Critical Care Nephrology · Two-Week Intensive · Lecture 7 of 9

Acid-Base Disorders & DKA

From physiology to the bedside — and the dialysis connection

40 minutes Nephrology Fellows Week 2

Based on Koyner, Handbook of Critical Care Nephrology (2021) · NTUH Yunlin Branch

Learning objectives

By the end of this session you will be able to…

Read any gas with a six-step algorithm, and quantify every disorder present.
State what the Stewart model adds — and where it changes nothing.
Work up a high-gap acidosis, and defend or refuse alkali therapy from the trial data.
Run DKA and HHS on the 2024 consensus: potassium-first sequencing, β-OHB-based severity and resolution.
Recognise toxic alcohol, salicylate and metformin poisoning, and apply the EXTRIP thresholds.

Koyner Ch 25, 27–28, 37 · Ronco Ch 64–71 · Umpierrez GE et al. Diabetes Care. 2024;47:1257 (ADA/EASD/JBDS/AACE/DTS consensus)

Reading the gas

Six slides of arithmetic that exist to answer one question: how many disorders does this patient actually have?

Koyner Ch 25 · Ronco Ch 64–65

Henderson at the bedside, and the six steps you never skip

pH = 6.1 + log₁₀(HCO₃⁻ ÷ (0.03 × PaCO₂)) → [H⁺] ≈ 24 × PaCO₂ ÷ HCO₃⁻
[H⁺] ≈ 10^(9 − pH): 7.4 = 40 · 7.2 = 63 · 7.0 = 100 Use it to verify a gas is internally consistent.

Acidaemia or alkalaemia? A normal pH with both values abnormal = two disorders.

Which variable moved with the pH? That one is primary.

Compensation inside the band? Outside = second disorder.

Albumin-corrected anion gap — every gas.

Delta ratio — does the gap rise match the HCO₃⁻ fall?

Osmolal gap if the gap is unexplained.

Koyner Ch 25 (Rodby) · Berend K, de Vries APJ, Gans ROB. N Engl J Med. 2014;371:1434 · Treger R et al. Clin J Am Soc Nephrol. 2010;5:390 · Ronco Ch 65

The six compensation coefficients — and three gases that use them

Metabolic acidosis: expected PaCO₂ = 1.5 × [HCO₃⁻] + 8 (±2) — Winter's
Metabolic alkalosis: expected PaCO₂ = 0.7 × [HCO₃⁻] + 20 (±5)
Respiratory: ΔHCO₃⁻ = +1 acute / +4 chronic per 10 mmHg PaCO₂ rise; −2 / −5 per 10 fall Respiratory compensation completes in 12–24 h; renal takes 3–5 days.

Gas	Rule	Verdict
pH 7.33 · PaCO₂ 70 · HCO₃⁻ 36	Chronic predicts 36	Chronic respiratory acidosis
pH 7.19 · PaCO₂ 70 · HCO₃⁻ 26	Acute predicts 27	Acute respiratory acidosis
pH 7.12 · PaCO₂ 25 · HCO₃⁻ 8	Winter's 18–22	Two disorders

Koyner Ch 25, Table 25.2 · Ronco Ch 65, 70 · Berend K et al. N Engl J Med. 2014;371:1434

The anion gap, the albumin correction, and the delta ratio

AG = Na⁺ − (Cl⁻ + HCO₃⁻); normal 10; AG_corr = AG + 2.5 × (4.0 − albumin)
Delta = (AG_corr − 10) ÷ (24 − HCO₃⁻) Albumin 2.0 lowers the gap by 5 — enough to hide a lactate of 5.

Ratio	Meaning	Pairing
< 0.8	Second normal-gap acidosis	Lactate + diarrhoea
0.8 – 1.6	Pure high-gap acidosis	DKA or lactic acidosis
> 1.6	Coexisting alkalosis	DKA with vomiting

Where the numbers fail

Gap falsely low: hypoalbuminaemia, lithium, paraproteins
Ratio unreliable in lactic acidosis and renal failure
Poor screen for lactate <4 — measure it

The finding only step 5 makes

ΔHCO₃⁻ zero or negative = high-gap acidosis plus hidden metabolic alkalosis.

Ronco Ch 64 (Table 64.1), Ch 65 · Kraut JA, Madias NE. Clin J Am Soc Nephrol. 2007;2:162 · Figge J et al. Crit Care Med. 1998;26:1807 · Koyner Ch 25, Table 25.3

The osmolal gap: a time-limited window on the toxic alcohols

Calc osm = 2 × Na⁺ + glucose/18 + BUN/2.8; gap = measured − calculated. Per 100 mg/dL: methanol 31 · EG 16 · ethanol 22 mOsm/kg Normal gap usually ≤10. Gap >25–30 with unexplained acidosis → fomepizole.

The osmolal gap tracks the parent alcohol; the anion gap tracks its acid metabolite
Early = big osmolal gap, normal AG; late = the reverse; mid-crossover both look normal
Measure by freezing-point depression
Other osmoles: mannitol, glycerol, propylene glycol

The trap

A normal osmolal gap never excludes ingestion — poor sensitivity, wide reference interval. If it fits, treat.

Koyner Ch 27 · Ronco Ch 64, 71 · Aabakken L et al. Hum Exp Toxicol. 1994;13:131 · Kraut JA, Xing SX. Am J Kidney Dis. 2011;58:480

Add ethanol/4.6 to the calculated osmolality if the patient has drunk. Isopropanol contributes 17 mOsm/kg per 100 mg/dL. Some references quote the normal gap as −9 to +19. Convert a gap to a level with that alcohol’s own factor — methanol 3.2, ethylene glycol 6.2 mg/dL per mOsm/kg. As the parent alcohol is metabolised the osmolal gap falls and the anion gap rises, but the anion gap always under-reads: the ceiling is 1 mEq/L of formate per mOsm/kg (2 for oxalate), and both anions are themselves cleared — the worked case later in this deck sits at an osmolal gap of 43 with only ~21 mEq/L of unmeasured anion. Early the gap is all osmolal, late all anion; neither alone excludes the diagnosis. Lactate, ketones and uraemia raise the osmolal gap only trivially. Send levels alongside treating.

Stewart properly: three independent variables, and what they explain

SID_a = (Na⁺ + K⁺ + Ca²⁺ + Mg²⁺) − (Cl⁻ + lactate⁻); SIG = SID_a − SID_e
Independent: SID, A_TOT, PaCO₂ — HCO₃⁻ is an output, not an input.

Why 0.9% saline acidifies

Na⁺ 154, Cl⁻ 154 → SID = 0; infusing it dilutes SID. Nothing acidic was added.

Three things it explains cleanly

Hypoalbuminaemic alkalosis — albumin 2.0 ≈ 5–6 base excess
Unmeasured anions of sepsis — a raised SIG
Acetazolamide — Na⁺ excreted in excess of Cl⁻

Limits and synthesis

SIG and corrected AG agree in most patients
Keep Henderson as the working language; correct for albumin, read the chloride

Ronco Ch 64 (Fig 64.1, Table 64.1), Ch 65 (Ring & Kellum) · Kellum JA, Elbers PWG. Stewart's Textbook of Acid-Base. 2nd ed. 2009 · Seifter JL. N Engl J Med. 2014;371:1821 · Story DA et al. Br J Anaesth. 2004;92:54

The algorithm end to end — worked gas

32 F · T1DM · stopped insulin · 4 days vomiting

pH 7.28PaCO₂ 24HCO₃⁻ 11Na⁺ 138K⁺ 3.4Cl⁻ 92Alb 4.2Glu 640β-OHB 6.1Lactate 1.4

1 pH 7.28 → acidaemia; Henderson ✓
2 HCO₃⁻ moves with pH → primary metabolic acidosis
3 Winter's 24.5 ± 2; measured 24 → appropriate
4 AG = 35
5 ΔAG/ΔHCO₃⁻ = 1.9 → coexisting alkalosis
6 No gap needed — β-OHB accounts for the anion
Implied pre-illness HCO₃⁻ = 36

Three findings, one gas — and the management consequence

High-gap ketoacidosis + chloride-depletion alkalosis. Treat the gap and replace chloride as KCl; the bicarbonate will fall further as the alkalosis corrects.

Koyner Ch 25, 37 · Ronco Ch 64, 69

Metabolic acidosis

The gap tells you which list to open; the list tells you which treatment actually works — and which one only moves a number.

Koyner Ch 25 · Ronco Ch 66–68

GOLD MARK — the ICU high-anion-gap list

Letter	Cause	Confirming finding
G — Glycols	Ethylene glycol; propylene glycol	Osmolal gap, oxalate crystalluria
O — Oxoproline	Chronic paracetamol in malnourished patients	Unexplained gap; stop the drug
L — L-lactate	Shock, sepsis, thiamine deficiency	Measure it — the AG misses lactate <4
D — D-lactate	Short bowel, overgrowth	Encephalopathy after carbohydrate; normal L-lactate
M — Methanol	Windshield fluid, illicit spirits	Osmolal gap, visual loss
A — Aspirin	Salicylate	Tinnitus; a mixed picture
R — Renal failure	Uraemic acidosis	Gap usually <20
K — Ketoacidosis	Diabetic, alcoholic, starvation, SGLT2i	β-hydroxybutyrate

Mehta AN, Emmett JB, Emmett M. Lancet. 2008;372:892 · Ronco Ch 64 (Table 64.2), Ch 66, 71 · Koyner Ch 25

Lactate is a stress signal, not an oxygen probe

The mechanism

Stress drives aerobic glycolysis — lactate made with adequate O₂.

Prognosis by mechanism

Mortality: lactic 56%, SIG 39%, hyperchloraemic 29%

Cause	Mechanism	Action
Type A	Dysoxia — shock, or ischaemia at a normal BP	Perfusion
Metformin (MALA)	Complex I inhibition; lactate >10–15	IHD if lactate >20
Thiamine deficiency	PDH is thiamine-dependent	Thiamine 200–500 mg
Propofol infusion syndrome	Impaired fatty-acid oxidation	Stop
Linezolid	Mitochondrial protein synthesis	Stop; reversible
Malignancy	Tumour aerobic glycolysis	Tumour response

ANDROMEDA-SHOCK

CRT vs lactate: mortality 34.9% vs 43.4%. Trend it, don't titrate to it.

Ronco Ch 67 · Gunnerson KJ et al. Crit Care. 2006;10:R22 · Hernández G et al. JAMA. 2019;321:654 · Calello DP et al. Crit Care Med. 2015;43:1716 (EXTRIP metformin) · Koyner Ch 25, 28

Alkali therapy: the trial, the arithmetic, and the isovolaemic route

BICAR-ICU (n = 389)

pH ≤7.20, bicarbonate vs none: composite 66% vs 71%; day-28 KRT 35% vs 52%.

BICAR-ICU · AKIN 2–3 (n = 182)

Primary 70% vs 82%; mortality 46% vs 63%; KRT 51% vs 73%.

BICARICU-2 (n = 627) · JAMA 2025

90-day mortality 62.1% vs 61.7%; KRT 35% vs 50%.

Deficit = (24 − HCO₃⁻) × 0.5–0.8 × wt — don't replace it
CVVH at 35 mmol/L → 120 mmol/h

When base is defensible

pH ≤7.10 with shock, or AKIN 2–3

Pitfall — the bicarbonate reflex

Not a vasopressor. Each ampoule generates CO₂, drops ionised Ca²⁺; organic acidoses pre-load an overshoot alkalosis.

Jaber S et al. Lancet. 2018;392:31 · Jung B et al. JAMA. 2025;334:2000 · Chua HR, Schneider A, Bellomo R. Ann Intensive Care. 2011;1:23 · Kraut JA, Kurtz I. Am J Kidney Dis. 2001;38:703 · Koyner Ch 25, 37

Ketoacidosis beyond classical DKA

Alcoholic ketoacidosis

Malnourished drinker, binge then abstinence, vomiting. High NADH/NAD⁺ shunts acetoacetate to β-OHB, so the nitroprusside strip under-reads.

Treat: saline with dextrose, thiamine before glucose. No insulin needed.

Starvation ketosis

Mild: gap rarely >15–18. Fasting, hyperemesis, post-op, glucose-free CKRT solutions.

Treat: feed. A rising gap on CKRT means checking the prescription.

Euglycaemic DKA — the ICU trap

Ketoacidosis with glucose <250. SGLT2 inhibitors are the modern cause: glycosuria removes the glucose signal, not the ketogenesis.

Treat: insulin and dextrose together from the start.

Koyner Ch 25, 37 · Ronco Ch 66 · Fralick M, Schneeweiss S, Patorno E. N Engl J Med. 2017;376:2300 · Umpierrez GE et al. Diabetes Care. 2024;47:1257

Normal-gap acidosis: two urine tests, then three tubular lesions

UAG = U_Na + U_K − U_Cl — a proxy for urinary NH₄⁺; negative = extrarenal, positive = renal Extrarenal: diarrhoea, fistulae, ureteral diversion. UAG is falsely positive with hippurate and β-OHB — use the urine osmolal gap instead.

Type	Defect	K⁺	U pH	ICU causes	Treatment
1 · distal	α-intercalated cell can't hold the H⁺ gradient	Low	>5.5	Amphotericin B, Sjögren, lithium	K-citrate
2 · proximal	Failed HCO₃⁻ reclamation	Low	<5.5	Fanconi: ifosfamide, tenofovir	Alkali + K⁺
4 · hyperkalaemic	Aldosterone deficiency or resistance	High	<5.5	Diabetic nephropathy, RAAS blockade	Stop the drug; loop

Ronco Ch 68 (Corey & Eckstein), Ch 55 · Koyner Ch 25 · Kim GH et al. J Am Soc Nephrol. 1996 (urine osmolal gap) · Soleimani M, Rastegar A. Am J Kidney Dis. 2016 core curriculum

Alkalosis & the respiratory axis

Metabolic alkalosis is the commonest disorder nobody is consulted about — until the patient cannot be weaned.

Koyner Ch 25 · Ronco Ch 69–70

Metabolic alkalosis: generate, maintain, then correct

Generation alone is never enough. It persists only if something maintains it: chloride depletion, hypokalaemia, hyperaldosteronism, reduced GFR.

Urine Cl⁻	Cause	Correction
<20 — responsive	Gastric loss; diuretics; post-hypercapnic	NaCl or KCl by deficit
>20 — resistant	Aldosteronism · Bartter/Gitelman · alkali load	Stop the driver; spironolactone

Uninterpretable during an active diuretic.

Chloride with the right cation — NaCl or KCl by deficit

Stop generating it — reduce the diuretic, add a PPI

Acetazolamide if volume overload forbids chloride

Rescue pH ≥7.60: central HCl, or KRT.

Ronco Ch 69 · Luke RG, Galla JH. J Am Soc Nephrol. 2012;23:204 · Faisy C et al. JAMA. 2016;315:480 (DIABOLO) · Libório AB et al. Intensive Care Med. 2015;41:479 · Koyner Ch 25

Respiratory disorders — and the renal reading of them

Respiratory acidosis — the pump fails

Drive: sedatives, brainstem lesion, obesity–hypoventilation
Pump: Guillain–Barré, myasthenia; hypophosphataemia weakens the diaphragm
Lung: COPD, asthma, ARDS, auto-PEEP

Respiratory alkalosis — drive is too high

Hypoxaemia — exclude first
Centre stimulation: sepsis, pain, fever, salicylate
Hepatic failure: HCO₃⁻ 18–20 — not acidosis
Consequences: cerebral vasoconstriction, low ionised Ca²⁺

Permissive hypercapnia, and how the kidney reads it

Lung protection gives PaCO₂ 50–80 and pH 7.20–7.30 — accepted, not treated. Avoid it with raised ICP.

Two pitfalls that look opposite but are the same error

(a) Fast PaCO₂ normalisation strands renal HCO₃⁻ — post-hypercapnic alkalosis, failed wean. (b) Cirrhotic HCO₃⁻ 14 is compensation, not acidosis.

Ronco Ch 70 · ARDS Network. N Engl J Med. 2000;342:1301 · Laffey JG, Kavanagh BP. N Engl J Med. 2002;347:43 · Koyner Ch 3, 25

DKA & HHS

Three drugs — volume, potassium, insulin — in the right order. The 2024 consensus changed the thresholds, not the order.

Koyner Ch 37 · ADA/EASD 2024 consensus

One hormone deficit, two syndromes — and the 2024 criteria

Insulin deficitAbsolute (DKA) or relative (HHS)

→

LipolysisHormone-sensitive lipase disinhibited

→

KetogenesisGlucagon lifts the brake on CPT-1

→

Osmotic diuresisWater, Na⁺, K⁺, Mg²⁺, PO₄ all lost

2024 consensus criteria — "D-K-A"

D: glucose ≥200 · K: β-OHB ≥3.0 · A: pH <7.3 or HCO₃⁻ <18
Severity: severe = pH <7.0, HCO₃⁻ <10, or β-OHB >6
HHS: glucose ≥600, eff osm >300, β-OHB <3.0

Measurement discipline

Measure β-OHB — the strip detects acetoacetate only
Venous gases are adequate
Eff osm = 2 × Na⁺ + glucose/18
~30% of DKA is mixed DKA–HHS

Umpierrez GE et al. Diabetes Care. 2024;47:1257 (ADA/EASD/JBDS/AACE/DTS consensus) · Koyner Ch 37 (Topf et al.) · Kitabchi AE et al. Diabetes Care. 2009;32:1335

The protocol: volume, potassium, insulin — in that order

VolumeDeficit ~5–7 L. 500–1000 mL in hour 1. Balanced crystalloid.

→

PotassiumDeficit 3–5 mmol/kg. K⁺ <3.5 → hold insulin; 3.5–5.0 → 10–20 mmol/L

→

Insulin0.1 U/kg/h fixed-rate; skip the bolus.

→

DextroseAdd it at glucose <250.

Fluid choice: SMART · SCOPE-DKA

SMART: MAKE30 14.3% vs 15.4%. SCOPE-DKA: resolution at 24 h 69% vs 36%.

Pitfall — insulin before potassium, and insulin before volume

Insulin shifts K⁺ within minutes — at K⁺ 3.1 it turns a survivable emergency into a cardiac arrest. Fluid, then potassium, then insulin.

Umpierrez GE et al. Diabetes Care. 2024;47:1257 · Semler MW et al. N Engl J Med. 2018;378:829 (SMART) · Zampieri FG et al. JAMA. 2021;326:818 (BaSICS) · Ramanan M et al. Intensive Care Med. 2021;47:1248 (SCOPE-DKA) · Koyner Ch 37

Ending the infusion, and the complications to pre-empt

Resolution and transition

Resolution: β-OHB <0.6 and (pH >7.30 or HCO₃⁻ ≥18)
Hyperglycaemia corrects ~6 h, ketosis ~12 h — the insulin must outlast the sugar
Basal SC 1–2 h before stopping the infusion

Complications, in the order they bite

Hypokalaemia — the commonest preventable death
Hypoglycaemia — dextrose at 250
Hypophosphataemia — replace if <1
Thrombosis — prophylaxis
AKI in ~50%; half resolves in 24 h

Why the bicarbonate lags

Once the gap closes most patients sit in a hyperchloraemic normal-gap acidosis, because ketoanions were excreted as salts. Check β-OHB, not HCO₃⁻.

Umpierrez GE et al. Diabetes Care. 2024;47:1257 · Koyner Ch 37 · Adrogué HJ et al. JAMA. 1989;262:2108 · Fisher JN, Kitabchi AE. J Clin Endocrinol Metab. 1983;57:177 · Orban JC et al. PLoS One. 2014;9:e110925

HHS, and the three populations that break the protocol

Feature	DKA	HHS
Tempo	Hours–days; acid	Days–weeks; osmolality
Glucose · β-OHB · pH	≥200 · ≥3.0 · <7.3	≥600 · <3.0 · ≥7.3
Fluid deficit	~5–7 L	9–12 L
Insulin	0.1 U/kg/h once K⁺ secured	0.05 U/kg/h after volume
Mortality	<1%	~5–20%

Sodium: adjust for orientation only

Adjusted Na⁺ = measured + 0.016 × (glucose − 100). High or normal → 0.45% NaCl.

Dialysis-dependent ESKD is a different disease

No osmotic diuresis, so no large K⁺ or PO₄ deficits, and often fluid overloaded. Insulin largely alone; delete standing K⁺ orders.

Pregnancy and SGLT2i

Pregnant patients ketose faster — treat at glucose <200. SGLT2i glycosuria persists: long infusion, generous dextrose.

Umpierrez GE et al. Diabetes Care. 2024;47:1257 · Koyner Ch 37 · Katz MA. N Engl J Med. 1973;289:843 · Hillier TA et al. Am J Med. 1999;106:399 · Schaapveld-Davis CM et al. Clin Diabetes. 2017;35:202

Does this patient need an ICU bed and an infusion at all?

SC rapid-acting analogues in mild–moderate DKA

8 RCTs, 415 patients: SC lispro/aspart q1–2 h vs IV — no difference in resolution, stay or hypoglycaemia.

Reasonable for a subcutaneous ward protocol

Mild–moderate: pH >7.0, HCO₃⁻ ≥10
Alert, tolerating oral fluid, haemodynamically stable
SC analogue 0.1–0.2 U/kg load, then 0.1 U/kg q1–2 h

Keep the IV infusion and the ICU bed

Severe DKA or altered mentation
Shock, hypoxaemia, vasopressors
Pregnancy, ESKD, mixed DKA–HHS
K⁺ <3.5

Pitfall — protocol without a brain

Order sets encode a typical patient and fail silently when the gap has closed but β-OHB has not.

Andrade-Castellanos CA et al. Cochrane Database Syst Rev. 2016;CD011281 · Karslioglu French E et al. meta-analysis, J Diabetes Complications. 2024 · Umpierrez GE et al. Diabetes Care. 2024;47:1257

Poisonings, cases & wrap-up

Two poisonings where the nephrologist's phone call changes the outcome — and where waiting for the level is the error.

Koyner Ch 27–28 · Ronco Ch 71 · EXTRIP

Methanol vs ethylene glycol: same enzyme, different target organ

Methanol — "wood alcohol"

Windshield fluid, illicit spirits
ADH → formaldehyde → formic acid
Target: optic nerve and putamen — blurred vision
Latency 12–24 h — early it looks like drunkenness
Folinic acid 50 mg q4–6 h

Ethylene glycol — antifreeze

ADH → glycolate → oxalate
Target: the kidney — calcium oxalate, AKI, crystals
Hypocalcaemia with a long QT
Thiamine 100 mg + pyridoxine 50–100 mg

Read the two gaps as a clock

High osmolal gap with a normal AG = early; block ADH now. High AG with a closing osmolal gap = late; dialysis moves up.

Koyner Ch 27–28 · Ronco Ch 71 · Hoffman RS et al. Goldfrank's Toxicologic Emergencies. 11th ed. 2019 · Barceloux DG et al. J Toxicol Clin Toxicol. 2002;40:415

Fomepizole, dialysis, and the other dialysable toxins

Block ADH now — don't wait for the level. Fomepizole 15 mg/kg IV load, then 10 mg/kg q12h. Start on suspicion.

Ethanol fallback: sedates, causes hypoglycaemia, erratic kinetics.

Haemodialysis — EXTRIP methanol: coma or visual deficit; pH ≤7.15; AG >24; methanol (mg/dL): >70 with fomepizole; >60 with ethanol; >50 without ADH blockade.

Intermittent HD, not CKRT. Redose fomepizole q4h during dialysis.

Fomepizole alone is often enough for ethylene glycol with early blockade.

Others: lithium (HD if >4.0 mmol/L with kidney impairment); metformin (lactate >20).

Roberts DM et al. Crit Care Med. 2015;43:461 (EXTRIP methanol) · Decker BS et al. Clin J Am Soc Nephrol. 2015;10:875 (EXTRIP lithium) · Calello DP et al. Crit Care Med. 2015;43:1716 (EXTRIP metformin) · Koyner Ch 27–28

Salicylate: read the gas, and think twice before intubating

Typical gas: pH 7.44 · PaCO₂ 18 · HCO₃⁻ 12 · AG = 26 Respiratory alkalosis and high-gap metabolic acidosis. An alkalaemic pH with a gap of 26 is salicylate until proved otherwise.

Mechanism, presentation, pitfalls

Medullary stimulation → tachypnoea; uncoupling → lactate, hyperpyrexia
Tinnitus, nausea, agitation → coma
Chronic toxicity in the elderly mimics sepsis

Why intubation can kill

A minute ventilation of 30–40 L/min keeps salicylate ionised and out of the brain. Conventional ventilation collapses that.

Alkalinise: target urine pH 7.5–8.0 — it fails without aggressive K⁺
Haemodialysis (EXTRIP): altered mental status, hypoxaemia, or a level >100

Juurlink DN et al. Ann Emerg Med. 2015;66:165 (EXTRIP salicylate) · Koyner Ch 28 · Ronco Ch 71

Case 1 · 44 M, found confused in a garage, no history available

Tachypnoeic, GCS 11, pupils sluggish, no focal signs. Ketones negative. Ethanol undetectable. Creatinine 1.1 mg/dL.

pH 7.16PaCO₂ 20HCO₃⁻ 7Na⁺ 142K⁺ 4.6Cl⁻ 106Albumin 2.0Glucose 108BUN 22Lactate 3.2Measured osm 341

Run all six steps. What is the diagnosis, what does the albumin do to your gap, and what do you order in the next ten minutes?

Think 60 seconds · full working on the next slide

Koyner Ch 27 · Ronco Ch 64, 71 · Aabakken L et al. Hum Exp Toxicol. 1994;13:131 · Kraut JA, Xing SX. Am J Kidney Dis. 2011;58:480

Case 1 — the working, and the decision fork

1 pH 7.16 → acidaemia; Henderson ✓
2 HCO₃⁻ moves with pH → primary metabolic acidosis
3 Winter's appropriate
4 AG = 34 (albumin-corrected)
5 ΔAG/ΔHCO₃⁻ = 1.4 → single high-gap
6 Osmolal gap 43; lactate explains only 3 of 24

Diagnosis

Toxic alcohol — mid-crossover, both gaps open. Examine the fundi.

Orders now

Fomepizole 15 mg/kg IV
Folinic acid, thiamine, pyridoxine
Gas, gap and osmolality hourly
Call for intermittent HD

The fork

A pH of 7.32 with a normal AG → fomepizole alone. What goes to the machine is the metabolite already made.

Roberts DM et al. Crit Care Med. 2015;43:461 · Koyner Ch 27–28 · Ronco Ch 64, 71

Case 2 · 24 F, type 1 diabetes, insulin pump failure, 12 h of vomiting

Kussmaul respiration, dry mucous membranes, HR 128, BP 96/58, alert. The admitting team has written for an insulin bolus plus infusion and 1 L of 0.9% saline.

pH 7.17PaCO₂ 17HCO₃⁻ 6Na⁺ 130K⁺ 3.1Cl⁻ 96Albumin 4.4Glucose 520Creatinine 1.8β-OHB 7.4Lactate 1.9

The infusion is about to be connected. What do you change, what do you calculate first — and what severity grade is this?

Hands up — then justify the order of the three drugs

Umpierrez GE et al. Diabetes Care. 2024;47:1257 (ADA/EASD/JBDS/AACE/DTS consensus) · Koyner Ch 37 · Kitabchi AE et al. Diabetes Care. 2009;32:1335

Case 2 — the first hour, and the arithmetic behind it

Henderson ✓; Winter's appropriate.
AG = 28; ratio 1.0 → pure high-gap ketoacidosis
Adjusted Na⁺ = 147 — translocational
HCO₃⁻ 6 with β-OHB 7.4 → severe DKA

The corrected sequence

Hold the insulin (K⁺ 3.1)
500–1000 mL crystalloid in hour 1
KCl ~10 mmol/h; insulin once K⁺ ≥3.5
Insulin 0.1 U/kg/h, no bolus
Stop on β-OHB <0.6 with the acid–base criteria met
No bicarbonate at pH 7.17

What else to anticipate

Cr 1.8 is largely haemodynamic
Expect a residual hyperchloraemic acidosis
Find the precipitant; a raised lipase is usually not pancreatitis
Thrombosis prophylaxis

Umpierrez GE et al. Diabetes Care. 2024;47:1257 · Koyner Ch 37 · Chua HR et al. Ann Intensive Care. 2011;1:23 · Orban JC et al. PLoS One. 2014;9:e110925

Case 3 · 68 M, day 5 after oesophagectomy, cannot be weaned

High-output nasogastric suction (1.5 L/day), furosemide infusion running for oedema. Shallow spontaneous breaths, three failed weaning trials, mild carpopedal spasm.

pH 7.52PaCO₂ 48HCO₃⁻ 38Na⁺ 141K⁺ 2.9Cl⁻ 90Albumin 2.4Mg 0.6 mmol/LIonised Ca 1.02Urine Cl⁻ 45

The urine chloride is 45 — does that make this chloride-resistant? And is the PaCO₂ of 48 a second disorder?

Think 45 seconds · working on the next slide

Ronco Ch 69 · Koyner Ch 25 · Luke RG, Galla JH. J Am Soc Nephrol. 2012;23:204

Case 3 — why the test lied, and why he will not wean

Henderson ✓; expected PaCO₂ 46.6 ± 5, measured 48 — appropriate compensation
AG = 17 corrected — hidden high-gap acidosis
ΔHCO₃⁻ = −14 → alkalosis masking acidosis

Diagnosis and management

Chloride-depletion alkalosis (gastric loss + loop diuretic) with appropriate hypoventilation, plus a small unmeasured-anion acidosis seen only with the albumin correction
Urine Cl⁻ 45 uninterpretable — furosemide still acting
Magnesium first (0.6), then KCl
Reduce furosemide; add PPI; acetazolamide if overloaded

Why the wean fails

PaCO₂ 48 is compensation working — it will fight every weaning trial. Alkalaemia raises calcium-albumin binding: ionised Ca 1.02 + Mg 0.6 explain carpopedal spasm and weak diaphragm. Fix chloride, magnesium, potassium.

Ronco Ch 69 · Koyner Ch 25 · Luke RG, Galla JH. J Am Soc Nephrol. 2012;23:204

Poll 1 — name the disorder

Post-operative patient, 6 L of 0.9% saline in 12 hours, extubated and comfortable. pH 7.29 · PaCO₂ 30 · HCO₃⁻ 14 · Na⁺ 140 · Cl⁻ 114 · K⁺ 4.0 · albumin 4.4 · lactate 1.1.

A. High-anion-gap metabolic acidosis · B. Normal-anion-gap metabolic acidosis with appropriate respiratory compensation · C. Metabolic acidosis plus a primary respiratory alkalosis · D. Primary respiratory alkalosis with renal compensation · E. Normal-gap acidosis plus an occult high-gap acidosis masked by hypoalbuminaemia

Hands up — then name the single number that rules out each wrong answer

Ronco Ch 64 (Table 64.1), Ch 65 · Kraut JA, Madias NE. Clin J Am Soc Nephrol. 2007;2:162 · Koyner Ch 25, Table 25.3

Poll 2 — the decision, not the diagnosis

62 F, type 2 diabetes on metformin, 3 days of diarrhoea, now hypotensive on noradrenaline. Anuric 8 hours. pH 6.93 · HCO₃⁻ 4.5 · Na⁺ 138 · Cl⁻ 100 · albumin 2.8 · K⁺ 6.2 · lactate 22 · creatinine 5.4 · β-OHB 0.4.

A. 150 mmol NaHCO₃ and reassess · B. CVVHDF with a 35 mmol/L bicarbonate effluent · C. Urgent intermittent haemodialysis · D. Escalate vasopressors; alkali only below pH 6.85 · E. Bicarbonate now, CKRT later

Vote first — then defend your choice against the trial data

Calello DP et al. Crit Care Med. 2015;43:1716 (EXTRIP metformin) · Ronco Ch 67 · Koyner Ch 25, 28

Answer C. This is metformin-associated lactic acidosis with AKI and hyperkalaemia: EXTRIP recommends extracorporeal treatment for lactate >20, pH ≤7.0, shock or reduced consciousness, and explicitly prefers intermittent HD with a bicarbonate buffer initially because metformin is small, minimally protein-bound and cleared far faster by IHD; switch to CKRT afterwards for the rebound. A is inadequate — an alkali requirement of this size approaches 100 mmol/h, which is not deliverable by drip. B is the common answer and is too slow for this degree of acidaemia and hyperkalaemia, though it is right after the first IHD run. D ignores that BICAR-ICU's benefit was concentrated in AKIN 2–3, exactly this patient. E delays the definitive treatment. Note the corrected gap: AG = 138 − (100 + 4.5) = 33.5, +3 for albumin 2.8 → 36.5, and lactate 22 accounts for most but not all of it — the delta ratio is (36.5 − 10) / (24 − 4.5) ≈ 1.36, consistent with predominantly high-anion-gap lactic acidosis; this ratio does not support an additional normal-gap component. 2 min.

Five arguments you will have on rounds — and where the evidence sits

Stewart vs Henderson

Equipoise on language, not physiology — both are internally valid.

Bicarbonate in lactic acidosis

Partly settled. BICAR-ICU positive in AKIN 2–3; BICARICU-2 less KRT, no survival gain.

Bicarbonate in DKA

Settled enough — no randomised benefit at pH ≥6.9.

Lactate as a resuscitation target

Contested. An excellent prognostic marker, a poor endpoint.

Permissive hypercapnia

Accept the pH, watch the kidney — but avoid it with raised ICP.

Jaber S et al. Lancet. 2018;392:31 · Surviving Sepsis Campaign 2026, Crit Care Med. 2026;54:725–812 · Hernández G et al. JAMA. 2019;321:654 · Umpierrez GE et al. Diabetes Care. 2024;47:1257 · Ronco Ch 65, 67, 70

Key takeaways

Six steps, every gas: pH → primary → compensation → corrected gap → delta ratio → osmolal gap.
Compensation is a prediction. Outside the band means a second disorder.
Stewart: correct for albumin, read the chloride.
Lactate is often adrenergic glycolysis, not dysoxia — trend it, don't titrate.
Alkali: BICAR-ICU positive in AKIN 2–3; BICARICU-2 less KRT, no survival gain; reserve for pH ≤7.10.
DKA 2024: β-OHB defines it. Fluid → potassium → insulin; hold insulin at K⁺ <3.5.
Unexplained gap plus osmolal gap = fomepizole now; an alkalaemic pH with a gap = salicylate.

References & further reading

Where to go deeper

Umpierrez GE, et al. Hyperglycaemic crises consensus (ADA/EASD). Diabetes Care. 2024;47:1257.
Jaber S, et al. BICAR-ICU. Lancet. 2018;392:31.
Hernández G, et al. ANDROMEDA-SHOCK. JAMA. 2019;321:654.
Jung B, et al. BICARICU-2. JAMA. 2025;334:2000.
Ramanan M, et al. SCOPE-DKA. Intensive Care Med. 2021;47:1248.
Berend K, et al. Acid-base assessment. NEJM. 2014;371:1434.
Kraut JA, Madias NE. Anion gap. CJASN. 2007;2:162.
EXTRIP: methanol (Crit Care Med. 2015;43:461) · salicylate (Ann Emerg Med. 2015;66:165).

Full list: (1) Umpierrez et al. ADA/EASD/JBDS/AACE/DTS consensus (Diabetes Care 2024;47:1257–1275); (2) BICAR-ICU (Lancet 2018;392:31–40), BICARICU-2 (JAMA 2025;334:2000–2010); (3) ANDROMEDA-SHOCK (JAMA 2019;321:654–664); (4) SMART (NEJM 2018;378:829–839), BaSICS (JAMA 2021;326:818–829); (5) SCOPE-DKA (Intensive Care Med 2021;47:1248–1257); (6) Berend et al. (NEJM 2014;371:1434–1445), Seifter (NEJM 2014;371:1821–1831); (7) Kraut & Madias anion gap (CJASN 2007;2:162–174), GOLD MARK (Lancet 2008;372:892); (8) EXTRIP methanol (Crit Care Med 2015;43:461–472), salicylate (Ann Emerg Med 2015;66:165–181), metformin (Crit Care Med 2015;43:1716–1730), lithium (CJASN 2015;10:875–887); (9) SC insulin for DKA (Cochrane 2016;CD011281); (10) Koyner Handbook Ch 25, 27–28, 37; Ronco Critical Care Nephrology 3e Ch 64–71.

Critical Care Nephrology · Two-Week Intensive

Thank you

Questions & discussion — bring me a gas you could not explain and we will run the six steps together.

Next: Lecture 08 — Kidney Replacement Therapy I: Access, Modalities & Timing