This is Gas No 2 in the series. The aim is to provide practice in interpreting blood gas results in a clinical context. Most of the case scenarios will have an emphasis on Anaesthetic or Intensive Care practice but interesting gas results on other patients that I come across may be included. Most of the results are ones you could typically come across in your daily practice so the emphasis is not on finding strange or extreme results though some unusual results may be included. Perhaps you have come across an 'interesting gas' - please feel free to send this along to me at kbrandis@bigpond.net.au. The style of analysis used here is that presented in my on-line text "Acid-Base pHysiology".
Please consider the following clinical scenario and then consider the question & discussions that follow. The answer to the Puzzle from Gas of the Week No 1 is at the bottom of the page.
A 19 year old pregnant insulin dependent diabetic female was admitted with a history of polyuria and thirst. She now felt ill and presented to hospital. There was a history of poor compliance with medical therapy. On examination, she was afebrile, chest was clear, peripheral circulation was adequate with BP 105/65. Perioral herpes was present.
Na 136, K 4.8, Cl 101, bicarbonate 10, glucose 19.0, urea 8.1 & creatinine 0.09 (all in mmol/l).
Urinalysis results: 2+ ketones, 4+ glucose
Therapy was started in the Emergency Department and the patient was subsequently transferred to the Intensive Care Unit
[1] What is the diagnosis in this patient?
[2] Is there evidence of any other acid-base disorders?
[3] What acid-base pattern may develop during therapy?
What do you think of this history and set of results? What is the diagnosis?
Using the approach outlined last week, I would approach the case like this:
Firstly: The Initial Clinical Assessment
The diagnosis here is obvious from the history: diabetic ketoacidosis (DKA). The underlying problem seems to be poor compliance but other precipitants should be actively searched for, especially infection. There is little clinical evidence of a respiratory infection and no comment is given regarding skin infection. A suitable urine specimen should be sent for microscopy and culture. The findings as regards underlying factors in acute diabetic ketoacidosis are:
[1] poor compliance in known insulin-dependent diabetic patients (30% of cases)
[2] treatment non-compliance (20%)
[3] a new diagnosis of diabetes (25%)
[4] and in 25% of cases, no precipitants are identified.
Why do you say the diagnosis is 'obvious' as you haven't even commented on the gas results yet? Aren't you getting a little bit ahead of yourself?
Absolutely not. The diagnosis can be made on the history, examination and some simple tests alone. In essence, it is a spot diagnosis made used 'pattern recognition' if you like. If you can't make this one, then its back to medical school! The pattern here is this: 'A known insulin dependent diabetic presents feeling unwell with increased urine production and simple bedside urine testing reveals ketones and glucose.' That situation makes the diagnosis.
Good. But why then bother with blood gases and other investigations at all?
For several reasons:
[1] Firstly: to confirm the diagnosis & assess its biochemical severity
[2] Secondly: to check for other or complicating acid-base disorders that may be present (eg dehydration may have resulted in poor tissue perfusion and a lactic acidosis, vomiting may have added a component of metabolic alkalosis)
[3] Thirdly: investigations to seek the underlying reason for the acute presentation (eg infection, or various serious medical conditions such as a myocardial infarct or pancreatitis) - these often are indicated by the history
[4] Finally: to monitor the response to therapy.
In the pre-insulin era, DKA had a mortality of virtually 100% but now with 'good' medical care using a low insulin regime the mortality rate is probably less than 2%. In my experience, the mortality today in the patients that reach hospital, is mostly in that subgroup of patients with an underlying serious medical illness as the precipitant of the acute DKA or which occurs as a consequence of the DKA (eg myocardial infarction).
You have made some good points. Can you continue with your assessment of the gases?
Its not the gases that I am really attempting to assess here. That is just a part of the overall assessment of the patient. I'm still on the first step: the initial clinical assessment.
The diagnosis then is DKA with probable non-compliance with insulin therapy as the cause: I would ask the patient further about this. An infection screen is essential and you should be prepared to be as aggressive as the history & examination suggest. Some patients will warrant lumber puncture but there is little reason to perform that here at present. I think blood cultures and urine cultures are useful to collect at this stage.
Also, the simple lab results reported so far contain additional useful information. For example:
[1] The anion gap (calculated as [Na] - [Cl] - [HCO3]) is elevated at 25 -> this is consistent with the presence of a high anion gap metabolic acidosis such as DKA
[2] Plasma glucose is 19.0 -> this is elevated as is the usual finding in DKA, and is well above the renal threshold for glucose explaining the presence of glucose in the urine
[3] There is no evidence of azotaemia -> this (and the absence of any information in the history) exclude metabolic acidosis due to renal failure
[4] The [K+] is 4.8 here. Patients with DKA may have significant hyperkalaemia on presentation and then may develop hypokalaemia after commencement of therapy as K+ moves back intracellularly. A baseline ECG must be performed on admission. As these patients primarily have a total body potassium deficit they typically require K+ replacement once therapy has decreased [K+] below 5mmol/l.
Good. I notice you are not just looking at the gases in isolation as an intellectual exercise but are using an approach to the whole patient.
Moving now to the second step: Making the Acid-Base Diagnosis using the 6 point approach I outlined last week.
1. pH: A pH of 7.26 is an acidaemia so a net acidosis must be present
2. The pCO2 & HCO3 pattern: Both the pCO2 & the [HCO3] are decreased, and both significantly so. This suggests the presence of either a metabolic acidosis or a respiratory alkalosis. A net acidaemia is present so clearly this pattern here confirms the presence of a metabolic acidosis.
A respiratory alkalosis is unlikely - it is not supported as the primary diagnosis because of the acidaemia and also because the [HCO3] is so low. The limit of compensatory decrease in [HCO3] in response to a primary decrease in pCO2 (resp alkalosis) is about 12 to 15 mmol/l so the value of 7.1mmol/l is too low.
The acid-base diagnosis so far is of an acute metabolic acidosis.
3. Checking for clues.: I have partly discussed this already, but looking at this specifically. The main general information to look for in a patient with metabolic acidosis are the 'Anion gap' and the plasma chloride level. As mentioned above, the anion gap is elevated at 25mmol/l.
As a guideline, if the anion gap is between 20 & 29mmol/l there is a 67% chance of a metabolic acidosis, but if the anion gap is 30 mmol/l or more, then a high anion gap acidosis is invariably present. The plasma chloride level in normal anion gap acidosis is typically raised but the rise may not be so marked if hyponatraemia is present.
Now the specific clues that are sought relating to aetiology: [glucose], [lactate], urea & creatinine levels. A lactate level is not provided. The glucose is elevated which is the expectation in most cases of DKA
The working diagnosis: An acute metabolic acidosis with hyperglycaemia: diabetic ketoacidosis but not so far completely excluding a second acid base diagnosis. The lactate level was not done here so a contribution from lactic acidosis cannot be fully excluded.
4. Assessing the Compensatory Response: This is assessed for a metabolic acidosis by calculating the 'expected pCO2'; that is, what the pCO2 should be in compensation for acute metabolic acidosis. This value is then compared against the 'actual pCO2'.
'Rule 5' is used for assessing the compensation:
Expected pCO2 = ( 1.5 x [HCO3] ) + 8 (+/- 2)
Now, plugging in the measured values:
Expected pCO2 = (1.5 x 7.1) + 8 (+/-2) = 18.6mmHg
(Actually, this is really done as a quick bit of mental arithmetic: one and a half times 7 then add 8 which is about 18.5 in this case. You don't need a calculator.)
This is only a few mmHg away from the actual values, so the conclusion is that the amount of respiratory compensation is essentially at the predicted value. Respiratory compensation commences early with a metabolic acidosis but may take as long as 12 to 24 hours to reach a stable maximum level. This compensation is never complete as regards return of extracellular pH completely to normal
5. Formulating the acid-base diagnosis: An acute metabolic acidosis (diabetic ketoacidosis) with maximal respiratory compensation. No evidence of a second acid-base disorder but a component of lactic acidosis (due poor muscle perfusion) cannot be entirely excluded as a lactate level was not done.
6. Confirmatory Tests: Other then a lactate level, none are necessary.
Now the final step.
Yes. The clinical diagnosis: Acute severe diabetic ketoacidosis. There is no evidence of any respiratory acidosis or alkalosis.
As I said earlier this is such an obvious diagnosis, you can perform all the steps in your head very quickly, certainly a lot quicker than it has taken me to explain to you. My main concerns are to perform the appropriate investigations as quickly as possible, and initiate therapy. We follow a general management protocol for patients with this diagnosis. This includes supplemental oxygen by mask, fluids, low-dose insulin infusion and K+ replacement (as the level falls) backed by specific treatment of any underlying cause (eg antibiotics for infection). Monitoring in a High dependency Unit or similar area and use of a 'Biochemistry Flowchart' to display results are important. Some cases receive phosphate replacement but this is guided by specific testing.
You mentioned several times about the missing lactate level. What is the significance of this?
Well, for 2 main reasons.
Firstly, this is important in how I approach a metabolic acidosis. The framework classification that I use and which I keep a mental picture of is:
1. High anion gap metabolic acidosis (HAGMA)
2. Normal anion gap metabolic acidosis (NAGMA) (also referred to as 'hyperchloraemic acidosis')
Now this is of course not entirely complete and there are qualifications. For example, a condition which normally causes a HAGMA may present as a NAGMA so you should be aware of this, and sometimes a NAGMA may not be hyperchloraemic for example. Nevertheless it is a great framework to work from. So the importance of the lactate level? Well here we have a HAGMA: this gives me 4 conditions to differentiate: ketoacidosis is positively selected based on history and the results of a few simple investigations so I now try to exclude the other 3 as well. Acidosis due to renal failure is excluded based on the normal urea & creatinine levels. Acidosis due to toxins is unlikely in the absence of a suggestive history. So this leaves me with lactic acidosis to exclude.
Now this is important because a patient with DKA is volume depleted and perfusion may be poor. The clinical ways to assess perfusion are urine output and peripheral perfusion (colour & temperature). Oliguria is not present because of the osmotic diuresis from the glucose so this is not useful. It is so easy just to assume you have the diagnosis and not even consider a lactic acidosis component but it should be done. A lactate level is easy to perform and can now be performed routinely on some automated blood gas machines. The argument seems to be well we know she must be volume depleted (polyuria) and are giving fluids anyway so it doesn't really matter but this is a wrong approach.
I must say I agree with you. But you said TWO reasons why you prefer a lactate level. What is the other?
I was getting to that. The acids in ketoacidosis are acetoacetic acid (AcAc) (pKa 3.58) and beta-hydroxybutyric acid (BOHB) (pKa 4.7) and these two acids are in equilibrium with each other. Now, if a significant lactic acidosis is present it shifts the equilibrium between these 2 components and alters the concentration ratio of AcAc:BOHB so that more BOHB is present and less AcAc is present. The reason for the shift has to do with NAD+ which I won't go into here.
So what is the point of this: the total amount of ketones is not altered just the relative amounts of the 2 'ketoacids?
Its important because it can result in a false negative result on testing of urine for ketones.
The common urine test uses a nitroprusside reagent which reacts with acetoacetate but not with beta-hydroxybutyrate. So a lactic acidosis cause less AcAc present and there may be insufficient to get a positive result on urine testing. Now, I'm relying on my urine test for ketones as vital confirmation.
This can also lead to an interesting situation. As the lactic acidosis improves with fluid therapy then the AcAc:BOHB ratio shifts in favour of AcAc and now the urine test becomes more positive (!) even though the patient is improving.
Still, if the blood sugar is high in a known diabetic you are suspicious even if the urine test is negative for ketones. But if this is the patient's first presentation with diabetes then such a result can cause delay in appropriate management.
You mentioned therapy including a low-dose insulin regime.
Well this has been a change in more recent times. Previously such patients were managed with high dose insulin regimes based on the idea of supposed 'insulin resistance'. This was associated with more marked biochemical derangements during therapy, for example regarding K+ shifts. Low dose regimes are superior and have now become a standard component of management.
Why use insulin at all?
Insulin is necessary to facilitate glucose uptake into cells. Normally, the concentration of glucose is higher in the extracellular fluid than in the intracellular fluid so glucose entry is a process of facilitated diffusion rather than active transport. It is carrier-mediated by a specific protein in the cell membrane.
The action of insulin peripherally is only on certain 'insulin sensitive' tissues. Of major importance are muscle and fats cells. These are important also because they contribute such a high proportion of total body mass. Insulin does not affect the entry of glucose in most other body cells, specifically it has no effect on red cells and in the brain.
Without insulin then, the muscle & fat cells are in effect starved of sufficient glucose.
Are you sure about that?
I was up until just now. But you are saying that I am wrong. I note that Stoelting (in "Pharmacology & Physiology in Anesthetic Practice" 2nd edition, p765) says 'Resting skeletal muscles are almost impermeable to glucose except in the presence of insulin". I remember being taught that hyperglycaemia in uncontrolled diabetics is due to impaired glucose entry into cells, which is consistent with what Stoelting says. Is this not correct?
Hyperglycaemia results predominantly from loss of insulin effects in the liver, not in the peripheral tissues. Insulin does indeed facilitate glucose uptake on the glucose transporters (GLUT-4) in muscle and fat cells, but the glucose levels in uncontrolled diabetes are high enough to result in normal or increased entry of glucose into these cells. The fat and muscle cells are certainly not 'starved' of glucose as you suggested. Researchers have bred mice who completely lack the insulin-sensitive glucose transporters (GLUT-4) in muscle and these mice do not have fasting hyperglycaemia or diabetes.
Fasting hyperglycaemia in diabetics who are not on insulin is due to over-production of glucose by the liver and is not due to underutilisation of glucose in peripheral tissues.
Perhaps you should read the article by Sonksen & Sonksen ('Insulin: understanding its action in health and disease', Brit J Anaes 2000, 85: 69-79).
Thank you.
You haven't mentioned the pO2 value which is 128mmHg.
The alveolar pO2 value can be calculated using the alveolar gas equation. Now assuming the patient is breathing room air, assuming R is 0.8 and using the measured pCO2 value:
Alveolar pO2 = Inspired PO2 - paCO2/R = [0.21 x (760-47)] - 16/0.8 = 129.7mmHg
Because the patient is hyperventilating to compensate for the metabolic acidosis, the arterial pCO2 is very low and this results in an increase in alveolar pO2.
The arterial pO2 is typically slightly lower (eg 5mmHg) than the calculated alveolar value because of 'venous admixture'. This term covers the slight effects of the small physiological shunt and ventilation-perfusion mismatching in slightly depressing the arterial pO2 in healthy people. The measured and actual values are so close here, it may be that the patient has been started oxygen by a face mask. If the actual pO2 value was higher than the value calculated for room air breathing then there has either been an error or the patient is breathing supplemental oxygen.
Good. We'll discuss oxygen and its transport in more detail next week, but let me leave you with a problem to think about.
I was looking through Lumb ("Nunn's Applied Respiratory Physiology" 5th ed. 2000, Butterworth-Heinemann) the other day and noticed the representation of the oxygen dissociation curve for normal adult haemoglobin in figs 11.9 & 11.10 on pages 266 & 267. As I looked at the curves I noticed 4 interesting points on this normal curve:
Now the problem here is that one of these points does NOT lie on the curve. So the problem to resolve is which one and why?
The above discussion is a purely hypothetical dialogue which is presented for educational purposes.