The concentration of protein in the cerebrospinal fluid (CSF) is about 0.5 per cent that of plasma. This difference in concentration is generally considered (1) to be dependent upon the blood-CSF barrier. The normal concentration gradient of protein from a low level in the ventricles, 6 to 15 mg. per cent, to an intermediate level in the cisterna magna, 15 to 25 mg. per cent, to the highest level in the lumbar sac, 20 to 45 mg. per cent (2), has not been satisfactorily explained. To study the nature of this gradient, two groups of experiments were performed, determining: 1) the exchange of radioiodinated serum albumin (RISA@) between the plasma and the various levels in the CSF, and 2) the relative concentration of the albumin, globulin and "prealbumin" constituents at these levels, employing zone electrophoretic analysis.In previous investigations in the dog (3), it has been shown that the plasma albumin exchanges rapidly with the protein in the CSF. Following intravenous injection, RISA® appeared in the cisternal fluid within 20 minutes, reached a maximum at about 20 hours, and subsequently disappeared at a slower rate than in the plasma. In the present study, RISA® was administered intravenously to infants with communicating hydrocephalus to determine if differences in the rates of RISA® appearance exist at different sites along the neuraxis. The demonstration of such rate differences could then be interpreted to indicate nonuniformity of the blood-CSF barrier to protein at these different levels. Similarly, the concentration of albumin in CSF would be expected to be 1 This work was aided in part by a grant from the National Multiple Sclerosis Society.2This paper was presented in part at the Annual Meeting of the American Neurological Association, June, 1955, Chicago, Ill. greatest in the region of maximal permeability of the barrier.
METHODSFor the isotopic studies, 12 infants, ranging in age from one to nine months, were studied. In all cases the diagnosis of communicating hydrocephalus was established by ventriculography and dye test or pneumoencephalography. Lugol's solution was administered 24 hours before the study and continued 10 days in order to block thyroidal uptake of the isotopic iodine. Three hundred to 500 microcuries of radioiodinated serum albumin 3 was administered intravenously, four to nine hours prior to combined ventricular, cisternal and lumbar puncture. Eight milliliters of CSF was removed from each of these sites, within a five minute interval. In each case the ventricular fluid was obtained first and the lumbar fluid last. Samples containing more than 100 red blood cells or 10 white blood cells per cubic milliliter were considered contaminated with blood proteins and were discarded. The CSF samples were dialyzed against 0.85 per cent NaCl for 24 hours to remove unbound radioiodine. Radioactivity was measured in duplicate samples with a well-type scintillation counter. The total protein of each sample was determined turbidimetrically using 5 per cent sulfosalicylic acid (4).Specif...