In Defense of Morphine (Part 2)

Yesterday, I introduced some of the history preceding the recent Thrombosis and Haemostasis study looking at morphine in patients experiencing ST-elevation myocardial infarctions (STEMIs). [For the curious, we’ll talk about that new article in Part 3. I have to work today, so I can’t guarantee that it’ll be up tomorrow, but soon!]

We started with a retrospective, observational study from 2005—one of the first to associate morphine with significant harms [1]. In that dredge of the CRUSADE data, patients experiencing acute coronary syndrome (ACS) who received morphine did significantly worse than those who didn’t. The numbers are alarming at first glance, showing that morphine administration conferred an increased risk of mortality, with an odd ratio of around 1.50 across most of the calculations.

One must never conflate correlation with causation.

There is a high risk of confounding with this sort of study, and the CRUSADE data missed an important variable: refractory ischemia. The registry gathered no information on which patients experienced recurrent or refractory chest pain (presumably due to ischemia). As a result, we cannot determine if the patients in the morphine group both received the medication, and did worse, in association with this third, outside factor.

I think it’s likely.

I’m reviewing these points because I believe they render the entire CRUSADE study useless. With such a large unmeasured confounder in play, we cannot trust any of the results. I wish there was something we could do to work around that, but there isn’t. Such are the pitfalls of retrospective registry studies. In an ideal world, this article would simply be forgotten and drift into obscurity as the non-evidence it is.

As you may have guess, we don’t live in an ideal world, and instead this study is continually cited in reference to morphine’s safety in ACS. A lot (let me know if that link doesn’t work, but the results are below anyway).

But what’s done is done, and many people now worry that morphine may harm patients with ACS. There are plenty of theories why it may be harmful, and some of them are interesting, but so far none of them fit with the effect-size measured in the CRUSADE article. The most popular at the moment is that morphine impairs absorption of P2Y12 inhibitors like clopidogrel. This effect is very real and is consistently demonstrable. There’s currently a rush to examine the mechanism; it’s almost certainly why this week’s article in Haemostasis and Thrombosis was performed. But there is a small issue…

It’s illogical.

I don’t mean that the mechanism is unproven, or the data is fuzzy. Rather, the whole explanation makes no sense when you look at the magnitude-of-effect it’s supposed to account for. It’s preposterous.

Thankfully, Dr. Rory Spiegel over at EMCrit/EMNerd has done a far better job examining this issue than I ever could [2] [3] [4]. If you want to have an intelligent discussion about this topic, you absolutely need to read at least reference #4: The Case of the Inconsequential Truth. I’ll summarize his major points here:

  • Opioids impair absorption of P2Y12 inhibitors, probably by decreasing intestinal motility and delaying gastric emptying.
    • Of note, it doesn’t turn the medications into an inert placebo—it makes them reach a slightly lower and significantly later peak concentration. They still function.
  • Investigators theorize that this reduced efficacy of a second antiplatelet agent (on top of aspirin) could explain the increased mortality seen in ACS patients who receive morphine.
  • HOWEVER, when you look at the trials that studied “upstream” P2Y12 inhibitors given before PCI, their beneficial effects were extremely small—some argue, negligible.
  • Those benefits were almost entirely measured as reductions in peri-procedural MIs (type 4 MIs), without significant reductions in mortality or morbidity.
  • If morphine is really impairing the function of these medications, doing so would only cut into that small benefit in periprocedural MIs.
    • If P2Y12 inhibitors do not decrease mortality, reducing their efficacy with morphine cannot, in turn, increase mortality. It’s like trying to multiply or divide by zero.

Based the science available to us, it is impossible for morphine’s effects on P2Y12 inhibitors to explain the increased mortality seen in the CRUSADE study. Either there is another mechanism at play, or morphine isn’t causing harms and we’re just seeing a spurious association due to confounding. As I argued in Part 1, the latter seems more likely to me.


Make sure you check out Part 3 in a day or two, when I’ll finally examine the recent study that inspired this entire line of posts.



What about fentanyl? I’ve seen many people who claim they are moving to fentanyl because it doesn’t have the same harms associated with it that morphine does. You might as well move to ketamine if you want a pain medicine with absolutely no outcome evidence against it in ACS.

Histamine release caused by parenteral morphine.

While I agree that fentanyl has a cleaner pharmacodynamic profile with less histamine release, vasodilation, or hemodynamic effects, it’s dishonest to cite morphine’s association with worse outcomes or effects on P2Y12 inhibitors as evidence for switching. The only reason fentanyl doesn’t carry the same baggage as morphine is that it has not been used as much in ACS. You can’t find any CRUSADE-style studies that show fentanyl causing harm because no one has done them. Likewise, until recently, morphine seemed unique in how it impaired P2Y12 inhibitor function, but that was only because it was the only opioid studied (although there was every reason to believe it was an opioid-class effect based on the proposed mechanism).

Enter: the recent PACIFY randomized clinical trial [5]. This study examined the effects of pre-PCI fentanyl on plasma concentrations of ticagrelor (another P2Y12 inhibitor) and platelet function. Guess what: it showed the same effect we’ve seen in morphine. Fentanyl delayed platelet inhibition with a lower area under the curve for the ticagrelor concentrations, but these differences narrowed after 4 hours.

I prefer fentanyl and I think you should too. There are a lot of things to like about it, but please, don’t make unsubstantiated claims that morphine is hurting patients; it’s a lazy, myopic misuse of the evidence.



  1.  Meine TJ, Roe MT, Chen AY, Patel MR, Washam JB, Ohman EM, Peacock WF, Pollack CV Jr, Gibler WB, Peterson ED; CRUSADE Investigators. Association of intravenous morphine use and outcomes in acute coronary syndromes: results from the CRUSADE Quality Improvement Initiative. Am Heart J. 2005 Jun;149(6):1043-9. doi: 10.1016/j.ahj.2005.02.010. PMID: 15976786.
  2. Spiegel, R. A Case of Identity.
  3. Spiegel, R. A Case of Identity Part Two.
  4. Spiegel, R. The Case of the Inconsequential Truth.
  5. McEvoy JW, Ibrahim K, Kickler TS, Clarke WA, Hasan RK, Czarny MJ, Keramati AR, Goli RR, Gratton TP, Brinker JA, Chacko M, Hwang CW, Johnston PV, Miller JM, Trost JC, Herzog WR, Blumenthal RS, Thiemann DR, Resar JR, Schulman SP. Effect of Intravenous Fentanyl on Ticagrelor Absorption and Platelet Inhibition Among Patients Undergoing Percutaneous Coronary Intervention: The PACIFY Randomized Clinical Trial. Circulation. 2018 Jan 16;137(3):307-309. Epub 2017 Oct 18. doi: 10.1161/CIRCULATIONAHA.117.031678. PMID: 29046319.

3 Trackbacks

Leave a Reply

Your email address will not be published. Required fields are marked *