I’ve started a Substack: The Dependent Variable. X is great for short-form discussion, but it’s not ideal for building a coherent archive. This will be the home for longer-form writing on cardiovascular physiology, haemodynamics, fluids, shock, ultrasound and critical care reasoning. All in one place, easier to follow and return to. First post: open.substack.com/pub/icmtea…

Can VExUS estimate CVP in ventilated children? A new prospective PICU study looked at 56 mechanically ventilated children with simultaneous invasive CVP and ultrasound of the IVC, hepatic, portal and intrarenal veins. The conventional categorical 0–3 VExUS-style score was only moderately good at identifying CVP >12 mmHg: • AUROC 0.74 • Sensitivity 45% • Specificity 98% They also tested a more granular 0–7 score: • Abnormal IVC = 1 point • Hepatic vein = 0–2 • Portal vein = 0–2 • Intrarenal vein = 0–2 This performed much better: • AUROC 0.94 • Sensitivity 82% • Specificity 91% Intrarenal venous Doppler alone had an AUROC of 0.92. The 0–7 score is not yet an established standard. It was proposed and tested in this paediatric study, although a very similar “extended VExUS” approach has also appeared in recent adult work. The useful physiological message is that a more graded description of venous waveform abnormalities may track RAP/CVP better than the usual coarse categorical system. And the bigger point remains: VExUS is principally a readout of the venous pressure/pulsatility environment, not a volume meter. Same CVP does not always mean the same waveform: intrathoracic pressure, abdominal pressure, tricuspid regurgitation, venous compliance and local tissue pressure all modify transmission. Carioca F de L, Franzon NH, Krzesinski LDS, et al. Pediatric Venous Excess Ultrasound Score (P-VExUS): a novel approach to assess central venous pressure in the PICU. Eur J Pediatr. 2026;185(11):855. doi.org/10.1007/s00431-026-0… Interested to hear the thoughts of @ThinkingCC @khaycock2 @NephroP @ross_prager

The heart supplies the energy for circulation. The flow that results depends on two constraints. The systemic pressure–volume state is one of them. If you’ve ever been confused by 'venous return' terminology, you are not alone. So I went back to first principles. Start with a single elastic container. Build V₀. Add volume. Add compliance. Then show how those pieces determine pressure, where blood resides, and why Pms is the pressure signature of the whole vascular pressure–volume state, not simply a measure of stressed volume. This is the foundation. Part 2 will turn the pump on and asks how that pressure–volume state constrains venous return. thedependentvariable.com/p/t…

When I started The Dependent Variable, one of my worries was that I might eventually run out of things I wanted to write about. I now have the opposite problem. Already outlined is an article exploring terms such as stressed volume, unstressed volume, compliance, capacitance, venous tone and mean systemic filling pressure. They are closely related. They are often used almost interchangeably. But they are not the same thing. Also mapped out is a six-part series on the way we often over-complicate haemodynamics by measuring or calculating ever more numbers: ScvO₂. The CO₂ gap. P(v−a)CO₂ / C(a−v)O₂. Pulse pressure, diastolic shock index, PPV and SVV. Dynamic arterial elastance: Eadyn = PPV/SVV. Ees/Ea and ventriculo-arterial coupling. For each of them, I want to start from zero (no prior knowledge assumed) work through what the number actually represents, where it comes from, what physiology sits underneath it, and whether it genuinely helps us make better clinical decisions. The aim is to take you from novice to expert without making the subject unnecessarily complicated. Lots more to come from thedependentvariable.com Stay tuned.

How to read a blood gas: Parts 1 and 2 explain the chemistry. But even if you haven’t read them, Part 3 will let you quickly work out what is driving the metabolic component of an acid–base disturbance. Anion gap, albumin, lactate, Story’s method and how opposing effects can hide inside a single base excess. Part 3 is out now and it's free to read. thedependentvariable.com/p/a…

The Normosol vs LR result was completely predictable from the chemistry. Lactate and acetate disappear rapidly. Gluconate persists. Once you count the anions that actually remain in plasma, the effective SID of the two fluids converges. Similar bicarbonate and BE was the obvious result. Looking for differences in mortality, dialysis or ventilator days was worse. There is no plausible route from this tiny compositional difference to major clinical outcomes. This is what happens when trials are designed without letting basic physiology veto bad questions. The chemistry is in my latest substack series.

Part 3 of the acid–base series is nearly ready. This is the bedside episode: anion gap, albumin, lactate, Story’s method, and how several metabolic effects can cancel each other out and leave a deceptively simple blood gas. The aim is a practical way to work out what is contributing to the result. If you missed the previous episodes then start with part 1 here: thedependentvariable.com/p/a…

This isn’t another article listing Stewart’s three “independent variables”. It tackles the awkward questions most explanations skip: If lower SID means more H⁺, where does that H⁺ actually come from? If water dissociates, what happens to the OH⁻? Why does low albumin alkalinise if albumin is a buffer? And how can Henderson–Hasselbalch and Stewart both be right? Often we learn the equations without ever quite joining the chemistry together. Part 2 of the series is out now. thedependentvariable.com/p/a…

Part 2 of the acid–base series will be out tomorrow on thedependentvariable.com This one is entirely about the chemistry behind Stewart: strong ions, weak acids, SID, electroneutrality, where the H⁺ comes from, and what “independent” and “dependent” variables actually mean. I wanted to take the time to explain the bits that are usually skipped over. Part 3 will then bring it all back to the bedside.

Writing Part 2 made me realise Stewart chemistry deserves its own episode. SID, weak acids, electroneutrality, where the H⁺ comes from, and what “dependent variable” actually means. Part 3 will finish the trilogy with how to use it at the bedside. Find part 1 here thedependentvariable.com/p/a…

Twenty years ago, someone handed me a photocopied article about Stewart’s approach to acid–base. I needed an explanation that started several steps earlier. So I’ve written one. Whose hydrogen is it anyway? From first principles to the disorders a normal pH can hide. thedependentvariable.com/p/a…

The Surviving Sepsis Campaign has shaped sepsis care for decades. But do some of its core recommendations now lag behind the physiology and the evidence? Rory Spiegel @TheEMNerd led this new dissenting opinion, which I was very pleased to contribute to. We challenge: • fixed 30 mL/kg fluid loading • chasing lactate with fluid • fluid responsiveness as a sufficient reason to give fluid • pressure-centric resuscitation • the omission of POCUS from modern haemodynamic assessment • rigid bundle-driven care The argument is not for less treatment. It is for more individualised, physiology-driven treatment of shock. authors.elsevier.com/c/1npdk… The 2026 Surviving Sepsis Campaign Guidelines: The dissenting opinion Spiegel et al. AJEM, just published.