To understand why this is the case, it's first necessary to understand that human plasma is normally composed of 93% plasma water and 7% proteins & lipids. Furthermore, it is necessary to understand that most clinical laboratories measure sodium using an indirect ion-selective electrode (ISE) which involves diluting the original blood sample in a 1:10 ratio and measuring whole plasma sodium based on the assumption that the sample is composed of 93% water. Thus, anything which increases the protein or lipid concentration of plasma will lead to an erroneously low sodium concentration. It occurs more often with hyperlipidemia than with hyperproteinemia, and one example would be in patients with familial hypercholesterolemia where their blood is highly lipemic.
One way to avoid measurements of pseudohyponatremia is to use a direct ion-sensitive electrode, which measures only the aqueous phase of an undiluted blood sample; however, in most labs this is not done routinely.
Pseudohyponatremia should also be differentiated from dilutional hyponatremia, in which an osmotic shift of water from cells to the vascular space following mannitol or IVIG infusions results in true (but hypertonic) hyponatremia.