Saturday, 11 February 2017

Should we give a lower dose of ketorolac for treating acute pain in the ED?

Ok, ketorolac is a little bit funny...

For some reason, it was never granted TGA approval in Australia for intravenous use although it is used all over the world in this manner. (Therefore it is perfectly fine for giving it IV off-label).

Second, it is probably the only drug where it is recommended to give a lower oral dose (10mg) than parenteral (30mg).

But now that 30mg IV dose is being challenged.

These authors from New York suggest the “ceiling dose” of ketorolac is probably 10mg and postulate that higher doses can cause more harm than good.

They performed a RCT of three different doses of IV ketorolac monotherapy; 10, 15 and 30mg. Inclusion criteria were adults with undifferentiated pain (most had acute flank, abdominal, musculoskeletal pain and headache). They had to have a numerical rating (NRS) of pain >5 not having had any previous pain medication.

The primary outcome was reduction in pain score at 30 minutes.

It was powered to find a 1.3 difference and as such 80 patients in each arm were enrolled.


The pain score dropped about 2.5 in both the 10 & 15mg doses of ketorolac. It dropped by 3.0 in the 30mg dose. This was not statistically significant.

The authors boldly conclude, “The results of our study provide a basis for changes in practice patterns and guidelines in ED care, supporting use of the 10mg intravenous ketorolac dose.

Unfortunately this conclusion is inappropriate.

Despite the decent design and blinding of this study, there were significant problems such that one cannot make any decent conclusions.

Any time one does not see a dose response with a particular drug then one must seriously consider if it is having any effect at all. This is pharmacology 101. I very much imagine that placebo (i.e. a ketorolac dose of zero) would have given similar results in this heterogeneous group of patients with these study methods.

I give ketorolac in patients in whom I think it will have a greater benefit. This is usually part of a cocktail of medications and not monotherapy. It works extremely well in patients with renal colic but less so in those with abdominal pain and headache. So extrapolating the results of this study to the individual patient is problematic.

A key foundation of this study is predicated on the theoretical harm with a higher dose of ketorolac. Otherwise this whole discussion is moot. I would seriously question if a single dose at the higher range would cause clinically important problems. In addition, we can pick and choose which patients get higher doses. An 85 year old 40kg woman with chronic renal failure would probably never get ketorolac in the first place.

In summary, we don’t know if a lower dose of ketorolac is appropriate at the moment. A single administration of ketorolac at the standard dose would be extremely unlikely to cause clinically significant problems.

It is always wise to be sceptical of small single centre studies. In the end they are often proven wrong. Nevertheless, this study could provide foundation for future study and clinical equipoise for inclusion of a placebo dose of ketorolac.

Watch this space but don’t change you dose of ketorolac just yet.

(Addendum: Please see lead authors reply to this post under the comments section)



  1. Dear Brian,
    Thank you for your post. While I agree with your points regarding the use of ketorolac for patients with renal colic, the better analgesic effect when ketorolac is combined with other analgesics, and a need for careful patient’s selection with respect to age and co-morbidities, several comments made in this blog deserve a response.
    The Authors did not “suggest” the analgesics ceiling dose of ketorolac to be a 10 mg, contrary, the used it as a base for their analgesics dose-finding study. This pharmacological phenomenon of NSAID’s class of analgesics has been extensively studied in patients with post-operative, malignant, dental and now acute pain in the ED for over 25 years (since the discovery of Ketorolac).
    The authors in this study did not “postulate” that higher doses will result in more side effects. The whole purpose of the study is to compare analgesic efficacy of 3 dosages of IV Ketorolac for short-term analgesia in the ED as it states in the beginning of the manuscript and at the conclusion. The study was not sufficiently powered or designed (no long-term follow up) to find dose-related serious side effects of different doses of IV ketorolac in the ED. However, several studies clearly demonstrated increased rates (not association) of post-operative bleeding of a single (supra-analgesic) dose of 30 mg IV or even 15 mg (Cawthorn 2012, Gallagher 1995, Rubin 2012), or increased bleeding time in healthy volunteers (Singer 2003). The harm is real, the data is there to support it, and it very much deserves the discussion and further (ED-based) investigations.
    The comment regarding an apparent lack of dose-response in our trial is inaccurate. The trial clearly demonstrated dose-analgesics response up to 2 h that was similar between three groups. The actual decrease of pain at 30 min, 60 min, 90 min and 120 min by 33%, 42% 48% and 51% respectively, it is a good response (statistically and clinically speaking) and cannot be ignored. Your pharmacology 101 reference does not address the fact that at 10 mg dose the dose -response curve of ketorolac sharply plateaus with respect to further analgesics response to 15 or 30 or 60 or 90 mg dosing regimens and subsequently curves back up with respect to increase in anti-inflammatory effect (higher the dose, more anti-inflammatory effect).
    It is solely my opinion that the use of placebo in pain studies is unethical. Furthermore, the studies on post-operative pain, migraine and acute pain in the ED showed clear superiority of IV Ketorolac to placebo. In addition, the very heterogenous population of patients in our study is as close of a resemblance of real patients who gets IV ketorolac in the ED’s across North America as it gets.
    Lastly, the authors conclusion that there is no need to exceed a 10 mg IV Ketorolac dose in the ED for managing variety of acute painful conditions that is based on analgesic ceiling dose, is appropriate and supports the prior work of EM Colleagues ( Catapano 1996, Raney 2000, Arora 2008, Bond 2013).

    Sergey Motov
    Co-author from New York

    1. Sergey,

      Thanks for your detailed reply. You make many excellent comments especially regarding previous study of ketorolac.

      Yes, it looks like the response from the different doses of ketorolac was quite similar over time. This is what makes me question if a dose of zero would have demonstrated similar results in this heterogenous patient population with the methods used in this study.

      I agree that there are significant ethical questions one must ask when entertaining the notion of using a placebo in studies of pain. But we know that pain does respond quite consistently to placebo. Some of the effect is even reversed with naloxone!

      I think if we were to truly sit back and examine our practice, we would find that we use “placebo” a lot more than we think. Is this bad? Interesting ethical question…