Glucose dynamics during continuous hemodiafiltration and total parenteral nutrition.

OBJECTIVE

To determine glucose balance during dextrose-free continuous hemodiafiltration with or without dextrose-containing ultrafiltrate replacement fluid and full nutritional support.

DESIGN

Prospective, nonrandomized, observational study.

SETTING

A 24-bed multiple trauma critical care unit in a level-I trauma center.

PATIENTS

Seventeen multiple trauma patients with multiple organ dysfunction syndrome requiring hemodialysis for acute renal failure.

INTERVENTIONS

Continuous hemodiafiltration effluent volume and glucose concentration were measured. Study days were classified according to whether dextrose was used in the ultrafiltrate replacement therapy. Use of dextrose in replacement therapy was determined clinically. Parenteral nutrition was not altered for potential glucose absorption from continuous hemodiafiltration. Ultrafiltrate replacement consisted of 5% dextrose in saline on 21 study days (D5YES) and dextrose-free solutions on 54 study days (D5NO).

RESULTS

The D5YES group received 316 +/- 145 g glucose/day from the ultrafiltrate replacement fluid, in addition to glucose in total parenteral nutrition (total glucose intake = 942 +/- 229 g/day in D5YES, 682 +/- 154 g/day in D5NO) (p < 0.05). Glucose loss in continuous hemodiafiltration effluent was 82 +/- 61 g/day in D5YES and 57 +/- 22 g/day in D5NO (P < 0.05), for a net glucose uptake of 8.1 +/- 2.1 mg/kg per min in D5YES and 5.4 +/- 1.5 mg/kg per min in D5NO (p < 0.05). Glucose loss was predictable when dialysate and ultrafiltrate replacement fluids were dextrose-free (R2 = 0.77), but less so when dextrose was used as ultrafiltrate replacement (R2 = 0.47).

CONCLUSION

Dextrose-free dialysate promotes glucose loss during continuous hemodiafiltration, but the loss is small and predictable. Use of a dextrose-containing ultrafiltrate replacement fluid results in a significant increase in glucose intake without a commensurate increase in glucose loss, and makes glucose loss in effluent less predictable.