The first objective in diabetic ketoacidosis is to restore the circulating volume and improve tissue perfusion. In any form of hypovolaemic shock the most efficient way of restoring circulating volume is to be use colloid solutions rather than crystalloids. At least three times the amount of crystalloid must be used to achieve the same effect. The historical reason for using isotonic saline in diabetic ketoacidosis is related not to its similarity to the fluid lost, but to its supposed efficiency in correcting the circulating volume. Excess crystalloid expands the interstitial space which results in pulmonary oedema, peripheral oedema and possibly cerebral oedema. Although currently difficult to define precisely in their more subtle forms, they all produce adverse pathophysiological effects. The fluid loss in diabetic ketoacidosis is equivalent to "half-normal" saline, a relatively hypotonic solution. As well as causing extensive oedema, resuscitation with isotonic saline can increase serum sodium and osmolarity while not providing free water to replace the intracellular losses.