May 23, 2009

Chemistries, Part III

This is a follow-up post of the series on Chemistries, Part I and Part II.

I get overwhelmed looking at something like this... so I always take a few deep breaths and approach it sequentially.

Starting with the admission data.
First, I scan it for anything that might KILL the patient. Check the Sodium, Potassium and Glucose levels.

Most alarming is the hyponatremia of 120! Is this something acute or chronic? A sudden drop of sodium may cause seizures and we may be observing a post-ictal state. However, as my astute classmate points out, by circling the blood glucose, we see that the low sodium may in fact be artifactual. When approaching a hyponatremia, we first look to see if the serum osmolality is altered -- Glucose, Alcohol, and Mannitol can cause HYPERtonic hypoNa. For every 100 above 100 of glucose (in this case, 879 or ~900) will draw out water into the intravascular space and dilute the present sodium down by 2.4 (in this case, 9*2.4 = 21.6) So the corrected sodium is actually 120+21.6 or 142. Whew, ok so that was not as scary as I first thought.

Even though we've got a high glucose, I try to stay on task so I don't miss anything. A low potassium. Combined with a low chloride and HCO3, I'd say that this is a volume depleted state with a metabolic acidosis. Normally, we would expect a high potassium with an acidosis as it drives a transcellular shift leaking out potassium. Also, a HYPERglycemic state will also lead to HYPERkalemia. The "paradoxical" hypoK so this is a very significant finding even though it may only be 0.2 below the lower limit of normal. It's not really a paradox, nor is it unexpected. Hang on for further exploration of this.

Let's address the acid-base disorder here.

On the surface, we've got someone with a HCO3 of 18 (which is <24), therefore there is a metabolic acidosis. However, we also note that the anion gap is 32 (which is significantly >>>12!) This is where the "delta-delta" comes into play.

There's all sorts of fancy equations of "it should be greater than this ratio" (d:d>2 = coexisting metabolic acidosis/alkalosis) or "the difference should be less than that" but let's keep things simple, the way my preceptor did.

AG of 32 -> 32-12 = 20. This means that there are 20mEq of unmeasured ions (presumably acid) floating around.
HCO3 of 18 <- for the HCO3 to drop by 20mEq (consumed by the acid in the anion gap, we would need to have an initial HCO3 of 20+18... 38 (>>24!) This markedly high HCO3 is a metabolic alkalosis.

There it is, folks, right there in the numbers: a metabolic alkalosis (likely from vomiting) and a metabolic acidosis (likely from ketone body production.) If you prefer, the d:d is 20:6 or 3.3.

Let's take a deviation in our chemistry analysis for a second....
If we note the time of admission labs as 0930, we can see that the ABG was drawn at 1353. The metabolic acidosis is GONE! It is quite a head-scratcher if this seemingly normal ABG (or only slightly metabolic alkalotic) is the first thing that you see.

Keep in mind though, that the underlying metabolic alkalosis cannot clear as quickly as the metabolic acidosis. We are seeing a cross-sectional sample of a resolving problem:
The tides of acidosis are pulling back, revealing the rocks of alkalosis that were already there, hiding under the surface.

The lesson: BE AWARE OF A WIDE ANION-GAP, even in a non-alkalemic/non-acidemic patient! ESPECIALLY in this seemingly benign situation, because the acid-base see-saw is balanced, but it won't stay there for long!

It's a shame that the ABG was not drawn at the same time as the Chem20.

Now, back to the chemistry analysis...

The differential for an anion-gap metabolic acidosis is classically remembered as MUDPILES: Methanol, Uremia, Diabetic ketoacidosis, Paraldehyde, Isoniazid, Lactic acid, Ethylene glycol, Salicylates. However, Paraldehyde is no longer used. GOLDMARK: Glycols, Oxoproline, L-/D-lactic acid, Methanol, ASA, Renal failure, Ketoacidosis is another mnemonic recommended by Dr. Topf.

Shifting further down the list of labs, we've got an increased creatinine. Sometimes ketones can artificially increase this. It may also reflect a 19:1.31... or ~15:1 prerenal azotemia at work. Decreased renal blood flow from a low volume state like dehydration accounts for this.

Now, the big money. A high glucose and a (+) "large" acetone clinches the diagnosis. As we suspected, from our systematic approach, a patient with hyperglycemia and very little endogenous insulin began to produce other sources of energy for the brain. Namely, proteins like beta-hydroxybutryate and acetoacetate which are acidic. These products are degraded spontaneously into acetone (the same stuff as nail polish remover, giving the patient a nice fruity breath.)

Whew. That's only the initial diagnosis.

We haven't even delved into the treatment and resolution of this common problem.

More to come.

No comments:

Post a Comment