Design: single-centre, open-label, randomised non-inferiority trial with a −1.5 mEq/L margin, intention-to-treat with multiple imputation
Population: 50 high-risk adults (25 per group) undergoing elective major abdominal surgery
Primary outcome: mean difference in standard base excess at skin closure
Effect: −0.42 mEq/L (95% CI −1.92 to 1.09); per-protocol −0.48 (−2.02 to 1.06); non-inferiority not established. Exploratory end-of-surgery lactate 0.9 vs 1.2 mmol/L (P = 0.031) with no difference in base excess or pH

Bicarbonate-buffered crystalloids avoid delivering exogenous lactate, which is an appealing property in a patient whose lactate you will be interpreting. A single-centre, open-label non-inferiority trial randomised 50 high-risk adults having elective major abdominal surgery to a bicarbonate-buffered solution or to compound sodium lactate as the intraoperative crystalloid, with standard base excess at skin closure as the primary endpoint and a non-inferiority margin of −1.5 mEq/L.

The estimate landed close to zero — a mean difference of −0.42 mEq/L — but the confidence interval ran from −1.92 to 1.09, crossing the margin, so non-inferiority was not established. The per-protocol analysis behaved the same way (−0.48, −2.02 to 1.06). An exploratory secondary finding showed lower end-of-surgery lactate with the bicarbonate-buffered fluid (median 0.9 vs 1.2 mmol/L, P = 0.031), with no corresponding difference in base excess or pH — which is what you would expect from a fluid that simply contains no lactate to measure.

The authors are unusually direct about what this does and does not support: the data establish neither interchangeability, equivalence, superiority nor inferiority, and because the trial was powered for a physiological endpoint, nothing at all can be said about organ dysfunction, complications, recovery or mortality. It is worth reading for that reason as much as for the result. Compound sodium lactate remains the default, and a lower measured lactate on a lactate-free fluid is an artefact of the measurement rather than a clinical gain.

Continue with compound sodium lactate; nothing here supports substitution.
Do not interpret a low lactate in a patient given a lactate-free fluid as evidence of better perfusion.
Note that a non-inferiority trial that fails to establish non-inferiority has not shown inferiority either.
Fifty patients powered for base excess says nothing about postoperative outcomes — do not extrapolate.
Record which crystalloid was used when lactate is being trended postoperatively.

Why it matters

It is a reminder that a fluid which removes the substance you are measuring will change the measurement without changing the patient.

The statistics, in plain English

Non-inferiority is established only when the whole confidence interval sits on the acceptable side of the margin. Here the point estimate (−0.42 mEq/L) is close to no difference at all, but the interval reaches −1.92, past the −1.5 margin, so the trial cannot exclude a difference large enough to matter. With 25 patients per group that was always a likely outcome — this is an inconclusive trial rather than a negative one, and the distinction matters because an inconclusive result is not evidence against the intervention.