Metabolic acidosis is common in septic shock, yet few data exist on its etiological temporal profile during resuscitation; this is partly due to limitations in bedside monitoring tools (base excess, anion gap). Accurate identification of the type of acidosis is vital, as many therapies used in resuscitation can themselves produce metabolic acidosis.Design:
Retrospective, cohort study.Setting:
Multidisciplinary pediatric intensive care unit with 20 beds.Patients:
A total of 81 children with meningococcal septic shock.Interventions:
None.Measurements and Results:
Acid–base data were collected retrospectively on 81 children with meningococcal septic shock (mortality, 7.4%) for the 48 hrs after presentation to the hospital. Base excess was partitioned using abridged Stewart equations, thereby quantifying the three predominant influences on acid–base balance: sodium chloride, albumin, and unmeasured anions (including lactate). Metabolic acidosis was common at presentation (mean base excess, −9.7 mmol/L) and persisted for 48 hrs. However, the pathophysiology changed dramatically from one of unmeasured anions at admission (mean unmeasured anion base excess, −9.2 mmol/L) to predominant hyperchloremia by 8–12 hrs (mean sodium-chloride base excess, −10.0 mmol/L). Development of hyperchloremic acidosis was associated with the amount of chloride received during intravenous fluid resuscitation (r2 = .44), with the base excess changing, on average, by −0.4 mmol/L for each millimole per kilogram of chloride administered. Hyperchloremic acidosis resolved faster in patients who 1) manifested larger (more negative) sodium chloride–partitioned base excess, 2) maintained a greater urine output, and 3) received furosemide; and slower in those with high blood concentrations of unmeasured anions (all, p < .05).Conclusions:
Hyperchloremic acidosis is common and substantial after resuscitation for meningococcal septic shock. Recognition of this entity may prevent unnecessary and potentially harmful prolonged resuscitation.