Blood Gas Calculator for Dogs & Cats Veterinary Acid-Base Interpreter
pH, pCO₂, HCO₃⁻, base excess, compensation, mixed and triple disorders, anion gap, delta ratio, semiquantitative Stewart analysis and strong ion gap
Which patient, and which sample?
Species and sample type set the reference intervals every value is judged against. Chronicity chooses the acute or chronic respiratory compensation rule.
Enter the blood gas values
pH and pCO₂ are required. Everything else unlocks a deeper read: Na⁺/Cl⁻ for anion gap and delta ratio, plus albumin, phosphate and lactate for the semiquantitative Stewart panel.
Reference ranges, lab normals & formulas
Acid-base interpretation
The verdict, how it was read, compensation, anion gap and the Stewart panel — rebuilt on every Interpret.
Your interpretation will appear here
Enter a pH and pCO₂, then press Interpret — add electrolytes, albumin, phosphate and lactate for the full Stewart read.
Clinical tool for veterinary professionals, not a diagnosis. Traditional interpretation uses published dog and cat reference intervals and DiBartola/de Morais compensation rules, which are best validated for arterial samples in dogs. Cats often compensate less predictably and venous pCO₂/HCO₃⁻ run higher than arterial, so those results are flagged as guides. The semiquantitative panel uses the Hopper & Haskins (JVECC 2008) dog and cat constants; edit the lab normals to match your own analyser. Strong ion gap uses Figge-Fencl and Constable equations and is shown for reference. Reference intervals vary by analyser and population. Blood gas numbers never replace clinical assessment, history and trend. Final decisions rest with the attending veterinarian.
How to interpret a blood gas in a dog or cat (the method this calculator uses)
Blood gas interpretation in small animals goes wrong in two predictable ways: calling a compensating patient "mixed", and missing a second disorder because the pH looks reassuring. This tool follows the same six steps an ECC clinician would, in the same order, and shows its working so you can disagree with it.
- Look at pH. Below the species and sample reference interval is acidaemia, above it is alkalaemia. A normal pH does not mean a normal patient: opposing disorders can cancel each other.
- Find the primary process. Match the pH direction to the variable that explains it. Low pH with low HCO₃⁻ is metabolic acidosis; low pH with high pCO₂ is respiratory acidosis. High pH with high HCO₃⁻ is metabolic alkalosis; high pH with low pCO₂ is respiratory alkalosis. If both variables push the same way, two primary disorders are additive.
- Check compensation. The body never over-compensates. For a metabolic disorder the lungs adjust pCO₂ by about 0.7 mmHg for every 1 mmol/L change in HCO₃⁻. For a respiratory disorder the kidneys adjust HCO₃⁻ by 0.15 (acute) or 0.35 (chronic) mmol/L per mmHg of pCO₂ rise, and by 0.25 (acute) or 0.55 (chronic) per mmHg of fall. Measured values outside the expected band mean a second disorder is present.
- Calculate the anion gap and correct it for albumin. AG = (Na⁺ + K⁺) − (Cl⁻ + HCO₃⁻). Every 1 g/dL of albumin below normal lowers the gap by roughly 3.7 to 4.2 mmol/L in dogs and cats, so hypoalbuminaemic patients hide high-gap acidoses unless you correct.
- Run the delta ratio when the gap is high. ΔAG divided by ΔHCO₃⁻ tells you whether the added anions fully explain the bicarbonate fall (ratio 0.8 to 2), whether a hyperchloraemic acidosis is riding alongside (ratio below 0.8), or whether a metabolic alkalosis is hiding underneath (ratio above 2). Corrected HCO₃⁻ = HCO₃⁻ + ΔAG shows the same thing as a number you can compare with the reference range.
- Partition the base excess (semiquantitative Stewart analysis). Using sodium, chloride, albumin, phosphate and lactate, standardised base excess is split into a free water effect, a chloride effect, an albumin effect, a phosphate effect and a lactate effect. Whatever is left over is the unmeasured anion effect (XA): ketones, uraemic anions, ethylene glycol metabolites, salicylate. This is where hyperchloraemic acidosis after saline resuscitation, or a lactic acidosis masked by hypoalbuminaemia, becomes visible when the traditional read says "mild" or "normal".
Normal blood gas values for dogs and cats
Typical published intervals for arterial and venous samples. Your analyser's own intervals take precedence, and the semiquantitative lab normals are editable inside the tool.
| Parameter | Dog arterial | Dog venous | Cat arterial | Cat venous |
|---|---|---|---|---|
| pH | 7.35–7.46 | 7.32–7.42 | 7.33–7.45 | 7.28–7.41 |
| pCO₂ (mmHg) | 32–43 | 38–46 | 26–36 | 33–45 |
| HCO₃⁻ (mmol/L) | 18–26 | 20–26 | 14–22 | 17–24 |
| Base excess (mmol/L) | −4 to +4 | −4 to +4 | −5 to +2 | −5 to +2 |
| PO₂ (mmHg) | 80–105 | 40–60 | 95–120 | 35–55 |
| Anion gap (mmol/L) | 12–24 | 13–27 | ||
Blood gas units: kPa vs mmHg
Analysers in the UK, Europe, Hong Kong and much of Asia report pCO₂ and pO₂ in kilopascals; North American and Australian machines and most textbooks use mmHg. Multiply kPa by 7.5 to get mmHg. The calculator has a mmHg/kPa toggle next to pCO₂ that converts both partial pressures and their reference ranges, and it will refuse a pCO₂ under 10 mmHg because that is almost always kPa typed into a mmHg field.
| Value | mmHg | kPa |
|---|---|---|
| Dog arterial pCO₂ range | 32–43 | 4.3–5.7 |
| Cat arterial pCO₂ range | 26–36 | 3.5–4.8 |
| Dog venous pCO₂ range | 38–46 | 5.1–6.1 |
| Cat venous pCO₂ range | 33–45 | 4.4–6.0 |
| Dog arterial PaO₂ range (room air) | 80–105 | 10.7–14.0 |
| Hypoxaemia threshold PaO₂ | < 80 | < 10.7 |
| Severe hypoxaemia PaO₂ | < 60 | < 8.0 |
Venous vs arterial blood gas in dogs and cats
A venous sample answers the acid-base question in almost every patient and is far easier to collect from a jugular or cephalic vein. Arterial sampling (dorsal pedal, femoral, auricular in larger dogs) is needed to assess oxygenation. Expected differences in a well-perfused animal are a venous pH about 0.03 to 0.05 lower, pCO₂ about 4 to 6 mmHg higher, and HCO₃⁻ about 1 to 2 mmol/L higher than arterial. Those offsets are built into the venous reference intervals above. Two cautions: in shock the arteriovenous CO₂ gap widens (a venous-arterial pCO₂ difference above about 6 mmHg is itself a marker of poor flow), and a venous PO₂ tells you about tissue extraction, not lung function. A "venous" PO₂ above 60 mmHg usually means an accidental arterial stick, which also invalidates the venous reference intervals for that sample.
Typical acid-base patterns by disease in dogs and cats
Starting points, not rules. Real cases mix and the calculator will show which components dominate.
| Condition | Usual pattern | What the Stewart panel typically shows |
|---|---|---|
| Diabetic ketoacidosis | High-AG metabolic acidosis; hidden metabolic alkalosis if vomiting; delta ratio often above 2 | Large negative XA (ketones), positive chloride effect from vomiting, positive albumin effect if hypoalbuminaemic |
| Feline urethral obstruction | High-AG metabolic acidosis, hyperkalaemia; respiratory compensation often blunted | Negative XA (uraemic anions), negative phosphate effect, lactate if hypoperfused |
| GDV / haemorrhagic shock | Lactic (high-AG) metabolic acidosis; respiratory alkalosis early, respiratory acidosis if decompensating | Negative lactate effect dominates; XA near zero once lactate entered |
| Gastric outflow obstruction / profuse vomiting | Hypochloraemic, hypokalaemic metabolic alkalosis; paradoxical aciduria | Large positive chloride effect; free water effect if hyponatraemic |
| Small-intestinal diarrhoea (e.g. parvovirus) | Normal-AG (hyperchloraemic) metabolic acidosis; delta ratio below 0.8 if lactate also raised | Negative chloride effect; positive albumin effect (protein loss) masking severity |
| Hypoadrenocorticism | Normal-AG metabolic acidosis with hyponatraemia and hyperkalaemia | Negative free water effect (dilution), mild negative chloride effect |
| Chronic kidney disease | Mild to moderate high-AG metabolic acidosis, usually compensated | Negative phosphate effect, negative XA (uraemic anions) |
| Large-volume 0.9% NaCl resuscitation | Normal-AG hyperchloraemic acidosis appearing after fluids | Negative chloride effect out of proportion to everything else |
| Hypoalbuminaemia (PLE, PLN, hepatic failure, sepsis) | Metabolic alkalosis on Stewart analysis; traditional read often "normal" | Positive albumin effect hiding concurrent chloride, lactate or XA acidoses |
| Brachycephalic obstructive airway crisis, laryngeal paralysis | Acute respiratory acidosis with hypoxaemia; metabolic acidosis if collapsed and hyperthermic | Lactate effect if struggling; otherwise Stewart panel unremarkable |
| Anaesthesia with spontaneous ventilation | Acute respiratory acidosis (hypoventilation), pCO₂ often 50 to 60 mmHg | Unremarkable unless fluids or pre-existing disease |
| Sepsis | Lactic acidosis with respiratory alkalosis; pH can be normal (offsetting mixed disorder) | Negative lactate and XA effects, positive albumin effect |
| Ethylene glycol toxicity | Severe high-AG metabolic acidosis with raised osmolar gap | Very negative XA that lactate does not explain |
| Furosemide therapy (congestive heart failure) | Hypochloraemic, hypokalaemic metabolic alkalosis | Positive chloride effect |
Sample handling and pre-analytical errors
The commonest "mixed disorder" is a bad sample. Air bubbles in the syringe let CO₂ escape and oxygen in, lowering pCO₂ and raising PO₂: expel air immediately and cap the syringe. Excess liquid heparin dilutes the sample and lowers pCO₂ and HCO₃⁻; use dry lithium-heparin syringes or a barely-coated hub. Delay lets cells metabolise: pCO₂ and lactate rise and pH falls, noticeably after about 15 minutes at room temperature, sooner with a high white cell count or polycythaemia; run it promptly or store on ice up to 30 minutes. A struggling, panting patient hyperventilates during collection and hands you a false respiratory alkalosis; note it and, in cats especially, interpret a low pCO₂ with that in mind. Prolonged venous stasis from a tight tourniquet raises lactate and potassium locally. Temperature correction matters little for acid-base decisions in most patients; the calculator assumes values reported at 37 °C. Finally, chemistry-analyser bicarbonate (TCO₂) from a separate serum tube can differ from the blood-gas HCO₃⁻ by several mmol/L; the calculator warns when the HCO₃⁻ you enter disagrees with the value implied by pH and pCO₂.
Semiquantitative constants (Hopper & Haskins, JVECC 2008)
| Effect (mEq/L) | Dog | Cat |
|---|---|---|
| Free water | 0.25 × (Na⁺ − 146) | 0.22 × (Na⁺ − 156) |
| Chloride | 110 − Cl⁻corr, where Cl⁻corr = Cl⁻ × 146 / Na⁺ | 120 − Cl⁻corr, where Cl⁻corr = Cl⁻ × 156 / Na⁺ |
| Albumin | 3.7 × (3.1 − albumin g/dL) | 3.7 × (3.1 − albumin g/dL) |
| Phosphate | 0.58 × (3.9 − phosphate mg/dL) | 0.58 × (5.0 − phosphate mg/dL) |
| Lactate | −1 × lactate mmol/L | −1 × lactate mmol/L |
| Unmeasured anions (XA) | Standardised base excess − sum of the five effects. A value below about −5 mEq/L means unmeasured anions are present. | |
Expected compensation in dogs (DiBartola / de Morais rules)
These are the canine rules the calculator applies. Feline data are limited and cats often under-compensate, so the tool flags feline results as guides.
| Primary disorder | Expected compensation | Tolerance used |
|---|---|---|
| Metabolic acidosis | pCO₂ falls 0.7 mmHg per 1 mmol/L fall in HCO₃⁻ | ±3 mmHg |
| Metabolic alkalosis | pCO₂ rises 0.7 mmHg per 1 mmol/L rise in HCO₃⁻ | ±3 mmHg |
| Acute respiratory acidosis | HCO₃⁻ rises 0.15 mmol/L per 1 mmHg rise in pCO₂ | ±2 mmol/L |
| Chronic respiratory acidosis | HCO₃⁻ rises 0.35 mmol/L per 1 mmHg rise in pCO₂ | ±2 mmol/L |
| Acute respiratory alkalosis | HCO₃⁻ falls 0.25 mmol/L per 1 mmHg fall in pCO₂ | ±2 mmol/L |
| Chronic respiratory alkalosis | HCO₃⁻ falls 0.55 mmol/L per 1 mmHg fall in pCO₂ | ±2 mmol/L |
Worked examples
Three cases you can type into the calculator to see the full read. Values are illustrative.
1. Dog, DKA with vomiting (arterial)
Acidaemia with low HCO₃⁻: metabolic acidosis. Expected pCO₂ 31 (28 to 34), measured 30, so respiratory compensation is appropriate. Albumin-corrected anion gap 39, delta ratio 2.6 and corrected HCO₃⁻ 35 reveal a hidden metabolic alkalosis from vomiting. The Stewart panel shows unmeasured anions of about −22 mEq/L (ketones) offset by a chloride effect of +13 and an albumin effect of +4. Traditional read alone would have called this a moderate acidosis; the numbers say the ketoacid load is severe and being masked.
2. Dog, post-resuscitation with 0.9% NaCl, hypoalbuminaemic (venous)
Every traditional value is inside the reference interval, so the verdict is "normal". The Stewart panel disagrees: chloride effect −10 (hyperchloraemic acidosis from saline) is being cancelled by an albumin effect of +6. Two real processes, invisible on pH and bicarbonate. Practical consequence: change to a balanced crystalloid.
3. Cat, urethral obstruction (venous)
Severe acidaemia with a high anion gap metabolic acidosis (uraemic anions, phosphate, lactate). Expected pCO₂ 33 (30 to 36); measured 40 is above the band, so the calculator flags a concurrent respiratory acidosis but reminds you that cats often fail to hyperventilate in response to metabolic acidosis. With K⁺ 8.5 the priority is the potassium, not the bicarbonate.
Common mistakes when interpreting blood gases in dogs and cats
- Using human reference intervals. Cats run a lower pCO₂ and HCO₃⁻ than dogs and humans; a feline HCO₃⁻ of 16 mmol/L can be normal. Human intervals over-diagnose feline acidosis.
- Ignoring the sample type. Venous pCO₂ is 4 to 6 mmHg higher than arterial. Judging a venous gas against arterial ranges manufactures a respiratory acidosis.
- Calling a compensating patient "mixed". Compensation pushes the counter-variable outside its reference interval on purpose. Compare it with the expected value, not the reference range.
- Not correcting the anion gap for albumin. A dog with albumin 1.8 g/dL and an anion gap of 20 has a corrected gap of about 25: a high-gap acidosis that would otherwise be missed.
- Treating the base excess as one number. SBE is the sum of five or six effects. A "mild" SBE of −3 can hide a −12 lactate effect and a +9 albumin effect.
- Chasing the bicarbonate with bicarbonate. Lactic acidosis needs perfusion, DKA needs insulin and fluids, hyperchloraemic acidosis needs a fluid change. Sodium bicarbonate is a last resort with specific indications.
- Forgetting chronicity. The acute and chronic respiratory rules differ more than two-fold. Choosing the wrong one creates or hides a mixed disorder.
Glossary
- Standardised base excess (SBE)
- The amount of strong acid or base needed to return extracellular fluid to pH 7.4 at pCO₂ 40 mmHg. Negative values indicate metabolic acidosis, positive values metabolic alkalosis. Reported by most analysers; calculated by the Van Slyke equation if not entered.
- Anion gap (AG)
- (Na⁺ + K⁺) − (Cl⁻ + HCO₃⁻). Estimates unmeasured anions but is lowered by hypoalbuminaemia and hypophosphataemia.
- Albumin-corrected anion gap
- AG adjusted upward by about 3.7 to 4.2 mmol/L for every 1 g/dL of albumin below normal, so hypoalbuminaemic patients are not under-called.
- Delta ratio
- ΔAG / ΔHCO₃⁻. Detects a second metabolic disorder co-existing with a high-anion-gap acidosis.
- Strong ion difference, apparent (SIDa)
- Sum of strong cations minus strong anions: Na⁺ + K⁺ (+ Ca₂⁺ + Mg₂⁺) − Cl⁻ − lactate. Determines pH together with pCO₂ and weak acids.
- Strong ion difference, effective (SIDe)
- HCO₃⁻ plus the charge carried by albumin and phosphate at the measured pH.
- Strong ion gap (SIG)
- SIDa − SIDe. Positive values indicate unmeasured strong anions such as ketones or ethylene glycol metabolites, corrected for albumin and phosphate.
- Total weak acids (Atot)
- Albumin plus phosphate, the non-volatile weak acids. Falling Atot (hypoalbuminaemia) is alkalinising; rising Atot (hyperphosphataemia) is acidifying.
- Free water effect
- The alkalinising effect of hypernatraemia (concentration) or acidifying effect of hyponatraemia (dilution) on the strong ion difference.
- Chloride effect
- The acidifying effect of hyperchloraemia or alkalinising effect of hypochloraemia after correcting chloride for the free water change.
- Unmeasured anion effect (XA)
- What remains of SBE after subtracting the free water, chloride, albumin, phosphate and lactate effects. Values below about −5 mEq/L indicate unmeasured anions.
- P:F ratio
- PaO₂ divided by FiO₂ (as a fraction). Below 300 indicates acute lung injury range, below 200 ARDS range, when clinically consistent.
Frequently asked questions about blood gas interpretation in dogs and cats
Can I use a venous blood gas instead of arterial in a dog or cat?
For acid-base status, yes. Venous pH runs about 0.03 to 0.05 lower and pCO₂ about 4 to 6 mmHg higher than arterial in a well-perfused patient, and the venous reference intervals in this calculator already account for that. Venous samples cannot assess oxygenation, and the arteriovenous gap widens in shock, so an arterial sample is still preferred for respiratory patients and for anyone with poor perfusion.
Why does the calculator flag "concurrent respiratory acidosis" in a cat that seems fine?
The compensation rules are derived from dogs. Cats often show blunted respiratory compensation for a metabolic acidosis, so a feline pCO₂ sitting slightly above the canine expected band may be normal feline physiology rather than a second disorder. The tool tells you this whenever cat is selected. Use the flag as a prompt to look at the patient, not as a diagnosis.
What is the difference between the anion gap and the strong ion gap?
The anion gap counts unmeasured anions but is distorted by albumin and phosphate, which are themselves anions. The strong ion gap (and the simpler semiquantitative XA used here) corrects for both, so it detects added anions such as ketones or ethylene glycol metabolites even in a hypoalbuminaemic patient whose conventional anion gap looks normal.
Why does albumin matter so much in acid-base interpretation?
Albumin is the main weak acid in plasma. A drop of 1 g/dL lowers the anion gap by roughly 3.7 to 4.2 mmol/L in dogs and cats and produces a metabolic alkalosis in its own right. Hypoalbuminaemic patients therefore hide acidoses: the pH and bicarbonate look acceptable while lactate or ketones are climbing. Entering albumin lets the tool correct the gap and quantify the effect.
What is a delta ratio and when should I use it?
The delta ratio compares how much the anion gap has risen (ΔAG) with how much bicarbonate has fallen (ΔHCO₃⁻). In a pure high-gap acidosis the two roughly match (ratio 0.8 to 2). A ratio below 0.8 means a hyperchloraemic acidosis is also present, for example diarrhoea or saline loading on top of DKA. A ratio above 2 means bicarbonate has not fallen as much as it should have, so a metabolic alkalosis, usually vomiting, is hiding underneath the acidosis. Use it whenever the anion gap is high.
Should I give bicarbonate for a metabolic acidosis?
Rarely, and never on the number alone. Lactic acidosis and DKA resolve with perfusion and insulin; bicarbonate generates CO₂ that a hypoventilating patient cannot exhale, shifts potassium and the oxygen dissociation curve, and can worsen intracellular acidosis. Consider it only for severe non-anion-gap acidosis (pH below about 7.1 to 7.2 with adequate ventilation) or specific indications such as hyperkalaemia, and re-check the gas after every step.
Why is a normal pH sometimes labelled a mixed disorder?
Compensation never returns pH fully to normal. When bicarbonate and pCO₂ are both clearly abnormal in opposite directions and the pH is normal, two primary disorders are cancelling each other, for example sepsis with lactic acidosis plus hyperventilation, or a vomiting patient who is also hypoventilating.
What is a hyperchloraemic acidosis and how do I recognise it?
An acidosis driven by excess chloride relative to sodium: the strong ion difference narrows and bicarbonate falls with a normal anion gap. Common causes are large volumes of 0.9% NaCl or hypertonic saline, diarrhoea, renal tubular acidosis and hypoadrenocorticism. The chloride effect in the semiquantitative panel makes it explicit; the practical fix is often just switching to a balanced crystalloid.
Built for real clinical use by ASK A VET™, the animal health platform. Talk. Track. Trust.
