sinaiem.org

I’m finally getting around to the some of the good articles on ED pain management that appeared this summer. The first was in the August 2006 Annals of EM, an article by Gallagher, Esses et al. entitled, “Randomized Control Trial of Morphine in Acute Abdominal Pain.” The authors tested the oft-repeated dictum that morphine affects diagnostic accuracy, measuring pain on a 0-100mm visual scale in a prospective double-blind random trial, giving 0.1 mg / kg of IV Morphine sulfate or placebo (ouch!) with an endpoint of “diagnostic accuracy” (ie, comparing the provisional diagnosis made by an emergency physician in 15 minutes after the agent is given, vs. diagnosis at six or more weeks of followup.)

They randomized 78 patients into the morphine arm, 73 into the placebo group. After fifteen minutes, the patients who got morphine changed their pain rating from 98 to 65, on average, whereas the placebo group went from 99 to 97. Diagnostic discordance occurred 11 times in each group, meaning that accuracy wasn’t affected by analgesia.

The dictum to avoid analgesia in abdominal pain dates to a 1921 proclamation by Sir Zachary Cope — a good example of emminence-based medicine. Maybe this warning was appropriate in the age before antibiotics and CT scanners, but we now have eleven trials in the last twenty years showing that Cope doesn’t cut it.

In this convenience sample, patients 21 and older with atraumatic abdominal pain of less than 48hrs duration, who consented to the study and were judged by an EP to need opiods, were enrolled. The excluded were those with flank pain, pregnant patients, those allergic to opioids, patients with sickle cell, and those who self-medicated prior to arrival. Also, no hypotensives.

They used a doc not involved in the patient’s care and not aware of any lab or imaging data to assess provisional diagnosis at 15 minutes. The patient’s vitals were checked q15-30 minutes for three hours, at which point the patient could get all the meds they needed.

The researchers worked hard to get a final diagnosis, searching med recs, path and imaging studies, and using standardized phone questionnaires at 6 weeks post-presentation. The study’s endpoint was accuracy, either an identical diagnosis or one trivially different (from a clinical standpoint). Discordance was defined as a difference from provisional to final that affected patient health status adversely, as defined by either of the clinician authors.

578 patients were screened, 160 participated (there were many who met exclusion criteria, but only 31 refused — which surprises me. I mean, if I showed up in an ED with abdominal pain and a researcher said there’s a 50% chance I won’t get pain meds for three hours, I’d refuse to consent).

Complications such as nausea / vomiting, hypotension were approximately equal in both groups. No naloxone was needed. The docs guessed right 63% of the time about who was getting what, better than 50% random chance.

One patient died in each group, and each cause (pancreatitis, mesenteric ischemia) was diagnosed within 90 minutes.

As stated above, diagnostic discordance occurred 11 times in each group. In the morphine group, 67 of the 78 patients had indentically or near-identical provisional and final diagnoses, for an 86% agreement. For the placebo, it was 64 / 75 for 85% concordance. Hence, there was virtually no difference between the two arms, with a 95% CI of -11% to +12% centered around 1.0% difference.

The authors noted some limitations to this study (and it’s discussed further online). One limitation is convenience sampling. Another is insanely high pain scores (98?? really?). Another limitation is the young age of the patients (average was 45 years old). Finally, the authors note that 85% of their patients were Latino or black, limiting general applicability.

 I would add that 15 minutes may be too soon to assume pain relief or attempt a diagnosis (so, it may be inappropriate physiologically but certainly from the standpoint of when consultants tend to show up). Also, this study makes no attempt to quantify or even mention the analgesic effects of reglan, IV fluids, or tylenol (which of these patients had a fever?).

A quarter of the patients (in both groups) had a final diagnosis of “nonspecific abdominal pain.” One third of the discordance was assigning this diagnosis to problems that in fact were more serious.

The authors note, in conclusion, that abdominal pain is the #1 complaint seen in ED’s, accounting for 7 million visits a year. Surveys suggest that EP’s still defer to Sir Cope’s dictum and withhold pain meds. But in the author’s admittedly non-rigorous review of seven trials of whether opioids influence diagnosis, plus theirs, they found six favor “no difference in accuracy” and one actually favoring morphine as improving accuracy. This holds true in pediatric cases (two small studies) and I’m sure the resident who lectured on this topic in conference last year has some more references. It’s worth noting that even Cope’s updated guide to the acute abdomen now states  ”there’s no evidence supporting the contention that morphine use is deleterious in any way.”

The accompanying commentary by Knopp and Dries notes that there are some patients who probably do have an alteration in their physical exam after getting pain meds, who might do better without. Anecdotal cases abound, for sure. But the controversy is not “give or withhold,” it’s what’s most effective? Others have studied patients whose RLQ or RUQ pain completely resolves with analgesia, and no missed appys or choles resulted. So, for now, prompt pain relief that permits effective diagnosis and care is the goal.

Posted on Thursday, October 26th, 2006 at 2:05 am by Nick. Filed under Pain Management, Journal Club.
You may post a comment.