excerpts from a controversial discussion on the critical care mailing list
Amit Ulrich wrote on the 9th november, 2008:
>
> There are some thoughts on Stewart's approach here:
> http://www.acid-base.com/strongion.php
> The main criticisms are quoted below:
>
> "Pros and Cons:
>
> 1) Understanding:
>
> Stewart's greatest contribution may be his focus on the
> importance of the factors controlling pH. [H+], [OH-] and [HCO3-] are
> merely dependent variables. This emphasis on the importance of the
> underlying causes rightly diminishes the importance of the bicarbonate ion.
> 2) Calculation:
>
> A major shortcoming lies in calculating a value for [SID] which
> depends upon accurate measurements of several variables. An acceptable
> level of error in the underlying measurements becomes less acceptable
> after subtraction. This is partly because the errors are summed and
> partly because any error now appears proportionately large against the
> resulting small value.
> 3) Standard Base Excess Accuracy:
>
> Standard base excess has been well validated both for accuracy
> and for clinical relevance through many years of familiarity and
> clinical correlation. Although [SID] yields a somewhat similar numerical
> value to Base Excess, it does so with less reproducibility and accuracy
> and it neglects the influence of hemolobin as a buffer. Moreover,
> additional effort is required to obtain the result; and there is a
> smaller database of clinical correlation."
>
>
> Best regards,
>
> Amit Ulrich Gupta
Rainer responds:
hello Amit Ulrich, hello all,
Amit Ulrich wrote:
> It dawned to me only after I had written my first mail that you must be
> a hard-core Stewardian ;-)
well, there is no arguing with natural laws, is there?
like mass equations and conservation of charge and mass are very close aproximations to reality at anything but cosmic or
quantum scales, aren't they?
for all who want to read more - there are two very good articles with clinical examples - which i can provide you with, btw:
- Peter Lloyd (he is dead by now, and the spreadsheet calculator no longer available for download, but i can send you a copy)
Using quantitative acid-base analysis in the ICU.
Crit Care Resusc. 2006 Mar;8(1):19-30.
- Fencl, Jabor, Kazda, Figge - Diagnosis of Metabolic Acid-Base Disturbances in Critically Ill Patients, Am J Respir Crit Care Med, 2000, Vol 162, 2246- 2251
i shall try and answer your three points, before reading Marek's posting:
point 1 and 2 are completely correct. the Na and Cl measurements are specially crucial.
if you do not have reliable Cl measurements, forget the Stewart approach as an analytical tool -
but as a tool for understanding what is happening, it will still be valid.
point 3, though, is different. let me take BE first, then haemoglobin.
the BE is meant to represent titratable acidity; as nobody, at least nowadays, titrates blood or plasma to find this value,
it has to rely on a set of standard assumptions about plasma composition - the influence of albumin, phosphate and the like
must be based on assumptions.
i often see patients after volume resuscitation - maybe coming from a ward or the operating theatre - with a BE of about -5mEq/l.
many are clinically well - they just happen to have been treated with relatively hyperchloric solutions, dextrose with NaCl,
H(a)es preparations in isotonic saline etc. BE is unable to distinguish between hyperchloraemic (or, what amounts to the same,
hyponatraemic) acidosis and metabolic acidosis caused by other anions.
Ian Stewart in may 2005 wrote a fine example about this on this list, i paste it at the bottom of this posting.
this list's own Kyle Gunnerson wrote an article about the clinical prognostic importance of the different types of metabolic acidosis:
Lactate versus non-lactate metabolic acidosis: a retrospective outcome evaluation of critically ill patients.
Crit Care. 2006 Feb;10(1):R22.
(freely available at: )
Kyle's article actually answers another of the criticisms proferred on the website quoted by you:
"lack of clinical correlation to validate benefit".
and i shall gladly admit that the clinical utility of the BE approach becomes equivalent to that of the Stewart approach,
once you correct for Cl and albumin levels. but why run into all these corrective troubles, when you can get an intellectually sound
explanation of the mechanisms and the analysis all in one?
from my high-school days with wide reading of anything from Freud to B.Russell i remember that somebody had actually come up with
Einstein's formula for correcting a fast-moving body's speed so it would not pass that of light - but he had no inkling about the
mechanism. THAT was what needed the stroke of genius.
(acid-base is, i admit, much less important than special relativity, but still .... )
and haemoglobin! how on earth can it influence pH on the other side of the erythrocyte membrane? (barring haemolysis .... ) ok - it
can modify PCO2 or it maybe makes some electrolytes move across that membrane. good. H+ moves across the membrane from the outside in,
to get mopped up by haemoglobin, you are bound to say, aren't you? and you are probably right, BUT that positive charge has to be
accompanied by a negative one - which do you choose? HCO3-? hardly! chances are, it is Cl-! or another positive one moves out, Na+
or K+. all of these are part of our measurements and of SID. so it does influence PCO2 and maybe SID - we already measured these,
so we do NOT NEED to measure the haemoglobin - its effects are already accounted for.
beginning of Ian's example:
sent to ccm-l on the 5th may, 2005
A woman is admitted to HDU after major gynaecological surgery with pelvic clearance.
On her first blood gas you have pH 7.28, pCO2 40 mmHg (or 5.3 kPa) and BE -6.7 mEq/l (or HCO3 17.3 if you prefer).
Everything was normal preoperatively. The registrar points out the acidosis, says that she is underperfused and wants to give some
fluid (such as a litre of saline).
However: other blood results reveal sodium 142 mmol/l, chloride 115 mmol/l and albumin 28 g/litre.
Again these were all normal preoperatively.
Work out the Na/Cl effect: Na - Cl - 38 = -11 [so you expect a BE of -11 from the Na/Cl effect]
Work out the albumin effect: (42 - Alb)/4 = +3.5
So overall you have just fully explained a base excess of -11 +3.5 = - 7.5
This compares pretty closely to the observed base excess of -6.7
She is NOT underperfused and in fact the problem is NaCl poisoning (6 litres intraoperatively by the anaesthetist).
So by doing these calculations you have avoided giving inappropriate therapy which would hae actually made things worse.
The million dollar question of course is So what? There is no doubt this so caled 'hyperchloraemic' acidosis occurs
[though as John Morgan has proven it can occur in the absence of hyperchloraemia so is more accuratey a 'narrow Na/Cl gap acidosis']
but does the acidosis matter, or have any adverse effects? I don't know the answer to that but would be interested in any insights.
Cheers, Ian
ending Ian's example