Nothing new today — filling in the ESICM fluid therapy guideline, all three parts

No new guidelines from tracked societies today. Eleven journals swept for 31 August – 2 September; the only two hits were research papers that matched a filter word, not guidelines. All societies clear.

Following the note added yesterday — that a watch looking back only 7–10 days cannot see landmark guidelines predating this archive — today goes to the ESICM fluid therapy guideline, which runs to three parts across 2024–2025 and is core critical care material this archive had never covered. All three abstracts are deposited, so this is first-hand rather than press-sourced.

Part 1 — the choice of resuscitation fluids

Intensive Care Medicine, 21 May 2024 · DOI 10.1007/s00134-024-07369-9 · PMID 38771364

GRADE methodology, international panel of clinicians and methodologists. All recommendations conditional:

Crystalloids rather than albumin for volume expansion — in critically ill patients generally (moderate certainty), in sepsis (moderate), in acute respiratory failure (very low), and perioperatively or in patients at risk of bleeding (very low).

Exceptions in both directions:

• Isotonic saline rather than albumin in traumatic brain injury (very low certainty)
• Albumin rather than crystalloids in cirrhosis (very low certainty)

Balanced crystalloids rather than isotonic saline in critically ill patients generally (low certainty), in sepsis (low certainty), and in patients with kidney injury.

Part 2 — the volume of resuscitation fluids

Intensive Care Medicine, 31 March 2025 · DOI 10.1007/s00134-025-07840-1 · PMID 40163133

• Sepsis and septic shock: up to 30 ml/kg of intravenous crystalloid in the initial phase, adjusted to clinical context with frequent reassessment — very low certainty
• Optimisation phase: an individualised approach (very low certainty). No recommendation for or against restrictive versus liberal strategies — explicitly “moderate certainty of no effect”
• Haemorrhagic shock: restrictive strategy after blunt trauma (moderate certainty) and penetrating trauma (low certainty); non-traumatic haemorrhagic shock guided by haemodynamic and biochemical parameters (ungraded best practice)
• Left-sided cardiogenic shock: fluid resuscitation as primary treatment is not recommended
• Cardiac tamponade: fluids cautiously until definitive treatment

Part 3 — fluid removal at the de-escalation phase

Intensive Care Medicine, 19 August 2025 · DOI 10.1007/s00134-025-08058-x · PMID 40828463

From 13 RCTs, three conditional recommendations:

• De-escalation of fluid therapy over no de-escalation after the acute resuscitation phase (low certainty)
• Protocolised fluid removal by diuretics over usual care after the acute phase — moderate certainty, the strongest evidence in the whole three-part set
• Against routine ultrafiltration or extracorporeal fluid removal in the absence of another indication for renal replacement therapy (low certainty)

Limited evidence to comment on fluid removal in specific patient cohorts.

Interpretation

The certainty ratings are the story. Across three documents and roughly a decade of trials, almost everything is conditional and most of it rests on low or very low certainty evidence. The 30 ml/kg figure — embedded in sepsis bundles, quality measures and audit worldwide — is graded very low certainty here. That is not the panel being timid; it is an accurate description of what the trials support, and it sits awkwardly against how prescriptively that number gets applied.

The one moderate-certainty recommendation is about getting fluid out, not putting it in: protocolised diuresis beats usual care after the acute phase. Of everything in these three parts, that is the recommendation with the firmest ground under it, and it is the one least likely to be protocolised in most units.

The explicit non-recommendation deserves attention too. On restrictive versus liberal strategy in the optimisation phase, the panel does not say “insufficient evidence” — it says moderate certainty of no effect. That is a substantive finding masquerading as a non-answer, and it follows CLASSIC and CLOVERS, both of which found no difference. The debate that has consumed the field for a decade may simply have no answer at the population level, which points the same way as the phenotype-driven arguments in the β-blockade review and the cardiogenic shock criteria.

Perioperatively, Part 1’s conditional recommendation for crystalloid over albumin in the perioperative period and in patients at risk of bleeding is worth knowing, as is the cirrhosis exception — a population where the reflex against colloid is wrong.

Part 1 · Part 2 · Part 3

---

Two major cardiothoracic guidelines were also identified today and are not covered here, because no abstract is deposited for either and neither full text was reachable: the 2025 ESC/EACTS Guidelines for the management of valvular heart disease (Eur Heart J, 1 November 2025, DOI 10.1093/eurheartj/ehaf194, PMID 40878295) and the EACTS/STS Guidelines for Diagnosing and Treating Acute and Chronic Syndromes of the Aortic Organ. Both are on the outstanding list. There is also an EJCTS paper specifically on the surgical implications of the valvular guidelines (DOI 10.1093/ejcts/ezag158) that would be the better read for your practice if it can be obtained.