When you get a call that your patient just spiked a fever there's an immediate disconnect. The nurse wants you to “order something for it” (meaning a medication such as Tylenol to bring it down) while you are immediately thinking etiology (Has the patient acquired an infection? Do I need to get cultures and an xray stat?). The nurse wants an order for an antipyretic. That's not at the top of your list. You may even feel you're being maneuvered into ordering something that could be harmful. That's when you need to be armed with the evidence.
I while back I reviewed this topic in a couple of posts  . More recently Mill Hill Ave Command addressed the topic and I thought it would be a good time to revisit it here. The general conclusion of the Mill Hill Command post was that lowering of body temperature with antipyretics or via external means is not supported by evidence. I agree although as I pointed out in my previous posts the exceptions must be kept in mind: hyperthermia syndromes and heat stroke, acute cerebrovascular disease and coma post cardiac arrest. As I also pointed out there was one study showing benefit from external cooling (not antipyretics!) in patients with septic shock on pressors, in terms of decreased pressor requirement and decreased mortality. That is just one study with methodologic concerns and the findings need to be replicated.
There may be other exceptions where physiologic rationale and clinical judgment suggest the use of antipyretics, e.g. fever driving tachycardia in a patient with heart failure or myocardial ischemia, though not supported by high level evidence.
Finally, since my original posts there has been a point-counterpoint with rebuttals published in Chest. It's an interesting exchange that probes the literature on both sides. I think the naysayers won the debate.