CLINICAL STUDIES

A MICRODIALYSIS TECHNIQUE FOR ROUTINE

MEASUREMENT OF MACROMOLECULES IN THE
INJURED HUMAN BRAIN
Jan Hillman, M.D., Ph.D.
Department of Neurosurgery,
University Hospital,
Linköping, Sweden
Oscar Åneman, M.D.,
Ph.D.
Department of Neurosurgery,
University Hospital,
Linköping, Sweden
Chris Anderson, M.D.,
Ph.D.
Department of Dermatology,
University Hospital,
Linköping, Sweden
Florence Sjögren, Ph.D.
Department of Dermatology,
University Hospital,
Linköping, Sweden
Carina Säberg, R.N.
Department of Neurosurgery,
University Hospital,
Linköping, Sweden
Pekka Mellergård, M.D.,
Ph.D.
Department of Neurosurgery,
University Hospital,
Linköping, Sweden
Reprint requests:
Jan Hillman, M.D., Ph.D.,
Neurosurgical Department,
University Hospital,
S-581 85 Linköping, Sweden.
OBJECTIVE: To evaluate a new intracerebral microdialysis catheter with a high-cutoff
membrane and its potential for the study of macromolecules in the human brain.
METHODS: Paired intracerebral microdialysis catheters were inserted in 10 patients
who became comatose after subarachnoid hemorrhage or traumatic brain injury and
were then treated in our neurosurgical unit. The only differences from the routine use
of microdialysis in our clinic were the length (20 mm) and cutoff properties of the
catheter membranes (100 kD) and the perfusion fluids used (standard perfusion fluid,
3.5% albumin, or Ringer-dextran 60). Samples were weighed (for net fluid fluxes) and
analyzed at bedside (for routine metabolites) and later in the laboratory (for total
protein and interleukin-6). The in vitro recovery of glucose, glutamate, and glycerol
were also investigated under different conditions.
RESULTS: Even brief perfusion with standard perfusion fluid resulted in a significant
loss of volume from the microdialysis system. For albumin and Ringer-dextran 60 fluid,
recovery was comparable to standard settings. Interleukin-6 (highest value close to
25,000 pg/ml) was sampled from all catheters, and total protein was analyzed from
catheters perfused with Ringer-dextran 60 (average concentration, 234 ␮g protein/ml).
There were detectable patterns of variations in the concentration of interleukin-6,
seemingly related to concomitant variations in intracerebral conditions. In the present
study, no direct comparison was made with the standard CMA 70 catheter (CMA
Microdialysis, Stockholm, Sweden), but in vivo, the measured mean concentrations of
glucose, glycerol, lactate, and pyruvate were comparable to those previously reported
from standard catheters. In vitro, the recovery of metabolites was better when using
Ringer-dextran 60 compared with albumin.
CONCLUSION: Microdialysis catheters with high-cutoff membranes can be used in
routine clinical practice, allowing for sampling and analysis of cytokines and other
macromolecules.
KEY WORDS: Cerebral microdialysis, Head trauma, Intensive care, Interleukin-6, Monitoring, Protein,
Subarachnoid hemorrhage
Neurosurgery 56:1264-1270, 2005 DOI: 10.1227/01.NEU.0000159711.93592.8D www.neurosurgery-online.com

Email: Jan.Hillman@lio.se
Received, February 17, 2004.
Accepted, January 20, 2005.
M
icrodialysis has become a major tool
in the study of brain biochemistry in
both experimental settings and clin-
ical practice (2, 8, 20). Until recently, only cath-
eters with membrane cutoff properties (20 kD)
permitting recovery of small tissue molecules
were commercially available for use in hu-
mans (9, 12). Even so, such catheters have
profoundly influenced the study of ischemic
and other injuries to the human brain through
sequential analysis of neurotransmitters and
metabolites and, in addition, have been of
clinical use in the neurointensive care unit
(NICU) for patient monitoring (4, 10, 19, 21).
Proteins act as important regulators of cel-
lular responses to injury. It seems likely that
cerebral microdialysis in humans, with stud-
ies of macromolecules, such as cytokines, neu-
rotrophic factors, or enzymes, would add a
new and important dimension to the under-
standing of brain injury and subsequent re-
parative processes (11, 15, 22). Microdialysis

1264 | VOLUME 56 | NUMBER 6 | JUNE 2005 www.neurosurgery-online.com

 HUMAN INTRACEREBRAL MICRODIALYSIS
membranes with high cutoff properties, allowing sampling of
macromolecules, have been used successfully in animal exper-
iments and in human dermis (6, 17). We are aware of only one
previous report using such technology for studies of proteins
in the human brain (23). Because the method presented there
seems to have limitations for routine clinical use, the aim of
the present study was to investigate the potential for a simpler
method. With minimal changes of the routines used for mi-
crodialysis in our NICU for several years, we investigated
properties of a new microdialysis catheter with a high-cutoff
membrane in a series of seriously injured patients.
PATIENTS AND METHODS
Intracerebral microdialysis using a modified type of cathe-
ter (see below) was performed in 10 patients treated in our
NICU. In this unit, microdialysis with conventional catheters
has been used routinely for several years as one of several
methods in a multimodality monitoring system designed to
meet the therapeutic challenge of each individual patient.
Ethical approval was granted by the Ethical Committee of the
University Hospital of Linköping.
Catheter Placement and Microdialysis Setup
Insertion of the catheters was performed according to our
normal routines, except that for this study, pairs of catheters
(instead of single catheters) were used. The catheters were
inserted via burr holes or through a craniotomy at the end of
open surgery. The positioning of catheters followed our pre-
viously established routines for microdialysis monitoring. Ac-
cording to these routines, catheters are most often placed in
the frontal lobe, 1 to 2 cm anterior to the coronary suture,
although in some patients, catheters are placed in tissues
considered “at risk,” e.g., in the proximity of a contusion.
Irrespective of locality, the two catheters were always posi-
tioned in close proximity to each other to allow for compari-
son of their properties in areas with comparable biochemistry.
All catheters were provided by CMA Microdialysis, Stock-
holm, Sweden, and were identical to the CMA 70 catheter
routinely used in our clinic, except that the membrane length
was 20 mm (instead of 10 mm) and had cutoff properties at
molecular weights of 100,000 (instead of 20,000). Perfusion of
the microdialysis system was standardized at 0.3 ␮l/min,
using a commercially available perfusion pump (CMA 106;
CMA Microdialysis). The outflow hydrostatic pressure of the
perfusion system was set at the zero midcranial reference level
by attaching the perfusate collecting vials to the bandage on
the patient’s head. In a few early patients, the microdialysis
system was initially perfused with standard perfusion fluid
(CMA Microdialysis) for a few hours. Subsequently, all cath-
eters were perfused with either Ringer-dextran 60 (RD60)
(Braun Medical AB, Stockholm, Sweden) or human albumin
solution (Albumin Immuno, 35 mg/ml; Baxter Medical AB,
Stockholm, Sweden), one perfusate fluid in each of the paired
catheters.
NEUROSURGERY
OF MACROMOLECULES
In Vivo Assays
Samples were analyzed at bedside for glucose, glutamate,
glycerol, lactate, pyruvate, and urea using the CMA 600 ana-
lyzer (CMA Microdialysis). The perfusates collected during
the initial 2 to 3 hours were discarded, and thereafter, consec-
utive 4- to 6-hour samples were collected continuously and
were weighed at a 0.1-␮g precision level to estimate net trans-
membrane fluid fluxes. The same samples were then frozen
(⫺70 C°) for later analysis in the laboratory for total protein
content and the cytokine interleukin-6 (IL-6). Protein was mea-
sured with a DC Protein Assay kit from Bio-Rad (Stockholm,
Sweden). This kit is based on the Bradford method, which
measures basic and aromatic amino acids. A sample volume of
5 ␮l was used for analyzing the total protein concentration.
IL-6 was determined by use of enzyme-linked immunosorbent
assay Quantiglo kits, a chemiluminescence-based method. The
IL-6 kit allows detection in the range of 0.3 to 3000 pg/ml.
(Samples with concentration values falling outside this range
at the first analysis were diluted and reanalyzed.) The assays
were performed according to the manufacturer’s instructions.
Standards and samples (when possible) were analyzed in
duplicate.
In Vitro Assays
Microdialysis catheters were placed in standard solutions of
glucose, glycerol, and glutamate provided by CMA. RD60 or
human albumin solution was perfused through the catheters
at a flow rate of 1 and 0.3 ␮l/min. Samples were assayed on
the CMA 600 analyzer (see above).
RESULTS
The 10 patients investigated had experienced either a trau-
matic brain injury or severe aneurysmal subarachnoid hem-
orrhage. On admission and during the monitoring, they were
comatose, with a Glasgow Coma Scale score of 9 or less.
Microdialysis continued from 1 to 10 days. In 1 patient, per-
fusion with 3.5% albumin was accompanied by strong net
fluid absorption. This was most likely because of a membrane
failure, related to technical problems during insertion of the
catheter. Microdialysis was discontinued, and the data from
this patient were not included in the material, which therefore
refers to the observations in 9 patients.
Pilot observations made it clear that even a brief perfusion
(4–8 h) with standard perfusion fluid resulted in a significant
loss of volume from the system, which had been suspected
already by visual inspection of the collecting microvial. The
idea of using standard perfusion fluid for comparison with
other types of perfusates was therefore abandoned right at the
beginning of the study.
For the two other types of perfusate, fluid recovery was
comparable to what is observed during standard settings of
clinical microdialysis (18–20). Data on transmembrane fluid
balance for the different perfusion solutions are given in Table
1. For RD60, the average fluid recovery was 94.0%, with little
VOLUME 56 | NUMBER 6 | JUNE 2005 | 1265
HILLMAN ET AL.
TABLE 1. Fluid recovery from microdialysis catheters perfused
with Ringer-dextran 60 or 3.5% albumin (paired observations
in nine patients, n ⴝ 114)a
Ringer-dextran 60 Albumin 3.5%
Mean 94% 100.1%
95% CI 92.5–95.5% 90.7–109.2%
a
CI, confidence interval.
variation between different patients. For 3.5% albumin, the
average fluid recovery was 100.1%, with the spreading of data
between different patients being somewhat larger compared
with RD60.
Total protein content was not analyzed regularly, and only
in samples collected from catheters perfused with RD60. In
samples in which analyses were made, protein concentration
varied between 50 and 1520 ␮g protein/ml, the average con-
centration being 234 ␮g protein/ml (213–254 ␮g protein/ml;
95% confidence interval). Figure 1 illustrates variations in total
protein over time in two patients (Patients A and B).
During the course of monitoring, IL-6, which was selected
as a model molecule of measurable regulatory proteins in the
human brain, was sampled from all patients. IL-6 was also
detected in all the catheters perfused with albumin. The IL-6
concentration varied between patients and over time in any
given patient. Initial values were seldom below 1000 pg/ml.
The highest value for IL-6 was close to 25,000 pg/ml.
Two examples of the variation of IL-6 concentration over time
(and between different patients) are given in Figures 2 and 3.
Figure 2A illustrates the temporal profile of IL-6 in extracted
extracellular fluid in a patient (Patient C) who had experienced a
severe traumatic brain injury. The microdialysis catheters were
positioned in an area close to a major contusion soon after injury.
As shown in Figure 2B, high levels of glycerol and a high lactate-
FIGURE 1. Graph showing temporal profiles of total protein concentra-
tion in microdialysis fluid sampled from two patients (Patients A and B)
with subarachnoid hemorrhage.
1266 | VOLUME 56 | NUMBER 6 | JUNE 2005
to-pyruvate ratio were ob-
served, indicating threatening
ischemic conditions. IL-6 in-
creased rapidly and then
gradually declined (Fig. 2A),
parallel to reduction of the
ischemic stress to the tissue
(Fig. 2B).
Figure 3A illustrates the IL-6
profile in a patient (Patient D)
with subarachnoid hemor-
rhage. At the end of the first
week of observation, this pa-
tient developed clinical signs
of vasospasm. Transcranial
Doppler showed increased
flow velocities, with a com-
puted tomographic scan de-
picting areas of lowered den-
sity. Focal ischemia was also
FIGURE 2. Graphs showing tem-
indicated by increases in ex- poral profile of intracerebral IL-6
tracellular glycerol and in the (A) and the corresponding concen-
lactate-to-pyruvate ratio (Fig. tration of glycerol and the lactate/
3B). As illustrated in Figure pyruvate (L/P) ratio (B) in a patient
3A, IL-6 concentration (Patient C) who had experienced
showed a sharp increase as severe traumatic head injury.
ischemia developed. Threatening ischemia was accompa-
The measured mean me- nied by rapidly increasing IL-6, fol-
tabolite concentrations of lowed by a gradual decline (see
Results).
some important metabolites
(glucose, glutamate, glyc-
erol, lactate, and pyruvate) were comparable to what would
be expected from measurements with standard catheters (e.g.,
CMA 70), as summarized in Table 2. Clearly, the spread of
these values depends on the types of patients surveyed. For
our present purpose, it was more important to compare the
recovery rates of catheters perfused with different types of
fluids (albumin and RD60, respectively). Thus, the relation-
ship (quota) between the concentration of each metabolite in
the two different types of perfusion fluid was measured at
each sampling time. In Table 2, “Q” is the average value for all
such paired comparisons. When comparing the measurements
of metabolites in this way, the overall ratio of recovery be-
tween albumin- and RD60-perfused catheters was 1.06. As
also shown in Table 2, glutamate and glucose showed higher
recovery in albumin, whereas the recovery of lactate and
pyruvate was slightly lower in albumin.
The results from the in vitro recovery studies of some me-
tabolites are presented in Table 3. In vitro, the recovery for the
metabolites was better when using RD60 than when using
albumin. As expected, increasing the perfusion flow rate from
0.3 to 1.0 ␮l/min was accompanied by a lower degree of
recovery. Albumin showed a low recovery percentage regard-
less of the flow rate used. The highest recovery was seen when
RD60 was used as the perfusion fluid at 0.3 ␮l/min.
www.neurosurgery-online.com
 HUMAN INTRACEREBRAL MICRODIALYSIS OF MACROMOLECULES

DISCUSSION related to the fourth power of its radius, and the overall
hydrodynamic properties of the membrane are described by
The present study shows the filtration coefficient. The intraluminal and tissue hydro-
that microdialysis catheters static pressures are balanced against each other, and the col-
with high-cutoff membranes, loid osmotic pressure of the perfusate is balanced against the
allowing for sampling of cyto- colloid osmotic pressure of the surrounding extracellular fluid
kines and other macromole- (2, 20). Several approaches to balancing transmembrane fluid
cules, can be used in routine fluxes in microdialysis catheters with large pore membranes
clinical practice. The in vitro have been described for various tissues (1, 17, 23). However,
recovery data and the fact that these methods have shortcomings entailing limitations for
a significant amount of pro- their application in neurosurgical patients, particularly if the
teins could be recovered from goal is to use microdialysis as a clinical routine in an NICU
the extracellular fluid in neu- setting.
rosurgical patients clearly In the present approach, we chose to use widely available
point to the potential of such equipment favored by many neurosurgical clinics as part of
probes for extending the ex- their patient monitoring equipment. The outflow pressure of
ploration and monitoring of the microdialysis system was kept at the level of the intrace-
human brain biochemistry. rebral catheter, i.e., close to the zero level, thus avoiding the
With a standard mem- problems of a “push and pull” technique. Instead, the focus
brane, only very small net fil- FIGURE 3. Graphs showing tem- was on the choice of perfusion fluid and, in particular, the
tration of standard perfusion poral profile of intracerebral IL-6 colloid osmotic pressure of the perfusate. For routine applica-
(A) and the corresponding concen-
fluid occurs to the brain at tion, a commercially available pharmaceutical product would
tration of glycerol and the lactate/
the perfusion rate of 0.3 ␮l/ pyruvate (L/P) ratio (B) in a patient be favored as perfusate. It is therefore encouraging that our
min (2, 8, 20). However, with (Patient D) with subarachnoid hem- data show that both 3.5% human albumin solution and RD60
the larger pores of the mem- orrhage (SAH). At the end of the effectively counterbalance the outward hydrostatic force over
brane used in the present first week, the patient developed the membrane segment, thereby minimizing net fluid loss.
study, there is an excessive clinical and Doppler-verified vaso- For both albumin and RD60, fluid recovery was comparable
loss of standard perfusion spasm, with an accompanying to that observed with standard perfusion fluid, with less in-
fluid to the surrounding tis- increase in IL-6 concentration. terpatient variation experienced with RD60. In the present
sue, as was also observed in study, no direct comparison was made with the standard
our pilot experience (see above). This kind of “leakage” has catheter (e.g., CMA 70), but the recovery of metabolites
been a major obstacle in attempts to develop a clinical routine seemed to be comparable to what has been reported previ-
microdialysis technique with high-cutoff membrane catheters ously for standard catheters (18, 19).
for the human brain (9, 23). An important finding of the study was that significant
The transmembrane fluid balance mechanism in a microdi- amounts of cytokines and other proteins could be sampled
alysis catheter is comparable to that of a single capillary. from every patient. With the possible exception of the S-100
Resistance to fluid flux through a single pore is inversely protein marker for brain injury (13), at present, proteins have
no defined role in clinical
multimodality monitoring of
the injured brain. However,
TABLE 2. Measurements of five different metabolites and comparison of recovery rates during it may well prove that some
microdialysis in nine neurointensive care unit patients with paired catheters, perfused with Ringer-
a macromolecular mediators of
dextran 60 or 3.5% albumin
early tissue injury or some
Metabolite Q (alb/RD60) 3.5% albumin RD60 mediators of early reparative
Glucose (mmol/L) 1.32 (1.04 –1.61) 1.9 (1.7–2.2) 1.7 (1.5–2.0) processes in injured tissue
can become sensitive indica-
Glutamate (␮mol/L) 1.53 (1.13–1.92) 149 (74 –224) 142 (69 –216) tors of the progression of in-
jury or the effects of therapy.
Glycerol (␮mol/L) 0.99 (0.88 –1.10) 216 (155–279) 236 (180 –292)
Among such macromole-
Lactate (mmol/L) 0.91 (0.85–1.01) 6.6 (5.9 –7.2) 7.8 (7.0 – 8.6) cules are chemokines and cy-
tokines (with powerful
Pyruvate (␮mol/L) 0.90 (0.85– 0.95) 230 (209 –251) 275 (246 –304)
proinflammatory and anti-
a
The relationship (quota) between concentrations of each metabolite in the respective perfusion fluid was measured at inflammatory actions); en-
each sampling interval, Q being the average value for all such paired comparisons (values given are mean and 95% zymes, e.g., cathepsins (likely
confidence interval; n ⫽ 114). alb, albumin; RD60, Ringer-dextran 60.
to play a role in apoptosis);
and a host of different neuro-

NEUROSURGERY VOLUME 56 | NUMBER 6 | JUNE 2005 | 1267

HILLMAN ET AL.

3. Billman GF, Hughes AB, Dudell GG, Waldman E, Adcock LM, Hall DM,
Orsini N Jr, Koska AJ, Van Marter LJ, Finer NN, Kulhavy JC, Feld RD,
TABLE 3. In vitro recovery of glucose, glutamate,
Widness JA: Clinical performance of an in-line, ex vivo point of care mon-
and glycerola itor: A multicenter study. Clin Chem 48:2030–2043, 2002.
% Recovery 4. Bullock RZ, Myseros JS, Marmarou A, Woodward JJ, Young HF: Evidence
Perfusion flow for prolonged release of excitatory amino acids in severe human head
Substance Albumin
(␮l/min) RD60 No. No. trauma. Ann N Y Acad Sci 765:290–295, 1995.
(3.5%) 5. Faden AI: Neuroprotection and traumatic brain injury: Theoretical option or
realistic proposition. Curr Opin Neurol 15:707–712, 2002.
Glucose 0.3 99.6 2 84.2 2 6. Fassbender K, Schneider S, Bertsch T, Schlueter D, Fatar M, Ragoschke A,
Kuhl S, Kischka U, Hennerici M: Temporal profile of release of
1.0 73.4 5 48.5 7 interleukin-1␤ in neurotrauma. Neurosci Lett 284:135–138, 2000.
7. Gozal E, Gozal D, Pierce WM, Thongboonkerd V, Schein JA, Brittain KR,
Glutamate 0.3 90.6 2 80.0 2 Guo SZ, Klein J: Proteomic analysis of CA1 and CA3 regions of hippocam-
pus and differential susceptibility to intermittent hypoxia. J Neurochem
1.0 71.8 5 42.5 7
83:331–345, 2002.
8. Hamani CL, Dujovny M: Microdialysis in the human brain: Review of its
Glycerol 0.3 97.7 2 83.8 2
applications. Neurol Res 19:281–288, 1997.
1.0 83.8 5 53.3 7 9. Hillered L, Persson L: Neurometabolic monitoring of the acutely injured
human brain. Scand J Clin Lab Invest Suppl 229:9–18, 1999.
a
RD60, Ringer-dextran 60. Both RD60 and 3.5% albumin were used as 10. Hillered L, Persson L, Ponten U, Ungerstedt U: Neurometabolic monitoring of the
perfusion buffers at both flow rates of 0.3 and 1.0 ␮l/min. ischemic brain using microdialysis. Acta Neurochir (Wien) 102:91–97, 1990.
11. Holmin SS, Biberfeld P, Mathiesen T: Intracerebral inflammation after hu-
man brain contusion. Neurosurgery 42:291–298, 1998.
12. Hutchinson PJ, O’Connell MT, Al-Rawi PG, Maskell LB, Kett-White R, Gupta AK,
trophic factors likely to be instrumental in posttraumatic brain Richards HK, Hutchinson DB, Kirkpatrick PJ, Pickard JD: Clinical cerebral
microdialysis: A methodological study. J Neurosurg 93:37–43, 2000.
repair processes (5, 15, 16, 22). New techniques for analysis, 13. Ingebrigtsen T, Romner B: Biochemical serum markers of traumatic brain
e.g., proteomics, are likely to add considerably to this list (7). injury. J Trauma 52:798–808, 2002.
A potentially useful clinical marker, the cytokine IL-6, was 14. McGlennen R: Miniaturization technologies for molecular diagnostics. Clin
sampled from all patients. The normal extracellular concen- Chem 47:393–404, 2001.
15. Morganti-Kossman MC, Rancan M, Stahel PF, Kossmann T: Inflammatory
tration of IL-6 in the extracellular space of the human brain is
response in acute traumatic brain injury: A double-edged sword. Curr Opin
not known. The highest concentration measured in the present Crit Care 8:101–105, 2002.
study approached 25,000 pg/ml, but in most measurements, 16. Rothwell NJ, Hopkins SJ: Cytokines and the nervous system: Part II—
the concentration was considerably lower, not seldom being Actions and mechanisms of action. Trends Neurosci 18:130–136, 1995.
less than 1000 pg/ml. More interesting, however, was the 17. Sjögren FS, Andersson C: Technical prerequisites for in vivo microdialysis deter-
mination of interleukin-6 in human dermis. Br J Dermatol 146:375–382, 2002.
observation illustrated in Figures 2, A and B, and 3, A and B, 18. Ståhl NM, Hallström Å, Ungerstedt U, Nordström CH: Intracerebral microdialysis
namely, the indications that a detectable pattern of variations in clinical practice: Baseline values for chemical markers during wakefulness,
in IL-6 related to different intracerebral conditions may exist. anesthesia and neurosurgery. Neurosurgery 47:701–710, 2000.
19. Ståhl NU, Ungerstedt U, Nordström CH: Brain energy metabolism during
controlled reduction of cerebral perfusion pressure in severe head injuries.
Intensive Care Med 27:1215–1223, 2001.
CONCLUSION 20. Ungerstedt U: Principles and application for studies in animal and man.
J Intern Med 230:365–373, 1991.
To analyze the protein content, it is currently necessary to 21. Unterberg AW, Sakowitz OW, Sarrafzadeh AS, Benndorf G, Lannksch R: Role of
bring the sampled fluid to a laboratory. This definitely limits the bedside microdialysis in the diagnosis of cerebral vasospasm following aneurysmal
usefulness of bedside sampling of macromolecules. However, in subarachnoid hemorrhage. J Neurosurg 94:740–749, 2001.
the near future, one can expect development of new technologies 22. Wilcockson DC, Campbell SJ, Anthony DC, Perry VH: The systematic and
local acute phase response following acute brain injury. J Cereb Blood Flow
(e.g., laboratory chip-based) to bring effective methods for the Metab 22:318–326, 2002.
bedside analysis of proteins and other macromolecules into the 23. Winter CD, Iannotti F, Pringle AK, Trikkas C, Clough GF, Church MK: A
clinical setting (3, 14). It may prove possible to find alternative microdialysis method for the recovery of IL-1␤, IL-6 and nerve growth factor
indicators of impending threats to the cerebral tissue that can be from human brain in vivo. J Neurosci Methods 119:45–50, 2002.
used on-line to select therapy tailored to the individual patient’s
needs. The present study clearly indicates that microdialysis Acknowledgments
could play a role in such development. We thank CMA Microdialysis, Stockholm, Sweden, for providing the cathe-
ters used in this study. We have no financial or other affiliation with any
manufacturer of the products used in the study.
REFERENCES
COMMENTS
1. Anderson C, Svensson C, Sjögren F, Andersson T, Wardell K: Human in
vivo microdialysis technique can be used to measure cytokines in contact
reactions. Curr Probl Dermatol 23:121–130, 1995.
2. Benveniste H, Huttemeier PC: Microdialysis: Theory and application. Progr
Neurobiol 35:195–215, 1990.
W hen a neurosurgeon thinks of cerebral microdialysis, he
or she most likely envisions patients who suffer from
severe traumatic brain injury or aneurysmal subarachnoid

1268 | VOLUME 56 | NUMBER 6 | JUNE 2005 www.neurosurgery-online.com

 HUMAN INTRACEREBRAL MICRODIALYSIS
hemorrhage. Part of the reason why champions of microdi-
alysis have gravitated toward such conditions relates to the
nature of the substances that can be measured with this tech-
nique. Limitations on the size of the molecules that can be
assayed have contributed to an emphasis on such molecules
such as glucose, lactate, pyruvate, and certain neurotransmit-
ters. Alterations in concentrations of these substances are as-
sociated with ischemia and other injuries to the brain.
In this article, Hillman et al. describe a simple yet important
study of the effectiveness of a microdialysis membrane with a
larger pore size. These catheters enable the analysis of mole-
cules up to 100 kD, as opposed to the cutoff of 20 kD for
standard catheters. By replacing the standard perfusion fluid
with either albumin or a Ringer’s-dextran solution, the in-
creased loss of fluid through the larger pores is offset by the
higher oncotic pressure of the perfusate.
This demonstration of the technical feasibility of analyzing
proteins and other larger molecules may help to extend the
horizons of cerebral microdialysis. As the authors discuss in
their conclusion, new technologies may facilitate bedside anal-
ysis of proteins and other macromolecules that have been
collected with the catheters described here. Further advances
in microdialysis methodology by the authors and by other
groups may improve our ability to care for our most critically
ill patients.
Alex B. Valadka
Houston, Texas
T he development of microdialysis has contributed signifi-
cantly to our understanding of brain pathophysiology.
Because of the cut-off properties of the catheters, however, it
has only been possible to recover relatively small molecules.
Here, Hillman et al. present an elegant study of 100 kD
cut-off catheters. The investigation was done in 10 patients
who were experiencing coma after subarachnoidal hemor-
rhage or traumatic brain injury. One 100 kD cut-off catheter
was placed parallel to a control catheter in each patient.
The authors show that even brief perfusion with the stan-
dard perfusion fluid caused significant loss of volume,
whereas fluid recovery with albumin and RD-60 was satisfy-
ing. Extracellular proteins (including interleukin-6 [IL-6])
were sampled from all catheters, and the authors noted pat-
terns of variations in the concentration of IL-6, apparently
related to pathophysiological changes. Concentrations of glu-
cose, glycerol, lactate, and pyruvate were comparable to those
recorded by standard catheters. The new catheters will allow
recovery of an expanded fraction of the large molecules that
we know from studies in preclinical models are important for
neuronal death or recovery.
Iver A. Langmoen
Stockholm, Sweden
T he authors report the use of a microdialysis catheter that
allows sampling of macromolecules from human brain
tissue. They used IL-6 as the test macromolecule and deter-
NEUROSURGERY
OF MACROMOLECULES
mined that the behavior of the assay seemed adequate to track
changes in this molecule in response to threatened ischemia in
patients with head injury or vasospasm owing to subarach-
noid hemorrhage. The authors further demonstrated that the
catheter under study was able to collect small molecules of
interest with reasonable recovery rates. A major difference
between the standard catheter and the test catheter techniques
was the use of either an albumin or Ringer’s-dextran perfusate
solution. These were better able to counterbalance the fluid
loss owing to larger pore size and hydrostatic effects. Metab-
olite recovery was better at lower perfusion rates.
This article is well written and the information is of value to
those using microdialysis techniques in humans. The ability to
evaluate changes in a variety of proteins in the human brain in
clinical settings is a major advance that will open new avenues
of understanding of pathophysiology and normal brain func-
tion. This is a valuable contribution to our literature.
Charles J. Hodge, Jr.
Syracuse, New York
T his is an important article because, to my knowledge, it
represents the first published experience with the new
CMA 100 microdialysis catheter (CMA Microdialysis, Stock-
holm, Sweden), which allows time-dependent, relatively non-
invasive, in vivo measurement of small molecular weight
proteins in the living human brain. This has never before been
possible.
This newly designed catheter has not yet been approved for
human use in the United States. This article is a demonstration
that in nine patients and over about 114 microdialysis mea-
surements, this technique seemed to be safe. Hopefully, as
more experience is obtained with these catheters, they will
replace the conventional ones and allow parallel estimation of
proteins and peptides, as well as 3-carbon based substrates.
There is a revolution of interest in proteomics in the brain after
a variety of brain insults, and this new catheter may make it
possible to serially measure proteins in different parts of the
living human brain. The authors have shown that the cyto-
kine, IL-6, seems to fluctuate in broad agreement with the
other microdialysis analytes. The theoretical possibilities with
these techniques are enormous. For example, it is not known
when the signal for apoptotic cell death is expressed in the
human brain after traumatic brain injury, although immuno-
histochemical studies on excised human contusion material
has shown that this process is prominent. With this new
microdialysis technique, it would be possible to measure pro-
apoptotic cytokines and allow early pre-treatment strategies
with, for example, IL-1 receptor antagonists, such as anakinra.
The authors have convincingly shown that the best dialy-
sate perfusion fluid for these studies is probably Ringer’s-
dextran 60. Normal saline or artificial cerebrospinal fluid is
not a suitable perfusion fluid with these probes, because, as
the authors have shown, much of the fluid migrates across the
membrane to the brain.
VOLUME 56 | NUMBER 6 | JUNE 2005 | 1269
HILLMAN ET AL.

Unfortunately, these experiments were quite badly designed. Al- fusion fluid that they have evaluated. Nevertheless, this data is still
though the authors implanted a standard CMA-20 microdialysis an important first start using a new technique, which may have
probe and the new CMA 100 probe together, they did not make great potential for the future.
direct cross comparisons for all the analytes of interest. The reasons
for this are not clear. Similarly, the authors do not make direct cross M. Ross Bullock
comparisons using this 2-probe design with the three types of per- Richmond, Virginia

1270 | VOLUME 56 | NUMBER 6 | JUNE 2005 www.neurosurgery-online.com