Without evidence of benefit, an intervention should not be presumed to be beneficial or safe.

- Rogue Medic

These authors read far too much into their limited study – Part I


There is a new study that looks at prehospital fentanyl. It starts out well, it is even randomized, blinded, and prospective, but it loses focus and draws conclusions that are not remotely justified by the study. Starting out well –

In one study looking at patients transported to EDs by ambulance with painful extremity injuries, patients who did not receive analgesics during transport waited approximately 75 min after arrival to the hospital before receiving any analgesic medication (2).[1]

This is a real problem.

Does prehospital fentanyl decrease that delay, or in other ways improve outcomes?

The authors cite several possible benefits – takes effect more quickly, is metabolized more quickly, causes less nausea, and causes less hypotension. However, they seem to ignore the last one – less likely to produce hypotension. In prehospital treatment, this is important.

They do not have many scene transports (only 16%) and they have doctors on the helicopters, so the authors’ flight service appears to take a conservative approach to treatment. On the other hand, their study protocol makes me look like Scrooge.

Each milliliter of study drug contained either 4 mg of morphine or 50 mg of fentanyl. The dosing was kept below the recommended loading doses of each narcotic to aid in blinding and allow for repeat dosing in small aliquots in the concentrations available in our pharmacy.[1]

However, that non sequitur plan appears to have worked even worse than we would expect.

Why not start out with a double dose?

The dose of morphine used in many studies is 1.0 mg/kg followed by doses of 0.05 mg/kg.

The repetition of the dose every 5 minutes should have made this not a problem, but the study protocol had a ridiculous limitation.

Each patient was allowed a maximum of five doses of the study drug.[1]

That is a maximum of 20 mg morphine or 250 µg fentanyl.

Depending on the weight of the patient, I can give more morphine or fentanyl on standing orders than these doctors can. If I reach my standing orders maximum, I can call medical command for orders to give more.

The authors even acknowledge this in their discussion.

Our initial doses correspond to 0.05 mg/kg of morphine and 0.71 mg/kg of fentanyl for a 70-kg adult, whereas others have found that sufficient pain relief required doses of 1.0 mg/kg and 1.6 mg/kg, respectively (7,20).[1]

What is the point of arbitrarily limiting the total dose?

The point certainly does not have anything to do with good patient care, since there is no evidence that any such arbitrary limit in any way improves outcomes.

If anyone knows of any research to suggest that this limitation improves outcomes, please provide it, but I do not know of any justification for this limit.


An average (mean) of 3 doses of morphine and 3.3 doses of fentanyl. That is 12 mg morphine and 165 µg fentanyl. These are not large doses, but the results show that they only had very limited success in managing pain.

Although they could start treating at a pain level of 1 out of 10, they still stopped at 5 out of ten – a pain level that would be appropriate for me to start treating with morphine or fentanyl, or to continue treating with morphine or fentanyl. I treat moderate to severe pain.

But it gets worse.

Of patients in the study, 57.5% received analgesics before being enrolled, without a significant difference between either arm. Although the medication name was recorded, the dose was not always recorded by the flight crew. Average doses of morphine, Dilaudid, and fentanyl were 4 mg, 1 mg, and 100 mg, respectively, and were similar in both arms, based on available data.[1]

Most of these patients had already received an average of one dose of study drug (or the equivalent amount of hydromorphone [Dilaudid]) prior to being entered in the study, so the pain management should have just been a continuation of treatment, even though the dose they already received had been inappropriately low for a starting dose.

Patients who reported any pain score other than zero were then given the study drug in a 1-mL intravenous bolus. Patients were then reassessed every 5 min (normal flight protocol, with automated monitor and clinical evaluation) during transport with a complete set of vital signs (including pulse oximetry) and another numeric pain score. During each reassessment, a 1-mL bolus of the study drug was given for any pain score > zero.[1]

Except that does not appear to be even close to what happened.

Although the study protocol called for administration of medication every 5 min, a mean of only three doses was given despite a mean patient care time of 40 min.[1]

A change of 13 mm on the visual analog scale and a corresponding change of 1.3 on the NPS have been generally accepted as a clinically significant change in pain relief (8,12,13). We found that both morphine and fentanyl at repeated study doses provided clinically significant pain relief by decreases in mean pain scores of 2.2 and 2.5, respectively.[1]

Yes, the difference was clinically significant, but we can do much better. Here are the details of the pain levels –

The mean pain score at the beginning of enrollment was 8.0 ± 2.0 in the morphine arm and 8.0 ± 1.8 in the fentanyl arm. The mean final pain score was 5.8 ± 2.7 in the morphine arm and 5.5 ± 2.4 in the fentanyl arm. The median initial pain score was 8, with an IQR of 3. There was no difference between the two groups. The median final pain score was 5, with an IQR of 3.5, with no significant difference between the two groups (Table 2).[1]

Starting morphine pain level from 6 to 10 out of 10.   Starting fentanyl pain level from 6.2 to 9.8 out of 10.

Ending morphine pain level from 3.1 to 8.5 out of 10.   Ending fentanyl pain level from 3.1 to 7.9 out of 10.


This is not even cutting the pain in half – this is only cutting the pain a little.

that is still important, but it could be much better.

61.5% of morphine patients and 69% of fentanyl patients had a significant improvement in pain level (≥2 according to their predefined criteria).

There were no incidences of pruritis or vomiting in either group. There were no episodes of hypotension in either group.[1]

To be continued in Part II.


[1] The Effectiveness and Adverse Events of Morphine versus Fentanyl on a Physician-staffed Helicopter.
Smith MD, Wang Y, Cudnik M, Smith DA, Pakiela J, Emerman CL.
J Emerg Med. 2012 Jul;43(1):69-75.
PMID: 21689900 [PubMed – in process]

There is one unusual aspect to the study that does not appear to affect the outcome, but raises questions about how many obstacles to research we create, when the obstacles may not be valid.

The study was fully reviewed by our Institutional Review Board, and given that both treatment arms are considered acceptable practice with equal risk, informed consent was not deemed necessary for this study. Upon completion of participation, each patient was given a verbal and written debriefing of his or her study involvement.[1]

Smith MD, Wang Y, Cudnik M, Smith DA, Pakiela J, & Emerman CL (2012). The Effectiveness and Adverse Events of Morphine versus Fentanyl on a Physician-staffed Helicopter. The Journal of emergency medicine, 43 (1), 69-75 PMID: 21689900