Every Bag Is a Question — Fluids on Autopilot

The attending stops at the foot of the bed. "What are you running on her?"

You glance at the pump. LR at 125. You say it. The attending waits.

You realize the question wasn't what. It was why.

The patient is POD 2 from a small bowel resection. The new ileostomy has put out 1.4 liters since yesterday. You pull up the morning labs. Sodium 132. Potassium 3.2. Creatinine bumped two-tenths. Urine output thin overnight. You wrote the fluid order on admission and never went back to look.

Trainees treat IV fluids like a default setting. Click the order set, ride the default the whole admission, only think about fluids when something goes sideways. By then you're not prescribing — you're reacting.

The real issue isn't which bag to hang. It's that every bag is answering a question, and you stopped asking it three days ago.

There are only three reasons to give IV fluid: resuscitation, maintenance, or replacement. Most intern errors come from doing the wrong one — usually maintenance when the patient needs replacement, or maintenance long after the patient started eating.

Resuscitation is the easy one. Hypotension, tachycardia, lactate climbing, urine output trailing. Bolus of balanced crystalloid, twenty to thirty mL/kg, then reassess. Not "open it up, check back in four hours." Reassess. Then decide whether the next liter helps or hurts.

A bolus is a question, not a setting on the pump. If you don't reassess, you're not resuscitating — you're just adding water.

A word on the bag. LR is the workhorse for surgical patients — sodium 130, a little potassium and calcium, a buffer that becomes bicarbonate in the liver. Normal saline runs a chloride of 154, higher than any human has ever had, and too much of it earns you a hyperchloremic acidosis you have to explain on rounds. NS has a role — hypochloremic alkalosis from NG losses, hyponatremia, hyperkalemia, head injury. Make NS earn the order.

Maintenance is for the patient who can't drink yet — NPO, ileus, intubated. Replace what they would've taken in by mouth: free water, a little sodium and potassium, a little glucose. D5 ½NS with 20 of KCl. The moment they start eating, back it down.

The 4-2-1 rule gives you a maintenance floor — four mL/kg/hr for the first ten kilos, two for the next ten, one for everything above. That number assumes your patient is sitting quietly with no NGT, no fever, no ostomy putting out 800 mL since midnight. That patient doesn't exist on your surgical floor.

Maintenance fluid running three days into a regular diet is how a 70 kg patient becomes a 76 kg patient with crackles and a sodium of 128.

Replacement is the bucket trainees miss. Anything leaving the body that isn't urine — NG suction, high-output ostomy, biliary drain, pancreatic fistula, secretory diarrhea — is on maintenance plus whatever's going out the tube. What's leaving tells you what to put back.

Gastric losses through an NG pull chloride, hydrogen, sodium, and potassium. Leave them unreplaced and the patient drifts into a hypokalemic, hypochloremic alkalosis — which is the K of 3.2 and the chloride that quietly dropped. Replace mL for mL with NS plus 20 of KCl. One of the few times saline earns the order.

Lower GI losses — high-output ileostomy, secretory diarrhea, biliary and pancreatic drains — go the other direction. Bicarbonate-rich, potassium-rich, headed toward a non-anion-gap acidosis with a rising creatinine that everyone will call "AKI" before anyone looks at the ostomy bag. Replace with LR. Check a magnesium and phosphorus when volume is high — nobody draws them until the patient is symptomatic.

NG losses pull the patient alkalotic. Ostomy and diarrhea pull the patient acidotic. Match the loss to the fluid that fixes the direction the patient is heading.

A 200 mL/hr ileostomy is not a maintenance problem. It's a 4.8 liter problem over 24 hours. Pull the I/Os when you write the order. Not the order set. The numbers.

I see this scenario on rounds all the time. POD 3 patient on LR at 125 since the OR. The ostomy is quiet day one, then opens up. Usually someone catches it. Occasionally nobody does — by the third morning she's a kilogram and a half up, lung bases wet, puffy fingers, with tight rings and sats drifting from 96 to 91. The order was the problem. Not the surgery. Not the patient. The order. She gets diuresed and she's fine, but that order got written on autopilot, and somebody paid for it in oxygen.

The labs are a late sign. The output sheet is the early warning. The kidney doesn't text you back — the lab is a slower copy of what your hands and eyes already saw.

Before rounds, give yourself a minute per patient. Does the patient need IVFs at all? Volume status. Can they drink. What are they losing. What's the potassium doing. What's the kidney telling you.

Then say it out loud — even if only to yourself in the workroom: "This patient is NPO, made 600 mL of urine overnight, had 1.1 liters from the NG tube, potassium is 3.1, chloride is down, creatinine is up slightly. I'd continue isotonic fluid, replace the NG losses with NS plus KCl, replete the potassium and magnesium, and repeat labs this afternoon."

That's not perfect. It's coherent. Coherent gets taught. Random gets corrected.

The patient who looks puffy on day three didn't get there from one bad order. He got there from twelve good-enough orders nobody went back to read.

---

What to Read Next: Post 14 — Why No Bowel Movement Yet? The team will ask every morning. The answer they want isn't a number — it's whether the gut is waking up.