Start with the familiar Boston question: does the respiratory compensation fit? Then
use Stewart Light to ask what the metabolic SBE is made of: strong-ion/chloride effects,
albumin effects, and residual unmeasured ions.
Stewart Light decomposes multiple metabolic processes.
Traditional bicarbonate/SBE-plus-compensation reasoning is often clear when one process
dominates. ICU acid-base problems are frequently less transparent: chloride shifts,
hypoalbuminemia, lactate, unmeasured ions, fluids, diuresis, and chronic CO2 retention can
move in opposite directions and make the total SBE look deceptively simple.
This tool keeps Boston-style compensation assessment visible, then partitions measured
SBE into chloride/SID, albumin/weak-acid, and residual unmeasured-ion components using
the Stewart Light method proposed by
Duška et al.
in the original manuscript.
Inputs
Blood gas and chemistry
Results
Interpretation
Orientation
What base excess means — and when it is misleading
SBE near zero can mean true metabolic normality or equal-and-opposite metabolic
processes, but there are several types of processes that can result in the same HCO3 /
SBE. Stewart Light is useful because it differentiates these by separating chloride/SID,
albumin, and residual unmeasured-ion effects.
Example Metabolic Processes
These are fixed teaching examples, not values calculated from the current inputs.
A normal total does not guarantee a simple process.
Step 1
Clinical context, pH severity, and Boston rules
Start with the clinical picture, pH direction and severity, and expected respiratory
compensation. For metabolic acidosis, this includes the Boston/Winter check of expected
PaCO2.
Enter values and calculate to review respiratory compensation, then inspect what the
metabolic SBE is made of.
Boston compensation view
pH → H+
Steps 2-4
Partition measured SBE
Base excess partitioning separates the measured metabolic SBE into SID/chloride,
albumin/weak-acid, and residual unmeasured-ion components. Segments can point in
opposite directions, so a near-normal total may reflect offsetting processes rather than
one normal metabolic state.
Step 2: Strong-ion effect.SBE_SID = Na - Cl - SID reference; outside pH 7.30-7.50, use
SID reference = 35 + 15 * (7.40 - pH), equivalent to 1.5 mmol/L per
0.10 pH unit, upward in acidemia and downward in alkalemia. Negative values suggest
strong-ion acidosis; positive values suggest strong-ion alkalosis.
Step 3: Weak-acid effect.SBE_Alb = 0.3 * (40 - albumin g/L); hypoalbuminemia contributes a
positive alkalinizing component.
Step 4: Remainder.SBE_UI = SBE - SBE_SID - SBE_Alb; if lactate is available, it is split
from the residual unmeasured-ion component.
Calculate a case to see component direction and magnitude.
Synthesis
Synthesize the Boston and Stewart Light views
Put the Boston compensation assessment together with the SID, albumin, and
unmeasured-ion components. Clinically decide whether a positive SBE_SID is consistent
with a primary chloride-depletion process or an expected renal response to chronic
hypercapnia.
Stewart Light decomposition
Context checks
Reconcile the partition with lactate, anion-gap context, chronic hypercapnia
suspicion, lab timing, and the overall clinical picture.
Anion gap context
Follow-up considerations
Why can the total look normal even when components are not?
This simplified calliper view shows the sodium-chloride gap as the SID anchor, then
separates weak-acid and residual unmeasured-ion effects. This is an educational bedside
simplification, not a full physicochemical reconstruction. Use it as a mental model for
how chloride/SID, albumin, and unmeasured ions can push SBE in different directions.
Calculate a case to see the conceptual SID, albumin, bicarbonate, and UI spaces.
Decomposition with physicochemical approach
Boston compensation map
This Boston compensation map is a visual cross-check for expected respiratory
compensation, not a replacement for the primary Boston-rule assessment.
Laboratory caveats
Use values from the same timepoint and specimen context whenever possible.
Blood gas results can be affected by air bubbles and delayed processing.
Compensation equations are approximations, not absolute truths.
Normal anion-gap ranges vary by laboratory method.
A normal osmolal gap does not reliably exclude toxic alcohol exposure.