First-year PCCM fellow — I missed early shock physiology on nights and can’t stop replaying it.
Just need to vent a bit about a recent ICU case that has been weighing on me.
I took signout from the day fellow on a late ICU transfer from the hospitalist service. The reported issues were pancreatitis, EtOH withdrawal, and euglycemic DKA. The day team was still working on admission orders, and the note when I came on. Initial labs were notable for K 7.9 and lactate 9, which were both downtrending. He was started on an insulin infusion and some fluids. At the time, he was not overtly in shock and was maintaining his BP.
As the night went on, his potassium only minimally improved — from about 6 pm to 3 am, it stayed around 7.5. He had made only ~50 cc of urine during that time (200cc about 12 hours earlier). At that point, it seemed increasingly likely he would need CRRT. I tried a furosemide stress test near the end of my shift with 160 mg given what appeared to be intrinsic AKI, but it failed, which, honestly, was not surprising.
When I signed out in the morning, I told the day team that nephrology needed to be called first thing for dialysis planning. About an hour later, he was intubated for worsening lethargy and started on pressors. CRRT was eventually started, though not until later that afternoon.
When I came back the following night, he was profoundly ill.
The thing bothering me is that I did not more aggressively interpret the hyperkalemia and lactate as early shock physiology. I anchored too heavily on the initial signout diagnosis of DKA/EtOH withdrawal/pancreatitis and attributed a lot of the metabolic derangements to that. In retrospect, his beta-hydroxybutyrate was barely elevated, and the persistent severe hyperkalemia plus worsening acidosis should have forced me to step back and reconsider the whole picture much earlier.
My attending told me this was not some catastrophic miss like ignoring someone actively peri-intubation on HFNC, but said it was “borderline,” which honestly felt fair.
Looking back, the biggest issue was cognitive offloading. I trusted the rushed signout too much and mentally categorized him early, which made me focus my energy elsewhere in the unit instead of continually reassessing why this patient kept getting worse. The warning signs were there. I just did not synthesize them appropriately enough in real time.
Thankfully, the patient is now turning the corner on day 3. But this case has been a pretty painful reminder that shock physiology can declare itself metabolically long before the blood pressure drops, and that severe persistent hyperkalemia/lactic acidosis deserves a much broader differential than “it’s just DKA.” I don't think I have had this bad of a miss since the start of fellowship, and I wonder if I had gotten this patient as a straight ER call for me to evaluate without any previous anchoring, I would have done things differently.