Latest / Interprofessional Critical Care Network (ICCN) Podcast / The clotting score already in your labs
Transcript
- 0:02SPEAKER_01: Day 2 of septic shock.
- 0:04SPEAKER_01: The platelet count has slid from 210 to 118.
- 0:07SPEAKER_01: The INR reads 1.5, and norepinephrine is still running.
- 0:11SPEAKER_01: The evening dose of subcutaneous heparin is due. Give it? Hold it?
- 0:15SPEAKER_01: Or does this patient need more?
- 0:17SPEAKER_01: This is the ICCN Daily Brief for Monday, November 2, 2026.
- 0:21SPEAKER_00: And every value that question turns on is already in the chart.
- 0:25SPEAKER_00: Today's article is about a clotting score built from labs you already drew and why
- 0:30SPEAKER_00: the heparin part of the story is still a hypothesis.
- 0:33SPEAKER_01: Start with the score itself.
- 0:35SPEAKER_01: What is sepsis-induced coagulopathy?
- 0:38SPEAKER_00: Sepsis switches on coagulation in nearly every patient.
- 0:41SPEAKER_00: The SIC criteria were built to catch the earlier, still compensated phase before overt DIC.
- 0:48SPEAKER_00: They were introduced in 2017, adopted by the ISTH two years later,
- 0:53SPEAKER_00: and they use only a platelet count, an INR, and the SOFA organ score.
- 0:58SPEAKER_00: So what did this new analysis do with it?
- 1:00SPEAKER_00: Tsao and colleagues built their main cohort from MIMIC4.
- 1:03SPEAKER_00: ICU stays at a single Boston academic center from 2008 to 2022
- 1:09SPEAKER_00: and tested what they found in eICU,
- 1:11SPEAKER_00: a database that pools many United States hospitals.
- 1:15SPEAKER_00: Adults aged 18 to 90 met sepsis-3 criteria within two days of reaching the ICU,
- 1:21SPEAKER_00: stayed at least 48 hours, and had at least two days with a calculable score.
- 1:26SPEAKER_00: That left 20,168 MIMIC-4 patients and 5,829 from eICU.
- 1:33SPEAKER_01: And they scored it every day.
- 1:35SPEAKER_00: For each of the first seven ICU days, they took the highest SIC score of the day.
- 1:40SPEAKER_00: And a latent class model sorted patients by the shape of that score over the week.
- 1:44SPEAKER_00: The main endpoint was death in hospital up to day 28.
- 1:49SPEAKER_01: Where does heparin come in?
- 1:51SPEAKER_00: As an exploratory analysis with a landmark design,
- 1:54SPEAKER_00: exposure meant subcutaneous prophylactic unfractionated heparin
- 1:59SPEAKER_00: in the first 48 hours.
- 2:01SPEAKER_00: And follow-up started at hour 48.
- 2:03SPEAKER_00: So the clock ran only after exposure was counted.
- 2:07SPEAKER_00: That left 16,860 MIMIC for patients
- 2:11SPEAKER_00: and 4,595 in EICU,
- 2:14SPEAKER_00: with overlap weighting to balance more than 20 measured variables.
- 2:18SPEAKER_00: Give me the shapes first.
- 2:20SPEAKER_00: Three courses fit best.
- 2:22SPEAKER_00: In Mimic 4, 31.9% of patients stayed persistently high across the week,
- 2:30SPEAKER_00: 10.6% rose to a peak around days 3 to 4 and then fell,
- 2:34SPEAKER_00: and 57.5% resolved from an intermediate start.
- 2:38SPEAKER_00: The same three shapes appeared in EICU at 49.1%,
- 2:44SPEAKER_00: 7.2%, and 43.7%.
- 2:46SPEAKER_01: And mortality tracked the shape.
- 2:48SPEAKER_00: In Mimic 4,
- 2:50SPEAKER_00: roughly 27.6% of the persistently high group died in hospital by day 28,
- 2:56SPEAKER_00: against 16.7% for rising then falling and 11.9% for resolving.
- 3:04SPEAKER_00: In EICU, the resolving, rising then falling,
- 3:08SPEAKER_00: and persistently high groups came in at 13%, 22.5%, and 28%.
- 3:14SPEAKER_00: Adjusted for age, sex, comorbidity, and SOFA,
- 3:18SPEAKER_00: the persistently high course still carried more risk than the resolving one,
- 3:23SPEAKER_00: a subdistribution hazard ratio of 1.65 in MIMIC-4,
- 3:27SPEAKER_00: with a 95% confidence interval from 1.52 to 1.78, and 1.73 in EICU.
- 3:34SPEAKER_01: Now, the heparin numbers.
- 3:36SPEAKER_00: In Mimic 4, patients given prophylactic heparin were less likely to die
- 3:39SPEAKER_00: in the four weeks after the landmark.
- 3:42SPEAKER_00: The subdistribution hazard ratio was 0.908,
- 3:46SPEAKER_00: with a 95% confidence interval from 0.839 to 0.983.
- 3:52SPEAKER_00: In plain terms, 15.72% against 17.06%,
- 3:57SPEAKER_00: an absolute difference of 1.34 points.
- 4:00SPEAKER_00: In EICU, the estimate pointed the same way, at 0.877,
- 4:05SPEAKER_00: but the interval ran from 0.744 to 1.034, so it did not reach significance.
- 4:13SPEAKER_00: The authors call the heparin findings hypothesis-generating and say
- 4:17SPEAKER_00: a subgroup signal seen in MIMIC-4 cannot guide anticoagulation decisions.
- 4:21SPEAKER_01: Let me push on the gradient.
- 4:23SPEAKER_01: The persistently high group died more?
- 4:25SPEAKER_01: Was that the trajectory or who they were on day one?
- 4:28SPEAKER_00: Much of it was who they were.
- 4:30SPEAKER_00: Mean SOFA was 8.81 against 5.5 in the resolving group,
- 4:35SPEAKER_00: mean platelets 149 against 236, and mean INR 2.13 against 1.4.
- 4:42SPEAKER_00: A gradient-boosted model built from first-day data picked out
- 4:46SPEAKER_00: the persistently high course with an AUROC
- 4:49SPEAKER_00: of 0.909 internally and 0.835 in EICU.
- 4:54SPEAKER_00: Even so, after adjustment, the persistently high course still carried more risk
- 4:59SPEAKER_00: than the resolving one.
- 5:00SPEAKER_01: And the courses are defined after treatment started.
- 5:03SPEAKER_00: Which is why they describe prognosis, not a lever.
- 5:07SPEAKER_00: And in routine care, the patients who do not get prophylactic heparin
- 5:11SPEAKER_00: in the first two days are often the ones with very low platelets, active bleeding,
- 5:16SPEAKER_00: a pending procedure, or liver failure.
- 5:18SPEAKER_00: Waiting balances what was measured.
- 5:21SPEAKER_00: It cannot balance a clinician's reasons that never reach the database.
- 5:25SPEAKER_01: What else is missing?
- 5:27SPEAKER_00: Low molecular weight heparin was left out of the exposure definition.
- 5:31SPEAKER_00: The landmark excludes everyone who died or left before hour 48,
- 5:36SPEAKER_00: and the randomized layer is thin and neutral.
- 5:39SPEAKER_00: A 2024 meta-analysis found only three trials, 426 patients in all,
- 5:44SPEAKER_00: with a risk ratio for mortality of 0.86
- 5:48SPEAKER_00: and a confidence interval from 0.6 to 1.24.
- 5:52SPEAKER_00: An interval that wide is compatible with real benefit and with real harm.
- 5:57SPEAKER_01: Take it to the unit.
- 5:59SPEAKER_01: What goes on morning rounds?
- 6:01SPEAKER_00: The score in one line.
- 6:03SPEAKER_00: Nurses or advanced practice providers can present platelets, INR, organ score total,
- 6:09SPEAKER_00: and the direction since yesterday.
- 6:11SPEAKER_00: A platelet count from 100 to under 150 adds one point, and under 100 adds two.
- 6:17SPEAKER_00: An INR above 1.2 up to 1.4 adds one, and above 1.4 adds two.
- 6:24SPEAKER_00: A sofa organ component of one adds one, and two or more adds two.
- 6:28SPEAKER_00: A total of four or more is positive.
- 6:31SPEAKER_00: And when the score will not come down?
- 6:34SPEAKER_00: Treat it as a prompt for a differential, not a diagnosis.
- 6:38SPEAKER_00: Look for heparin-induced thrombocytopenia, microangiopathy, liver disease, and dilution.
- 6:44SPEAKER_00: Pharmacists. Keep prophylaxis on VTE and bleeding grounds.
- 6:48SPEAKER_00: Confirm the dose, the timing, and any missed doses.
- 6:52SPEAKER_00: A modest platelet decline is a reason to reassess, not an automatic reason to stop.
- 6:57SPEAKER_00: A steep or unexplained one is a reason to check for hit.
- 7:01SPEAKER_00: ICCN calls that a practice suggestion, not a tested rule.
- 7:07SPEAKER_00: Report what laboratory values miss, oozing at line sites, new petechiae or bruising,
- 7:13SPEAKER_00: blood in the endotracheal secretions or gastric aspirate,
- 7:16SPEAKER_00: and any held or missed prophylaxis those belong in the handoff,
- 7:20SPEAKER_00: along with the score today and yesterday and the platelet direction.
- 7:24SPEAKER_01: And the ICCN position for whoever is leading the plan.
- 7:27SPEAKER_00: Score SIC every day.
- 7:30SPEAKER_00: Keep prophylactic heprin decisions on their usual grounds, VTE risk against bleeding risk.
- 7:36SPEAKER_00: Do not withhold or escalate anticoagulation because of an SIC number alone
- 7:42SPEAKER_00: and do not start therapeutic anticoagulation for sepsis-induced coagulopathy outside
- 7:48SPEAKER_00: a clinical trial.
- 7:49SPEAKER_00: The Teaching Pearl SIC is a screen built from three numbers you already have.
- 7:54SPEAKER_00: It tells you who is in trouble.
- 7:56SPEAKER_01: It has not yet told anyone what to give.
- 7:59SPEAKER_00: The score earns closer attention and a differential.
- 8:02SPEAKER_00: It does not earn a drug.
- 8:04SPEAKER_00: So what changed today?
- 8:06SPEAKER_00: Three conversations.
- 8:08SPEAKER_00: On rounds, present the SIC score and its direction.
- 8:11SPEAKER_00: With your pharmacist, keep prophylaxis on VTE and bleeding risk,
- 8:15SPEAKER_00: and ask about HIIT when platelets fall fast.
- 8:19SPEAKER_00: And when a database study says a familiar drug lowers sepsis mortality,
- 8:23SPEAKER_00: ask who did not receive it.
- 8:25SPEAKER_01: The full article with the scoring thresholds and the presenter slides is on the ICCN substack.
- 8:30SPEAKER_01: The link is in the show notes.
- 8:32SPEAKER_01: ICCN publishes one article a day for ICU clinicians.
- 8:36SPEAKER_01: It is $8 a month or $80 for the year with a seven-day free trial.
- 8:40SPEAKER_01: This has been the ICCN Daily Brief.
- 8:42SPEAKER_01: Back tomorrow at 6.30.