Cisterno-lumbar gradient of complement fractions in geriatric patients with suspected normal pressure hydrocephalus.

Background: The complement system is a functional link between the innate and adaptive immune system and present in all compartments of the body. The composition of the cerebrospinal fluid (CSF) differs between the ventricular, cisternal and lumbar space. Usually, concentrations of blood-derived CSF proteins increase from ventricular to lumbar fractions.

Methods: In 20 geriatric patients with suspected normal pressure hydrocephalus (NPH) [13 women, 7 men, age 80.5 (75/85) years; median (25th/75th percentile)] a lumbar spinal tap of 40 ml was performed, and 10 ml of serum was drawn. CSF, sequentially collected in 8 fractions of 5 ml (1st fraction: lumbar CSF; 8th fraction: cisterna magna-near CSF), was analyzed for complement protein C3, and the activation products C3a and sC5b-9 by enzyme immunoassay.

Results: The concentrations of the complement factors measured in fractions 1 and 8 of each individual patient were strongly correlated: C3 (Spearman's rank correlation coefficient r = 0.75, p = 0.0002); C3a (r = 0.93, p < 0.0001); sC5b-9 (r = 0.64, p = 0.002). CSF complement concentrations were lower in the cistern-near fraction 8 than in the lumbar fraction 1 (C3: p = 0.005; C3a: p = 0.0009; sC5b-9: p = 0.0003, Wilcoxon signed rank test). The concentrations of complement factors in CSF were two orders of magnitude lower than those in serum. C3 levels in the lumbar CSF strongly correlated with the lumbar CSF/serum albumin concentration quotient (Q) as a measure of the functionability of the blood-CSF barrier and the velocity of CSF flow (r = 0.84, p < 0.0001) suggesting diffusion of C3 from blood to CSF. The lumbar and cistern-near concentrations of C3a did not significantly correlate with Q (r = 0.26) pointing to a local conversion of C3 to C3a. The lumbar concentrations of sC5b-9 moderately correlated with Q (r = 0.62, p = 0.004). Plotting the CSF/serum quotient of C3 and sC5b-9 versus the Q revealed an approx. 50% local synthesis of C3, but a strong production of sC5b-9 in the CNS.

Conclusions: The increase of the complement concentrations from cisternal to lumbar CSF and the strong correlation of C3 with Q suggest that (1) a substantial portion of complement C3 in CSF originates from blood and (2) the complement system is mildly activated in the CSF of NPH patients.