Monday, May 2, 2022

Simplifying Opioid Conversions

by Drew Rosielle (@drosielle)

A Satirical Monologue in One Act:

“Ok, 3rd year resident, let’s talk about rotating opioids. What do I mean by ‘rotating’ opioids? It’s just therapeutically switching one opioid with another. It’s um, like, a turnstile, I guess? Anyway--first let’s look at this equianalgesic table. Do you know what equianalgesia means? No? It is the concept that different opioids have the same analgesic power but at different milligram doses due to different potencies. That is, the idea that, say, 50 mg of oral morphine has the same analgesic power as 10 mg of hydromorphone. So 50 mg of oral morphine is equivalent to 10 mg of oral hydromorphone.

"Ok, let’s look at this table, every entry on this grid is equianalgesic, meaning that the 30 mg here of oral morphine has the same analgesic power as the 1.5 mg of IV hydromorphone. So, let’s say we have a patient on 100 mg oral morphine. How much hydromorphone is that “equivalent” to? Ok, so what we have to do is set up a cross multiplication problem…here’s how you do that.

[We’ll skip the painful talk-through of the cross multiplication.]

"Ok, so it looks like the answer is 5 mg of hydromorphone IV. Now we have to adjust that dose for incomplete cross-tolerance. Do you know what that is? No? Ok, so incomplete-cross tolerance is the idea that our patient on 100 mg of morphine is somewhat tolerant to the morphine. They are therefore also presumably somewhat tolerant to any other opioid agonist like hydromorphone. The problem is that the tolerance they have to morphine may not fully apply to the tolerance they have to the hydromorphone—the morphine tolerance “incompletely” transfers so to speak to the patient’s hydromorphone tolerance. So, what that means is that even though this table says 100 mg of PO morphine is equal to 5 mg IV hydromorphone, for safety’s sake we need to reduce that 5 mg somewhat to determine what we actually put the patient on. That’s what we mean by reducing for incomplete cross-tolerance.

"If you have 5 minutes, it’s actually an interesting story, because no one is really sure that ‘incomplete cross-tolerance’ is even physiologically a phenomenon, like on the cellular level, our field just calls it that because that’s what it’s always been called. In fact, the idea of ‘incomplete cross-tolerance’ only makes sense if you think there actually is a firm “equianaglesic” potency relationship between morphine and hydromorphone that applies to most patients. In fact, no one actually believes that and the broad consensus is that there is a pretty significant range of ‘equianalgesia’ between any 2 opioids in different patients. I.e., in one patient, 100 mg of oral morphine may end up providing equal analgesia to 12 mg IV hydromorphone, in another patient, 4 mg of IV hydromorphone, etc. It’s a little bit of a crap-shoot.

"Actually, it’s probably far worse even with fentanyl, we’re talking about a far, far wider range of individual responses, and don’t even get me started talking about methadone and buprenorphine [self-knowing chuckle here]!

"So, back to incomplete cross-tolerance. If you’re going to argue that it would just be easier to say ‘dose reductions due to safety’ or ‘dose reductions due to individual variability in responding to opioids’, you’d be right. That’s much easier to understand, and probably more accurate, but maybe it makes us feel like we’re the keepers on secret knowledge to say incomplete cross-tolerance? Who knows?

"Regardless, this is where it gets really interesting, because to properly account for ‘incomplete cross-tolerance’ we have to decide how much less opioid we need to put someone on after we do the initial calculation. Yes, there are 2 calculations: the equianalgesic one, then a 2nd incomplete cross-tolerance one in which we reduce the first amount by a further amount, most people recommend 25-50%.

"How do you choose how much you actually dose reduce? Well, there’s about 10 patient factors you need to balance when making that decision. Yes, I said 10, but don’t worry, if you do this for a couple years it’ll become second nature to you [self-knowing chuckle]!

"Ok, so let’s talk about these 10 patient factors….”

[End]

Dear friends and colleagues, I think we can all do better than this, and here is my current thinking as to a better approach.

In my last post I outlined why I think equianalgesic tables (EATs) are broken and proposed we collectively move instead to using Conversion Tables (CTs) when teaching others how to switch opioids. Briefly, this is because CTs engender much simpler and clearer math; and we can easily adjust the conversion factors in CTs based on emerging data, without having to rejigger every ‘equianalgesic’ relationship which EATs force us to do. My offer from the last blog post still stands: it would be ideal if there could be a consensus process to generate a consensus CT that our professional community can broadly adopt. While I like the CT I made, I don’t think any one person should be setting the conversion ratios alone, even if it is standing on the shoulders of giants! I’m not an investigator but am willing to lend any support I can to this. It is also possible the best option is for organized medicine to lead this, and I’ve been chatting with some folks about this a little already. I’m not sure where it will lead, nor if AAHPM has any appetite for that, we’ll see.

Regardless, I’ve been thinking and talking with others a lot about the entire enterprise of how we teach and talk about opioid switching and I have a lot of ideas that go beyond CTs for potential ways we can improve our approach to this. This post is simply to capture those ideas, none of which I’m claiming are independently mine.

1) We should never utter the phrase “incomplete cross-tolerance” again.

I hope this is not controversial, I’ve yet to meet anyone in the “Incomplete Cross-Tolerance: What a Super Helpful Concept!” camp, but its use continues to pervade discussions of opioid switching. My fundamental objections to incomplete cross-tolerance: it is a concept that requires an entire paragraph of explanation (!!) to someone unfamiliar with it, plus it may not even be a ‘real’ thing, plus there are readily available, broadly understood, completely accurate alternatives to it (e.g., dose reductions for safety, dose reductions due to individual variability, etc.). While the monologue that started this post is satire, I’m sure I’ve uttered every phrase in it for real at one point or the other in the past 20 years. I know I am not alone.

2) We should talk about ‘methods for safe opioid switching’ and not equianalgesia. Equianalgesia is unnecessarily jargony and I don’t think it even exists on the population level.

Incomplete cross-tolerance only makes sense if we think equianalgesia itself is a firm and helpful concept. If you believe that we can announce, without qualification, that 20 mg of IV hydromorphone is ‘equianalgesic’ to 100 mg of oral morphine then maybe it makes sense to talk about incomplete cross-tolerance. But I don’t know of anyone who thinks that “100 mg of oral morphine is equivalent to 20 mg of IV hydromorphone” in any sort of widely generalizable, arbitrary way. I think all of us operating in this realm may acknowledge that across 100 patients maybe the averages end up close to 100:20, but who cares? Honestly, that factoid is a DISTRACTION because the way EAT switching has currently been set up has encouraged us to first focus on this fake-and-based-in-less-science-than-we’d-want notion of equianalgesia, then prompted us to do this secondary work-around-caused-only-by-the-fact-that-we-set-a-ratio-in-the-first place. Why do it this way? Why not just switch at a ratio that works most of the time, instead of nearly none of the time?

The clinical task we are faced with is not an abstract exercise in the variable potencies of different opioid molecules averaged across the human population, instead it’s a practical task of How do I safely switch this patient’s opioids? Safely means two things here: that the patient does not have significant sedation from the switch nor a major worsening of their pain. I do think what I’m arguing here is more than semantics—that going from equianalgesia to ‘methods of safe switching’ is not just me renaming it. It is a significant conceptual frame shift in this critical and common.

Using a concept like ‘equianalgesia’ promotes this idea that there’s a “right answer” to the question of “If I’m switching someone from 100 mg of morphine in a day to IV hydromorphone how much hydromorphone do I prescribe?”

My proposal is there isn’t a right answer. It is not a helpful goal or concept and we need to move onto the more salient task of switching safely. That framework focuses us on a practical clinical task. From talking to colleagues, many of us who prescribe methadone and buprenorphine long ago moved on from any idea that it was our task to somehow divine the ‘equianalgesic’ dose of methadone from another opioid. Instead, we have a protocol to make the switch, the protocol works (‘is safe’), all is well with the world. My proposal here is that this is how we conceptualize all opioid conversions.

3) We must start emphasizing that the absolute dose of opioids involved matters for safe switching.

Stepping back from the finer points of what conversion ratios we should be using for opioids—honestly, if what you’re talking about is doses in the range, say, of 40 mg of oral morphine, I don’t really think it matters what ratios you use to switch. Frankly, it probably doesn’t matter much if you reduce further for safety or not.

Consider someone is on 40 mg a day of morphine PO. You want to put them on IV hydromorphone:

-- Classic EAT: divide by 20 to get 2 mg of IV hydromorphone

-- DOC2 Table: divide by 12.5 to get 3.2 mg of IV hydromorphone

If you reduce further for safety, presumably you’ll still get an answer about 1.2 mg apart between the 2 tables. Honestly, I don’t think the difference there is much of a safety concern. Both are fine. No opioid- tolerant patient became sedated or had a pain crisis because of a difference of 1.2 mg of IV hydromorphone over 24 hours; of course, the ‘fine-tuning’ might be a bit different, but as long as a patient had close follow-up the analgesic outcomes would be about the same too.

But what if the patient was on 400 mg a day of oral morphine? Personally, I would not be so cavalier about saying there is no relevant difference between 20 and 32 mg of IV hydromorphone in a day.

Yet, if you read pretty much all the opioid conversion stuff out there, they all treat these opioid switching ratios as fixed across all ranges of opioid doses. They suggest the same approach for 400 mg of morphine as 40 mg of morphine. To me, this is shocking, and profoundly unsafe. I’ve never met a single person who treats opioid switching ratios as linear across all opioid doses, meaning they would be more ‘conservative’ in their switching for someone going from 400 mg of morphine than 40 mg. I believe I am merely naming here a clinical practice which is already widespread! Yet, for the last twenty years of journal articles, books, national presentations, and local presentations that I’m aware of, this is not splattered in flashing neon ALL CAPS at the top of every table. It is not a core part of our basic nor specialty teaching on this topic. I genuinely don’t understand it.

Fundamentally, I’m arguing that we’ve heaped HUGE amounts of attention on the complex subtleties and nuances of reducing for safety when it comes to opioid switching, and virtually no attention to the nuances of switching at higher dose ranges. I think we got this wrong.

All this leads me to a few practical proposals:

1) “We”* stop using EATs and instead use simple “going from X to Y” conversion tables (see this post for more details).

2) We frame this practice as methods of ‘safe switching’ and do away with ‘equianalgesia’ (at least as a core concept in teaching others about opioid switching) and ‘incomplete cross-tolerance’ (that we can discard completely).

3) We promulgate a conversion table without need for additional dose reductions in most patient scenarios. I appreciate this is in some ways the most radical thing I’m proposing here, and certainly it is the one I’m most tentative about. But I suspect that a lot of our need for “reducing for incomplete cross-tolerance” has come from the fact that in order for any of our EATs to make sense they’ve needed to have on them conversion ratios that are dangerously aggressive. E.g., the for the Classic EAT I would divide the PO morphine dose by 20 to get the IV hydromorphone dose and, truth be told, it would be rare I would dose reduce beyond that: it was a good safe ratio much of the time in my experience. However, the reverse conversion, multiplying the IV hydromorphone dose by 20 to get the PO morphine dose, I regularly, heavily dose reduced (often by 50%). My argument is that to a large extent our need to create that 2nd step of dose-reduction really comes from the fact that we locked ourselves into aggressive conversion ratios with EATs and if we had just a simple “Going from X to Y” conversion table with pre-set ‘safe’/conversative ratios, we can side step this incredibly complex beast we’ve created (the need to reduce for safety / “incomplete cross tolerance”).

How would we do this: my idea is that we create a conversion table which folks broadly agree they’d be comfortable handing to a first month medical intern and telling them: “Young doctor, within these parameters, just use this.” If this idea freaks you out, remember for decades we’ve handed them inscrutable equianalgesic tables with even more inscrutable and hand-waving suggestions about reducing for incomplete cross-tolerance and expected them to know what to do. My hypothesis here is that the method I’m proposing is safer and requires far less needless work for our generalist colleagues.

As an example, we’d create a conversion table for a hypothetical 70-year-old with an eGFR of 50 and mild frailty—something like that. The idea being that most of the conversions happening out there are for patients like that or more physiologically robust, so the table would be built-in safety for a broad range of patients. I think it could be close to my draft Conversion table (Dropbox, Google Drive), but some of the conversions would be more conservative for sure.

The table would need to be explicitly labeled with the dose limits (E.g., this is for patients on 100 mg or less a day of oral morphine, 20 mg a day or less of IV hydromorphone, etc. This would avoid unnecessary calculations of oral morphine equivalents (OMED, OMDD, OMD, or any variation on those abbreviations). What those ‘dose caps’ are would require, I imagine, vigorous discussion, and I don’t strongly believe the, e.g., 100 mg a day morphine cap is actually the right one.

The table and associated teaching strongly emphasize the structure of care and close follow up is what’s most important for switching opioids safely. Don’t switch and walk away, switch and follow up!

Importantly, I’m proposing this is a conversion table for ‘generalists’ who switch opioids. I think the role of palliative and pain specialists is to have a deep knowledge of all this, including the history of opioid switching, and the nuances of dose reductions for safety, etc. All our fellows (and all our colleagues) should read Demystifying Opioid Conversions 2 Ed (aka DOC2) if they haven’t already! However, that book is not for generalists, they aren’t reading it and we can’t expect them to read it. To be clear—I’m not suggesting we jettison that book or the accumulated knowledge of opioids switching, I’m more proposing there are simpler and arguably a little safer way to teach our generalist colleagues about this, keeping in mind that to date we’ve made it so difficult for them most of them are using online calculators of various levels of reliability!

This is just a proposal which has come out of one person’s brain. I very much believe that ideally this is best decided by consensus and committee (in the absence of good clinical data for which there are zero). So, it is meant as a starting point for discussion. Ideally, I’m hoping others will join me and I am talking with CAPC about this and plan on with AAHPM too. I think the new data from MD Anderson which lead to the change in the DOC2 Table are a great opportunity for us as a field to collectively look at all this and try to do something that is far simpler, and at least a little safer. If you’re interested, let me know!

(This post was updated May 3, 2022 to correct an error in the opioid calculations!)

For more Pallimed posts about opioids.
For more Pallimed posts by Dr. Rosielle click here.


Drew Rosielle, MD is a palliative care physician at the University of Minnesota M Health Fairview in Minneapolis. He founded Pallimed in 2005. You can occasionally find him on Twitter at @drosielle.



Pallimed | Blogger Template adapted from Mash2 by Bloggermint