Endotoxin in storage medium of human corneal grafts and clinical course after penetrating normal-risk keratoplasty.

PURPOSE: It is well known that endotoxins in storage medium may stimulate cytokine production and expression of adhesion molecules as well as endothelial damage in human corneal grafts. It has been supposed that endotoxin exposure of corneal grafts may, therefore, cause immune reactions and lead to reduced endothelial cell count after penetrating keratoplasty. It was the purpose of this prospective study to evaluate if this hypothesis is true. METHODS: A consecutive series of 274 samples of sterile organ culture storage medium from 274 human corneal grafts was collected between August 1998 and February 1999 and tested for endotoxin using Limulus amebocyte-lysate assay (LAL) after 7 days of organ culture. Threshold endotoxin level was set at 1.0 U/ml. A total of 161 grafts were transplanted and 113 were discarded. Within the 161 corneas transplanted, 62 were grafted to normal-risk patients and 99 to high-risk patients. Only normal-risk keratoplasty patients were included in the study and followed for at least 10 months. Immune reactions, graft failures, and postoperative endothelial cell counts were recorded. RESULTS: The mean endotoxin level in organ culture medium of all transplanted grafts was 1.07+/-2.96. Mean endotoxin level in organ culture medium of discarded grafts was 1.68+/-5.76, with 71 samples being below and 42 above the threshold of 1.0 U/ml called endotoxin-negative and endotoxin-positive, respectively. In all 36 culture medium samples from the 62 grafts transplanted to the group of normal-risk keratoplasty patients were endotoxin-negative and 26 endotoxin-positive. An influence of endotoxin levels on incidence of immune reactions, graft failure, and postoperative endothelial cell counts could not be revealed in patients with normal-risk keratoplasty. CONCLUSION: Low endotoxin levels in storage medium neither seem to promote immune reactions nor to contribute to postoperative chronic endothelial cell loss in normal-risk keratoplasty patients.

Strategies for PCR based detection of Burkholderia pseudomallei DNA in paraffin wax embedded tissues.

Recently, several cases of melioidosis imported to Europe have been reported. The diagnosis of the acute or chronic infection remains challenging. This report describes an optimised protocol for fast and reliable DNA preparation for use in two different polymerase chain reaction (PCR) assays, namely: (1) a seminested PCR assay targeting a genus specific sequence of the ribosomal protein subunit 21 (rpsU) gene and (2) a nested PCR assay targeting the gene encoding the filament forming flagellin (fliC). Various strains of Burkholderia spp, strains of closely related genera, and spleen tissue samples of experimentally infected mice were investigated. The combination of PCR and sequencing of the amplicons resulted in high sensitivity and specificity. These procedures may allow rapid, sensitive, and reliable detection of B pseudomallei DNA in routinely formalin fixed and paraffin wax embedded samples, thus providing a safe diagnostic tool and avoiding the cultivation of a risk group 3 agent. In addition, this method could be useful for retrospective histopathological investigations.

The cell wall-associated serine protease PrtA: a highly conserved virulence factor of Streptococcus pneumoniae.

The surface-associated subtilisin-like serine protease PrtA was identified by screening a genomic expression library from Streptococcus pneumoniae using a convalescent-phase serum. In Western blot analysis two forms of PrtA were detected in whole cell lysate and a truncated form only in culture supernatant suggesting that PrtA is produced as a precursor protein, translocated to the cell surface, truncated, and released into the surroundings. A 5' fragment of the gene was found highly conserved among 78 pneumococcal isolates of clinical relevance. Immunogenicity of PrtA, limited genetic variation, and the involvement in pneumococcal virulence demonstrated in in vivo experiments might identify PrtA as a promising candidate for a protein based vaccine.

Effect of deficiency of tumor necrosis factor alpha or both of its receptors on Streptococcus pneumoniae central nervous system infection and peritonitis.

Tumor necrosis factor alpha (TNF-alpha) and TNF-beta are key mediators in bacterial inflammation. We therefore examined the role of TNF-alpha and its two receptors in murine pneumococcal central nervous system infection. TNF-alpha knockout mice and age- and sex-matched controls and TNF receptor (p55 and p75)-deficient mice and heterozygous littermates were infected intracerebrally with a Streptococcus pneumoniae type 3 strain. Mice were monitored until death or were killed 36 h after infection. Bacterial titers in blood, spleen, and brain homogenates were determined. Leukocyte infiltration and neuronal damage were assessed by histological scores. TNF-alpha-deficient mice died earlier than the controls after intracerebral infection although overall survival was similar. TNF-alpha deficiency did not inhibit leukocyte recruitment into the subarachnoid space and did not lead to an increased density of bacteria in brain homogenates. However, it caused a substantial rise of the concentration of S. pneumoniae cells in blood and spleen. Spleen bacterial titers were also increased in p55- and p75-deficient mice. TNF receptor-deficient mice showed decreased meningeal inflammation. Neuronal damage was not affected by either TNF-alpha or TNF receptor deficiency. In a murine model of pneumococcal peritonitis, 10(2) CFU of S. pneumoniae produced fatal peritonitis in TNF-alpha-deficient, but not wild-type, mice. Early leukocyte influx into the peritoneum was impaired in TNF-alpha-deficient mice. The lack of TNF-alpha or its receptors renders mice more susceptible to S. pneumoniae infections.

Pneumolysin is the main inducer of cytotoxicity to brain microvascular endothelial cells caused by Streptococcus pneumoniae.

In pneumococcal meningitis it is assumed that bacteria cross the blood-brain barrier (BBB), which consists mainly of cerebral endothelial cells. The effect of Streptococcus pneumoniae on the BBB was investigated with an in vitro BBB model using a human brain microvascular endothelial cell line (HBMEC) and primary cultures of bovine brain microvascular endothelial cells (BBMEC). Within a few hours of incubation with pneumococci, rounding and detachment of the HBMEC were observed, and the transendothelial electrical resistance of the BBMEC monolayer decreased markedly. An S. pneumoniae mutant deficient in pneumolysin did not affect the integrity of the endothelial cell monolayer. Neither cell wall fragments nor isolated pneumococcal cell walls induced changes of endothelial cell morphology. However, purified pneumolysin caused endothelial cell damage comparable to that caused by the viable pneumococci. The cell detachment was dependent on de novo protein synthesis and required the activities of caspase and tyrosine kinases. The results show that pneumolysin is an important component for damaging the BBB and may contribute to the entry of pneumococci into the cerebral compartment and to the development of brain edema in pneumococcal meningitis.

D- and L-lactate in rabbit and human bacterial meningitis.

Increased total CSF lactate is an important indicator differentiating bacterial from aseptic meningitis. Bacteria can produce D- and L-lactate; mammalian cells produce only L-lactate. We measured D- and L-lactate production of Streptococcus pneumoniae, Staphylococcus aureus, Neisseria meningitidis and Escherichia coli in vitro, of S. pneumoniae and E. coli in rabbit experimental meningitis and of various common pathogens in CSF from patients with bacterial meningitis. Despite marked in vitro production of D-lactate by S. aureus (maximum: 4.59 mmol/l; i.e. 34.9% of total lactate), N. meningitidis (4.62 mmol/l; i.e. 98.1%) and E. coli (3.14 mmol/l; i.e. 97.2%), minimal amounts were measured in human S. aureus (0.38 mmol/l; i.e. 1.3% of total lactate) or N. meningitidis (0.28 mmol/l; i.e. 3.9%) and experimental E. coli meningitis (0.75 mmol/l; i.e. 4.4%). In only 9 of 54 human CSF samples did D-lactate exceed 0.15 mmol/l. S. pneumoniae did not produce significant amounts of D-lactate in vitro (maximum: 0.55 mmol/l; i.e. 2.7% of total lactate), in experimental meningitis (0.18 mmol/l; i.e. 3%) or in human cases of meningitis (0.28 mmol/l; i.e. 1.9%). In conclusion, increased total CSF lactate in meningitis consists mainly of L-lactate and originates predominantly from host cells. CSF D-lactate is of limited diagnostic value.

Antigenicity, expression, and molecular characterization of surface-located pullulanase of Streptococcus pneumoniae.

A putative pullulanase-encoding gene from Streptococcus pneumoniae was identified by screening a genomic expression library with human convalescent-phase serum. The 3,864-bp gene encoded a 143-kDa protein. Surface location and pullulanase activity of the protein, designated SpuA, was demonstrated. SpuA was present in all investigated pneumococcal isolates of different serotypes. The spuA 5' end was highly conserved among clinical isolates except for a 75-bp region. The properties of SpuA reported here indicate that this novel immunogenic surface protein might have potential as a vaccine target.

Induction of necrosis and apoptosis of neutrophil granulocytes by Streptococcus pneumoniae.

Apoptosis followed by macrophage phagocytosis is the principal mechanism by which neutrophil granulocytes (PMN) are removed from the site of inflammation. To investigate whether Streptococcus pneumoniae causes apoptosis of PMN, we exposed PMN to viable and heat-killed pneumococci and purified pneumococcal cell walls (PCW). The occurrence of PMN cell death was quantified by flow cytometry using annexin V/propidium iodide labelling of the cells. Intracellular histone-associated DNA fragments were quantified by ELISA. The presence of apoptosis was confirmed by in situ tailing. Exposure of PMN to viable pneumococci caused necrosis of the cells. The pneumococcal cytotoxin pneumolysin, the bacterial production of hydrogen peroxide, and PCW contributed to necrosis. Heat-killed pneumococci accelerated the process of apoptosis observed in cultivated non-stimulated PMN in vitro. These results demonstrated that pneumococci induce PMN cell death. Depending on the intensity of the stimulus, PMN necrosis and apoptosis were observed.

A review on melioidosis with special respect on molecular and immunological diagnostic techniques.

Melioidosis is an 'emerging' tropical disease which causes diagnostic problems in endemic and especially in nonendemic areas e.g. Germany when imported. In the last decade, many efforts have been made to develop new molecular and immunological techniques for diagnostic use. However, an actual comprehensive review does not exist. Apart from classical microbiological procedures (microscopy, culture and biochemical identification) efforts have been made to identify Burkholderia pseudomallei using specific antibodies and PCR. The direct antigen detection can be done within a few hours leading to diagnosis days before cultural proof. ELISA and immunoblot techniques were examined for their ability to replace indirect hemagglutination and immunofluorescence test in serology. The diagnostic value of serological procedures for early detection of melioidosis is limited, however. The rapid and reliable diagnosis of melioidosis is required for an adequate onset of therapy. But no evaluated test kit based on the detection of specific antibodies, specific antigens, or on the amplification of species-specific DNA sequences is commercially available up to now. Even PCR testing--primers can easily be ordered by many gene technology companies--can not be recommended as B. pseudomallei DNA for positive controls is not available. Therefore, only microscopy and biochemical identification systems like the API 20NE can be used in the routine laboratory up to now. At this point, it has to be stressed that B. pseudomallei is a level 3 agent in many countries e.g. Germany and that only laboratories with high containment are allowed to handle it. In all cases the help of an experienced reference laboratory is adviced.

Detection of 23 immunogenic pneumococcal proteins using convalescent-phase serum.

A genomic expression library of Streptococcus pneumoniae was screened with a convalescent-phase serum for immunoreactive proteins. Six known and 17 unknown pneumococcal proteins were detected. Five of the known proteins were surface-located virulence factors, and eight of the unknown proteins were putative membrane proteins.

Clinical outcome in pneumococcal meningitis correlates with CSF lipoteichoic acid concentrations.

Lipoteichoic and teichoic acids are components of the cell wall of Streptococcus pneumoniae. A recently developed enzyme immunoassay was used in patients with pneumococcal meningitis to investigate lipoteichoic and teichoic acid concentrations in CSF at the first lumbar puncture in relation to the clinical outcome determined by the Glasgow Outcome Scale. Lipoteichoic and teichoic acid concentrations in CSF were significantly associated with neurologic sequelae and mortality in S. pneumoniae meningitis.

Rifabutin for experimental pneumococcal meningitis.

Rifabutin is a lipophilic antibacterial with high in vitro activity against many pathogens involved in bacterial meningitis including pneumococci. Resistance to beta-lactam antibiotics in pneumococci is not associated with a decreased sensitivity to rifabutin (30 strains from Germany with intermediate penicillin resistance; MIC range of penicillin: 0.125-1 mg/l, MIC of rifabutin: < 0.008-0.015 mg/l). Rifabutin at doses of 0.625, 1.25, 2.5, 5 and 10 mg/kg/h i.v. was investigated in a rabbit model of meningitis using a Streptococcus pneumoniae type 3 (MIC/MBC of rifabutin: 0.015/0.06 mg/l). The bacterial density in CSF at the onset of treatment was 7.3 +/- 0.6 log CFU/ml (mean +/- SD). Rifabutin decreased bacterial CSF titers in a dose-dependent manner [delta log CFU/ml/h (slope of the regression line log CFU/ml vs. time) at a dose of 0.625 mg/kg/h: -0.16 +/- 0.06 (n = 3), at 1.25 mg/kg/h: -0.20 +/- 0.12 (n = 4), at 2.5 mg/kg/h: -0.24 +/- 0.04 (n = 4), at 5 mg/kg/h: -0.31 +/- 0.10 (n = 8), and at 10 mg/kg/h: -0.29 +/- 0.10 (n = 5)]. At high doses rifabutin was as active as ceftriaxone at 10 mg/kg/h (delta log CFU/ml/h: -0.29 +/- 0.10, n = 10). Two and 5 h after initiation of therapy, CSF TNF-alpha activities were lower with rifabutin 5 mg/kg/h than with ceftriaxone (medians 2 vs. 141 U/ml, p = 0.005 at 2 h; median 51 vs. 120 U/ml 5 h after initiation of therapy, p = 0.04). This did not result, however, in a decrease of indicators of neuronal damage. In conclusion, intravenous rifabutin was bactericidal in experimental pneumococcal meningitis. Provided that a well-tolerated i.v. formulation will be available it may qualify as a reserve antibiotic for pneumococcal meningitis, in particular when strains with a reduced sensitivity to beta-lactam antibiotics are the causative pathogens.

Trovafloxacin delays the antibiotic-induced inflammatory response in experimental pneumococcal meningitis.

This study evaluates the ability of the new fluoroquinolone trovafloxacin to attenuate the inflammatory burst known to occur after initiation of antibiotic treatment in pneumococcal meningitis. After exposure to trovafloxacin or ceftriaxone for 3 h in vitro, Streptococcus pneumoniae was injected intracisternally (i.c.) into rabbits every 3 h over 9 h (n = 6 for each antibiotic). Ceftriaxone-treated S. pneumoniae induced consistently higher CSF leucocyte counts (median 2568/microL versus 543/microL at 6 h; P = 0.03; 4560/microL versus 2207/microL at 18 h; P = 0.03) than trovafloxacin-treated bacteria. Meningitis induced in rabbits by i.c. injection of live S. pneumoniae was treated with equal doses of trovafloxacin or ceftriaxone i.v. (ten per group). The bactericidal rates of both antibacterial agents in CSF were almost identical. In comparison with ceftriaxone, trovafloxacin resulted in lower tumour necrosis factor (TNF) and interleukin 1beta (IL-1beta) CSF levels 2 h after the initiation of treatment (TNF levels, median 26 U/mL versus 141 U/mL; P = 0.02; IL-1beta levels 455 pg/mL versus 1399 pg/mL; P = 0.02). Twelve hours after initiation of therapy, however, TNF and IL-1beta were higher in trovafloxacin-treated animals (TNF, 61 U/mL versus 7 U/mL; P = 0.001; IL-1beta, 4320 pg/mL versus 427 pg/mL; P = 0.006). The increase in CSF lactate was less during trovafloxacin therapy than with ceftriaxone (median: 2.0 mmol/L versus 4.0 mmol/L; P = 0.03). In conclusion, S. pneumoniae treated in vitro with trovafloxacin induced less CSF leucocytosis than ceftriaxone-treated S. pneumoniae. After i.c. inoculation of live S. pneumoniae, trovafloxacin therapy delayed, but did not inhibit, the release of the proinflammatory cytokines TNF and IL-1beta, probably by slowing the liberation of bacterial cell wall components into the subarachnoid space.

Limited efficacy of pentoxifylline as anti-inflammatory agent in experimental pneumococcal meningitis.

Dexamethasone appears to show some adverse side-effects as adjunctive anti-inflammatory agent in bacterial meningitis. For this reason, we tested the anti-inflammatory and neuroprotective effect of pentoxifylline administered 15 min before starting antibiotic treatment with ceftriaxone (n = 10) versus antibiotic therapy alone (n = 9) in the rabbit model of pneumococcal meningitis. Pentoxifylline lowered the medians of leucocyte density, tumour necrosis factor-alpha (TNF-alpha) and lactate in the cerebrospinal fluid (CSF), but only leucocyte migration into the subarachnoid space was significantly inhibited 8 h after initiation of therapy (P = 0.01). CSF protein, brain water content, and the entry of ceftriaxone into CSF were not influenced by pentoxifylline. The density of neuronal apoptoses in the dentate gyrus was slightly lower in animals receiving pentoxifylline than in those treated with ceftriaxone only. The median concentration of neuron-specific enolase in CSF was lower in the pentoxifylline-treated group, but the difference was not significant. In conclusion, pentoxifylline showed some anti-inflammatory activity in pneumococcal meningitis, but the substance failed significantly to reduce neuronal damage.

Elimination of blood-derived macrophages inhibits the release of interleukin-1 and the entry of leukocytes into the cerebrospinal fluid in experimental pneumococcal meningitis.

Parameters of inflammation during pneumococcal meningitis were determined in rabbits after monocyte elimination by dichloromethylene diphosphonate (Cl(2)MDP)-containing mannosylated liposomes in comparison with untreated controls. Monocyte depletion reduced the migration of white blood cells into the cerebrospinal fluid (CSF) (medians: 42 versus 2146/mm3 at 18 h, 323 versus 7413/mm3 at 24 h p.i., p < 0.01). CSF IL-1beta concentrations were lower in depleted animals (379 versus 3282 pg/ml, 24 h p.i., p < 0.01), whereas TNF-alpha concentrations were not different. Monocyte-depleted animals lost body temperature during the experiment carried out in anaesthesia (p = 0.01) indicating that macrophages are necessary for thermogenesis during meningitis.

Anti-inflammatory treatment influences neuronal apoptotic cell death in the dentate gyrus in experimental pneumococcal meningitis.

Apoptotic neuronal death and the increase of neuron-specific enolase (NSE) in cerebrospinal fluid (CSF) were studied in a rabbit model of experimental pneumococcal meningitis after treatment with antimicrobial (ceftriaxone) and antiinflammatory agents (dexamethasone, monoclonal antibodies against the beta-subunit of beta 2-integrins [anti-CD18 mAb]). Twenty-four hours after infection, apoptotic cell death was found solely in the granular cell layer of the dentate gyrus. Neurons with DNA fragmentation were quantified with the in situ tailing (IST) reaction. Dexamethasone and anti-CD18 mAb inhibited the NSE increase in CSF significantly (p = 0.003, p = 0.011). After administration of dexamethasone the density of apoptotic neurons was significantly higher than in control animals receiving only ceftriaxone (p = 0.044). The median of the density of apoptotic neurons was lower in the dentate gyrus in animals receiving anti-CD18 mAb and ceftriaxone vs those receiving only ceftriaxone, although the difference did not reach statistic significance (p = 0.058). In conclusion, apoptotic cell death occurs in the dentate gyrus during the early phase of bacterial meningitis. The extent was influenced by antiinflammatory therapy. The systemic administration of glucocorticoids increased the quantity of apoptotic neurons in the dentate gyrus but reduced overall neuronal damage as indicated by low levels of NSE concentration in CSF.

Activity of fosfomycin in a rabbit model of experimental pneumococcal meningitis.

Fosfomycin is an antibacterial substance of low molecular weight and negligible binding to plasma proteins exhibiting in-vitro activity against most pathogens involved in bacterial meningitis including pneumococci. Due to these properties the drug has been recommended for therapy of central nervous system (CNS) infections. For this reason, fosfomycin at doses of 10, 40, 80 and 160 mg/kg/h iv, was investigated in the rabbit model of pneumococcal meningitis. Bacterial counts in cerebrospinal fluid (CSF) before, and 2, 5 and 8 h after initiation of therapy were quantitated by plating on blood agar. Fosfomycin concentrations in serum and CSF were determined by the agar well diffusion method. The MIC and MBC of fosfomycin for the Streptococcus pneumoniae type 3 strain used was 4 and 32 mg/L, respectively. The MIC of ceftriaxone was 0.016 mg/L. In vitro, both drugs showed an additive effect (fractional inhibitory concentration index = 0.75). In vivo at each dose tested, fosfomycin was less active than ceftriaxone (means +/- S.D.): delta log cfu/mL/h at 10 mg/kg/h + 0.130 +/- 0.062 (n = 2), at 40 mg/kg/h -0.217 +/- 0.185 (n = 3), at 80 mg/kg/h -0.270 +/- 0.121 (n = 3), at 160 mg/kg/h -0.331 +/- 0.118 (n = 3) vs -0.647 +/- 0.193 at 10 mg/kg/h ceftriaxone (n = 3). CSF penetration of fosfomycin as estimated by the CSF-to-serum concentration ratio at 8 h was 0.55 +/- 0.22 (n = 11). For bactericidal activity CSF concentrations of at least ten times the MIC were necessary. Coadministration of both drugs (1 mg/kg/h ceftriaxone + 40 mg/kg/h fosfomycin) tended to be more active than either drug alone (in-vivo drug interaction = 1.3). In conclusion, fosfomycin at very high doses reduced bacterial counts in CSF. However, fosfomycin CSF concentrations usually observed in patients with meningitis receiving fosfomycin were not bactericidal in this model. At all doses tested the bactericidal rate was lower than that of ceftriaxone. Fosfomycin is therefore unsuitable as a single agent, but may be used as a reserve antibiotic in combination with a newer cephalosporin for pneumococcal meningitis unresponsive to conventional therapy.

[Bacterial CNS infections in adults in Southern Lower Saxony. A retrospective study of the Göttingen Neurologic University Clinic]. / Bakterielle ZNS-Infektionen bei Erwachsenen in Südniedersachsen. Eine retrospektive Untersuchung an der Neurologischen Universitätsklinik Göttingen.

All 155 patients with suspected bacterial central nervous system (CNS) infections treated from 1986 to 1991 at the Department of Neurology, University of Göttingen, were evaluated in a retrospective study. According to the clinical symptoms presented at admission, 7 cases were classified as encephalitis, 44 as meningitis, 15 as radiculitis, 19 as ventriculitis, 61 as meningoencephalitis and 9 as meningoradiculitis. In 78% of these cases, the causative bacteria were either isolated from cerebrospinal fluid (CSF), or other relevant sources (blood, wound swabs) or identified by serological methods; (all cases of CNS borreliosis, and 3 of the 5 cases of listeriosis were identified by means of the last mentioned method). CNS infections caused by staphylococci and Borrelia burgdorferi were most frequent, followed by those due to pneumococci, meningococci and other streptococci. CNS infections caused by Mycobacterium tuberculosis, Listeria monocytogenes, Haemophilus influenzae and enterobacteriaceae were less frequent. In comparison to the CNS infections due to other bacteria, the pneumococcal and meningococcal meningitic infections were associated with more pronounced CSF alterations (on the average, there were higher white blood cell counts, and higher CSF protein and lactate). Pneumococci predominated in older patients and those with an impaired immune system, or infections of organs neighboring the CNS. Meningococci were most frequent in young and previously healthy individuals. All patients with CNS listeriosis had predisposing conditions. Meningococcal meningitis was either fatal or resolved with or without minimal neurological deficits. Infections caused by staphylococci or pneumococci were associated with a high percentage of neurologic sequelae.

Pharmacokinetic quantification of the exchange of drugs between blood and cerebrospinal fluid in man.

Various parameters which may be useful in quantification of drug transit from blood into CSF and vice versa after a short duration infusion are compared here by recalculating previously published data from our group. Due to the slower entry into and elimination from the CSF compartment as compared to the central compartment, the ratio of drug concentrations in CSF and serum sampled at the same time increase with time after an infusion. Therefore, concentration quotients of simultaneously drawn blood and CSF are inadequate to characterise CSF penetration. The ratio of the areas under the concentration-time curves in a body fluid and serum (AUCbody fluid/AUCs) is an established measure to quantify overall penetration from the central into a peripheral compartment. AUCCSF/AUCs is closely correlated with the quotient of the maximum CSF and serum concentrations (CmaxCSF/CmaxS) (rs = 0.87, n = 42, P < 0.001) and with the rate constant of distribution in CSF (CLin/VCSF) (rs = 0.80, n = 42, P < 0.001). Since CmaxCSF/CmaxS depends on the mode of drug administration, it is suggested that AUCCSF/AUCs be used to quantify overall drug transit into CSF. CLin/VCSF is of use when CSF can only be sampled once, or when the velocity of the transit of a drug into CSF is to be described. The CSF exit rate constant (CLout/VCSF) characterises elimination from CSF independent of the elimination from serum and may be applied to estimate the formation rate of CSF; in the present study it averaged 20 ml/h.