9.3 Complex & Mixed Acid-Base Disorders: Boston vs Copenhagen vs Stewart

Systematic Arterial Blood Gas (ABG) Interpretation

Acid-base derangements in the ICU are rarely isolated, monolithic events. Critically ill patients frequently harbor two or three concurrent acid-base processes simultaneously (e.g., septic shock with lactic acidosis, hypoventilatory respiratory failure, and contraction alkalosis from nasogastric suctioning). Systematic ABG analysis must follow an unvarying, sequential five-step algorithm:

Loading diagram...

Step-by-Step Diagnostic Framework

1. Step 1: Evaluate Arterial pH:
   • Normal pH: $7.35\text{ to }7.45$;
   • Acidemia: $\text{pH} < 7.35$;
   • Alkalemia: $\text{pH} > 7.45$;
   • Note: A normal pH does not establish normal acid-base physiology; it frequently signifies a severe mixed disorder (e.g., severe HAGMA balanced by severe primary metabolic or respiratory alkalosis).
2. Step 2: Identify the Primary Disturbance:
   • Look at $\text{PaCO}_2$ (normal $35-45\text{ mmHg}$) and $[\text{HCO}_3^-]$ (normal $22-26\text{ mEq/L}$);
   • If pH is acidemic ($< 7.35$): primary metabolic acidosis if $[\text{HCO}_3^-] < 22\text{ mEq/L}$; primary respiratory acidosis if $\text{PaCO}_2 > 45\text{ mmHg}$;
   • If pH is alkalemic ($> 7.45$): primary metabolic alkalosis if $[\text{HCO}_3^-] > 26\text{ mEq/L}$; primary respiratory alkalosis if $\text{PaCO}_2 < 35\text{ mmHg}$.
3. Step 3: Calculate Expected Physiological Compensation (The Boston Rules): Physiological compensation never fully normalizes arterial pH. If measured compensation deviates from predicted values, a secondary independent acid-base disorder is confirmed:
   • Metabolic Acidosis (Winter's Formula): $\text{Expected } \text{PaCO}_2 = 1.5 \times [\text{HCO}_3^-] + 8 \pm 2$
     • If measured $\text{PaCO}_2 > \text{expected}$: concurrent respiratory acidosis (respiratory muscle exhaustion, hypoventilation, CNS depression);
     • If measured $\text{PaCO}_2 < \text{expected}$: concurrent respiratory alkalosis (pain, sepsis, early salicylate toxicity, hepatic encephalopathy).
   • Metabolic Alkalosis: $\text{Expected } \text{PaCO}_2 = 40 + 0.7 \times ([\text{HCO}_3^-] - 24) \pm 5 \quad (\text{or } \Delta\text{PaCO}_2 = 0.7 \times \Delta[\text{HCO}_3^-])$
     • The respiratory center compensates by hypoventilating, but physiological hypoxemic drive rarely permits $\text{PaCO}_2$ to rise above $55\text{ to }60\text{ mmHg}$ in spontaneous breathing.
   • Respiratory Acidosis:
     • Acute ($< 24\text{ hours}$, cellular buffering only): For every $10\text{ mmHg}$ rise in $\text{PaCO}_2$, $[\text{HCO}_3^-]$ increases by $1\text{ mEq/L}$ ($\text{pH}$ falls $\approx 0.08$);
     • Chronic ($> 48-72\text{ hours}$, renal compensation via $\text{H}^+$ secretion): For every $10\text{ mmHg}$ rise in $\text{PaCO}_2$, $[\text{HCO}_3^-]$ increases by $3.5\text{ to }4.0\text{ mEq/L}$ ($\text{pH}$ falls $\approx 0.03$).
   • Respiratory Alkalosis:
     • Acute: For every $10\text{ mmHg}$ decrease in $\text{PaCO}_2$, $[\text{HCO}_3^-]$ decreases by $2\text{ mEq/L}$ ($\text{pH}$ rises $\approx 0.08$);
     • Chronic: For every $10\text{ mmHg}$ decrease in $\text{PaCO}_2$, $[\text{HCO}_3^-]$ decreases by $4\text{ to }5\text{ mEq/L}$ ($\text{pH}$ rises $\approx 0.03$).

---

The Serum Anion Gap and Albumin Correction

The serum anion gap represents the concentration of unmeasured anions in plasma relative to unmeasured cations, derived from the law of electroneutrality:

$\text{Anion Gap (AG)} = [\text{Na}^+] - ([\text{Cl}^-] + [\text{HCO}_3^-])$

• Normal reference range: $10\text{ to }12\text{ mEq/L}$ (or $8-12\text{ mEq/L}$). Some centers include potassium ($[\text{Na}^+ + \text{K}^+] - [\text{Cl}^- + \text{HCO}_3^-]$), shifting normal to $12-16\text{ mEq/L}$; standard critical care practice omits potassium.

Mandatory Albumin Correction

Circulating serum albumin accounts for approximately $75\%$ to $80\%$ of the normal unmeasured anion pool. Each albumin molecule carries approximately 19 negative charges at physiological pH ($7.40$). Consequently, in hypoalbuminemia (ubiquitous in sepsis, cirrhosis, malnutrition, and nephrotic syndrome), the "baseline" normal anion gap is markedly depressed.

---

High Anion Gap Metabolic Acidosis: The GOLD MARK Framework

When corrected AG exceeds $12\text{ mEq/L}$, a High Anion Gap Metabolic Acidosis (HAGMA) is present. The contemporary, subspecialty-standard mnemonic GOLD MARK covers modern ICU etiologies:

Loading diagram...

• G — Glycols:
  • Ethylene glycol (antifreeze): metabolized by alcohol dehydrogenase to glycolic and oxalic acid; causes acute tubular necrosis with calcium oxalate envelope-shaped crystalluria and acute renal failure.
  • Propylene glycol: vehicle for IV medications (lorazepam, diazepam, esmolol, nitroglycerin, phenobarbital); causes high osmolal gap HAGMA with proximal renal tubular injury.
• O — Oxoproline (5-Oxoproline / Pyroglutamic Acid): Occurs in critically ill, septic, malnourished females receiving chronic therapeutic doses of acetaminophen (paracetamol). Glutathione depletion uncouples the gamma-glutamyl cycle, shunting gamma-glutamylcysteine into 5-oxoproline, causing profound unexplained HAGMA without an elevated osmolal gap.
• L — L-Lactate: Normal $< 2.0\text{ mmol/L}$; severe lactic acidosis $> 4.0\text{ mmol/L}$. Partitioned into:
  • Type A: Tissue hypoperfusion and anaerobic metabolism (septic shock, cardiogenic shock, hemorrhagic shock, mesenteric ischemia, severe hypoxemia).
  • Type B: Impaired cellular clearance or mitochondrial poisoning without overt macro-hypoperfusion (liver failure, metformin toxicity, malignancies [Warburg effect], thiamine deficiency [pyruvate dehydrogenase failure], nucleoside reverse transcriptase inhibitors [NRTIs]).
• D — D-Lactate: Produced by colonic lactobacilli fermenting unabsorbed carbohydrates in patients with short bowel syndrome or extensive small bowel resection. Standard laboratory assays measure only L-lactate; D-lactic acidosis presents with profound encephalopathy, ataxia, confusion, and unexplained HAGMA with a "normal" reported L-lactate.
• M — Methanol: Found in windshield wiper fluids; metabolized to formaldehyde and formic acid, causing optic papillitis, retinal edema, "snowstorm" visual loss, and bilateral putaminal necrosis.
• A — Aspirin (Salicylates): Uncouples oxidative phosphorylation and directly stimulates the medullary respiratory center. Classically produces a mixed Primary Respiratory Alkalosis and High Anion Gap Metabolic Acidosis.
• R — Renal Failure (Uremia): In KDIGO Stage 3 AKI or advanced CKD, impaired glomerular filtration fails to excrete unmeasured organic and inorganic acid anions (sulfates, phosphates, urate, hippurate).
• K — Ketoacidosis: Accumulation of acetoacetate and beta-hydroxybutyrate:
  • Diabetic Ketoacidosis (DKA);
  • Alcoholic Ketoacidosis (AKA): Occurs in chronic alcoholics after a binge followed by protracted vomiting, starvation, and sudden alcohol cessation; characterized by marked ketonemia with low/normal blood glucose;
  • Starvation Ketoacidosis (SKA): Mild ketoacidosis ($[\text{HCO}_3^-]$ rarely $< 18\text{ mEq/L}$) from prolonged carbohydrate deprivation.

---

The Delta-Delta ($\Delta\text{AG} / \Delta[\text{HCO}_3^-]$) and Delta Gap Analysis

In pure, uncomplicated HAGMA, every milliequivalent of unmeasured acid anion added to the circulation titrates and consumes exactly one milliequivalent of bicarbonate buffer ($1:1$ stoichiometric titration). Therefore, the increase in anion gap above normal ($\Delta\text{AG}$) should equal the decrease in serum bicarbonate below normal ($\Delta[\text{HCO}_3^-]$).

Mathematical Formulation

$\Delta\text{AG} = \text{Observed AG} - 12$ $\Delta[\text{HCO}_3^-] = 24 - \text{Observed }[\text{HCO}_3^-]$ $\text{Delta-Delta Ratio} = \frac{\Delta\text{AG}}{\Delta[\text{HCO}_3^-]} = \frac{\text{Observed AG} - 12}{24 - [\text{HCO}_3^-]}$

Loading diagram...

Clinical Interpretation

1. Ratio $< 0.4\text{ to }0.8$ (Mixed HAGMA and NAGMA): Bicarbonate has dropped much further than can be accounted for by the addition of unmeasured anions. This proves the coexistence of a Normal Anion Gap (Hyperchloremic) Metabolic Acidosis (e.g., patient with diabetic ketoacidosis who also has severe secretory diarrhea, or septic shock resuscitated with large volumes of $0.9\%$ normal saline).
2. Ratio $1.0\text{ to }1.6$ (Pure HAGMA): Uncomplicated single disturbance; unmeasured anions account entirely for bicarbonate depletion.
3. Ratio $> 1.6\text{ to }2.0$ (Mixed HAGMA and Metabolic Alkalosis): Bicarbonate is significantly higher than expected for the degree of anion gap elevation. This establishes a concurrent metabolic alkalosis (e.g., patient with lactic acidosis from mesenteric ischemia who is actively vomiting or receiving nasogastric suction) or pre-existing chronic compensated respiratory acidosis (elevated baseline baseline bicarbonate in severe COPD).

---

Normal Anion Gap Metabolic Acidosis (NAGMA) and the Urine Anion Gap

NAGMA (hyperchloremic metabolic acidosis) occurs when serum bicarbonate is lost and replaced milliequivalent-for-milliequivalent by serum chloride, maintaining electroneutrality and preserving an anion gap between $10\text{ and }12\text{ mEq/L}$.

Etiologic Partitioning: GI vs. Renal Loss

• Gastrointestinal Bicarbonate Loss: Severe watery diarrhea, enterocutaneous fistulae, biliary/pancreatic drainage, surgical ureterosigmoidostomy;
• Renal Tubular Acidosis (RTA): Inability of the nephron to excrete net acid or reabsorb filtered bicarbonate;
• Exogenous Chloride Loading: Rapid infusion of large volumes of $0.9\%$ normal saline.

The Urine Anion Gap (UAG)

When the cause of NAGMA is not clinically obvious, calculating the Urine Anion Gap differentiates gastrointestinal bicarbonate loss from renal tubular dysfunction:

$\text{Urine Anion Gap (UAG)} = [\text{U}_{\text{Na}}] + [\text{U}_{\text{K}}] - [\text{U}_{\text{Cl}}]$

Loading diagram...

• Physiological Basis: In metabolic acidosis, healthy kidneys dramatically upregulate ammonium ($\text{NH}_4^+$) excretion (up to $200-300\text{ mEq/day}$). Ammonium is excreted primarily paired with chloride ($\text{NH}_4^+\text{Cl}^-$). Because $\text{NH}_4^+$ is an unmeasured cation, high urinary ammonium forces urinary chloride ($[\text{U}_{\text{Cl}}]$) to rise substantially above the sum of $[\text{U}_{\text{Na}}]$ and $[\text{U}_{\text{K}}]$.
• Negative UAG (typically $-20\text{ to } -50\text{ mEq/L}$): Reflects robust renal ammoniagenesis and excretion of unmeasured $\text{NH}_4^+$. Confirms intact renal response to GI bicarbonate loss (diarrhea).
• Positive UAG (typically $+20\text{ to } +50\text{ mEq/L}$): Demonstrates failure of renal $\text{NH}_4^+$ excretion despite severe systemic acidemia. Confirms Distal Renal Tubular Acidosis (Type 1 RTA) or Type 4 RTA (hypoaldosteronism).

---

The Stewart Physicochemical Approach: Strong Ion Difference (SID)

Traditional acid-base frameworks (the Boston and Copenhagen schools) view bicarbonate and hydrogen ions as primary drivers of acid-base balance. In 1983, Peter Stewart introduced a quantitative, physicochemical framework based on fundamental physical laws of solutions: electroneutrality, conservation of mass, and dissociation equilibria of weak electrolytes.

The Three Independent Variables

Under Stewart's principles, pH and $[\text{HCO}_3^-]$ are dependent variables that change instantaneously and passively in response to alterations in three strictly independent variables:

1. Arterial $\text{PaCO}_2$: The volatile, respiratory independent variable controlled by alveolar ventilation;
2. Strong Ion Difference (SID): The electrical charge difference between fully dissociated strong cations and strong anions in plasma;
3. Total Non-Volatile Weak Acids ($A_{\text{tot}}$): The total concentration of non-dissociated and dissociated weak acids, consisting almost entirely of serum albumin and inorganic phosphate.

Apparent Strong Ion Difference (SIDa)

Strong ions are completely dissociated at physiological pH. SID represents the net positive charge of these ions:

$\text{SIDa} = ([\text{Na}^+] + [\text{K}^+] + [\text{Ca}^{2+}] + [\text{Mg}^{2+}]) - ([\text{Cl}^-] + [\text{Lactate}^-])$

In simplified clinical bedside practice, SID is approximated using the predominant strong ions: $\text{Simplified SID} \approx [\text{Na}^+] - [\text{Cl}^-]$

• Normal Plasma SID: $38\text{ to }42\text{ mEq/L}$ (typically $\approx 40\text{ mEq/L}$);
• Strong Ion Acidosis: A reduction in SID ($< 38\text{ mEq/L}$) forces water molecules to dissociate into $\text{H}^+$ and $\text{OH}^-$ to satisfy electroneutrality, increasing $[\text{H}^+]$ and driving metabolic acidosis. The quintessential cause is hyperchloremia;
• Strong Ion Alkalosis: An elevation in SID ($> 42\text{ mEq/L}$) decreases $[\text{H}^+]$, driving metabolic alkalosis. The quintessential cause is hypochloremia (vomiting, loop diuretics).

The Physicochemical Mechanism of 0.9% Normal Saline Acidosis

Stewart physiology provides the definitive thermodynamic explanation for "dilutional / normal saline acidosis":

• Normal plasma has a sodium of $140\text{ mEq/L}$ and chloride of $100\text{ mEq/L}$, yielding a baseline SID of $40\text{ mEq/L}$.
• Unbalanced crystalloid ($0.9\%$ Normal Saline) contains $154\text{ mEq/L}$ of $\text{Na}^+$ and $154\text{ mEq/L}$ of $\text{Cl}^-$. Its Strong Ion Difference is $0\text{ mEq/L}$ ($154 - 154 = 0$).
• When large volumes of normal saline are infused into a patient, the addition of a solution with $\text{SID} = 0$ dilutes plasma SID toward zero (lowering plasma SID from $40\text{ mEq/L}$ down to $30-32\text{ mEq/L}$). By thermodynamic necessity, this drop in SID drives auto-ionization of water, producing hyperchloremic metabolic acidosis.
• Furthermore, high tubular chloride delivery to the macula densa triggers tubuloglomerular feedback, inducing renal afferent arteriolar vasoconstriction, dropping renal cortical perfusion, and predisposing to acute kidney injury.
• Balanced Crystalloids (Lactated Ringer's, Plasma-Lyte): Contain physiological chloride concentrations ($98-109\text{ mEq/L}$) and contain metabolizable organic anions (lactate, acetate, or gluconate) that are rapidly consumed by the liver and muscle. Their effective SID mirrors plasma ($28-50\text{ mEq/L}$), preserving plasma SID and largely avoiding hyperchloremic acidosis. Clinically, SMART (2018) found a small reduction in the composite of death, new RRT, or persistent kidney dysfunction, but the larger BaSICS (2021) and PLUS (2022) trials found no mortality difference; the 2026 Surviving Sepsis Campaign still suggests balanced crystalloids over 0.9% saline for sepsis. Hypotonic balanced fluids such as lactated Ringer solution are often avoided in traumatic brain injury, based on trial subgroup data rather than the sepsis guideline.

---

Step-by-Step Acid-Base Algorithm and Worked Cases

---

Metabolic Alkalosis: Generation, Maintenance and Urine Chloride

Metabolic alkalosis needs a generating process and a maintaining process, because healthy kidneys excrete excess bicarbonate quickly.

• Generation: loss of acid (vomiting, nasogastric suction, mineralocorticoid excess), an alkali load (sodium bicarbonate, citrate from massive transfusion or regional citrate anticoagulation, acetate in parenteral nutrition), contraction around a fixed bicarbonate pool (loop or thiazide diuretics), and posthypercapnic alkalosis after chronic hypercapnia is corrected too quickly on the ventilator.
• Maintenance: volume and chloride depletion, hypokalemia, ongoing mineralocorticoid activity, and reduced GFR.

Urine sodium is misleading in active vomiting, because bicarbonaturia forces sodium into the urine despite volume depletion. Urine chloride is the better marker.

ICU-specific therapy:

• Acetazolamide ($250\text{--}500\text{ mg}$ IV) increases bicarbonate excretion in volume-overloaded patients who cannot receive saline; it worsens hypokalemia.
• Reduce or stop the citrate load and correct hypokalemia and hypomagnesemia.
• For severe, symptomatic alkalemia (pH above about $7.55$ with arrhythmias or seizures) that does not respond, use dilute hydrochloric acid ($0.1\text{--}0.2\text{ N}$) through a central line, or a low-bicarbonate dialysate.
• On the ventilator, lower minute ventilation gradually in chronic hypercapnia to avoid posthypercapnic alkalosis.

---

Low or Negative Anion Gap

A low anion gap is usually an albumin effect, but it can signal toxic or hematologic disease:

• Hypoalbuminemia: each $1\text{ g/dL}$ fall in albumin lowers the gap by about $2.5\text{ mEq/L}$.
• Cationic paraproteins: IgG multiple myeloma.
• Unmeasured cations: lithium toxicity, and severe hypercalcemia or hypermagnesemia.
• Bromide or iodide exposure: many analyzers read these halides as chloride ("pseudohyperchloremia"), producing a very low or negative gap.
• Laboratory error, including severe hyperlipidemia.

---

Primary Respiratory Disorders: Common ICU Causes

Salicylate toxicity typically combines a primary respiratory alkalosis with a high anion gap metabolic acidosis. Intubating such a patient without matching their high minute ventilation can cause fatal acidemia.