One thing I love about Emergency Medicine is that no two cases of the same disease are ever exactly alike. Yes this can be anxiety provoking, but I prefer to think of it as akin to great jazz tunes – it nostalgically reminds you of other versions of a song you’ve heard before, but the players and the riffs are distinctly unique.
A case in point, the hot joint has presented some interesting challenges to me on a number of cases. Most recently, a thirty something intravenous drug user with several days of increasing pain, redness, swelling of her left shoulder appeared for all the world to be a septic joint. She had a great history, classic presentation, and initial labs showed and white count, ESR, and CRP through the roof. Now I love doing taps, but in our ED we involve orthopedics for many of them, particularly the shoulders. The talented orthopedic resident to my surprise was not interested in tapping it.
My initial response to him was that no matter what you tell me the pretest probability that this IV drug user has a septic joint is so high, no test other than an arthrocentesis is going to satisfy me. But the orthopedic resident had some interesting and valid concerns, and the more I thought about it the more I recognized that there were some interesting clues along the way that led away from a septic joint:
1. The pain began after direct inoculation with a needle into the shoulder (rather than by hematogenous spread of bacteria to the joint from an intravenous needle) so it seemed unlikely to me that the patient had jammed a needle all the way into her glenohumeral joint.
2. The patient really didn’t want to move the shoulder at all, and was guarding it gingerly, but in fact with gentle passive range of motion there was a few degrees of flexion and extension, in contrast the patient did not want to abduct at all, offering the possibility of an infected subacromial bursitis
The orthopedic resident was concerned that putting a needle into the joint through an infected abscess or bursitis would potentially seed a sterile joint with bacteria and make things worse, so we agreed on a quick initial ultrasound. This was read by radiology as a septic joint, with increased joint space fluid and surrounding reactive hyper-vascularity.
At this point we tapped the joint. Our posterior approach was a dry tap. Frustrating as this was, it was clear we were in the joint space, but there was nothing.
So what now. If this was a deltoid abscess, having orthopedics open up the joint would be a mistake. An MRI would have been nice but she actually had an old needles embedded in her shoulder from prior injections so that made radiology put their foot down on that one…
So a CT of the shoulder was done which again was read as a septic joint with fluid around the joint space. Remarkably the joint space itself was well-preserved on my read of the CT, which seemed odd to me. My only thought at the time was that the direct inoculation of the joint had made a tract anteriorly through which the pus was draining and surrounding the joint capsule externally.
Ultimately orthopedics took her to the OR. There they found a septic bursitis that had ruptured anteriorly and surrounded the joint capsule with pus and fluid. The integrity of the joint itself was well maintained. Ultimately, the patient did very well and went home several days after admission on antibiotics.
The take home points for me.
- Think about abscess or infected bursitis before sticking a needle into a joint. You could make things worse if you plunge a needle through infected tissue into a sterile joint space.
- Any inflammatory markers in this situation are utterly useless.
- Imaging can be falsely positive and may again lead to attempted arthrocentesis.
- Complex infections around a joint are still best served on orthopedics. There was some discussion about general surgery involvement for abscess drainage, but given the high likelihood of joint involvement in an equivocal case like this orthopedics is better equipped to debride and wash out around joint structures.
Until recently most of my thinking about red-hot and swollen joints are “what fun I get to tap it” and second I think, “when do I get to tap it”. This general teaching holds true for most cases, but I have recently been humbled by missed taps, indeterminate taps which turn out to be infectious not inflammatory, and the reverse, cases that got unnecessary wash-outs. So my belief that the hot joint is the last bastion of simple diagnostic procedures in the ED has been finally crushed. There you are, you either love jazz or you don’t…
Miles Davis said “Don’t just play what’s there, play what’s not there.” Meaning, know what to leave out. So: where did this the idea of getting an ESR AND a CRP come from? How can those tests help your patient in an acute setting, when you already know that the patient is in an inflammatory state? And, most definitely, not all hot joints need to be tapped: I have seen mildly edematous but definitely not infected knees and shoulders get very septic after an overzealous ER doc put a needle in. There can be wisdom in forbearance! That said, thanks for sharing an interesting case.
Thanks, love the Miles Davis quote, and thanks for all your comments. Good question about the ESR/CRP. I’ve reviewed the literature on this before, but the most recent paper I’ve seen is an evidence based review in Acad Emerg Med. 2011 Sep;18(9):1011. Where they have no utility essentially. Answering why I use them sometimes is a too long for this forum, but I’ll leave it simply by saying EBM has its limits, as do tests, and I don’t always agree with findings from evidence based reviews.
As for overzealous ED docs, I was being a little flippant at the end of the post, I should say that while I love procedures, there are many joints I don’t tap so your point about forbearance is well taken. My poorly made point was more that my fallback until now had always been if there is any doubt, tap. Since missing a septic joint is never good. This case brought some of that thinking into doubt for me. Thanks again and keep the comments coming!