It might be that the information on left ventricular end-diastolic pressure doesn't help clinicians after all. Or it might be that the side effects narrowly outweigh the benefits.
A recent conversation with an intensivist illustrated the point:
Fellow: "What's the #1 complication due to Swans?"
Fellow: "That happens, but it's not #1."
Me: "Piercing the heart and causing tamponade!"
Fellow: "That happens, but not #1."
Me: "Infection! Thrombus formation!"
Fellow: "No, no."
Fellow: "It's actually misinterpretation of the data. Overwedging, or mistaking PAP for LVEDP. It leads to errors in fluid management."
Which made this exchange on rounds, few days later, that much more clear to me:
Intensivist: "Why haven't we pulled the Swan from this patient? It's not telling us anything anymore, and it's just going to lead to problems. We need it out."
Resident: "But the surgeons want it in."
Intensivist: "OK, fine. It's their patient. But we can still protect against complications."
Intensivist: "Turn of the monitor readings from the Swan. This way, it stays in the patient, but can't cause misinterpretations and bad management."
Everybody was happy, and we moved on to the next bed.