A Really Tricky ABG?

aka Metabolic Muddle 007.1

A 30 year-old female was admitted to the ICU following a traumatic brain injury.  Her ABG was normal on arrival. Since admission her condition has deteriorated and she now has circulatory failure requiring 100mcg/min of noradrenaline IV.

This is her arterial blood gas now:


At this point we usually go through the case in Q&A fashion, with answers provided. Not this time folks!

Your challenge is to provide the most complete description and interpretation of the arterial blood gas data provided, taking the (sparse) clinical context into account. Use Occam’s Razor liberally, but don’t forget Hickam’s Dictum either!

The best (and earliest) answer — if worthy — will earn the answeree F.UCEM status complete with a virtual diploma… Take it away LITFLers! (You have 2 weeks 1 week from today… I changed it after seeing comments appear at a rate of 2 per hour over the first 10 hours!)

Addendum 7/8/2012:

The deadline for F.UCEM status is now closed.. the Judges are conferring!

Print Friendly


  1. monique says

    She has a low-normal anion gap acidosis (AG=6), most likely secondary to excessive resuscitation with normal saline IV therapy (severely hyperchloremic). In addition, she is not compensating adequately for her metabolic acidosis -- her expected CO2 is 25 -- is she ventilated? if so, she is being under ventilated. If not, then she may be hypoventilating due to reduced GCS due to progression of head trauma. Diuretic therapy and renal tubular acidosis may also be causing her normal anion gap acidosis. She may be lactemic secondary to poor perfusion/renal failure/shock. Her calcium is also low -- did she receive massive transfusion? It should be corrected to help her cardiac contractility. She needs correction of her respiratory acidosis by modifying ventilatory parameters, which should improve her acid-base status, and hence also improve her cardiac function and circulatory state. Avoid further normal saline or fluids with high chloride content. Underlying causes of lactemia/acidosis should be sought such as sepsis/concurrent toxin ingestion.

    • Do you really need to know? says

      Have you considered the effects of a low PaCO2 on the patients brain. If this patient is ventilated, which I strongly believe she is, one of the goals for neuro protection is to maintain a PaCO2 between 35-45. Therfore it would be better to correct the acidosis with continuous dialysis

  2. Jake says

    Metabolic acidosis with no respiratory compensation. Lack of respiratory compensation due to the ventilator settings, maybe dial up the RR?

    Primary component of the acidosis is the low SID (17); given the high Na + Cl this could be due to the fluid resuscitation as the others said, but I figured it would be more likely secondary to hypertonic saline given to reduce ICP. She also has a lactic component to the acidosis which is most likely due to shock, but could also be from seizures secondary to the TBI.

    Not sure about the K or the Ca… hypokalaemia maybe from catecholamine load or steroids… No idea about the calcium… massive transfusion?

  3. JarradHall says

    As a microbiologist, I’m going to take a wild stab.

    Damage is to the medulla oblongata and pituitary gland. This has altered breathing patterns, perhaps hyper-oxygenated?

    Pituitary gland has stimulated the adrenals altering kidney function. Potassium is being lost, sodium and chloride retained. Lactate as suggested from the kidneys, perhaps retention?

    Perhaps via similar mechanism affecting the thyroid causing calcium redeposition in bone? Low calcium and potassium contributing to heart problems.

  4. Peter Jemley says

    Acute renal failure as a consequence of severe head trauma. Acidemia shown by ABG is the result of renal failure, mimicking metabolic acidosis, instead of acidosis as a result of decreased pulmonary function, as O2 saturation is fine. Trauma of this type which results in acute renal failure has a dismal prognosis. Peritoneal dialysis is recommended and is best chance for regaining renal function.

  5. Andre Keeweedoc says

    Acidosis from multiple causes. Firstly hyperchloremia causing acidosis likley from fluid resus. Lactic acidosis secondary to poor perfusion with shock plus norad. her expected CO2 is about 25is but is being underventilated. wonder if cause of hypocalemia is dilutional did they recieve other fluids blood ect?
    Hickam shaves his balls with occams razor.

  6. MSQ says

    Grand mal seizure post head injury, causing lactic acidosis. Intracellular movement of lactic acid in exchange for chloride = hyperchloraemic metabolic acidosis

  7. Erin says

    Metabolic acidosis due to rapid infusion of sodium chloride to Mx MAP for existing HI. Likely dilution of HCO3 and infusion of Cl- resulting in net loss of base through decreased reabsorption in renal tubules. As mentioned above hyperchloremic acidosis.

  8. Mel says

    Its a normal anion gap (severe) metabolic acidosis with hyperchloremia,hypokalemia, minimally high sodium ( and hypocalcemia). Givent the context of traumatic brain injury and the patient now being on ionotropes, it is likely that the patient is producing large amounts of insipid urine and getting normal saline (large amounts) as the resus fluid given her TBI. The bicarb and potassium is possibly being lost with the diuresis. She could have other underlying reasons (preexistant RTA), could be losing it from diarrhoea etc.

  9. Kevin says

    NAG acidosis. In ICU setting it’s usually from aggressive fluid resuscitation with crystalloid but you always have to consider RTA and GI tract losses of bicarb. Yes as others have mentioned her PCO2 is not compensated, but with head trauma she’s likely on a ventilator and they are avoiding hyperventilation given her brain injury.

    She is hypernatremic which suggests she’s losing hypotonic fluids through GI tract (would go along with NAG acidosis), or due to brain injury she could have neurogenic diabetes insipidus requring the excessive amount of IV fluids and causing hypovolemic shock. So if I had to pin it all down on one thing I would say it is diabetes insipidus causing hypovolemic shock and the subsequent aggressive fluid resuscitation caused the non-anion gap acidosis.

    • Stephen Dick says

      I would check the patient’s Mg and Phosphate:

      Hypomagnesaemia is possible.
      Hyperphosphataemia (secondary to critical illness) with resultant hypocalcaemia (phosphate binds calcium).

      If they’d needed a transfusion then the citrate found in the blood could chelate the calcium also.

  10. Stephen Dick says

    Tricky indeed!

    Firstly, the pH of 7.1 indicates acidaemia and therefore an acidosis(es) of some description.

    The HCO3 is low, indicating the likely presence of a metabolic acidosis. If we then calculate the anion gap:
    [Na + K] -- [Cl + HCO3] = [146 + 3] -- [129+11] = 149 -- 140 = 9

    This is a normal anion gap metabolic acidosis -- or a hyperchloraemic metabolic acidosis.

    Using the Boston Rule to calculate the expected PCO2, we find:

    Expected PCO2 = 1.5 x [HCO3-] + 8 = 1.5 x 11 + 8 -- 24.5.

    The PCO2 of 35 is higher than expected, indicating the presence of a respiratory acidosis also -- i.e. two primary processes, I think.

    So we have a NAGMA and a concordant respiratory acidosis.

    Other results:
    Low calcium probably consistent with critical illness.
    High lactate is somewhat interesting -- could it be because of tissue hypoperfusion secondary to circulatory failure, as indicated in the vignette? Likely. The normal anion gap means that this is not likely to be the primary cause of the acidaemia, though, as lactic acidosis is typically a high anion gap acidosis.
    The hypokalaemia is a bit of a mystery, also. Redistribution into the cells is unlikely in the presence of significant acidemia; the normal BSL shown makes insulin administration certainly possible. Perhaps there is K+ depletion, either in the kidney (likely, perhaps secondary to acute kidney injury or even activation of the RAS to support circulation) or in the gut (less likely).

    Going back to the acid/base question -- there is a NAGMA with respiratory acidosis also.

    Of the known causes of NAGMA, extra chloride is the most likely given the serum Cl- and what we know about the case, I would assume the use of crystalloid in the treatment of circulatory failure, which could account for the drop in HCO3. Most of the others -- e.g fistulae, carbonic anhydrase inhibitors, diarrhoea -- are probably not relevant here. I suppose renal tubular acidosis could be present, especially with hypokalaemia, but not likely given the case vignette given.

    Respiratory acidosis is certainly compatible with CNS injury, traumatic in this setting, with subsequent hypoventilation. The possibility of impaired gas exhange also exists, but unfortunately we’re not told what the patient’s FiO2 is (I would assume above 21%), so it makes it difficult to use the alveolar gas equation to calculate the A-a and the money would be on CNS depression of the respiratory centre.

    I would suggest that this is a Normal Anion Gap Metabolic Acidosis and concordant respiratory acidosis. This is somewhat supported by calculation of the Delta Ratio:
    DR = increase in anion gap / decrease in HCO3- = 0/11 = 0.

    A delta ratio of <0.4 is highly suggestive of a hyperchloraemic metabolic acidosis.

    • Brian says

      Is it valid to use the delta ratio in a low-normal anion gap? I thought it tended to be restricted to HAGMAs. Just curious

      • Stephen Dick says

        That’s a fair point, actually.


        Some websites that I looked at say that it’s to be used for HAGMAs only to detect the presence of an additional acid/base disorder.

        The website above mentions for the use in metabolic acidoses (anion gap not specified) and from what I can gather from the resources given again it’s not specified.

        I suppose it does support the presence of a NAGMA here but we kind of already know that anyway by virtue of the elevated chloride.

        • says

          It is in acid-base problems like this one that a quantitative (Stewart) approach becomes particularly appealing. The delta ratio is low, so there is no high anion gap acidosis??? Except we have a lactate of 4.9, absolute evidence of a anion gap acidosis. Hend-Hass acid base works extremely well if…
          You calculate the actual anion gap using albumin (and not the arbirtary 12)
          You always perform a delta gap (not delta ratio)

          If instead, we look at this quantitatively then we can say
          SID is low at 17 which fills the base excess of -17 with 4 mmol of unaccounted for base, lactate adds another 4.9 of unaccounted for base, so 9 mmol so far of base we can find.

          A low albumin may account for some of this, but it would have to be near zero to account for all of it. Other causes of unmeasured cations are: Hypercalcemia, Hypermagnesemia, Hyperkalemia, Immunoglobulins, Bromide, Nitrates, Lithium Overdose, and of course the unmeasured cations of crit illness?

          As many have mentioned the last quant piece is an inappropriately high PaCO2.

          So the pt has a triple acid base disorder-mixed metabolic acidosis and resp acidosis.

          Now as to why, I leave that to folks much smarter than me.

          • Pete says

            just speculating but could the hyperchloraemia associated with hypertonic saline administration for head injury change the anion gap without perhaps affecting the underlying cause….i.e. could this be a WAGMA with the AG being masked by hypertonic saline use?

  11. anonymous. says

    Metabolic acidosis (normal anion gap), less than expected pco2 (exp 23-27), hypernatremia, hypocalcemia, hypokalemia prob related to diabetes insipidus, mild lactic acidosis fr inotrope use.

  12. joannao says

    Severe non-anion gap metabolic acidosis. Predicted respiratory compensation.
    Hyperchloraemia, possibly due to enthusiastic saline administration. Potassium is normal when corrected for low pH.
    Lactate high -- tissue hypoperfusion, due to hypotension/ seizure.
    Hypocalcaemia, hypernatramia and lactic acidosis could maybe be explained by Valproic acid which may have been given for seizure control post head injury.

  13. Hari says

    Patient has primary metabolic acidosis as well as respiratory acidosis, due to CO2 retention. Expected pCO2 to be 24.5. No anion gap.

  14. Andy Chapman says

    So there,s a hyperchloraemic reduced SID acidosis presumably secondary to Fluid administration with a low SID (hypertonic saline seems a likely culprit), although a type one RTA could cause a similar picture and explain the hypokalaemia (could check urinary pH). There is hyperlactaemia, probably secondary to the shock but other causes would include poor liver perfusion or liver damage from the trauma plus all the usual drugs (propofol possible cause in this situation: young patient/TBI/vasopressors). The unexplained bit is the low AG in the setting of a raised lactate suggesting a concurrent low AG acidosis : commonest cause would be hypoalbuminaemia, less likely would be hypophosphataemia,lithium, bromide and iodide. The hypocalacaemia could be post transfusion in the setting of trauma, pancreatitis, renal failure, chronic metabolic Alkalosis or even medullary thyroid carcinoma (which would give you hypocalcaemia, hypophosphataemia and hence a low AG acidosis). How about an albumin, phosphate and a urinary pH to sort the whole mess out? Occam, schmoccam.

  15. Brian says

    Normal-low anion gap metabolic acidosis -- severe (hyperchloremic metabolic acidosis)
    Lactic acidosis

    AG = (Na + K) -- (Cl + HCO3) = 9. Low measured AG in setting of significant lactic acidosis may be due to dilution from aggressive fluid resuscitation.

    Lactic acidosis may be due to hypo perfusion /grand mal seizure activity/ norepinephrine.

    Inadequate respiratory compensation -- superimposed respiratory acidosis.
    Winter’s formula PCO2 = (HCO3 x 1.5) + 8 +/- 2 = 22.5-26.5. Actual PCO2 35.
    Severe TBI pt in ICU likely to be ventilated with supplemental O2 hence elevated PO2.
    Practice would be to keep PCO2 in range of 30-35 to avoid vasoconstriction and cerebral ischemia. if patient is not ventilated, should be as hypoventilation likely due to decreased GCS and consequent unsafe airway. Ventilatory control important to prevent secondary injury.

    Minimal hypernatremia and moderate-severe hyperchloremia may be due to aggressive resuscitation with normal saline. Marked hyperchloremia raises the possibility of the use of hypertonic saline to reduce ICP in severe TBI refractory to mannitol.

    Moderate hypokalemia -- not uncommon in young trauma patients on presentation.
    Causes -- catabolic losses in trauma
    -- elevated aldosterone levels in hypovolemia -- may explain inequality -- rise in Na versus rise in Cl
    -- use of osmotic diuretic such as mannitol increases loss of K
    -- some radio contrast dyes may act as osmotic diuretic
    -- aggressive resuscitation with K free saline
    Adrenergic simulation by endogenous and exogenous catecholamines also causes intracellular K shift -- may be principal mechanism for hypokalemia seen early after head injury.

    Marked hypocalcemia -- possible causes in trauma patient -- massive transfusion, rhabdomyolysis. High citrate load or high phosphate load. Radiocontrast dyes may also increase citrate load.

    Circulatory failure in TBI patient is most likely due to extra-cranial trauma. Less likely, hypotension and shock may be neurogenic secondary to concomitant spinal cord injury.

  16. Dr John Dyett says

    This patient is critically injured and may have progressed to brain death. The ABG shows:
    1. Severe acidaemia is present
    2. Primary metabolic acidosis (significantly negative base excess, out of proportion to Lactate concentration), which is non-anion-gap; hyperchloraemic metabolic acidosis, secondary to hypertonic saline infusion (most likely used for ICP control); contributed to by hyperlactaemia secondary to profound shock as evidenced by vasopressor requirement.
    3. Mixed respiratory alkalosis (compensatory, likely iatrogenic in intubated/ventilated patient) and respiratory acidosis (incomplete compensation, Winter’s rule suggests pCO2 should be 25mmHg)
    4. Hypocalcemia secondary to likely blood product transfusion and citrate toxicity, combined with dilution
    5. Hypokalaemia secondary to dilution

  17. ExamCandidate says

    1. Ok -- Primary Metabolic Acidosis -- normal anion gap (apparent by calculation) but high base deficit and lactate suggest high anion gap is present. Concurrent respiratory acidosis also present.
    2. Patient in ICU -- previously well, now very unwell -- Iatrogenic cause -- smells of a poison.
    3. Urea not given but assuming it is normal, osmolarity appears high
    4. Severe metabolic acidosis with low bicarb, high lactate, hypocalcaemia, and increased serum osmolarity + cardiovascular collapse despite large dose vasopressors -- reminiscent of toxic alcohol poisoning.

    Overall likely cause -- Sodium load + dose of toxic alcohol in a brain injured patient
    =>Phenytoin Sodium iv push

    (Ddx: Severe Pancreatitis secondary to initial Trauma)

  18. says

    -Moderate metabolic acidosis (pH7.1) with low/normal AG (6 without K, 9 with K) likely due to marked hyperchloraemia, which in the context suggests use of hypertonic saline to manage acute elevation of ICP (calculated serum osmalrity > 299 once urea is known)
    - There is inadequate respiratory compensation as the PaCO2 is higher than expected (mid 20s). Patient either needs intubation & vent or, if already intubated (likely, as suggested by the supranormal PaO2) may be being kept at that PaCO2 for temporary ICP management while other therapies are being arranged, eg craniotomy.
    - Hyperlactataemia (4.9mmol/l) likely to be secondary to NorAdr infusion, which may also be contributing to the mild hypokalaemia (3.2mmol/l). Could also be generated by a seizure in setting of TBI, rhabdomyolysis in seting of trauma, or liver injury causing impaired clearance.
    - The hypokalaemia might also be the result of an insulin infusion targetting euglycaemia in TBI.
    - Low ionised calcium (0.8mmol/L) -> Are there other injuries that prompted a MBT? Is there a crush injury or a long lie with rhabdomyolysis resulting in lactate acidosis with hypokalcaemia (usually see hyperKalaemia though)

    That might do for now.

    Take care

    - Can ignore the Temp of 37C on the ABG. That’s just the machine operating temp.

  19. says

    Oooh, one other thing; and this is just for fun … By reverse engineering the Alveolar Gas Equation you know that she is on an FiO2 of no less than 0.3, assuming her lungs are normal and the Aa gradient is negligible. If her lungs are not normal then the FiO2 will be higher again.

    Ta ta!

  20. susan says

    Is she on an ativan infusion for sedation? Ativan comes in propylene glycol diluent and can account for some of the abnormalities seen…especially the possiblity of toxic alcohol ingestion.

  21. teedub says

    1.) Hypertonic Saline --> Hyperchloremic Acidosis.
    2.) In Shock (herniated?) + on levophed --> Lactic Acidosis
    3.) Low HCO3 due to #1 and #2.
    4.) TBI --> Mechanically ventilated to goal pCO2 --> uncompensated ventilation.
    5.) Hypokalemia due to dilution from hypertonic saline.
    6) Low iCal from critical illness (and/or blood product if she had other injuries…hard to know).
    +/- 7.) Susan makes a great point. Propofol infusion syndrome (which i guess would have a higher lactate) or barbituate tox due to sedation might have a similar metabolic look.

  22. David Hersey says


    Excellent stuff.

    The anion gap is normal (AG=9 by my calculations).

    These have been suggested by posters:

    Hypercholermic metabolic acidosis: Cl- 129
    Renal loss of HCO3- : ATN Type 4
    Toxic ingestion of ethylene glycol, propylene glychol

    I would like to add:

    Propofol Infusion Syndrome?
    Pump failure? I have seen some SAH and TBI develop what was called a ‘stunned myocardium’ from a CNS injury.

    Thank you,

  23. Samuel Bruchfeld says

    I have nothing more to offer than a hunch and some pseudophysiological babbling but that’ll have to do for now (who knows, I might even learn something for next time).

    Let’s see, we have acidemia with metabolic acidosis and insufficient respiratory compensation. Apparently we have a normal anion gap and concurrent hyperlactaemia (kind of contradictory, no? maybe we need more lactate!) with some screwed up electrolytes.

    The electrolytes are in line with what you’d expect with (TBI-caused) Diabetes Insipidus and “standard” fluid resus (that is, isotonic fluid resus), the other electrolytes are simply “flushed out”. High lactate is due to poor perfusion (low CO times very high TPR yields 1) Nice MAP 2) Sucky perfusion).

    Circulatory failure might be due to: 1) Multiorgan failure 2) TBI-hypertension, 3) ACTH-deficiency --> Glucocorticoid deficiency, 4) Electrolyte deficiency.

    Onfortunately think the explanation closest at hand is brain death with post-morten DI and multiorgan failure (how’s that for Occam’s razor), in particular liver failure causing hypoalbuminemia thus screwing up anion gap reasoning. However I think that the other pathologies mentioned above might just provide explanation enought -- so too speak focal, instead of global, organ failure.

    Love your blog!


  24. susan says

    I’ve been thinking about some of the math done here. And after preaching last night to an intern and a second year resident that you can’t use the ABG bicarb to do any of the math because that’s what the computer of the ABG machine does (at least here in the US). So really you can’t say if there’s an anion gap or not.

    I really like the idea of propofol infusion syndrome especially in the setting of sedation for head trauma.

    Has it been a week yet? I’m rather impatient.

  25. Ian says

    So we have a young patient with a severe head injury who is presumably intubated and ventillated who has subsequently developed respiratory failure.

    1. She is acideamic
    2. There is a metabolic acidosis (HCO3 = 11).
    Expected CO2 = (1.5 x 11) + 8 = 24.5
    Actual CO2 is 35 --> she is underventilated, ie -- there’s a respiratory acidosis also
    Back to the met acidosis -- what the AG? AG = (146 + 3.2) -- (129+ 11) = 9.2
    So, we have a normal AG met acidosis. Causes include GI losses(not apparent with clinical setting), renal losses (acetazolamide, Addison’s disease, renal tubular acidosis [type I with the low K], or too much damned chloride)
    3. Let’s look at the lactate -- it’s high, so why is there no HAGMA? Well, lets check the delta ratio. Actually I’m no sure if there is a point is using the delta ratio when the calculate AG is low anyway -- but hey, someone teach me if I’m wrong…. delta AG / delta HCO2 = (16-9.2)/(24-11) = 0.52
    0.52 suggests a HAGMA component -- liklyl the lactic acidosis
    4. And now the electrolytes. Probably just to much saline given (including hypertonic as her head has gone off). Also, could consider Conn’s syndrome (high Na, low K, low Ca -- but the patient is acidaemic rather than alkalaemic). aLso as mentioned above there may be a underlying renal tubular acidosis type I

    So, I reckon:

    Intubated and underventilated patient who’s received lots of saline and now has circulactory failure contributing to a type 1 renal tubular acidosis. She’s also shocked with a modestly elevated lactate.

    Change her IV fluids, increase ventilation a tad, give her blood or a more balanced IV solution (eg. 150ml of 8.4 HaHCO3 in 850 sterile water), top up to Ca, pat her on the head, and everything should sort it self out in the wash.

  26. Morris says

    Well there are only 3 possible ways she can be acidotic -- ketones, lactate, and/or toxic alcohol metabolites; if the patient had a normal ABG on arrival and then deteriorated rather quickly my first guess is that she may have ingested a toxic alcohol; the patient is a trauma patient after all, and etoh and trauma go hand in hand.

  27. raj says

    hyperchloremic acidosis secondary to fluid resusitation (normal Salini)
    hypokalemic RTA

    HAGMA severe lactic acidosis
    alcohol toxicity
    nor adrenalin induced

    respiratory Acidosis

    head injury
    excess fluid resusitation
    Critical illness