Neurotoxicology 69 (2018) 11–16

Contents lists available at ScienceDirect

Neurotoxicology
journal homepage: www.elsevier.com/locate/neuro

Full Length Article

CSF sodium at toxic levels precedes delirium in hip fracture patients T

a,b,⁎ b,c d,e f e,g
Bjørnar Hassel , Espen Mariussen , Ane-Victoria Idland , Gry T. Dahl , Johan Ræder ,
Frede Frihagenh, Jens Petter Bergi, Farrukh A. Chaudhryj, Torgeir B. Wyllerd,e, Leiv O. Watned,j
a
Department of Neurohabilitation and Complex Neurology, Oslo University Hospital, University of Oslo, Oslo, Norway
b
Norwegian Defence Research Establishment (FFI), Kjeller, Norway
c
Norwegian Institute for Air Research (NILU), Kjeller, Norway
d
Oslo Delirium Research Group, Department of Geriatric Medicine, Oslo University Hospital, Oslo, Norway
e
Institute of Clinical Medicine, University of Oslo, Oslo, Norway
f
Department of Anesthesiology, Diakonhjemmet Hospital, Oslo, Norway
g
Department of Anesthesiology, Oslo University Hospital, Oslo, Norway
h
Division of Orthopaedic Surgery, Oslo University Hospital, Oslo, Norway
i
Department of Medical Biochemistry, Oslo University Hospital, Oslo, Norway
j
Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway

A R T I C LE I N FO A B S T R A C T

Keywords: Delirium is an acute state of confusion and a fluctuating level of consciousness. It is precipitated by physical
Sodium illness or trauma, such as pneumonia, heart infarction, or hip fracture. Delirium is common among elderly
Potassium hospitalized patients, and as many as 50% of hip fracture patients may develop delirium. Delirium may pre-
Delirium cipitate dementia, but recent studies indicate that delirium is caused by unknown neurotoxic mechanisms that
Cerebrospinal fluid
are different from those that are associated with dementia. Experimental studies have shown that high extra-
Choroid plexus
Serum
cellular levels of sodium are neurotoxic. We sampled lumbar cerebrospinal fluid (CSF) from hip fracture patients
during hip surgery and analyzed metal ions that influence neuronal function. Eight patients who developed
delirium after surgery had 21% higher CSF sodium than 17 patients who did not develop delirium (median value
175 mmol/L; range 154–188, vs. 145 mmol/L (112–204; p < 0.008) or 39 patients who underwent elective
surgery under spinal anesthesia without developing delirium (145 mmol/L; 140–149; p = 0.0004). Seven pa-
tients who had developed delirium before CSF sampling had a median CSF sodium of 150 mmol/L (144–185;
p = 0.3). CSF potassium was also 21% higher in patients who developed delirium (p = 0.024), but remained
within the physiological range. Serum sodium and potassium were normal in all patient groups. This study, on a
small sample of patients, confirms the neurotoxic potential and clinical importance of high extracellular levels of
sodium in the brain. High CSF sodium would likely affect cerebral function and could precipitate delirium;
further, it could interact with dementia-specific mechanisms to precipitate dementia development.

1. Introduction
Neuronal function is critically dependent on the extracellular con-
centration of cations, especially sodium and potassium ions (Hodgkin
and Huxley, 1952; Antonio et al., 2016; Morland et al., 2016), and an
increase in the concentration of these ions may have profound effects
on both neuronal function and survival, as recently seen in experi-
mental studies (Antonio et al., 2016; Morland et al., 2016). An increase
in the extracellular sodium concentration in the brain, as reflected in
the cerebrospinal fluid (CSF) level, has previously been seen in mi-
graine patients during a migraine attack (Harrington et al., 2006) and
in children who suffered neurological damage from dehydration (Habel
and Simpson, 1976), illustrating that the CSF sodium concentration
may be of great clinical importance. Delirium is an acute state of con-
fusion, inattention and fluctuating level of consciousness (Inouye et al.,
1990, 2014). It is precipitated by a physical condition, such as infection
or trauma, and it is a common occurrence among elderly patients. For
instance, 40–50% of acutely hospitalized hip fracture patients, may
develop delirium (Dasgupta and Dumbrell, 2006; Bruce et al., 2007;
Juliebø et al., 2009; Inouye et al., 2014). It is not known how the brain
dysfunction in delirium results mechanistically from a systemic or
peripheral disorder, e.g. a hip fracture. Some suggestions have been
forwarded, among them neuroinflammation and altered mono-
aminergic or cholinergic neurotransmission in the brain (for review, see

⁎
Corresponding author at: Department of Neurohabilitation and Complex Neurology, Oslo University Hospital, University of Oslo, 0450 Oslo, Norway.
E-mail address: bjornar.hassel@medisin.uio.no (B. Hassel).

https://doi.org/10.1016/j.neuro.2018.08.010
Received 27 March 2018; Received in revised form 20 August 2018; Accepted 21 August 2018
Available online 24 August 2018
0161-813X/ © 2018 Elsevier B.V. All rights reserved.
B. Hassel et al.
Maldonado, 2013). Delirium may precipitate dementia, or worsen a
pre-existing dementia (Saczynski et al., 2012; Davis et al., 2017), and so
one could suspect that the two conditions were two aspects of the same
neurodegenerative condition. However, they differ clinically, as de-
mentia is not marked by a fluctuating level of consciousness, which is
the case in delirium, and recent evidence points to unique, but un-
known, neuropathologic mechanisms underlying delirium that are dif-
ferent from those that cause dementia (Davis et al., 2017).
To investigate whether ionic changes in the extracellular milieu of
the brain could participate in the pathology of delirium we obtained
lumbar cerebrospinal fluid (CSF) from hip fracture patients undergoing
hip surgery under spinal anesthesia. Some of the patients went on to
develop delirium after surgery. This provided an opportunity to look for
alterations in CSF ionic concentrations in delirium-prone patients.
Patients who were overtly delirious at the time of surgery were in-
cluded to see whether changes in pre-delirious patients were main-
tained in the overtly delirious state.
2. Methods
2.1. Participants and sampling of CSF and blood
This prospective cohort, observational study was approved by the
Regional Ethics Committee for the Southern and Eastern parts of
Norway. All participants (N = 71) gave informed, written consent.
Thirty-two patients had a hip fracture within the last 90.1 h before
surgery. Patients were excluded from the study if the fracture was
caused by a high energy trauma (e.g. a car accident), or if they were
terminally ill. Patients who were diagnosed with dementia were also
excluded, because a previous study had shown that very few patients
with dementia did not develop any symptoms of delirium after hip
surgery (Watne et al., 2016), so statistical comparisons of groups would
become difficult. Trial inclusion took place in the emergency room at
Oslo University Hospital by the orthopedic surgeon on call. All patients
underwent hip surgery under spinal anesthesia. Thirty-nine patients
served as controls. These patients underwent elective (non-acute) sur-
gery for various complaints, including orthopedic, gynecological, and
urological conditions.
CSF was sampled under sterile conditions from the stainless steel
cannula used for lumbar spinal anesthesia immediately before admin-
istration of the anesthetic agent. CSF was collected in a polypropylene
tube and centrifuged. Supernatants were stored at −80 °C. Blood and
serum were collected before and after surgery in standard sample tubes
by venipuncture.
2.2. Measurement of metal ions
For measurements of metal ions in CSF, samples were diluted 1:100
in 5 mL ultrapure HNO3, 0.65% (vol/vol; Merck; distilled in-house).
Sodium, potassium, magnesium, and calcium ions were analyzed on an
inductively coupled plasma mass spectrometer (ICP–MS; Thermo X-
series II ICP-MS; ThermoFisher Scientific) as described (Dahlberg et al.,
2015). An internal concentration standard containing scandium (45Sc),
rhodium (103Rh), indium (115In), and lutetium (175Lu) was added to
each sample to correct for any drift in signal intensity. The elements of
interest were quantified with the use of a four point standard curve
(1–1000 ppm). Reference solutions (Rain-97 or Bigmoose-02, and
Battle-02, TM 23.4, TMDA 61.2, and TMDA 53.3; Analytical Reference
Material, Environment Canada, Gatineau, Quebec, Canada) were ana-
lyzed in addition to in-house made standards. CSF samples were ana-
lyzed in random order in one single run. Samples were analyzed twice,
and average values from the two analyses are reported. Blank values
were < 1% of sample values for all four cations measured.
Serum was analyzed with respect to sodium and potassium by in-
direct potentiometry and ion-selective electrodes, using a Modular
P800 (Roche; Basel, Switzerland). Hemoglobin, albumin, creatinine,
12
Neurotoxicology 69 (2018) 11–16
urea, C-reactive protein, and leukocyte count were analyzed in whole
blood or serum with standard hospital methods for clinical samples.
Serum and whole blood were analyzed immediately after sampling as
part of the hospital routine.
2.3. Evaluation of delirium and dementia
The hip fracture patients were examined for delirium prior to sur-
gery and daily until day 5 after surgery with the Confusion Assessment
Method (CAM; Inouye et al., 1990). CAM scores were based on inter-
view with the patient, including tests of cognition, attention, and
alertness (digit span test, orientation and delayed recall), information
from close relatives and nurses, and evaluation of hospital records for
the current admission (Watne et al., 2016). A geriatrician and a trained
research nurse performed the assessments.
A geriatrician and a geriatric psychiatrist determined whether pa-
tients had dementia prior to the hip fracture. Diagnosis was based on
the Informant Questionnaire on Cognitive Decline in the Elderly
(IQCODE; Jorm et al., 1994; 2004), cognitive tests, including Mini-
Mental State Examination (MMSE; Folstein et al., 1975), clock drawing
test (Shulman, 2000), the 10 word test from the Consortium to Establish
a Registry for Alzheimer’s Disease battery (CERAD; Welsh et al., 1994),
and The Clinical Dementia Rating scale (CDR; Hughes et al., 1982).
Next of kin was interviewed regarding pre-hip fracture Activities of
Daily life (Barthel ADL Index; Wade, 1992).
The patients who underwent elective surgery were tested for de-
lirium prior to surgery. Delirium status after surgery was determined
from review of the patients’ medical charts.
2.4. Data presentation and statistics
Data are given as concentration values. CSF values were not nor-
mally distributed according to the D'Agostino and Pearson normality
test; therefore data are reported as median values and full range. In box-
and-whisker plots the box represents the middle two quartiles and the
whiskers represent range. Groups were compared statistically with the
Kruskal-Wallis all-pairwise test with Dunn’s method to control for
multiple comparisons. Paired testing (pre- vs. post-surgery blood and
serum values) was done with the paired Student’s t-test. P-values <
0.05 were considered significant.
3. Results
3.1. Patient characteristics
Seventeen hip fracture patients did not develop delirium during the
hospital stay; these are referred to as’ non-delirious’ (median age 85
years). Eight patients (median age 82 years) developed delirium within
2 days after hip surgery; these are referred to as ‘pre-delirious’ at the
time of CSF sampling. Seven patients (median age 84.5 years) had de-
veloped delirium prior to surgery; these are referred to as ‘overtly de-
lirious’ at the time of CSF sampling. The patient groups were not sig-
nificantly different with respect to age, sex distribution, time from
hospital admission to surgery, or use of medication that could affect
CNS function or (possibly) CSF production: diuretics, antihypertensive
drugs, opioids, tranquilizers, hypnotics, levothyroxine, corticosteroids,
cardiac glycosides, or antidiabetics (Table 1).
Thirty-nine patients who underwent elective surgery served as
controls; 17 were women, and 22 were men. Median age was 67.5 years
(range 64–93). None of these patients developed delirium.
3.2. CSF concentrations of sodium, potassium, magnesium, and calcium
Patients who underwent elective surgery had a median CSF sodium
concentration of 145 mmol/L (range 140–149; Fig. 1a), which corre-
sponds to normal levels (Dahlberg et al., 2015). Non-delirious hip
B. Hassel et al.
Table 1
Age, sex ratio, time from admission to surgery, drugs, and CSF concentrations of glucose, total protein, albumin, and leukocytes for 32 hip fracture
patients. Seventeen patients did not develop delirium during the hospital stay (‘No delirium’), 8 developed delirium after CSF sampling and surgery
(‘Pre-delirium’), and 7 had developed delirium prior to CSF sampling and surgery (‘Overt delirium’). The patients did not have dementia. Data are
median values and full range in parentheses. 1: Drugs with possible CNS effects include hypnotics, opioid analgesics, tranquilizers, diuretics, anti-
hypertensive drugs, corticosteroids, levothyroxine, and cardiac glycosides.
No delirium (N = 17)
Age (years) 85 (60–93)
Female/male 11/4
Time from admission to surgery (hours) 23.5 (12.8–90.1)
No. of drugs with possible CNS effects1 1.5 (0–6)
CSF glucose (mmol/L) 4.1 (3.4–5.1)
CSF total protein (g/L) 0.447 (0.233–0.595)
CSF albumin (g/L) 0.277 (0.120–0.389)
CSF leukocytes (cells/nL) 3 (0–27)
fracture patients also had a median CSF sodium concentration of
145 mmol/L, but the variation was greater (range 112–204). In pre-
delirious patients, median CSF sodium concentration was 30 mmol/L
(21%) higher than in elective patients or non-delirious hip fracture
patients: 175 mmol/L (range 154–188). In overtly delirious patients,
CSF sodium was not significantly different from that found in non-de-
lirious hip fracture patients (median value 150 mmol/L; range
144–185).
The median CSF concentration of potassium was 21% higher in pre-
delirious patients than in non-delirious hip fracture patients (2.9 mmol/
L vs. 2.4 mmol/L; Fig. 1b). CSF potassium was also significantly higher
in overtly delirious patients, the median value being 16% higher than in
non-delirious hip fracture patients. CSF magnesium was higher in non-
delirious and pre-delirious hip fracture patients than in patients un-
dergoing elective surgery, but the hip fracture patient groups were not
different from each other (Fig. 1c). CSF calcium was not significantly
different between the patient groups (Fig. 1d). CSF concentrations of
glucose, protein, and leukocytes were not different among the groups of
hip fracture patients (Table 1).
Patients who underwent elective surgery had a lower median age
(67.5 years; range 64–93) than hip fracture patients. However, these
patients fell into two distinct age categories, those above 77 years and
those below 69 years. Nineteen patients were above 77 years. They had
a median age of 83 years (range 77–93), which was similar to the age
range of the hip fracture patients (see Table 1). Their CSF sodium was
145 mmol/L (range 140–148); CSF potassium was 2.6 mmol/L
(2.4–2.9), CSF magnesium was 0.96 mmol/L (range 0.89–1.04), and
CSF calcium was 0.90 mmol/L (range 0.75–1.11). These values were
highly similar to those of the group of elective surgery patients as a
whole, and statistical comparisons with hip fracture patients yielded
essentially the same results as when the group as a whole was compared
(data not shown). Thus, differences between patients undergoing elec-
tive surgery and hip fracture patients in terms of CSF ion levels could
not be explained by age differences between the groups.
3.3. Serum and whole blood values
Serum sodium and potassium values were similar in the hip fracture
patient groups (Table 2). There was a small, but significant reduction in
serum sodium after surgery in the non-delirious patients.
There were no significant differences between patient groups with
respect to whole blood or serum levels of hemoglobin, albumin, crea-
tinine, urea, CRP, or leukocytes (Table 3). However, for all patients
groups there was a significant increase in CRP after surgery, reflecting
an inflammatory response to surgery. Further, all patients groups had a
significant reduction in serum albumin after surgery, reflecting perio-
perative fluid treatment. Among patients who did not develop delirium
there was also a reduction in hemoglobin and creatinine after surgery.
13
Neurotoxicology 69 (2018) 11–16
Pre-delirium (N = 8) Overt delirium (N = 7)
82 (76–101) 84.5 (80–88)
5/3 5/2
20.8 (12.6–39.8) 29.5 (13–50.8)
3 (0–5) 1 (0–5)
4.1 (3.3–6.5) 4.3 (3.2–5.6)
0.395 (0.275–0.755) 0.360 (0.184–0.654)
0.205 (0.103–0.518) 0.194 (0.085–0.440)
5 (4–14) 2 (0–15)
4. Discussion
4.1. CSF sodium is high in delirium-prone hip fracture patients
We show here that pre-delirious patients, who developed delirium
within 2 days after surgery for hip fracture, had a higher median CSF
concentration of sodium and potassium prior to delirium development
than patients who did not develop delirium. The increase was specific
for sodium and potassium, the major cation electrolytes in CSF; a si-
milar increase was not seen for magnesium or calcium. The increase in
CSF sodium and potassium could not be explained by increased serum
levels, because serum sodium and potassium were similar among the
patient groups. The increase in CSF sodium was not found in overtly
delirious patients, suggesting that this increase could be a transient
phenomenon that is involved in the triggering, but not the main-
tenance, of the delirious state.
An increase in CSF sodium without a similar increase in serum so-
dium has previously been seen in migraine patients during a migraine
attack (Harrington et al., 2006) and in children who suffered neurolo-
gical damage from dehydration (Habel and Simpson, 1976), under-
scoring that elevated CSF sodium concentration may have important
clinical consequences, and that CSF sodium may increase independently
of serum sodium. Interestingly, in the present study, serum sodium was
seen to decrease after surgery only in the hip fracture patients that did
not develop delirium. Further studies are needed to elucidate whether a
reduction in the serum concentration of sodium may help prevent de-
lirium development. An obvious limitation of the present study is the
low sample size. Further studies are needed to examine the relationship
between delirium development and CSF sodium to see whether CSF
sodium may have a causal role in the precipitation of delirium. Further
studies are also needed to see whether an increased CSF sodium con-
centration may underlie delirium development in clinical conditions
other than hip fracture.
4.2. An increase in CSF sodium may damage brain function
A high extracellular concentration of sodium may have a number of
adverse effects on the brain. First, it may cause a reduction in the ex-
tracellular concentration of neurotransmitters glutamate and GABA
through increased uptake into brain cells, because this uptake is driven
by the sodium gradient across cell membranes (Lehmann, 1989;
Morland et al., 2016); the result would be reduced glutamatergic and
GABAergic neurotransmission. Second, increased CSF sodium may lead
to a reduction in intracellular free calcium concentration in neurons
(Morland et al., 2016), affecting intracellular signal transduction.
Third, increased CSF sodium may cause a reduction in neuronal glucose
metabolism and ATP formation that may cause neuronal death
(Himmelseher et al., 2001; Morland et al., 2016). Fourth, a high con-
centration of sodium implies hyperosmolarity, which causes water to
B. Hassel et al. Neurotoxicology 69 (2018) 11–16

Fig. 1. CSF levels of sodium, potassium, magnesium, and calcium. Patients underwent hip surgery under spinal anesthesia, and CSF was sampled prior to anesthesia.
CSF was analyzed for a) sodium, b) potassium, c) magnesium, and d) calcium by inductively coupled plasma mass spectrometry. ‘Elective patients’ (N = 39; white
boxes) did not have hip fracture, but underwent elective (non-acute) surgery under spinal anesthesia; they did not develop delirium. Hip fracture patients (grey
boxes) underwent acute hip surgery. These were grouped into those who never developed delirium (N = 17; ‘No delirium’; light grey boxes), those who developed
delirium after CSF sampling and surgery (N = 8; ‘pre-delirium’; medium grey boxes), and those who had developed delirium prior to CSF sampling and surgery
(N = 7; ‘Overt delirium’; dark grey boxes). Boxes represent the middle two quartiles together with the median value; the whiskers represent minimum and maximum
values. Group differences were analyzed with Kruskal-Wallis all-pairwise analysis and Dunn’s correction for multiple comparisons. Asterisks: difference from ‘No
delirium’ hip fracture group; *: p < 0.05; **: p < 0.01. #: difference from patients undergoing elective surgery; #: p < 0.05; ##: p < 0.01; ###: p < 0.001;
####: p < 0.0001.

Table 2
Serum concentrations of sodium and potassium. Hip fracture patients underwent acute hip surgery for hip fracture under lumbar spinal anesthesia. CSF was sampled
prior to induction of anesthesia. Seventeen patients did not develop delirium (‘No delirium’), 8 patients developed delirium after CSF sampling and surgery (‘Pre-
delirium’), and 7 patients had developed delirium prior to CSF sampling and surgery (‘Overt delirium’). The patients did not have dementia. Data are mmol/L,
median and (in parentheses) minimum and maximum values. Asterisk: difference between pre-operative and postoperative values; *: p = 0.043; Student’s paired t-
test. There were no significant difference in serum concentrations of sodium or potassium between patients who did not develop delirium and those who did.
No delirium (N = 17) Pre-delirium (N = 8) Overt delirium (N = 7)

Pre-op Post-op Pre-op Post-op Pre-op Post-op

+
s−Na 140 (128–143) 137 (132–141)* 139 (135–142) 139 (135–143) 137 (128–143) 136 (128–142)
s−K+ 4.3 (3.3–5.5) 4.1 (3.7–6.0) 4.0 (3.3–4.3) 4.0 (3.8–4.7) 4.5 (3.5–4.8) 4.3 (3.2–4.9)

shift from the intracellular to the extracellular fluid compartment. The
resulting increase in the extracellular space may reduce synaptic ac-
tivity, as previously shown in brain slices (Traynelis and Dingledine,
1989). In vitro, these effects were pronounced when the extracellular
concentration of sodium was increased by 25–30 mmol/L (Traynelis
and Dingledine, 1989; Morland et al., 2016), similar to the median
increase of 30 mmol/L in CSF sodium seen in pre-delirious hip fracture
patients in the present study.
We focus here on CSF sodium, because the level in delirium-prone
patients was markedly elevated. Although CSF potassium was elevated
to the same extent percentwise, the increase was less conspicuous in
absolute terms. The median CSF potassium level of 2.9 mmol/L in pre-
delirious patients was significantly higher than the median value of
2.4 mmol/L in non-delirious patients, but it was still within normal
values for CSF potassium (Dahlberg et al., 2015).
Delirium may precipitate dementia (Saczynski et al., 2012). Recent
data suggest that delirium is triggered by unknown mechanisms that
are different from those that underlie dementia development, such as

14
B. Hassel et al. Neurotoxicology 69 (2018) 11–16

Table 3
Whole blood or serum values for hemoglobin, albumin, creatinine, urea, C-reactive protein (CRP), and leukocytes. Pre-operative (Pre-op) and post-operative (Post-
op) values are given for hip fracture patients who did not develop delirium (‘No delirium’), those who developed delirium after surgery (‘Pre-delirium’), and those
who had developed delirium prior to surgery (‘Overt delirium’). Patients did not have dementia. Data are median and (in parentheses) minimum and maximum
values. Asterisks: difference between pre-operative and postoperative values; *: p < 0.05; **: p < 0.01; ***: p < 0.001; Student’s paired t-test. There were no
significant difference between patients who did not develop delirium and those who did with respect to any of the parameters in the table, pre- or post-operative.
Units are: hemoglobin (g/dL), albumin (g/L), creatinine (μmol/L), urea (mmol/L), leukocytes (cells/nL), CRP (mg/L).
No delirium (N = 17) Pre-delirium (N = 8) Overt delirium (N = 7)

Pre-op Post-op Pre-op Post-op Pre-op Post-op

Hemoglobin 12.6 (10.5–14.1) 10.1 (7.9–12.0)*** 11.0 (8.3–13.5) 10.2 (8.7–11.1) 10.7 (8.0–21.9) 10.1 (7.3–15.9)
Albumin 41 (32-47) 32 (25-36)*** 33 (27-47) 31 (21-37)** 40 (35-44) 31 (28-34)**
Creatinine 66 (39–193) 63 (32–166)* 59 (53–132) 68 (48–218) 68 (48–218) 54 (41–240)
Urea 7.1 (2.0–14.9) 5.5 (2.6–10.2) 9.0 (4.5–15.0) 7.9 (4.4–15.0) 7.3 (4.7–11.7) 5.4 (3.6–14.3)
Leukocytes 10.0 (2.4–17.5) 10.2 (1.9–13.2) 10.2 (4.0–17.6) 9.5 (3.8–18.2) 10.7 (8.0–21.9) 10.1 (7.3–15.9)
CRP 1.6 (1–113) 137 (10–306)*** 8.9 (0.6–224) 133 (0.6–250)* 9.4 (1–172) 145 (52–253)***

beta-amyloid plaques, vascular pathology, or Lewy bodies (Davis et al.,
2017). A high CSF sodium level may be such a delirium-triggering
mechanism; a similar high CSF sodium level is not seen in Alzheimer’s
disease patients, in whom CSF sodium is normal (Vitvitsky et al., 2012).
However, given the many adverse effects of high extracellular sodium
on nerve cells (Lehmann, 1989; Himmelseher et al., 2001; Morland
et al., 2016), a high CSF sodium level could interact with dementia-
specific mechanism to trigger dementia development. Support for this
possibility comes from findings of beta-amyloid pathology in non-de-
mented patients who develop delirium (Xie et al., 2014; Idland et al.,
2017; Cunningham et al., 2018).
4.3. How CSF sodium may increase prior to delirium development: a
hypothesis
The mechanism behind an increase in CSF sodium in spite of a
normal serum concentration of sodium is not known. Below we offer a
hypothesis to explain this phenomenon, provide a link between trauma
or surgery on the one hand and delirium development on the other, and
explain the delirium-preventive effect of alpha2 adrenergic agonists
(Shukry et al., 2005; Su et al., 2016).
Sodium is transferred from blood to the CSF by sodium-potassium
pumps at the choroid plexus (Damkier et al., 2013; Hladky and Barrand,
2014). Water moves passively from blood to CSF according to the os-
motic gradient (created by sodium) and the hydrostatic pressure dif-
ference between blood and CSF (Damkier et al., 2013; Hladky and
Barrand, 2014). The CSF sodium concentration could increase if the
hydrostatic pressure was reduced, limiting water movement, while so-
dium transfer from blood to CSF was maintained. Hydrostatic pressure
in the choroid plexus capillaries would be reduced by increased sym-
pathetic activity in precapillary arterioles. The choroid plexus is richly
innervated by sympathetic fibers arising in the superior cervical ganglia
(Lindvall and Owman, 1981). Studies in experimental animals have
shown that stimulation of the superior cervical ganglia may reduce CSF
production by as much as 30% (Lindvall and Owman, 1981). Sympa-
thetic activity is high in hip fracture patients (Bäcklund et al., 1999;
Kudoh et al., 1999; Oehmke et al., 2008) and could cause constriction
of precapillary arterioles in the choroid plexus, reducing the capillary
hydrostatic pressure and water movement, while sodium transfer by the
sodium-potassium pumps is maintained. Such a mechanism could ex-
plain how CSF sodium may increase, while serum sodium remains
normal. Further, this mechanism could explain the delirium-preventive
effect of sympatholytic alpha2 adrenergic agonists (Shukry et al., 2005;
Su et al., 2016). Alpha2 adrenergic agonists would reduce precapillary
sympathetic activity in the choroid plexus, restoring the capillary hy-
drostatic pressure and water movement into the cerebral ventricles,
thus preventing an increase in CSF sodium concentration.
In several studies an increase in the CSF concentration of various
metabolites, proteins or inflammatory markers have been found in hip
fracture patients who went on to develop delirium after CSF sampling
and surgery (Cape et al., 2014; Westhoff et al., 2015; Hall et al., 2016;
Watne et al., 2016). A reduction in CSF formation could contribute to
the increased concentration of metabolites and proteins in the CSF by
reducing their dilution. This is not to say that the increased con-
centration of, for instance cytokines, due to diminished CSF production
is without biological importance. On the contrary, an increased con-
centration of neuroactive substances would be of biological con-
sequence irrespective of the cause of their increase.
Non-delirious hip fracture patients had a great variation in CSF
sodium compared to patients undergoing elective surgery. Some of
these patients had very high CSF sodium values (up to 204 mmol/L),
suggesting that although high CSF sodium may be associated with a
susceptibility to delirium, this association is not absolute. Other pa-
tients had CSF sodium values well below normal levels. The latter
finding, although unexplained in our study, points to several mechan-
isms behind the dysregulation of CSF production in hip fracture pa-
tients.
Conflict of interest
The authors declare that there are no conflicts of interest.
Transparency document
The Transparency document associated with this article can be
found in the online version.
Acknowledgements
This study was supported by The Norwegian Health Association
(Grant #1513).
References
Antonio, L.L., Anderson, M.L., Angamo, E.A., Gabriel, S., Klaft, Z.J., Liotta, A., Salar, S.,
Sandow, N., Heinemann, U., 2016. In vitro seizure like events and changes in ionic
concentration. J. Neurosci. Methods 260, 33–44.
Bäcklund, M., Lepäntalo, M., Toivonen, L., Tuominen, M., Tarkkil, P., Pere, P., Scheinin,
M., Lindgren, L., 1999. Factors associated with post-operative myocardial ischaemia
in elderly patients undergoing major non-cardiac surgery. Eur. J. Anaesthesiol. 16,
826–833.
Bruce, A.J., Ritchie, C.W., Blizard, R., Lai, R., Raven, P., 2007. The incidence of delirium
associated with orthopedic surgery: a meta-analytic review. Int. Psychogeriatr. 19,
197–214.
Cape, E., Hall, R.J., van Munster, B.C., de Vries, A., Howie, S.E., Pearson, A., Middleton,
S.D., Gillies, F., Armstrong, I.R., White, T.O., Cunningham, C., de Rooij, S.E.,
MacLullich, A.M., 2014. Cerebrospinal fluid markers of neuroinflammation in de-
lirium: a role for interleukin-1β in delirium after hip fracture. J. Psychosom. Res. 77,
219–225.
Cunningham, E.L., McGuinness, B., McAuley, D.F., Toombs, J., Mawhinney, T., O’Brien,
S., Beverland, D., Schott, J.M., Lunn, M.P., Zetterberg, H., Passmore, A.P., 2018. CSF
Beta-amyloid 1-42 concentration predicts delirium following elective arthroplasty

15
B. Hassel et al.
surgery in an observational cohort study. Ann. Surg. https://doi.org/10.1097/SLA.
0000000000002684. [Epub ahead of print].
Dahlberg, D., Ivanovic, J., Mariussen, E., Hassel, B., 2015. High extracellular levels of
potassium and trace metals in human brain abscess. Neurochem. Int. 82, 28–32.
Damkier, H.H., Brown, P.D., Praetorius, J., 2013. Cerebrospinal fluid secretion by the
choroid plexus. Physiol. Rev. 93, 1847–1892.
Dasgupta, M., Dumbrell, A.C., 2006. Preoperative risk assessment for delirium after
noncardiac surgery: a systematic review. J. Am. Geriatr. Soc. 54, 1578–1589.
Davis, D.H., Muniz-Terrera, G., Keage, H.A., Stephan, B.C., Fleming, J., Ince, P.G.,
Matthews, F.E., Cunningham, C., Ely, E.W., MacLullich, A.M., Brayne, C.,
Epidemiological Clinicopathological Studies in Europe (EClipSE) Collaborative
Members, 2017. Association of delirium with cognitive decline in late life: A neu-
ropathologic study of 3 population-based cohort studies. JAMA Psychiatry 74,
244–251.
Folstein, M.F., Folstein, S.E., McHugh, P.R., 1975. "Mini-mental state". A practical method
for grading the cognitive state of patients for the clinician. J. Psychiatr. Res. 12,
189–198.
Habel, A.H., Simpson, H., 1976. Osmolar relation between cerebrospinal fluid and serum
in hyperosmolar hypernatraemic dehydration. Arch. Dis. Child. 51, 660–666.
Hall, R.J., Watne, L.O., Idland, A.V., Raeder, J., Frihagen, F., MacLullich, A.M., Staff, A.C.,
Wyller, T.B., Fekkes, D., 2016. Cerebrospinal fluid levels of neopterin are elevated in
delirium after hip fracture. J. Neuroinflammation 13, 170.
Harrington, M.G., Fonteh, A.N., Cowan, R.P., Perrine, K., Pogoda, J.M., Biringer, R.G.,
Hühmer, A.F., 2006. Cerebrospinal fluid sodium increases in migraine. Headache 46,
1128–1135.
Himmelseher, S., Pfenninger, E., Morin, P., Kochs, E., 2001. Hypertonic-hyperoncotic
saline differentially affects healthy and glutamate-injured primary rat hippocampal
neurons and cerebral astrocytes. J. Neurosurg. Anesthesiol. 13, 120–130.
Hladky, S.B., Barrand, M.A., 2014. Mechanisms of fluid movement into, through and out
of the brain: evaluation of the evidence. Fluids Barriers CNS 11, 26.
Hodgkin, A.L., Huxley, A.F., 1952. Currents carried by sodium and potassium ions
through the membrane of the giant axon of Loligo. J. Physiol. (Paris) 116, 449–472.
Hughes, C.P., Berg, L., Danziger, W.L., Coben, L.A., Martin, R.L., 1982. A new clinical
scale for the staging of dementia. Br. J. Psychiatry 140, 566–572.
Idland, A.V., Wyller, T.B., Støen, R., Eri, L.M., Frihagen, F., Ræder, J., Chaudhry, F.A.,
Hansson, O., Zetterberg, H., Blennow, K., Bogdanovic, N., Brækhus, A., Watne, L.O.,
2017. Preclinical amyloid-β and axonal degeneration pathology in delirium. J.
Alzheimers Dis. 55, 371–379.
Inouye, S.K., van Dyck, C.H., Alessi, C.A., Balkin, S., Siegal, A.P., Horwitz, R.I., 1990.
Clarifying confusion: the confusion assessment method. A new method for detection
of delirium. Ann. Intern. Med. 113, 941–948.
Inouye, S.K., Westendorp, R.G., Saczynski, J.S., 2014. Delirium in elderly people. Lancet
383, 911–922.
Jorm, A.F., 1994. A short form of the Informant Questionnaire on Cognitive Decline in the
Elderly (IQCODE): development and cross-validation. Psychol. Med. (Paris) 24,
145–153.
Jorm, A.F., 2004. The Informant Questionnaire on cognitive decline in the elderly
(IQCODE): a review. Int. Psychogeriatr./IPA 16, 275–293.
Juliebø, V., Bjøro, K., Krogseth, M., Skovlund, E., Ranhoff, A.H., Wyller, T.B., 2009. Risk
factors for preoperative and postoperative delirium in elderly patients with hip
fracture. J. Am. Geriatr. Soc. 57, 1354–1361.
16
Neurotoxicology 69 (2018) 11–16
Kudoh, A., Ishihara, H., Matsuki, A., 1999. Response to surgical stress in elderly patients
and Alzheimer’s disease. Can. J. Anaesth. 46, 247–252.
Lehmann, A., 1989. Effects of microdialysis-perfusion with anisoosmotic media on ex-
tracellular amino acids in the rat hippocampus and skeletal muscle. J. Neurochem.
53, 525–535.
Lindvall, M., Owman, C., 1981. Autonomic nerves in the mammalian choroid plexus and
their influence on the formation of cerebrospinal fluid. J. Cereb. Blood Flow Metab. 1,
245–266.
Maldonado, J.R., 2013. Neuropathogenesis of delirium: review of current etiologic the-
ories and common pathways. Am. J. Geriatr. Psychiatry 21, 1190–1222.
Morland, C., Pettersen, M.N., Hassel, B., 2016. Hyperosmolar sodium chloride is toxic to
cultured neurons and causes reduction of glucose metabolism and ATP levels, an
increase in glutamate uptake, and a reduction in cytosolic calcium. Neurotoxicology
54, 34–43.
Oehmke, M.J., Podranski, T., Mann, M., Frickey, N., Kuhn, D.F., Hempelmann, G., 2008.
Perioperative concentrations of catecholamines in the cerebrospinal fluid and plasma
during spinal anesthesia. Acta Anaesthesiol. Scand. 52, 487–492.
Saczynski, J.S., Marcantonio, E.R., Quach, L., Fong, T.G., Gross, A., Inouye, S.K., Jones,
R.N., 2012. Cognitive trajectories after postoperative delirium. N. Engl. J. Med. 367,
30–39.
Shukry, M., Clyde, M.C., Kalarickal, P.L., Ramadhyani, U., 2005. Does dexmedetomidine
prevent emergence delirium in children after sevoflurane-based general anesthesia?
Paediatr. Anaesth. 15, 1098–1104.
Shulman, K.I., 2000. Clock-drawing: is it the ideal cognitive screening test? Int. J. Geriatr.
Psychiatry 15, 548–561.
Su, X., Meng, Z.T., Wu, X.H., Cui, F., Li, H.L., Wang, D.X., Zhu, X., Zhu, S.N., Maze, M.,
Ma, D., 2016. Dexmedetomidine for prevention of delirium in elderly patients after
non-cardiac surgery: a randomised, double-blind, placebo-controlled trial. Lancet
388, 1893–1902.
Traynelis, S.F., Dingledine, R., 1989. Role of extracellular space in hyperosmotic sup-
pression of potassium-induced electrographic seizures. J. Neurophysiol. 61, 927–938.
Vitvitsky, V.M., Garg, S.K., Keep, R.F., Albin, R.L., Banerjee, R., 2012. Na+ and K+ ion
imbalances in Alzheimer’s disease. Biochim. Biophys. Acta 1822, 1671–1681.
Wade, D., 1992. Measurment in Neurological Rehabilitation. Oxford University Press,
Oxford.
Watne, L.O., Idland, A.V., Fekkes, D., Raeder, J., Frihagen, F., Ranhoff, A.H., Chaudhry,
F.A., Engedal, K., Wyller, T.B., Hassel, B., 2016. Increased CSF levels of aromatic
amino acids in hip fracture patients with delirium suggests higher monoaminergic
activity. BMC Geriatr. 16, 149.
Welsh, K.A., Butters, N., Mohs, R.C., Beekly, D., Edland, S., Fillenbaum, G., Heyman, A.,
1994. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part
V. A normative study of the neuropsychological battery. Neurology 44, 609–614.
Westhoff, D., Witlox, J., van Aalst, C., Scholtens, R.M., de Rooij, S.E., van Munster, B.C.,
de Jonghe, J.F., Houdijk, A.P., Eikelenboom, P., van Westerloo, D.J., van de Beek, D.,
van Gool, W.A., Koenderman, L., 2015. Preoperative protein profiles in cerebrospinal
fluid in elderly hip fracture patients at risk for delirium: a proteomics and validation
study. BBA Clin. 4, 115–122.
Xie, Z., Swain, C.A., Ward, S.A., Zheng, H., Dong, Y., Sunder, N., Burke, D.W., Escobar, D.,
Zhang, Y., Marcantonio, E.R., 2014. Preoperative cerebrospinal fluid β-Amyloid/Tau
ratio and postoperative delirium. Ann. Clin. Transl. Neurol. 1, 319–328.