To determine which laboratory studies are most predictive of the fluid deficit in acutely dehydrated children, we studied a convenience sample of 40 children requiring intravenous fluid resuscitation. Nine laboratory studies (serum BUN/cr, total serum CO2, serum uric acid, serum anion gap, urine anion gap, venous pH, venous base deficit, urine specific gravity, and fractional excretion of sodium) were individually assessed in simple linear regression models with fluid deficit as the dependent variable. Only the serum BUN/cr and serum uric acid were significantly associated with increasing fluid deficit (r = 0.52, P = 0.0005 and r = 0.35, P = 0.03, respectively). The sensitivities and specificities of these two laboratory studies for the detection of > 5% fluid deficit were poor. Conventional laboratory studies used to assess dehydration in children are poorly predictive of fluid deficits.