Original Paper
Eur Neurol 2009;62:293–297
DOI: 10.1159/000235808
Repetitive Lumbar Punctures as Treatment
for Normal Pressure Hydrocephalus
T.S. Lim S.W. Yong S.Y. Moon
Department of Neurology, Ajou University School of Medicine, Suwon, South Korea
Key Words
Lumbar punctures ⴢ Normal pressure hydrocephalus ⴢ
Shunt operation
Abstract
Selected normal pressure hydrocephalus (NPH) patients
cannot be treated by shunt operation because of the pro-
cedure’s high complication rate. We have treated cases in
which prolonged clinical improvement of NPH was experi-
enced after one or two lumbar punctures (LPs). We eval-
uated the predictors of prolonged improvement of NPH
symptoms by repeated LP. Thirty-one NPH patients were ret-
rospectively evaluated (age 72.5 8 5.8 years). Gait distur-
bance, urinary incontinence, and cognitive impairment were
semiquantified. We divided the patients into three groups
(non-responders, temporary responders, and prolonged re-
sponders) according to their responses after LP. We analyzed
the characteristics of the groups. Gait disturbance (p = 0.046)
and urinary incontinence (p = 0.040) scores and total NPH
symptom score (p = 0.007) after cerebrospinal fluid drainage
were more significantly improved in prolonged responders
than in temporary responders. On multiple logistic regres-
sion analyses, total NPH score improvement was the only
predictor of the prolonged responders (p = 0.03, odds ratio
0.148). Our study showed that some NPH patients could
maintain favorable courses for at least 1 year after LP without
shunt operation. Repeated LP could be an alternative treat-
ment in selected NPH patients.
Copyright © 2009 S. Karger AG, Basel
© 2009 S. Karger AG, Basel
0014–3022/09/0625–0293$26.00/0
Fax +41 61 306 12 34
E-Mail karger@karger.ch Accessible online at:
www.karger.com www.karger.com/ene
Received: March 2, 2009
Accepted: July 13, 2009
Published online: August 27, 2009
Introduction
Normal pressure hydrocephalus (NPH) is character-
ized by slowly progressive gait disturbance, cognitive de-
cline, and urinary disturbance. It can be diagnosed by
characteristic clinical features and laboratory findings,
especially brain imaging showing enlarged ventricles
with relatively less significant periventricular white mat-
ter changes [1]. However, in the clinical setting, making
an NPH diagnosis is not straightforward. It is necessary
to rule out other possible causes and refer to response to
lumbar punctures (LP) and imaging findings to make a
diagnosis and plan for the treatment. Dementia from
NPH is treatable by either a ventriculoperitoneal shunt or
ventriculoatrial shunt. The response to the shunt opera-
tion can be predicted by an LP or external cerebrospinal
fluid (CSF) drainage prior to the operation [2]. However,
it is not simple to select candidates for the shunt operation
due to combined pathology and poor general medical
conditions in the aged population. In addition, according
to the long-term follow-up studies, about 6–8% of pa-
tients experience serious and permanent neurological
deficits such as hemiparesis or even death. Various com-
plications such as subdural hematoma, intracranial hem-
orrhage, or shunt infection were reported in more than
30% of patients [3]. These problems underscore the need
for other treatment options.
A few studies reported that patients who underwent
an LP once or twice without the shunt operation showed
prolonged symptomatic improvement [4]. However, there
So Young Moon, MD, PhD
Department of Neurology, Ajou University School of Medicine
San 5, Woncheon-dong, Yeongtong-gu
Suwon, Kyungki-do 443-721 (South Korea)
Tel. +82 31 219 5175, Fax +82 31 219 5178, E-Mail symoon.bv@gmail.com
 Fig. 1. MRI images showing dilated ventricles with various degrees of white matter changes in three represen-
tative prolonged reponders (a, b, c).
have been no studies that evaluated if repetitive LP can
be an alternative treatment to the shunt operation in
NPH patients. In this study, we reported patients with
NPH whose symptomatic response to repetitive LP was
maintained for more than 1 year without the need for a
shunt operation and analyzed what was the predictor of
the prolonged symptomatic response to repetitive LP in
patients with NPH.
Patients and Methods
Patients
Among patients who visited the Department of Neurology at
the Ajou Medical Center, Suwon, South Korea, from January 2001
to December 2007, we recruited 31 patients who met the criteria
for NPH and were observed for at least 1 year after admission
treatment. All patients had brain MRIs and an LP. The clinical
criteria for NPH included the following: (1) insidious onset, age
of onset 1 40, disease duration 13–6 months, progressive course,
and no other neurologic, psychiatric, or medical condition apart
from the disease symptoms; (2) MRI showing ventricular en-
largement not entirely attributable to cerebral cortical atrophy; (3)
gait disturbance with or without cognitive deficit or urinary dis-
turbance, and (4) no evidence for the increased intracranial pres-
sure (70–245 mm H2O) [5]. After making a diagnosis of NPH, one
or two LPs were performed to drain 30–50 ml of the CSF. All par-
ticipants were fully informed that there were two therapeutic op-
tions of repetitive LP or shunt surgery and agreed to be treated by
repetitive LP at least for 1 year. The study was approved by the
Institutional Review Board of the hospital, and informed consent
was obtained in accordance with the principles of the Declaration
of Helsinki.
294 Eur Neurol 2009;62:293–297
Evaluation of NPH Symptoms
An NPH scale modified from Larsson et al. [6] and Krauss et
al. [7] that assessed gait (1 = normal, 2 = walk without any assis-
tive device but insecure, 3 = walk with cane, 4 = walk with bi-
manual support (walker), 5 = walk aided by an assistant, 6 =
wheelchair-bound), urinary disturbance (0 = normal, 1 = spo-
radic incontinence or urge phenomena, 2 = frequent incontinence
or urge phenomena, and 3 = no or minimal control of bladder
function) and cognitive deficit (0 = normal, 1 = minimal attention
or memory deficits, 2 = considerable attention or memory deficits
but oriented to situational context, and 3 = not or only margin-
ally oriented to situational context) was used to characterize and
grade the clinical syndrome. Patients were evaluated both before
and 6 h after the LP.
Classification of Patients
After 1 year of observation, patients were grouped into non-
responders, temporary responders, and prolonged responders.
Non-responders were patients who showed neither improvement
in any scores nor subjective improvement after the LP. Temporary
responders were patients whose score in any category improved
by at least one point or who reported subjective improvement but
whose improvement was not sustained up to 3 months. Prolonged
responders were patients whose improvement corresponded to
that mentioned above but whose improvement was sustained up
to 3 months.
MRI
All MR images of patients were reviewed by three neurologists
(authors) with visual inspection (fig. 1a–c). The Evans index and
white matter changes were assessed using MRI. The Evans index
was defined as the maximal frontal horn ventricular width di-
vided by the transverse inner diameter of the skull and signifies
ventriculomegaly if it is 60.3 [5]. White matter change was eval-
uated by the method designed by Mantyla et al. [8]. Both peri-
ventricular (0–6) and deep (0–4) white matter changes were as-
sessed [9].
Lim/Yong/Moon
 Table 1. Demographic and clinical characteristics of the patient groups

Non- Transient Long-term Total

responders responders responders
(n = 6) (n = 12) (n = 13) (n = 31)

Men:women 5:1 8:4 8:5 21:10

Age, years 72.886.7 71.685.8 73.285.9 72.585.8
Duration, days 3348353 5218458 4448381 4538400
Evans index 0.3480.03 0.3480.02 0.3380.02 0.3480.02
PVWMH 2.0081.0 2.8381.8 3.6181.7 3.0081.71
DWMH 1.1681.16 2.2581.28 2.0781.18 1.9681.25
White matter changes 3.1682.13 5.0883.02 5.6982.65 4.9682.79
CSF protein, mg/dl 42.3813.2 36.5811.4 38.4810.8 38.4811.3
CSF glucose, mg/dl 68.0813.7 78.2821.6 70.1811.8 72.8816.7
Anti-parkinsonian drugs, n 3 7 6 16
Pre-gait score 2.080.00 2.8381.33 3.1581.21 2.8081.19
Pre-urinary score 1.1680.40 1.3380.77 1.5380.77 1.3880.71
Pre-cognition score 0.8380.40 1.0880.79 1.1580.80 1.0680.72
Pre-NPH score sum 4.0080.63 5.2582.34 5.8481.72 5.2581.93
Post-gait score 2.0080.00 2.2580.86 1.8480.55 2.0380.65
Post-urinary score 1.1680.40 1.0880.66 0.6980.48 0.9380.57
Post-cognition score 0.8380.40 0.9180.51 0.7680.43 0.8380.45
Post-NPH score sum 4.0080.63 4.2581.86 3.3081.10 3.8081.42

PVWMH = Periventricular white matter hyperintensity; DWMH = deep white matter hyperintensity;
CSF = cerebrospinal fluid; NPH = normal pressure hydrocephalus.

Statistical Analysis To look for predictors to be able to discriminate be-

First, using the Mann-Whitney test, we compared demo-
graphic features, CSF findings, Evans index, white matter chang-
es, medications taken, total and categorized scores for the NPH
before and after the LP, and the amount of the improvement in
the scores after the LP between temporary responders and pro-
longed responders. We also used logistic regression analysis to
search for independent predictors to discriminate between the
two groups. p ! 0.05 was regarded as statistically significant. The
statistical analyses were performed using commercially available
software (SPSS, Version 12.0).
Results
tween temporary and prolonged responders, we used
multiple logistic regression analysis. We categorized sev-
eral variables as follows: age of onset (61–70, 71–80, 181),
disease duration (!1 year, 61 year), scores for white mat-
ter changes (!3, 63), amount of CSF protein (!40 mg/dl,
640 mg/dl), and changes in the total NPH scores (!2,
62). The analysis showed that the amount of changes in
the total NPH scores was the only independent predictor
to discriminate two groups (odds ratio 0.148 (95% confi-
dence interval 0.026–0.860), p = 0.03).

We recruited 31 patients (21 men and 10 women). The Discussion

mean age of onset was 72.5 and the mean disease dura-
tion was 453 days. These features are summarized in ta-
ble 1. The patients consisted of 6 non-responders, 12 tem-
porary responders and 13 prolonged responders. We sta-
tistically compared temporary and prolonged responders
using the Mann-Whitney test. Among the variables test-
ed, the amount of changes in total NPH scores, gait score
changes, and urination score changes after LP were sig-
nificantly different (table 2).
Although a few studies have reported that patients
showed prolonged response to LP, these studies have not
been given full attention. In addition, good response to
the LP, which is revealed to be an independent predictor
of the prolonged responders in this study, is also the pre-
dictor of favorable outcome of the shunt operation [10].
Therefore, repeated LP has not been systematically stud-
ied since the shunt operation was introduced in 1960s.

Lumbar Punctures for Normal Pressure Eur Neurol 2009;62:293–297 295

Hydrocephalus
Table 2. Comparison of demographic and clinical characteristics pressure according to the Pascal’s law. If the ventricles are
between transient responders and long-term responders slightly shrunken by CSF removal, the elasticity of the
ventricles can recover along with symptomatic improve-
Transient Long-term p
responders responders value ment for a certain period. It seems that prolonged re-
(n = 12) (n = 13) sponders to LP did not acquire irreversible damage to the
brain although patients showed symptoms. However, if
Men:women 8:4 8:5 0.852 appropriate CSF removal is not done at the threshold,
Age, years 71.685.8 73.285.9 0.503 irreversible damage to the cerebrum occurs and can
Duration, days 5218458 4448381 0.852
Evans index 0.3480.02 0.3380.02 0.650 produce non-responders or temporary responders to
PVWMH 2.8381.8 3.6181.7 0.295 the LP.
DWMH 2.2581.28 2.0781.18 0.728 We should accept that this study has several limita-
White matter changes 5.0883.02 5.6982.65 0.611 tions. This study was retrospective and could not quan-
CSF protein, mg/dl 36.5811.4 38.4810.8 0.689 tify each symptom accurately. Each symptom was evalu-
CSF glucose, mg/dl 78.2821.6 70.1811.8 0.437
Anti-parkinsonian drugs, n 7 6 0.825 ated every 3 months at the outpatient clinic. Therefore,
Pre-gait score 2.8381.33 3.1581.21 0.437 we could not assess exactly when patients started to wors-
Pre-urinary score 1.3380.77 1.5380.77 0.470 en. In addition, some parts of the scales were somewhat
Pre-cognition score 1.0880.79 1.1580.80 0.810 subjective and vulnerable to patients’ other medical con-
Pre-NPH score sum 5.2582.34 5.8481.72 0.225 ditions including mood. Unfortunately, we did not evalu-
Post-gait score 2.2580.86 1.8480.55 0.347
Post-urinary score 1.0880.66 0.6980.48 0.247 ate mood with scales throughout the study although we
Post-cognition score 0.9180.51 0.7680.43 0.611 excluded patients with major depression. Furthermore,
Post-NPH score sum 4.2581.86 3.3081.10 0.270 many patients failed to be followed after just 1 year of ob-
Gait score improvement 0.5880.90 1.3080.94 0.046* servation. Future studies are needed to compare clinical
Urinary score improvement 0.2580.45 0.8480.68 0.040* courses according to the treatment modality (the shunt
Cognition score improvement 0.1680.38 0.3880.65 0.538
NPH score improvement 1.0080.95 2.5381.50 0.007*
operation/repeated LP) among non-responders, tempo-
rary responders, and prolonged responders.
PVWMH = Periventricular white matter hyperintensity; In conclusion, this study showed that good responders
DWMH = deep white matter hyperintensity; CSF = cerebrospinal to the LP should be observed for at least 3 months to see
fluid; NPH = normal pressure hydrocephalus. * p < 0.05. if their responses are prolonged. If they are prolonged and
patients agree, they can be observed without an immedi-
ate shunt operation, especially in cases where the patients
are not in an appropriate condition for the operation or
could have several degenerative diseases.
This study showed that prolonged responders had sig-
nificantly more improvement in the total NPH score, gait
score, and urination score. In addition, we found that the Acknowledgement
amount of the total NPH score change was an indepen-
This study was supported by a grant of the Korea Health 21
dent predictor to discriminate between temporary and
R&D Project, Ministry of Health and Welfare, and Family Af-
prolonged responders. Although it is impossible to dis- fairs, Republic of Korea (A050079).
tinguish responses to the LP from the natural course of
the NPH, the presence of patients who showed sustained
response with repeated LP could be evidence for the ther-
apeutic responses. In addition, it is difficult to exclude
any disease responsive to the anti-parkinsonian drugs.
However, there was no significant difference in the drug
history among the three groups.
At an early stage of NPH, patients can compensate for
abnormalities in CSF flow through both reduced produc-
tion and increased absorption of the CSF due to increased
intracranial pressure [11]. However, as the ventricles get
larger, they lose the elasticity with the even intracranial

296 Eur Neurol 2009;62:293–297 Lim/Yong/Moon

 References
1 Adams RD, Fisher CM, Hakim S, Ojemann
RG, Sweet WH: Symptomatic occult hydro-
cephalus with ‘normal’ cerebrospinal-fluid
pressure. A treatable syndrome. N Engl J
Med 1965;273:117–126.
2 Graff-Radford NR: Normal pressure hydro-
cephalus. Neurol Clin 2007;25:809–832.
3 Hebb AO, Cusimano MD: Idiopathic normal
pressure hydrocephalus: a systematic review
of diagnosis and outcome. Neurosurgery
2001;49:1166–1184.
4 Tsakanikas D, Relkin N: Normal pressure
hydrocephalus. Semin Neurol 2007; 27: 58–
65.
Lumbar Punctures for Normal Pressure
Hydrocephalus
5 Relkin N, Marmarou A, Klinge P, Berg- 9 Gyldensted C: Measurements of the normal
sneider M, Black PM: Diagnosing idiopathic ventricular system and hemispheric sulci of
normal-pressure hydrocephalus. Neurosur- 100 adults with computed tomography. Neu-
gery 2005;57:S4–S16. roradiology 1977; 14:183–192.
6 Larsson A, Wikkelso C, Bilting M, Ste- 10 Vanneste JA: Diagnosis and management of
phensen H: Clinical parameters in 74 con- normal-pressure hydrocephalus. J Neurol
secutive patients shunt operated for normal 2000;247:5–14.
pressure hydrocephalus. Acta Neurol Scand 11 Bateman GA: Pulse wave encephalopathy: a
1991;84:475–482. spectrum hypothesis incorporating Alz-
7 Krauss JK, Regel JP, Vach W, Jungling FD, heimer’s disease, vascular dementia and nor-
Droste DW, Wakhloo AK: Flow void of cere- mal pressure hydrocephalus. Med Hypoth-
brospinal fluid in idiopathic normal pres- eses 2004;62:182–187.
sure hydrocephalus of the elderly: can it pre-
dict outcome after shunting? Neurosurgery
1997;40:73–64.
8 Mantyla R, Erkinjuntti T, Salonen O, Aronen
HJ, Peltonen T, Pohjasvaara T, Standert-
skjöld-Nordenstam CG: Variable agreement
between visual rating scales for white matter
hyperintensities on MRI. Comparison of 13
rating scales in a poststroke cohort. Stroke
1997;28:1614–1623.
Eur Neurol 2009;62:293–297 297