Comparison of commonly used clinical indicators of hypovolaemia with gastrointestinal tonometry

Abstract

Objective: The gastrointestinal tonometer, which allows measurement of gastrointestinal mucosal CO₂ and subsequent derivation of gut intramucosal pH (pHi), has been demonstrated to be a sensitive predictor of outcome following major surgery. Current theory suggests that the origin of the low pH may be hypovolaemia. This study was designed to compare the temporal sequence of changes in tonometric readings with invasive blood pressure, stroke volume, heart rate, lactate and arterial blood gas measurements during progressive haemorrhage. Design: Observational healthy volunteer study. Setting: Intensive care unit at University College London Hospitals. Subjects: Six healthy, medically qualified volunteers. Interventions: After obtaining baseline measurements, the subjects were progressively bled 25 % (range = 21–31 %) of their blood volume over a period of 1 h in two approximately equal aliquots. Equilibration was allowed for 30 min following the bleed, after which further measurements were made and the blood was then retransfused over 30 min. Measurements and main results: There was no consistent change in any of the haemodynamic variables other than gastric intramucosal CO₂ : arterial CO₂ gap (PiCO₂− PaCO₂) after removal of the first aliquot of blood, although five of the six subjects also demonstrated a fall in pHi. After removal of the second aliquot of blood, PiCO₂− PaCO₂ gap and pHi continued to indicate a worsening gastric intramucosal acidosis; stroke volume, as measured by suprasternal Doppler, demonstrated a marked fall, while all other variables measured had not altered consistently or to such a degree as to elicit a clinical response or cause suspicion of a hypovolaemic state. On retransfusion, all variables returned towards baseline. Conclusions: This study demonstrates the value of tonometry as an early monitor of hypovolaemia and highlights the shortcomings of other more commonly measured clinical variables.