Since Babb and Scribner introduced the concept of middle molecule (MM), there has been a great deal of controversy about what kind of solutes should be considered as middle molecules. The debate concerning the evidence and potential "toxic" properties of some of the 300-2000 molecular weight substances is still open. Middle molecules cannot be identified from their chemical characteristics. Their molecular weight (MW) itself is controversial. As MM are extracted by gel permeation chromatographic technics, their MW is usually obtained from elution volume of the column. A discrepancy may exist between the mol wt as it is measured from the rejection rate on calibrated cellulose acetate membrane and as it appears from the elution volume in Sephadex G-15 chromatography. Moreover the permeability of membranes to MM during dialysis or hemofiltration depends upon the size of the molecule more than upon their MW. For example, hydrated molecules such as phosphate or sulfate ions have a much greater size than would be expected from their MW (Fig 1). Most of the tentative characterizations of MM have focused on the aminoacid content of peptides. Today it appears that MM have carbohydrate fractions. After almost ten years, the middle molecule hypothesis remains a concept mainly based on clinical observations and supported by bioassays. We will overview the present state of the art in this field and try to show the clinical issues which were reached from this approach.