cases. Forsberg M et al. Hum Exp Toxicol 2016 Nov 24 [Epub ahead of print]

Abstract

The authors’ goal was “to present a systematic review and case analysis of practically all published reports on humans treated with lipid rescue for LAST [local anesthetic systemic toxicity] or oral poisoning.”

The focus of the paper is on oral poisonings. The authors report that they identified 94 reported cases of oral poisoning with “alleged” positive response to lipid rescue therapy (LRT.) Two authors reviewed each case and rated causality (that is, LRT -> positive response) using a modified WHO-UMC scale. (As the authors note, the WHO-UMC scale was actually designed to determine causality in adverse drug reactions, not antidote response.) If the two authors disagreed as to the causality score, “consensus was obtained through discussion.”

The authors report that all 94 oral poisonings received scores of 2 (probable causality,) 3 (possible causality,) and 4 (unlikely causality.) in fact, 86% received a score of 3 or 4. They note that 91% of these cases received other treatment modalities at approximately the same time they received LRT.

The authors conclude:

“Considering the findings of the present study, the weak and contradictory scientific evidence for lipid rescue being an effective antidote and its increasingly reported adverse effects, it is reasonable to strictly limit its use in clinical practice. We would not recommend its use at all in cases of oral poisoning.”

There are several significant problems with this conclusion:

  1. Because of the nature of case reports, it is most often impossible to determine causality, especially using a scale that was not designed for this purpose. There are few methods describe as to how the authors assigned a score to each case. One could go back to original cases and make one’s own conclusions, but the reference list has only 22 citations and does not include all or most of the relevant papers. The paper says this information is contained in online appendices at http://het.sagepub.com/supplemental, but my efforts to reach that material through the link came back with either a blank page or an error message. This is really inexcusable. While it’s true that this paper was posted online before it appears in print, it should not in my opinion have been posted until the crucial supplemental material was available.
  2. It is completely unclear how the “discussions” of disputed cases were handled. Was one author more pessimistic and more persuasive than the other?
  3. Of course many of these patients received multiple, nearly simultaneous, treatment modalities. This certainly would make determining causality more uncertain, but not to the extent of concluding that LRT should not be used at all in oral poisoning.

I would add that the Cochrane Review states that a systematic review “summarizes the results of available carefully designed healthcare studies (controlled trials) and provides a high level of evidence on the effectiveness of healthcare interventions.” Case reports are not “carefully designed healthcare studies.” This is no systematic review, and this paper totally misappropriates the term.
Related posts:

Effect of lipid rescue therapy on laboratory tests

Excellent review of lipid rescue therapy

Lipid rescue therapy can interfere with critical lab values

 

One Comment:

  1. Leon Gussow Says:

    I received the following remarks from Dr. Guy Weinberg via e-mail. With his permission, I am posting them here as a comment on the Forsberg et al paper:

    “Thanks, Dr. Gussow for the very helpful critique of this paper. There are other shortcomings, including design and statistical flaws that I will address in a different forum. Meantime, I will comment on a couple of other issues I find problematic with Forsberg et al.

    The premise and results of the paper are connected by a circular argument that detracts from both. The authors start the Discussion saying, “Before evaluating the main findings of this study, one should consider the … immense impact of publication bias.” – an argument that seems to invalidate their use of case reports for the purpose of this study. Starting the Discussion by disparaging your key dataset indicates a particular degree of cynicism particularly given that their apparent goal is to devalue support for lipid resuscitation.

    That is, the title, “No support for lipid rescue….” seems more a goal of the study than a result. Did they truly find ‘No support for lipid rescue in oral poisoning:…” Really? No support? None? Anywhere? No support from experimental studies? No evidence of a beneficial effect in any case report? The title suggests to me the authors approached the project with considerable confirmation bias – the tendency to interpret data to confirm one’s preexisting beliefs.

    Did Forsberg et al disregard entirely the clinical judgment and observations of physicians who witnessed improvement with lipid resuscitation? Do they view their post facto interpretation of the effects of lipid resuscitation as a more accurate assessment? Were their findings just a forgone conclusion? Isn’t it really an opinion piece disguised as a ‘study’. This assertion is supported by the considerable and one-sided attention the authors devote to topics far afield from their data: ‘sliding of indications’ for lipid, experimental models of oral ingestion, adverse effects, etc. As Ioannidis and others point out, mislabeling an editorial as a systematic analysis doesn’t make it one.”

Leave a Comment:

Comments will be posted after review and approval by the editor. TPR reserves the right to delete a comment for any reason.