Latest / Interprofessional Critical Care Network (ICCN) Podcast / Every liter is a chloride dose: saline versus balanced crystalloids in the ICU
Transcript
- 0:00SPEAKER_01: A liter of 0.9% saline carries 154 millimoles of chloride.
- 0:05SPEAKER_01: A liter of plasmalite 148 carries 98 millimoles.
- 0:09SPEAKER_01: Inside a large randomized trial of those two fluids,
- 0:12SPEAKER_01: serum chloride rose about 2 millimoles per liter more with saline.
- 0:16SPEAKER_01: Real, but modest.
- 0:18SPEAKER_01: And kidney function shaped it.
- 0:20SPEAKER_01: This is the ICCN Daily Brief for Thursday, November 12, 2026.
- 0:25SPEAKER_00: If you follow the show, each morning's brief is waiting for you.
- 0:28SPEAKER_00: Today's article treats every liter of saline as a chloride dose and asks how much
- 0:33SPEAKER_00: of the morning chemistry panel the fluid order ever controlled.
- 0:37SPEAKER_01: Where do the numbers come from?
- 0:42SPEAKER_00: A secondary analysis of the BASICS trial by Zampieri and colleagues, published in 2026.
- 0:48SPEAKER_00: BASICS was a double-blind,
- 0:50SPEAKER_00: randomized trial in 75 Brazilian ICUs between 2017 and 2020.
- 0:56SPEAKER_00: More than 10,000 adults with a risk factor for acute kidney injury were randomized
- 1:01SPEAKER_00: to Plasmalite 148 or 0.9% saline, supplied in identical 500-milliliter bags.
- 1:09SPEAKER_01: And this analysis used only part of that trial.
- 1:12SPEAKER_00: Yes. It looked only at days one to three, when most study fluid was given,
- 1:17SPEAKER_00: and included the 3,482 participants, 33% of the trial,
- 1:22SPEAKER_00: who had a baseline serum chloride and at least one value in that window.
- 1:26SPEAKER_00: Chloride was not a protocol measurement, so it was drawn only when clinicians ordered it.
- 1:31SPEAKER_01: And there are three analyses.
- 1:34SPEAKER_00: First, a randomized comparison of the change in serum chloride between arms.
- 1:38SPEAKER_00: It carries causal weight, but for the assigned fluid as a whole,
- 1:42SPEAKER_00: since saline and plasma light differ in more than chloride.
- 1:45SPEAKER_00: Second, that difference per unit of chloride dose,
- 1:49SPEAKER_00: plus a look within the saline arm only, where volume was the clinician's choice.
- 1:53SPEAKER_00: That estimate is observational.
- 1:55SPEAKER_00: Third, a descriptive model that split the variation in serum chloride into its sources.
- 2:01SPEAKER_00: Give me the numbers.
- 2:02SPEAKER_00: By day two, among 2,721 patients with paired values,
- 2:08SPEAKER_00: serum chloride rose 2.15 millimoles per liter more with saline than with plasmolyte,
- 2:14SPEAKER_00: with a 95% confidence interval from 1.76 to 2.53.
- 2:20SPEAKER_00: By then, saline recipients had received 145.7 millimoles more study fluid chloride
- 2:26SPEAKER_00: in Yet average study fluid volume differed by only 27 milliliters between arms.
- 2:32SPEAKER_00: And per unit of dose?
- 2:34SPEAKER_00: Dividing the chloride rise
- 2:36SPEAKER_00: by the chloride dose gave 1.47 millimoles per liter per additional 100 millimoles,
- 2:42SPEAKER_00: with a confidence interval from 1.17 to 1.81.
- 2:47SPEAKER_00: Within the saline arm, the observational slope was far flatter.
- 2:51SPEAKER_00: Each extra 100 mmol was associated with only about 0.31 mmol per liter more rise.
- 2:59SPEAKER_00: The article calls that roughly a fifth of the randomized estimate.
- 3:03SPEAKER_00: Where did kidney function come in?
- 3:05SPEAKER_00: It changed the size of the gap.
- 3:07SPEAKER_00: At an EGFR of 30, the model difference between the fluids on day two was
- 3:12SPEAKER_00: about 3 millimoles per liter.
- 3:15SPEAKER_00: At an EGFR of 90, it was about 1.8 millimoles per liter.
- 3:20SPEAKER_00: The sensitivity model using creatinine pointed the same way,
- 3:23SPEAKER_00: but its interval included no difference.
- 3:26SPEAKER_00: How much did the fluid explain?
- 3:28SPEAKER_00: Very little.
- 3:29SPEAKER_00: Fluid accounted for 0.7% of the modeled variation
- 3:33SPEAKER_00: in serum chloride and 1.8%
- 3:36SPEAKER_00: when the randomized arm was added.
- 3:38SPEAKER_00: The patient's starting chloride alone accounted for about 47%.
- 3:44SPEAKER_00: Persistent differences between patients
- 3:46SPEAKER_00: that measured characteristics could not explain accounted for about 25%
- 3:52SPEAKER_00: and day-to-day fluctuation for about 27%.
- 3:55SPEAKER_00: What does this not show?
- 3:58SPEAKER_00: Outcomes. This analysis measured biochemistry over three days.
- 4:02SPEAKER_00: It did not test kidney outcomes or survival,
- 4:05SPEAKER_00: so it cannot tell us that a two millimoles per liter shift is clinically harmless
- 4:10SPEAKER_00: or harmful.
- 4:11SPEAKER_00: Nor can it tell us that chloride explains any benefit of balanced fluids.
- 4:16SPEAKER_00: The randomized effect belongs to the fluid as a whole.
- 4:19SPEAKER_01: Fluid explained under 1% of the variation.
- 4:22SPEAKER_01: Does that make the effect trivial?
- 4:24SPEAKER_00: The article says no.
- 4:25SPEAKER_00: A variance share describes how much of the scatter between patients a factor accounts for.
- 4:31SPEAKER_00: A randomized difference describes what happens when you change that factor for everyone.
- 4:36SPEAKER_01: Who is missing from the data?
- 4:38SPEAKER_00: About two-thirds of the trial.
- 4:40SPEAKER_00: Chloride was drawn in only a third of participants at clinician's discretion,
- 4:44SPEAKER_00: and the patients who were tested may differ from those who were not.
- 4:48SPEAKER_00: Just over half of the analytic cohort came to the ICU after planned surgery,
- 4:53SPEAKER_00: and all sites were in Brazil.
- 4:55SPEAKER_00: And the studies on hidden chloride?
- 4:57SPEAKER_00: Observational.
- 4:58SPEAKER_00: In a prospective study of 1,633 critically ill adults by Zhou and colleagues,
- 5:04SPEAKER_00: fluid creep, meaning drug doluents, flushes, and carrier infusions,
- 5:09SPEAKER_00: made up 48.4% of daily fluid volume.
- 5:13SPEAKER_00: Higher creep volumes were associated with more major adverse kidney events
- 5:17SPEAKER_00: at 30 days, which is not proof of harm.
- 5:20SPEAKER_00: And the two dolent switch studies the article cites compared eras,
- 5:23SPEAKER_00: and neither proves a mortality effect.
- 5:26SPEAKER_00: Bedside, what does the team do with this?
- 5:29SPEAKER_00: Do the chloride arithmetic once.
- 5:32SPEAKER_00: Swapping one liter of plasmalite 148 for one liter of saline adds 56 millimoles of chloride.
- 5:39SPEAKER_00: Using the BASICS randomized ratio,
- 5:41SPEAKER_00: the article works that out to an average rise of a little
- 5:44SPEAKER_00: under one millimoles per liter.
- 5:46SPEAKER_00: That is arithmetic from a cohort average, not a prediction for any individual patient.
- 5:51SPEAKER_00: And it will run larger when kidney function is reduced.
- 5:54SPEAKER_00: Then read the baseline chloride and the EGFR before the fluid choice.
- 5:59SPEAKER_00: Pharmacists and nurses?
- 6:01SPEAKER_00: Pharmacists audit the hidden load.
- 6:03SPEAKER_00: Count chloride from diluents, carrier lines, and flushes, not just boluses.
- 6:08SPEAKER_00: Review which drugs can be compatibly diluted in glucose 5% or a balanced solution.
- 6:14SPEAKER_00: And pair any default chains with a plan for glucose monitoring and insulin adjustment.
- 6:19SPEAKER_00: Nurses make creep visible.
- 6:21SPEAKER_00: Chart flushes and carrier volumes, which often go unrecorded.
- 6:25SPEAKER_00: And reading the result when it comes back?
- 6:28SPEAKER_00: Trend, do not react to one value.
- 6:30SPEAKER_00: A 2 to 3 millimole per liter change from yesterday may be noise.
- 6:34SPEAKER_00: Respiratory therapists, when a ventilated patient's bicarbonate drifts down
- 6:38SPEAKER_00: without another explanation,
- 6:40SPEAKER_00: flag the chloride trend and fluid history on rounds
- 6:43SPEAKER_00: before chasing minute ventilation.
- 6:45SPEAKER_00: Dietitians, count parenteral nutrition and electrolyte replacement in the Daily Review.
- 6:50SPEAKER_00: One boundary. Fluid choice in traumatic brain injury and other settings
- 6:54SPEAKER_00: where tonicity matters follow separate evidence and local protocols.
- 6:58SPEAKER_00: The teaching pearl.
- 7:00SPEAKER_00: Saline's chloride is a dose, the kidney sets its clearance,
- 7:04SPEAKER_00: and the patient's starting value predicts most of what you will see.
- 7:09SPEAKER_00: Order fluids as if all three were true.
- 7:12SPEAKER_00: What changed?
- 7:13SPEAKER_00: Saline-raised serum chloride by about 2 millimoles per liter more than plasmalite,
- 7:18SPEAKER_00: roughly 1.5 millimoles per liter per 100 millimoles,
- 7:23SPEAKER_00: with a larger gap when kidney function was reduced.
- 7:26SPEAKER_00: That is a real, randomized effect.
- 7:29SPEAKER_00: And it is small beside the patient's own starting chloride and day-to-day variability.
- 7:34SPEAKER_00: Treat chloride as a dose with a clearance, count the hidden sources,
- 7:38SPEAKER_00: and resist reading outcome claims into a biochemical result.
- 7:43SPEAKER_01: The full article with the bedside takeaways and the six things not to overassume is
- 7:47SPEAKER_01: on the ICCN substack.
- 7:49SPEAKER_01: The link is in the show notes.
- 7:51SPEAKER_01: ICCN publishes one article a day for ICU clinicians, $8 a month or $80 for the year,
- 7:57SPEAKER_01: with a seven-day free trial.
- 7:59SPEAKER_01: This has been the ICCN Daily Brief, back tomorrow at 6.30.