An Equation to Distinguish Acute Blood Loss from Dilutional Changes in Serum Hemoglobin Concentration in Patients Who Receive Intravenous Fluids

Background

Patients given intravenous (IV) fluids are often found to have a falling serum hemoglobin concentration due to a dilutional effect. This presents a diagnostic challenge since occult blood loss may also manifest as a falling serum hemoglobin concentration. To improve diagnostic clarity we developed an equation that predicts dilutional changes in serum hemoglobin with IV fluids. It estimates the changes in intravascular volume that occur after distribution of fluids between intravascular and extravascular spaces. The phenomenon of hemoglobin dilution with IV fluids is broadly recognized, though previous attempts at quantifying this relationship have been limited. Retrospective analyses of fluid resuscitation in sepsis have identified a weak relationship between IV fluids and hemoglobin dilution, but do not control for weight and body composition.

Methods

We derived our equation based on the following assumptions. Conceptually, it estimates a patient’s total body hemoglobin and dilutes it in the fraction of an intravenous fluid bolus that is expected to stay intravascular. It also utilizes a method of estimating blood volume that accounts for the observation that blood volume does not scale linearly with body mass index (Lemmens et al, 2006). To assess our equation’s ability to predict dilutional changes in serum hemoglobin we evaluated its performance on patients diagnosed with sepsis who received ≥1 L of IV cystalloid fluids. To assess its ability to discriminate between patients with and without blood loss who receive IV fluids, we applied it to patients who underwent vascular surgery and had at least 100 mL of estimated blood loss (EBL). For both groups, patients were excluded if they received blood products or colloid fluids.

Results

215 cases of sepsis were found to meet all inclusion criteria and lack exclusion criteria. This population received an average of 2.1 L of IV cystalloid fluids with a resulting fall in the serum hemoglobin concentration of -1.14 g/dL. Our equation predicted an average fall in hemoglobin concentration of -1.13 g/dL based on this population’s metrics. This is an average underestimation of the change in serum hemoglobin of -0.01 g/dL (95% CI -0.15 – 0.13) with a standard error of 0.07 g/dL. An intraclass correlation coefficient was calculated to be 0.89.

100 cases of vascular surgery were identified. This population had a median EBL of 150 mL. The average change in serum hemoglobin concentration was found to be -1.53 g/dL. Our equation, which accounts for hemodilution from fluids but not blood loss, predicted a decrease in serum hemoglobin concentration of -1.04 g/dL. The average difference between these was found to be -0.49g/dL, indicating an underestimation of the actual change in serum hemoglobin. To assess our equation’s ability to discriminate between cases with and without blood loss, we compared our two populations. The difference was found to be statistically significant with a p-value <0.001.

Conclusion

Our equation accurately predicts dilutional changes in serum hemoglobin with IV fluids, underestimating dilutional changes by only -0.01 g/dL with a standard error of 0.07 g/dL. It can discriminate between patients with and without blood loss to an EBL of only ≥100mL. We hope that this equation will provide diagnostic clarity to clinicians faced with the challenge of a patient who, having received IV fluids, is found to have a falling serum hemoglobin concentration.

Disclosures

The authors declare that they have no relevant disclosures.