Abstract
BACKGROUND: This study evaluated the prognostic significance of serial alactic base excess (ABE) measurements in patients with sepsis. METHODS: We conducted a retrospective cohort study including 521 adult patients with sepsis. Arterial blood gas analyses obtained at 0, 6, 12, 24, and 48 h were used to calculate ABE. Patients were classified into three trajectory groups: persistently low ABE (ABE < 0 at all time points), normalizing ABE (initial ABE < 0 with recovery to ≥0 by 48 h), and normal ABE (ABE ≥ 0 at admission). The primary outcome was 28-day mortality. RESULTS: Overall 28-day mortality was 35.9%. Non-survivors exhibited more severe metabolic derangement at admission, with higher lactate levels (median 4.22 [2.89-5.31] vs. 3.66 [2.31-4.82] mmol/L) and more negative base excess values (median -6.00 [-8.54 to -3.88] vs. -5.09 [-7.63 to -2.71] mmol/L) (both p < 0.001). Admission ABE was significantly lower in non-survivors and remained consistently reduced throughout the first 48 h (all p < 0.001). Patients with a persistently low ABE trajectory experienced the highest mortality compared with those whose ABE normalized (52.9% vs. 19.8%, p < 0.001). In multivariable Cox regression adjusted for age, sex, baseline eGFR, SOFA score, and APACHE II score, persistently low ABE independently predicted 28-day mortality (adjusted HR 2.539; 95% CI 1.510-4.267; p < 0.001). Furthermore, each 1 mmol/L increase in ΔABE over 48 h was associated with a 25% relative reduction in mortality risk (adjusted HR 0.750; 95% CI 0.633-0.889; p = 0.001). At ICU admission, ABE showed numerically higher discrimination for 28-day mortality than BE (AUC 0.671 vs. 0.639), although this difference did not reach statistical significance (p = 0.057). CONCLUSION: Serial ABE trajectories provide independent prognostic information in sepsis. Failure to normalize ABE within the first 48 h identifies a high-risk phenotype associated with markedly increased short-term mortality.