IV Fluids
Fluid is a drug. It has a dose, a route, an indication and a toxicity, and most people write for it as though it were tap water.
Three questions before you write the order
If you cannot answer all three, you are not ready to write the order.
Not "is the patient in hospital," and not "the order set has a line for it." Is there a deficit, an ongoing loss, or a perfusion problem that volume will fix? A patient eating and drinking needs nothing.
Driven by what you are correcting: intravascular volume, free water, or an electrolyte. Three problems, three bags.
"Normal saline at 100 an hour" with no stop date is how patients end up 9 liters positive on day four. Write a volume and an endpoint, or a rate and a time you will look again.
Resuscitation, maintenance and replacement are not the same thing
- Resuscitation fills an empty tank to restore perfusion. Given fast, in defined boluses, to a physiologic endpoint. Typically 250 to 500 mL of a balanced crystalloid over 15 minutes, then reassess.
- Maintenance covers obligate daily water and electrolyte turnover in someone who cannot drink. Slow, continuous, small.
- Replacement matches an identifiable ongoing loss: nasogastric output, an ileostomy, a drain, diarrhea, a burn. Roughly volume for volume, with something resembling what is being lost.
Conflate them and you get the classic disaster: the patient who needed a 500 mL bolus is put on maintenance saline at 125 an hour, is still underfilled at hour two, and is 6 liters positive by day three with the same blood pressure and new hypoxia. You gave a lot of fluid and never resuscitated anyone.
Where the fluid actually goes
Total body water is roughly 60 percent of body weight in an adult male and 50 percent in an adult female, less in the elderly and in obesity. Two thirds of it is intracellular, one third extracellular, and of that extracellular third about a quarter is plasma. In a 70 kg man: 42 L total, 28 L inside cells, 14 L outside, of which 3.5 L is plasma.
Isotonic saline fills the extracellular space and does not enter cells. Roughly a quarter of a liter, about 250 mL, stays intravascular once it equilibrates. The other 750 mL becomes interstitial edema.
Dextrose in water is worse. Once the dextrose is metabolized you are giving free water, which distributes across total body water, and under 100 mL per liter ends up in the plasma. D5W is not a resuscitation fluid. If you are bolusing it for hypotension, stop.
The crystalloids compared
| Fluid | Na | Cl | K | Buffer | Osmolarity | Typical use |
|---|---|---|---|---|---|---|
| Normal saline (0.9 percent) | 154 | 154 | 0 | None | 308 | Hypochloremic alkalosis, hyponatremia, head injury |
| Lactated Ringer's | 130 | 109 | 4 | Lactate 28 (also Ca 2.7) | 273 | Resuscitation, trauma, surgery, pancreatitis |
| Plasma-Lyte | 140 | 98 | 5 | Acetate 27, gluconate 23 (Mg 3) | 294 | Balanced resuscitation without calcium |
| D5W | 0 | 0 | 0 | None | 252 | Free water, drug diluent, hypoglycemia |
| Half normal saline (0.45 percent) | 77 | 77 | 0 | None | 154 | Hypernatremia with volume depletion |
Values vary by manufacturer. Lactated Ringer's contains calcium, which is why it stays out of a line running citrated blood. Its lactate is a buffer substrate, not a marker of anaerobic metabolism: it is metabolized to bicarbonate, does not meaningfully raise a lactate you are trending, and is not contraindicated in sepsis.
Why large volume saline causes an acidosis
Plasma chloride sits around 100. Saline delivers it at 154. Give several liters and the plasma chloride rises well above physiologic while the bicarbonate falls to compensate. The result is a hyperchloremic, normal anion gap metabolic acidosis.
This matters practically. The patient you resuscitated with 5 liters of saline now has a bicarbonate of 17, and everybody starts hunting for sepsis or ischemia. Calculate the gap. If it is normal and the chloride is 116, you caused it. Use the anion gap calculator before you order a toxicology screen.
High chloride loads also reduce renal cortical perfusion through tubuloglomerular feedback, which is the argument behind the next section.
Balanced crystalloid versus saline
The SMART trial in critically ill adults and the SALT-ED trial in emergency department patients both compared balanced crystalloids against saline. Both found a modest advantage for balanced solutions in major adverse kidney events, meaning death, new renal replacement therapy or persistent renal dysfunction. The absolute difference was small, roughly a percentage point, but consistent across very large numbers of patients.
Read that as a sensible default, not a mandate. Absent a reason to pick saline, reach for a balanced crystalloid. Saline is still right in three situations:
The vomiting patient, or the heavily diuresed one with a chloride of 88 and a bicarbonate of 38. Here the chloride load is the treatment, not the side effect.
Saline carries a sodium of 154 against Lactated Ringer's 130. If you are raising a sodium of 124, the higher sodium fluid is the one you want, at a rate that respects the correction limit.
Lactated Ringer's has an osmolarity around 273 and is relatively hypotonic to plasma. In a brain that has lost autoregulation, free water is the enemy. Use saline, or hypertonic saline where indicated.
Albumin and colloids
The theory was that oncotically active molecules stay intravascular and expand plasma volume more efficiently per liter. It holds better in a healthy endothelium than a septic one, where capillary leak lets albumin into the interstitium and it takes water with it.
Albumin has narrow but real indications: large volume paracentesis, spontaneous bacterial peritonitis and hepatorenal syndrome. It is also a second line expander in the patient still underfilled after a large crystalloid volume. It is expensive and not superior to crystalloid for routine resuscitation.
Starches are essentially abandoned in critical illness because of consistent signals of increased renal replacement therapy, coagulopathy and mortality. Finding one on a formulary is not an invitation.
One caveat overriding all of the above: albumin performed poorly in traumatic brain injury. Do not use it there.
Maintenance fluid and the 4-2-1 rule
The 4-2-1 rule gives an hourly rate: 4 mL/kg/hour for the first 10 kg, 2 for the next 10 kg, 1 for every kilogram above 20.
Worked through on an 80 kg adult
First 10 kg: 10 x 4 = 40 mL/hour
Next 10 kg: 10 x 2 = 20 mL/hour
Remaining 60 kg: 60 x 1 = 60 mL/hour
Total: 120 mL/hour, about 2.9 liters a day.
Now be suspicious of that number. Nearly 3 liters a day is a lot for someone lying in a bed. The rule came out of pediatric anesthesia and overestimates the needs of most hospitalized adults, particularly the elderly, the septic, and anyone whose stress response is driving antidiuretic hormone.
If the patient is eating and drinking, stop the fluid. If they will be eating tomorrow, they probably do not need it tonight. A saline lock and an oral diet is a legitimate order.
A sick inpatient has high antidiuretic hormone from pain, nausea, surgery, opiates or the illness itself, and cannot excrete free water normally. Run D5 half normal at 100 an hour into that patient and the sodium falls, sometimes dangerously, sometimes with seizures. If you are giving maintenance fluid, use an isotonic solution and check a sodium the next day.
Will this patient actually respond to fluid
Static numbers are poor predictors. A central venous pressure of 6 does not tell you a bolus will help, and a 14 does not tell you it will not. Dynamic measures are better: pulse pressure variation, stroke volume variation, a passive leg raise, a collapsing inferior vena cava on ultrasound. The full treatment, including what invalidates pulse pressure variation, is in the fluid responsiveness section of the lines page.
One point bears repeating: fluid responsive is not the same as needing fluid. About half of healthy people are responsive at any moment. The question is whether raising cardiac output helps this patient, and where the volume ends up.
The harms of too much fluid
Positive fluid balance is an independent marker of worse outcomes in critical illness. Some of that is confounding. Some is causal, and the mechanisms are not subtle.
- Pulmonary edema, and with it more hypoxia, more time on the ventilator and delayed extubation.
- Gut edema, with ileus, feeding intolerance and anastomotic leak in the surgical patient.
- Abdominal compartment syndrome, underdiagnosed after massive resuscitation. Rising airway pressures, falling urine output, tense abdomen. Measure a bladder pressure.
- Renal congestion. The kidney is an encapsulated organ draining into a congested venous system, so raising central venous pressure worsens renal function. Fluid overload makes acute kidney injury worse, not better.
Once the shock has resolved and the pressors are off, taking fluid back off is an active plan, not an afterthought. Stop the maintenance fluid, concentrate the drips, feed enterally, and diurese to a daily negative target. Write it as its own line. Patients who never get deresuscitated stay wet, stay hypoxic and stay in the unit.
Two worked plans
Septic patient, 75 kg, lactate 4.2, blood pressure 84/50
Phase one. Lactated Ringer's 500 mL over 15 minutes, then reassess perfusion, pressure and mental status. Repeat toward roughly 30 mL/kg over the first few hours while the patient keeps responding, deciding each bolus on a dynamic measure rather than running to a fixed number in someone with heart failure or on dialysis. Antibiotics do not wait for the fluid.
Phase two. If the pressure does not hold after adequate volume, start norepinephrine rather than giving a sixth liter. Vasoplegia is not treated with saline.
Phase three. Once stable, stop the fluid. No maintenance. Recheck a chemistry, look at the chloride and the gap, and plan the negative balance.
Heart failure patient, ejection fraction 25 percent, admitted with dyspnea and edema
Fluid order: none. Saline lock the line. This patient's problem is too much volume in the wrong place, not too little.
Intravenous furosemide dosed off their home dose, strict intake and output, daily weights, an explicit net negative target, and sodium and fluid restriction.
If they become hypotensive during diuresis, ask whether this is a low cardiac output state before reflexively bolusing. Volume into a congested failing heart makes the pressure worse and the lungs wetter. The exception is the genuinely over-diuresed patient, which is why you look at the whole picture.
Related: common night calls, and arterial lines and CVP for measuring what you are doing.
This is not medical advice. It is a teaching outline for clinicians and clinicians in training. The fluid choices, rates and volumes below are illustrative. Follow your institution's protocols, verify everything against a current reference, and use your own judgment.