Deciphering Acid-Base Disorders

Deciphering Acid-Base Disorders


AcidBaseDerangements in acid-base status are commonly discovered on routine emergency department evaluation and often suggest the presence of severe underlying disease. Many acute conditions can disrupt homeostatic mechanisms used to buffer and excrete acid, and these changes may necessitate immediate intervention. When you discover a patient with an abnormal pH, what is your approach to the diagnosis?

Large amounts of fixed and volatile acid are produced as normal byproducts of cellular metabolism. In addition to excretion of these byproducts via the lungs and kidneys, physiologic pH is maintained using carbon dioxide and bicarbonate as buffers. The balance of serum PCO2 and HCO3 directly impact pH, and disruption of this ratio results in primary acidemia or alkalemia, depending on the direction of the change. Notably, “acidemia” and “alkalemia” describe changes in serum pH, whereas “acidosis” and “alkalosis” refer to the clinical conditions underlying those changes.

These conditions can be either metabolic or respiratory in etiology, leading to 4 categories of primary acid-base disorders. In response to this primary change, compensatory mechanisms attempt to drive the pH toward normal by keeping the PCO2/HCO3 ratio constant. The degree of compensation is related to the chronicity of the primary disorder, but does not generally restore the pH to normal. To add to the complexity, patients can have more than one primary disorder. Features of each disorder are detailed as follows, which assumes ABGs are used:

Primary Disorder pH Primary Feature Acute Compensatory Change (predicted) Chronic Compensatory Change (predicted) Potential Causes
Metabolic Acidosis < 7.36  ↓ HCO3 ↓ PCO2 by 1.2 mmHg for each 1 mEq ↓ in HCO3 *

AG: toxic ingestion, uremia, renal failure, lactic acidosis**, DKA

Non-AG: RTA, K-sparing diuretics, diarrhea, NS infusion

Metabolic Alkalosis > 7.44 ↑ HCO3 ↑ PCO2 by 0.7 mmHg for each 1 mEq ↑ in HCO3 * Vomiting, contraction alkalosis from diuretic use, hypokalemia, hypomagnesemia
Respiratory Acidosis < 7.36 ↑ PCO2

↑ HCO3 1 mEq for each 10 mmHg ↑ in PCO2

↓ pH 0.08 units*** for each 10 mmHg ↑ in PCO2

↑ HCO3 4 mEq for each 10 mmHg increase in PCO2

↓ pH 0.03*** units for each 10 mmHg increase in PCO2

Hypoventilation (including such causes as CNS lesion, opiate overdose, COPD, diaphragmatic weakness)
Respiratory Alkalosis > 7.44 ↓ PCO2

↓ HCO3 2 mEq for each 10 mmHg ↓ in PCO2

↑ pH of 0.08 units*** for each 10 mmHg ↓ in PCO2

↓ HCO3 5 mEq for each 10 mmHg decrease in PCO2

↑ pH of 0.03*** units for each 10 mmHg decrease in PCO2

Hyperventilation (including such causes as anxiety, CNS lesion, early salicylate toxicity)

AG = anion gap

* Respiratory compensation becomes less effective for chronic metabolic disorders

** There is conflicting data regarding reliable elevation in AG from isolated lactic acidosis

*** For isolated respiratory processes only

Checklist: Evaluating Acid Base Disturbances

Acid-base physiology is complex and dynamic, but when you suspect an acid-base disturbance, the following checklist can guide your evaluation to the correct diagnosis1,2:

1. Collect all the data

  1. Historical features can guide your expectations about acid-base status. Recognizing such conditions as malperfusion, sepsis, COPD, DKA, anxiety or toxic ingestion can place the pH in context, particularly if a patient has a complicated mixed disorder. Prior laboratory values, including a blood gas, can also be helpful.
  2. Laboratory values required for full analysis of an acid-base disorder include serum electrolytes, creatinine, albumin and serum pH. Other adjunctive labs such as lactic acid may be helpful based upon clinical suspicion. With regard to pH, correlation between peripheral venous and arterial blood samples is well established; pH is approximately 0.03-0.05 units lower in peripheral venous samples, whereas PCO2 is 4-5 mmHg higher. This correlation is less reliable in patients with shock or other cardiopulmonary compromise. ABGs provide a direct measurement of pH and PCO2 and may be useful in patients with mixed disorders or hemodynamic instability. In addition, note that the bicarbonate value from a blood gas is extrapolated from the Henderson-Hasselbalch equation and may not accurately represent the serum bicarbonate.

2. Characterize the primary disorder

  • This is done by using pH, bicarbonate, and PCO2.  This assumes a normal blood gas at baseline with pH of 7.4, PCO2 of 40, and biacarbonate of 24.
  1. Acidemia with low bicarbonate indicates a primary metabolic acidosis. If a metabolic acidosis is discovered, calculate the anion gap. Note albumin is an assumed part of a normal AG, so a “normal” AG is decreased for patients with hypoalbuminemia.
  2. Acidemia with high bicarbonate indicates a primary respiratory acidosis.
  3. Alkalemia with a low PCO2 indicates a primary respiratory alkalosis.
  4. Alkalemia with a high PCO2 indicates a primary metabolic alkalosis.
  5. Normal pH in the presence of abnormal bicarbonate or PCO2 indicates the presence of a mixed disorder.

3. Determine degree of compensation

  1. Identify whether the acid-base disorder is simple or mixed. For metabolic processes, compensatory changes can be roughly estimated using 1:1 ratios of PCO2: HCO3.
  2. If compensation is appropriate, the patient has a simple disorder. Patients with an AG metabolic acidosis are a special case, and regardless of seemingly appropriate compensation, may have a second metabolic disorder (see below).
  3. If compensation is not appropriate, the patient has a mixed disorder. Characterizing a mixed disorder is complicated, though some guiding principle are helpful.
    1. Respiratory acidosis and respiratory alkalosis cannot coexist.
    2. For primary metabolic processes, if actual PCO2 is higher than predicted, there is a concomitant respiratory acidosis. If it is lower than predicted, there is a concomitant respiratory alkalosis. Underlying chronic respiratory acidosis can make interpretation of these values challenging.
    3. For primary respiratory processes, if actual HCO3 is higher than predicted, there is a concomitant metabolic alkalosis. If it is lower than predicted, there is a concomitant metabolic acidosis.

For patients with metabolic acidosis, calculate the AG and Delta Gap regardless of compensation

Seemingly normal respiratory compensation can mask mixed metabolic disorders (any combination of AG metabolic acidosis, non-AG metabolic acidosis and metabolic alkalosis) that will be revealed by calculating the “delta-gap” or “delta-delta”.

  • In order to maintain electroneutrality, any elevation in AG should equal the decrease in HCO3. The calculation of ΔAG/ΔHCO3 is referred to as the “delta gap” or “delta-delta”. If the decrease in HCO3 is more than the increase in AG, the patient has both an AG and a non-AG metabolic acidosis.  If the decrease in HCO3 is less than the increase in AG, the patient has both an AG acidosis and a concomitant metabolic alkalosis.

Sample Cases

For the following demonstrative examples, assume normal AG =12, normal PCO2= 40, normal bicarbonate = 24 and normal pH is 7.40. These cases are representative and are not taken from actual patient data.

Case 1

VBG 7.12 / 75 / 43, serum bicarbonate 27

Na 139, Cl 102, Cr 0.9

Collect data: 19M bib EMS for acute altered mental status.

Primary disorder: The patient is acidemic. Elevation in bicarbonate suggests a primary respiratory acidosis.

Determine compensation: Given the history, as well as only slight elevation in the bicarbonate, you suspect this is an acute process. To confirm this, you know that bicarbonate should rise 1 mEq and pH should decrease 0.08 units for each 10 mmHg increase in PCO2.  In this case, using data from a VBG, predicted bicarbonate is 27.5 and predicted pH is 7.12, suggesting appropriate compensation. This would be more precise if actual or extrapolated ABG values (pH 7.16, PCO2 71) were used, but that is not necessary for this case.

Diagnosis: Acute primary respiratory acidosis from narcotic overdose

Case 2

ABG 7.18 / 32 / 104, serum bicarbonate 18

Na 138, Cl 98, Cr 2.1 (normally 0.9), Albumin 3.9

Collect data: 56F with fever and vomiting

Primary disorder: The patient is acidemic. The decreased bicarbonate suggests a primary metabolic acidosis.

Determine compensation: For this decrease in bicarbonate, you would expect the PCO2 to decrease by approximately 7 mmHg. For this case, compensation appears nearly appropriate. Given the metabolic acidosis, you calculate the AG to be 22 and proceed to calculate a “delta-delta”. Here, the delta AG = (measured gap – normal gap) = 10, and delta bicarbonate =  (normal bicarbonate – measured bicarbonate) = 6. The bicarbonate has decreased less than the AG has increased. If the AG metabolic acidosis were the only disorder, you would expect the bicarbonate to be 14. Therefore, this suggests a superimposed metabolic alkalosis.

Diagnosis: Acute primary metabolic acidosis with superimposed primary metabolic alkalosis from urosepsis, renal insufficiency and vomiting

Case 3

ABG 7.12 / 65 / 104, serum bicarbonate 18 (from admission 3 weeks ago: 7.32 / 70, bicarbonate 38)

Na 144, Cl 94 (from admission 3 weeks go: Na 140, Cl 96)

Collect data: 70M with ESRD, COPD and DM presenting with cough. He missed HD today.

Primary disorder: The patient is acidemic. The decreased bicarbonate suggests a primary metabolic acidosis.

Determine compensation: For this decrease in bicarbonate, you would expect the PCO2 to decrease by approximately 7 mmHg. For this ABG in isolation, compensation appears to be inappropriate. However, based on the prior ABG, you determine that the patient has a compensated chronic respiratory acidosis from COPD with baseline bicarbonate of 38. Based on this information, compensation seems appropriate. In addition, given the metabolic acidosis, you calculate the AG to be 32 and proceed to calculate a “delta-delta”. Here, the delta AG = (measured gap – normal gap) = 20, and delta bicarbonate = (normal bicarbonate – measured bicarbonate) = 20, using 38 as the patient’s normal baseline bicarbonate. This is a primary pure AG metabolic acidosis.

Diagnosis: Acute primary metabolic acidosis from DKA and renal failure in the setting of a primary chronic respiratory acidosis from COPD

  • Note: Even if you did not have the prior baseline ABG, you could come to a similar conclusion, diagnosing the metabolic acidosis with inappropriate compensation. The elevated PCO2 would suggest respiratory acidosis. Calculation of delta gap would then suggest a third disorder of metabolic alkalosis. The old ABG confirms that the respiratory acidosis with compensatory metabolic alkalosis is chronic, but the triple disorder can be characterized nonetheless. If a prior ABG were normal and all three disorders were acute, this patient could have respiratory acidosis from hypoventilation in the setting of COPD, as well as infection with ESRD and vomiting (for example).
Adrogué H, Madias N. Secondary responses to altered acid-base status: the rules of engagement. J Am Soc Nephrol. 2010;21(6):920-923. [PubMed]
Adrogué H, Rashad M, Gorin A, Yacoub J, Madias N. Assessing acid-base status in circulatory failure. Differences between arterial and central venous blood. N Engl J Med. 1989;320(20):1312-1316. [PubMed]

Expert Peer Review

Acid-base analysis is a critical skill for Emergency Physicians but one that can only be mastered through deliberate practice. Every EM conference has a “Making Acid-Base Interpretation Easy” lecture but there’s truly no secret to it. So Todd is tackling a difficult topic here but does it extremely well. This is a great overview of acid-base basics based on Henderson-Hasselbalch approach. This is the system we all learned in medical school and the system most EPs use when they work. However, in recent years, EM critical care docs have challenged the utility of Henderson-Hasselbalch. Many critical care guys instead use the Strong Ion Theory (or Stewart approach) for acid-base interpretation. A comprehensive discussion of Strong Ion Theory is beyond the scope of this peer review. I would refer any readers who are interested in learning this technique to Scott Weingart’s EMCrit blog (episodes 44-46, 50 and 96) and to the referenced articles.

 A brief historical perspective is important. When the Henderson-Hasselbalch approach became the standard, we had more limited diagnostic capacity. It was difficult to measure things like albumin and lactate. Thus, there was created a “normal anion gap” which basically reflects albumin and lactate. Any gap larger than the normal and the clinician had to start looking for things to fill it (i.e. ketones, toxic alcohols). Fast forward a couple decades and we find that we routinely, and rapidly measure albumin and lactate. The Strong Ion Theory accounts for this. The main tenant here that is different is that it’s not bicarbonate that determines the pH. Those are dependent variables. It’s the strong ions (ions that dissociate fully in water like Na and Cl) that are important in the acid-base determination.

Does it really matter? This is debatable. Scott argues that it does but that if you are factoring in albumin while doing Henderson-Hasselbalch, your analysis will be quite similar. Dubin et al found that the Strong Ion Theory was superior to Henderson-Hasselbalch in that it identified a substantial number of patients with metabolic disturbances that were missed by the traditional method. It did not, however, offer a diagnostic or prognostic advantage.

The bottom line is that learning acid-base physiology is both central to EM practice and difficult. Deliberate practice is necessary for mastery.  Why not learn both techniques and see if it makes a difference yourself?\"


  • Kaplan LJ, Frangos S. Clinical review: Acid-base abnormalities in the intensive care unit. Critical Care 2005; 9(2): 198-203. PMID 15774078
  • Dubin A, Menises MM, Masevicius FD, Moseinco MC et al. Comparison of three different methods of evaluation of metabolic acid-base disorders. Crit Care Med 2007; 35: 1264-1270. PMID:17334252 
Anand Swaminathan, MD MPH
Assistant Residency Director and Assistant Professor of Emergency Medicine, Bellevue/NYU ; Faculty Editor of EM Lyceum

Todd A. Seigel, MD

Todd A. Seigel, MD

ALiEM Featured Contributor
Clinical Fellow in Critical Care Medicine
University of California, San Francisco (UCSF)
Todd A. Seigel, MD

Latest posts by Todd A. Seigel, MD (see all)


    hi i ask if you call delta Gap is the same as delta ratio ,
    one case i need interpretation :
    53 yof with crohn’s disease presented with diarrhea vomiting and fever
    pH 7.36
    PCO2 22
    HCO3 14
    Na 147
    CL 121

    • Todd Seigel

      Thanks for your comment!

      When I refer to the “delta-gap” or the “delta-delta”, I am ultimately referring to the ratio of ΔAG/ΔHCO3. Broadly speaking, a ratio 1 suggests an AG acidosis with concomitant metabolic alkalosis.

      For the case you have provided, I interpret the data such the patient has a primary metabolic acidosis with concomitant primary respiratory alkalosis. The AG by my calculation is 12, which to me suggests that the metabolic acidosis is entirely non-gap. I would not invoke the ratio as the AG is normal. Intuitively, this makes sense given diarrhea; in addition, given the values for Na/Cl, I may suspect a hyperchloremic metabolic acidosis secondary to resuscitation.

      • Dr RAFID IMRAN


        • Todd Seigel

          There is an inverse relationship between chloride and bicarbonate, particularly with exogenous chloride administration, and this part of the reason patients resuscitated with large amounts of NS will develop hyperchloremic metabolic acidosis. Outside of that, because chloride is accounted for in the anion gap, I am not certain a reliable 1:1 correlation can be made between them.

  • Shirah

    Use and love the EMCrit method/ stewart approach. I tried the memorize-the-values approach, but it was too easy for me to mix up numbers with a Real Sick patient! Love the Stewart approach; much more intuitive and I can suss out the underlying vs compensetory values along the way.

    • Michelle

      Excellent point. Here’s Scott’s link on the Stewart method:

      The key is to have a methodical, consistent approach to whenever you analyze the ABG. Thanks for the comment.

    • Todd Seigel

      Thanks for bringing that up. Though I have focused on the traditional method of evaluating acid-base disturbance, the Stewart model is another way to characterize these disorders based upon the concepts of strong ion difference, the total plasma concentration of nonvolatile acid, and PaCO2. Notably, another premise of the Stewart model is that serum bicarbonate does not alter blood pH, but rather is a dependent variable in a system modulated by changes in the variables above. I find pros and cons to each method, and data has not demonstrated benefit to one method over the other. EMCrit has a great podcast on the Stewart approach that can be found at

  • Dr Abd Elrhman Ali

    Hi , can I ask what is the least well compensated acid base disorder?