BACKGROUND: Acid-base derangements can be interpreted using the Stewart-Fencl approach, which includes calculation of the apparent strong ion difference (SID(app)), the effective SID (SID(eff)), and the strong ion gap (SIG). These calculations require the measurement of several variables. We hypothesized that the SID and SIG calculated by different analyzers would not be reproducible because of variability in the measured values. METHODS: In this prospective observational study conducted in a biochemistry laboratory, we analyzed 179 routine blood samples from consecutive patients over a 3-mo period using two automated blood chemistry analyzers, the LX20 (Beckman) and the Modular (Roche). Measured and calculated parameters from the two analyzers were compared. RESULTS: Although the correlation between measured values was satisfactory, there were large differences in the limits of agreement for calculated values (SID(app): 9.6 mEq/L, SID(eff): 6.4 mEq/L, and SIG: 11.7 mEq/L) and a weak correlation (SID(app): r(2) = 0.54 and SIG: r(2) = 0.12) between the analyzers. CONCLUSIONS: The results of the Stewart-Fencl approach for interpretation of acid-base status can vary according to the analyzer used. These differences may have important clinical and research implications..