doi:10.

1093/brain/awt255 Brain 2013: 136; 3395–3407 | 3395

BRAIN
A JOURNAL OF NEUROLOGY

Phenotypic spectrum of probable and

genetically-confirmed idiopathic
basal ganglia calcification
Gaël Nicolas,1,2,3,4,5 Cyril Pottier,1,3,4 Camille Charbonnier,4 Lucie Guyant-Maréchal,5
Isabelle Le Ber,6 Jérémie Pariente,7 Pierre Labauge,8 Xavier Ayrignac,8 Luc Defebvre,9
David Maltête,1,2,3 Olivier Martinaud,2,4 Romain Lefaucheur,2 Olivier Guillin,1,3,10

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

David Wallon,1,2,3,4 Boris Chaumette,10 Philippe Rondepierre,11 Nathalie Derache,12
Guillaume Fromager,12 Stéphane Schaeffer,12 Pierre Krystkowiak,13 Christophe Verny,14
Snejana Jurici,15 Mathilde Sauvée,16 Marc Vérin,17 Thibaud Lebouvier,18 Olivier Rouaud,19
Christel Thauvin-Robinet,20 Stéphane Rousseau,1,3,4 Anne Rovelet-Lecrux,1,3,4 Thierry Frebourg,1,3,5
Dominique Campion,1,3,4,21 Didier Hannequin1,2,3,4,5 and the French IBGC study group§

1 Inserm U1079, Rouen, France

2 Department of Neurology, Rouen University Hospital, Rouen, France
3 Normandie Univ, IRIB, Rouen, France
4 CNR-MAJ, Rouen, Lille and Paris Salpêtrière University Hospitals
5 Department of Genetics, Rouen University Hospital, Rouen, France
6 Inserm, UMR_S975, CRICM; UPMC Univ Paris 06, UMR_S975; CNRS UMR 7225, and AP-HP, Pitié-Salpêtrière Hospital, Centre de Référence des
Démences Rares, Paris, France
7 Inserm, imagerie cérébrale et handicaps neurologiques UMR 825, Centre Hospitalier Universitaire de Toulouse, Toulouse, France; Université de
Toulouse, UPS, imagerie cérébrale et handicaps neurologiques UMR 825, Centre Hospitalier Universitaire de Toulouse, Toulouse, France; and
Service de neurologie, pôle neurosciences, Centre Hospitalier Universitaire de Toulouse, Toulouse, France
8 Department of Neurology, Montpellier University Hospital, Montpellier, France
9 Department of Neurology A, Salengro University Hospital, and EA4559, Lille, France
10 University Department, Rouvray Psychiatric Hospital, Sotteville-lès-Rouen, France
11 Department of Neurology, Dreux Hospital, Dreux, France
12 Department of Neurology, Caen University Hospital, Caen, France and Inserm U1077, Caen, F-14000, France
13 Department of Neurology, Amiens University Hospital; EA 4559 - Laboratoire de Neurosciences Fonctionnelles et Pathologie (LNFP), University of
Picardy Jules Verne (UPJV), Amiens, France; SFR CAP-Santé (FED 4231), UFECAP
14 Department of Neurology, Angers University Hospital, School of Medicine; CNRS UMR 6214, Angers, France; Inserm U1083, Angers, France
15 Department of Neurology, Perpignan Hospital, Perpignan, France
16 Department of Neurology, Nancy University Hospital, Nancy, France
17 Department of Neurology, Pontchaillou Hospital, Rennes University Hospital, Rennes, France
18 Department of Neurology, Nantes University Hospital, Nantes, France
19 Department of Neurology, Dijon University Hospital, Dijon, France
20 Centre de Génétique et Centre de Référence Anomalies du Développement et Syndromes Malformatifs de l’interrégion Grand-Est, Hôpital
d’Enfants, Dijon University Hospital, Dijon, France
21 Department of Research, Rouvray Psychiatric Hospital, Sotteville-lès-Rouen, France
§
Details of the French IBGC study group can be found in the Acknowledgements.

Correspondence to: Gaël Nicolas, Inserm U1079,

Faculté de Médecine et de Pharmacie,
22, boulevard Gambetta,
76183 Rouen Cedex 1, France
E-mail: [email protected]

Received June 10, 2013. Revised July 9, 2013. Accepted July 24, 2013. Advance Access publication September 24, 2013
ß The Author (2013). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.
For Permissions, please email: [email protected]
3396 | Brain 2013: 136; 3395–3407 G. Nicolas et al.

Idiopathic basal ganglia calcification is characterized by mineral deposits in the brain, an autosomal dominant pattern of inheritance in
most cases and genetic heterogeneity. The first causal genes, SLC20A2 and PDGFRB, have recently been reported. Diagnosing idiopathic
basal ganglia calcification necessitates the exclusion of other causes, including calcification related to normal ageing, for which no
normative data exist. Our objectives were to diagnose accurately and then describe the clinical and radiological characteristics of idiopathic
basal ganglia calcification. First, calcifications were evaluated using a visual rating scale on the computerized tomography scans of 600
consecutively hospitalized unselected controls. We determined an age-specific threshold in these control computerized tomography scans
as the value of the 99th percentile of the total calcification score within three age categories: 540, 40–60, and 460 years. To study the
phenotype of the disease, patients with basal ganglia calcification were recruited from several medical centres. Calcifications that rated
below the age-specific threshold using the same scale were excluded, as were patients with differential diagnoses of idiopathic basal
ganglia calcification, after an extensive aetiological assessment. Sanger sequencing of SLC20A2 and PDGFRB was performed. In total, 72
patients were diagnosed with idiopathic basal ganglia calcification, 25 of whom bore a mutation in either SLC20A2 (two families, four
sporadic cases) or PDGFRB (one family, two sporadic cases). Five mutations were novel. Seventy-one per cent of the patients with
idiopathic basal ganglia calcification were symptomatic (mean age of clinical onset: 39  20 years; mean age at last evaluation:
55  19 years). Among them, the most frequent signs were: cognitive impairment (58.8%), psychiatric symptoms (56.9%) and movement
disorders (54.9%). Few clinical differences appeared between SLC20A2 and PDGFRB mutation carriers. Radiological analysis revealed that
the total calcification scores correlated positively with age in controls and patients, but increased more rapidly with age in patients. The

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

expected total calcification score was greater in SLC20A2 than PDGFRB mutation carriers, beyond the effect of the age alone. No patient
with a PDGFRB mutation exhibited a cortical or a vermis calcification. The total calcification score was more severe in symptomatic versus
asymptomatic individuals. We provide the first phenotypical description of a case series of patients with idiopathic basal ganglia calci-
fication since the identification of the first causative genes. Clinical and radiological diversity is confirmed, whatever the genetic status.
Quantification of calcification is correlated with the symptomatic status, but the location and the severity of the calcifications don’t reflect
the whole clinical diversity. Other biomarkers may be helpful in better predicting clinical expression.

Keywords: Fahr’s disease; calcification; SLC20A2; PDGFRB; ageing

Abbreviations: BGC = basal ganglia calcification; IBGC = Idiopathic BGC; TCS = total calcification score

Introduction studies (Oliveira et al., 2004; Kostic et al., 2011). Therefore, aetiolo-
gical assessment remains necessary; firstly to select patients for genetic
Basal ganglia calcification (BGC) was first described in 1850 screening and secondly to attribute a probable diagnosis of IBGC in
(Delacour, 1850) and is known to be caused by numerous condi- cases not related to a mutation in one of the two recently described
tions such as phosphocalcic metabolic disorders, mitochondrial genes. This type of assessment entails searching for known causes of
diseases, numerous hereditary and non-hereditary congenital syn- BGC, including calcification associated with normal ageing.
dromes and acquired conditions (Baba et al., 2005; Manyam, Reports have shown wide clinical diversity within and between
2005). BGC has also been reported in normal ageing and seems families. Calcification severity also seems to be diverse. Only one
to be common in older people (Simoni et al., 2008; Yamada et al., case series was published on different phenotypes associated with
2013). Nevertheless, thus far no published work has reported a the disease before the discovery of the first causative genes
method for determining when BGC can be attributed to normal (Manyam et al., 2001). Their identification opens the door to
ageing or to a pathological process. As BGC can be clinically silent, clinical and radiological studies of patients affected by genetically
its classification as a normal part of ageing can therefore not be confirmed IBGC. With the aim to describe the phenotype asso-
based exclusively on neuropsychiatric examination. ciated with definite IBGC as well as probable IBGC (with unknown
Idiopathic BGC (IBGC), also known as Fahr’s disease, is clinically genetic cause), we selected the patients bearing BGC with no
heterogeneous. Patients with calcification may exhibit neurological known cause following an etiological assessment, and included
and/or psychiatric symptoms with diverse severity and ages of only cases exhibiting calcifications rated above the age-specific
onset. Others can remain asymptomatic throughout life (Yamada threshold, determined by using data from controls. To rate the
and Hayashi, 2000; Manyam, 2005). For these reasons, the diagnosis calcifications on CT scans, we used a new visual rating scale in
is based on (i) the presence of BGC, which may or may not involve controls and IBGC patients.
other brain areas, regardless of the clinical status; and (ii) the exclusion
of the other causes of BGC. As a consequence, an extensive clinico-
biological aetiological assessment is necessary to diagnose probable
IBGC. Definite IBGC can now be diagnosed in patients bearing a cau-
Patients and methods
sative mutation in one of the two recently identified genes: SCL20A2
(Wang et al., 2012) and PDGFRB (Nicolas et al., 2013). Nevertheless, Visual rating of calcifications
these two genes do not account for all cases of IBGC, confirming the Axial view CT scans were analysed while blinded to identity, by two
genetic heterogeneity of the disease previously suggested by linkage investigators (G.N. and D.H.), separately and then jointly, using an
Phenotype of idiopathic basal ganglia calcification Brain 2013: 136; 3395–3407 | 3397

original visual grading system. Consensus was reached when the two age 550), evaluation of episodic memory and assessment of executive
investigators did not attribute the same score for each location. As this functions. Episodic verbal memory was assessed using Free and Cued
system was designed to improve the accuracy of IBGC diagnoses, we Recall Test (if age 550). Cueing was considered efficient when pa-
focused on brain parenchymal calcification. Calcifications were rated tients with impaired free recall achieved normal total recall (Sarazin
when there was a spontaneously hyperdense area with a Hounsfield et al., 2007; Godefroy et al., 2009). Executive functions were evalu-
value consistent with a calcification. The following locations were ana- ated using the Frontal Assessment Battery, Stroop Test, Trail Making
lysed: left and right lenticular nucleus, left and right caudate nucleus, Test, Modified Card Sorting Test and verbal fluencies. Patients were
left and right thalamus, left and right cerebral subcortical white matter, considered to suffer from cognitive dysexecutive syndrome according
left and right internal capsule only if independent of other calcifica- to normative data (Godefroy et al., 2010). Complementary neuropsy-
tions, cerebral cortex, left and right cerebellar hemisphere, vermis, left chological evaluation included, when available: Oral Naming Test,
and right midbrain, pons and medulla. The scores were attributed ac- Gestural Praxis Evaluation and Rey Figure Copy.
cording to specific definitions with visual examples (see Supplementary
material for a comprehensive description): 0 = absent calcification;
1 = punctate; 2 = faint; 3 = moderate; 4 = severe; 5 = severe and con- Genetic analyses
fluent. The scores were compiled into a total calcification score (TCS) DNA was extracted from peripheral blood. Sanger sequencing of
by addition. SLC20A2 and PDGFRB was performed in each proband. Segregation
of the novel variants (not retrieved in the dbSNP132 and EVS databases)
was studied when DNA from relatives was available. We concluded that

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

Control patients a variant was a disease-causative mutation if: it was not retrieved
To determine age-specific thresholds, we rated the CT scans of control among the dbSNP132 and EVS databases or our 173 in-house controls
patients hospitalized in our university department of neurology during (obtained by whole-exome sequencing); the variant was highly disrup-
a 6-month period. The study was approved by our local ethics com- tive (e.g. introducing a frameshift or altering the splice); or a missense
mittee. All CT scans were performed using the same protocol (General variant was predicted to alter the protein function by at least two soft-
Electric Lightspeed 64-slice scanner, 120 kVp, 280 mAs, 1.25 mm slice ware programmes (PolyPhen2, Mutation taster and SIFT) or by one
thickness, 0.9 mm interslice gap, non-contrast enhanced). We then software programme and the base was highly conserved across species
sorted the CT scans into three categories according to the age of (according to the PhyloP score) (Liu et al., 2011).
the control patients: 540.0 years, 40.0–59.9 years, 560.0 years of
age. A total of 200 CT scans were rated in each age category. The
main diagnosis or clinical context was collected for each control pa-
Statistics
tient. If a large area involving the location(s) to be scored was not Inter-rater agreement was assessed using Cohen’s weighted kappa
visible because of a pathological process (e.g. extensive haematoma or reliability test. Correlations between age and TCS were evaluated
tumour), the CT scans were not included in the study. Scans showing using a Kendall correlation test. Comparisons between the three
brain calcification clearly attributable to a previously known patho- groups of age in controls were performed using Fisher exact tests.
logical process (e.g. calcified tumour) were also not included. We We used t-tests to compare the mean scores between symptomatic
determined the threshold for each category of age to be the value and asymptomatic patients with IBGC and differences between the
of the 99th percentile of the TCS. mean ages of the SLC20A2 and PDGFRB mutation carriers. To account
for the combined effects of age and genotypes (definite IBGC with
SLC20A2 or PDGFRB mutation, and probable IBGC with unknown
Patients with idiopathic basal ganglia genetic cause) on the quantification scores (TCS and scores from
calcification each location), we adopted a negative binomial model, which allows
for over-dispersion of count data (Ver Hoef and Boveng, 2007; Zeileis
Patients were recruited from multiple centres and all patients and/or
et al., 2008).
relatives gave informed, written consent. Each individual who partici-
We used the statistical software R.
pated in the study underwent a semi-structured interview, a neuro-
logical examination and a CT scan (with diverse acquisition protocols).
The calcifications were reviewed and rated by the same investigators
as for the scans of the control patients, using our visual rating scale. In Results
accordance with previous studies, IBGC status was determined based
first on radiological criteria (Kostic et al., 2011; Nicolas et al., 2013).
The diagnosis of probable IBGC was retained if: (i) patients were af-
Control patients
fected by bilateral lenticular calcifications; (ii) the TCS was above the Inter-rater agreement was excellent (weighted kappa = 0.969). Of
age-specific threshold (599th percentile); and (iii) no cause was the CT scans analysed consecutively, 30 showed a large patho-
retrieved after an extensive aetiological assessment in the proband logical process and were thus impossible to rate. These 30 control
(Nicolas et al., 2013). Genetic analyses were performed in the pro-
patients were therefore not included in the study. Three more
bands first diagnosed with probable IBGC. Patients with a mutation
control patients were not included because of a known cause of
within SLC20A2 or PDGFRB were then considered to be affected by
brain calcification (unilateral thalamus calcification in a 45-year-old
definite IBGC, as well as their relatives bearing both mutation and
BGC. Patients with IBGC were considered symptomatic if they pre- patient, located in a sequelae of a toxoplasma infection, brainstem
sented at least one sign among the list provided in the previous case tumour calcification in a 47-year-old patient, and another calcified
series (Manyam et al., 2001). Symptomatic patients were evaluated tumour in the cerebral cortex of a 59-year-old patient). Finally,
with a neuropsychological test battery, which included, depending on 200 control patients were included in each group (Table 1). The
age: Mini Mental State Examination, Mattis Dementia Rating Scale (if main diagnosis or clinical context of the CT scans in the 600
3398 | Brain 2013: 136; 3395–3407 G. Nicolas et al.

control patients is provided in Supplementary Table 1. Ninety

(threshold)
percentile
TCS: 1st
control patients (15%) had a TCS 4 0. Among them, 87 had len-
ticular calcifications (50 unilateral, 37 bilateral), with a maximum

0
4
5
score of 3 for each lenticular calcification (Table 1). Five control
patients showed calcifications in areas other than lenticular nu-

locations
clei; they were all 460 years old (Supplementary Note 1). In

Other
the group of control patients who were 540, only one (0.5%),

0
0
0
aged 28, presented a calcification. In the group of patients be-
tween 40 and 60 years old, 29 (14.5%) had a TCS 4 0. In the

n / max score
calcification,
group of control patients over 60 years old, 60 (30.0%) had a
TCS 4 0.

Cortical

1 /3
There were significantly more patients with calcifications in the
older groups (P 5 0.0001 for each comparison, except that

0
0
between 40–60 years and 460 for which P = 0.0002) and there

calcification,
was a positive correlation between age and TCS (Kendall

hemisphere
Cerebellar

n / max
tau = 0.2873883, P 5 0.0001).

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

3 /1
score
The 99th percentile of the TCS were: 0 in 540 years, 4 in
40–60 years, and 5 in 460 years.

0
0
calcification,
Patients with idiopathic basal ganglia

n / max
Caudate
calcification

1 /3
score

0
0
Of the 53 index patients referred for bilateral lenticular calcifica-
tions, 15 were excluded from our study: three scored a TCS under

calcification,
the age-specific threshold (they did not have family history of

Lenticular

bilateral,
n / max
neuropsychiatric disorders and were diagnosed respectively with

14 / 6
23 / 6
score
stroke, vascular dementia and Alzheimer’s disease with compatible

0
CSF biomarkers) and 12 were attributed another cause following
the aetiological assessment [pseudohypoparathyroidism type 1a
(n = 2), type 1b (n = 1); hyperparathyroidism with renal insuffi- calcification,
unilateral,
Lenticular

n / max
Table 1 Demographic characteristics and calcification scores of the controls

ciency (n = 2), with parathyroid gland hyperplasia (n = 2); brain

15 / 3
34 / 3
1 /1
irradiation in toto  chemotherapy (n = 2); Down syndrome score
(n = 1); neurolupus (n = 1) and uncharacterized autoimmune
disease (n = 1)].
Maximum

Thus, the diagnosis of probable IBGC was first retained in 38

patients. Among them, nine belonged to a pedigree with a
TCS

1
6
6

demonstrative autosomal dominant inheritance of IBGC and six

calcifications
Presence of

belonged to families in which the number of relatives with avail-

29 (14.5)
60 (30.0)

able brain imaging was insufficient to ascertain a specific pattern

1 (0.5)
n (%)

of inheritance. Taking into account the affected relatives, the IBGC

Yes

series was comprised of 38 patients from the nine autosomal dom-

inant families, 11 patients from the six unspecified families, and 23
(males /
females)

75/125
101/99
108/92

with a sporadic presentation (total = 72).

Sex

Genetic analyses
29.6  6.4
50.7  5.8
75.1  8.2
age  SD

We previously reported two families with an SLC20A2 mutation

Mean

(of four and two affected individuals) (Nicolas et al., 2013). We

found four new SLC20A2 mutations in four cases with a sporadic
presentation (Table 2). None of the four new SLC20A2 mutations
40–60 years (n = 200)
540 years (n = 200)

560 years (n = 200)

have previously been reported by others (Schottlaender et al.,

2012; Wang et al., 2012; Hsu et al., 2013; Lemos et al., 2013;
Zhang et al., 2013). All relatives of an SLC20A2 mutation carrier
who also carried the mutation exhibited BGC and were thus
included.
Group

We previously reported two missense mutations within PDGFRB

in a large family (13 affected individuals, Family 870) and a case
Phenotype of idiopathic basal ganglia calcification Brain 2013: 136; 3395–3407 | 3399

Table 2 Mutations within SLC20A2 and PDGFRB in the definite cases of IBGC. Control exomes are from 173 individuals with non-neurological or psychiatric disease
with a sporadic presentation (Nicolas et al., 2013). In Family 870,

PhyloP

5.69

4.73
0.21
5.77

5.53

5.69
4.89
1.50
3.51
all relatives carrying the familial mutation presented BGC with
diverse severity (see brain imaging section) and were therefore

Deleterious (0.05)

Deleterious (0.00)

Deleterious (0.00)
Deleterious (0.00)
Deleterious (0.00)
Deleterious (0.01)
included. We found a new PDGFRB missense mutation in a

Tolerated (0.29)
patient with a sporadic presentation (Table 2).
All 25 patients with a mutation in SLC20A2 or PDGFRB were
therefore diagnosed with definite IBGC. A total of 47 patients had
SIFT

no mutation in the known IBGC-causative genes, and were diag-

NA

NA
nosed with probable IBGC (Table 3, ‘unknown’).

Probably damaging (1.000)

Possibly damaging (0.745)

Probably damaging (0.98)

Probably damaging (1.00)

Probably damaging (1.00)

Probably damaging (1.00)
Neurological and psychiatric symptoms
Most (70.8%) of our patients were symptomatic with a mean age
Benign (0.155)

of 54.9  18.8 years at last evaluation and clinical onset at a mean

Polyphen2

age of 39.2  20.0 years; 29.2% were asymptomatic (mean age

at the time of their last evaluation: 49.3  18.3 years, P = 0.94)
NA

NA

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

(Table 3). The most frequent symptoms were cognitive impair-
ment (58.8% of the symptomatic patients), psychiatric signs
Disease causing (0.99)

Disease causing (0.99)

Disease causing (0.99)
Disease causing (1.00)
Disease causing (0.99)
Disease causing (0.99)
Disease causing (0.95)

(56.9%) and movement disorder (54.9%) (Table 3).

Polymorphism (0.99)

The most frequently impaired cognitive field was memory (70%).

In patients with memory impairment, episodic memory disorder was
characterized by impaired encoding and retrieval with efficient
Mutation

cueing in 9 of 21 patients (43%). This pattern is consistent with

Taster

frontal subcortical dysfunction. Conversely, 10 patients (47%) had

NA

severe episodic memory impairment without efficient cued recall

EVS, control

[for 2 of 21 (10%), memory impairment was not assessed by free

dbSNP132,

(n = 173)

and cued recall test]. Sixty per cent of the patients with cognitive
exomes

Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent

impairment presented a dysexecutive syndrome. Among the symp-

tomatic patients with complementary neuropsychological evalu-
Ref 1 = Nicolas et al., 2013. Items to be included in our criteria for disease-causative mutations are in bold type.

ation, oral naming was pathological in 2 of 13, gestural praxis in

*The nomenclature of the mutations refers to transcripts NM_006749 (SLC20A2) and NM_2609 (PDGFRB).
p.Asn509LysFs*7

3 of 10, and Rey figure copy in 7 of 16, according to normative

p.Glu1071Val
p.Pro184Leu

p.Leu658Pro
p.Arg987Trp
p.Asn194Ser
p.Glu571Ser

p.Asp28Asn

data. One sporadic patient and two patients of two autosomal

Protein*

dominant families presented intellectual disability.

Psychiatric signs included mostly mood disorders (59%), ranging
NA

from one or recurrent depressive episode(s) to bipolar disorder,

and psychotic symptoms (28%) ranging from some positive
c.431-1G 4 T
c.1711G 4 A

c.1973T 4 C
c.2959C 4 T
c.3212A 4 T
c.581A 4 G
c.551C 4 T

c.1527delT
c.82G 4 A

signs to psychosis. Other psychiatric symptoms included personal-

c.DNA*

ity disorders, attention deficit hyperactivity disorder, uncharacter-

ized behavioural disturbances and somatoform symptoms. No
obsessional compulsive disorder was diagnosed. Interestingly,
Frameshift

psychosis was clinically undistinguishable from schizophrenia in

Mutation

Missense

Missense
Missense

Missense
Missense
Missense
Missense

five patients. Among them, three belonged to a pedigree with

Splice

autosomal dominant inheritance (unknown genetic cause) (Le

Ber et al., 2007). The other two had a sporadic presentation:
SLC20A2
SLC20A2
SLC20A2
SLC20A2
SLC20A2
SLC20A2
PDGFRB
PDGFRB
PDGFRB

first, a patient with a mutation within PDGFRB (age of psychosis

Gene

onset = 16 years), who later experienced two ischaemic strokes at

the age of 62; and second, a woman without known genetic
cause (age of psychosis onset = 35 years) presenting intellectual
Affected

disability, depressive elements and who insidiously developed

(n)

13
2
4
1
1
1
1

1
1

extrapyramidal, pyramidal sign and gait disorder with static cere-

bellar syndrome in her early 50 s. Three additional patients had
Family 1038 (Ref 1)
Family 752 (Ref 1)

Family 870 (Ref 1)

Case 1060 (Ref 1)

some psychotic signs: one patient, with a sporadic presentation

Case 1352 (new)

Case 1106 (new)

Case 413 (new)
Case 485 (new)
Case 424 (new)

and no known genetic cause, had positive signs (delusion, hallu-

identification

cinations), apathy and then a major depressive episode without

Individual
or family

disorganized speech or behaviour (Nicolas et al., 2012). The

other two developed few positive signs (visual hallucinations in
one and a transient delusion in the second).
3400 | Brain 2013: 136; 3395–3407 G. Nicolas et al.

Table 3 Demographic and clinical characteristics of patients with definite (SLC20A2 or PDGFRB mutation) and probable
(unknown genetic cause) IBGC

Definite IBGC Probable IBGC Total Manyam et al. (2001)

Gene SLC20A2 PDGFRB Unknown All Unknown
Cases (n) 10 15 47 72 99
Inheritance
Autosomal dominant, n families (cases) 2 (6) 1 (13) 6 (19) 9 (38) NA (73)
Familial, unspecified, n families (cases) 0 (0) 0 (0) 6 (11) 6 (11) NA (12)
Sporadic cases 4 2 17 23 14
Sex (male/female) 1.50 1.50 1.35 1.40 1.36
Asymptomatic, n (%) 3 (30.0) 9 (60.0) 9 (19.1) 21 (29.2) 32 (32)
Age at last evaluation (mean, SE) 67.0 (8.0) 43.3 (5.0) 49.4 (6.7) 49.3 (4.0) 32 (20)
Symptomatic, n (%) 7 (70.0) 6 (40.0) 38 (80.9) 51 (70.8) 67 (68)
Age at onset (mean, SE) 37.3 (8.9) 24.5 (8.6) 41.9 (3.0) 39.2 (2.8) NA
Age at onset (range) 10–77 6–21 3–80 3–80
Age at last evaluation (mean, SE) 67.0 (3.9) 46.2 (9.9) 54.1 (3.0) 54.9 (2.6) 47 (15)

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

Cognitive, n (%a) 5 (71.4) 2 (33.3) 23 (60.5) 30 (58.8) 26 (39)
Dysexecutive 3 2 13 18 (60.0b) NA
Memory 4 2 15 21 (70.0b) NA
Intellectual disability 1 0 2 3 (10.0b) NA
Uncharacterized/other 2 1 11 14 (46.7b) NA
Psychiatric, n (%a) 5 (71.4) 5 (83.3) 19 (50.0) 29 (56.9) 21 (31)
Psychotic signs 1 1 6 8 (27.6c) NA
Mood disorder 2 2 13 17 (58.6c) NA
Other 2 2 2 6 (20.7c) NA
Movement disorder, n (%a) 5 (71.4) 2 (33.3) 21 (55.3) 28 (54.9) 37 (55)
Akinetic-hypertonic 4 2 16 22 (78.6d) 21 (57b)
Tremor 1 1 5 7 (25.0d) 3 (8b)
Chorea 0 1 3 4 (14.3d) 7 (19b)
Dystonia 1 1 5 7 (25.0d) 3 (8b)
Athethosis 0 0 0 0 (0.0d) 2 (5b)
Orofacial Dyskinesia 1 0 2 3 (10.7d) 1 (3b)
Other 1 0 1 2 (7.1d) NA
Gait, n (%a) 3 (42.9) 1 (16.7) 18 (47.4) 22 (43.1) 12 (18)
Speech, n (%a) 1 (14.3) 2 (33.3) 17 (45.9) 20 (39.2) 24 (36)
Cerebellar, n (%a) 0 (0.0) 1 (16.7) 13 (35.1) 14 (27.5) 24 (36)
Pyramidal, n (%a) 1 (14.3) 1 (16.7) 10 (26.3) 12 (23.5) 15 (22)
Seizures, n (%a) 0 (0.0) 0 (0.0) 4 (10.5) 4 (7.8) 6 (9)
Gastro-intestinal, n (%a) 0 (0.0) 0 (0.0) 4 (10.5) 4 (7.8) 8 (12)
Genito-urinary, n (%a) 0 (0.0) 1 (16.7) 3 (7.9) 4 (7.8) 9 (13)
Sensory/pain, n (%a) 1 (14.3) 0 (0.0) 1 (2.6) 2 (3.9) 11 (16)

Uncharacterized cognitive impairments include patients with pathological global scores without available tests examining memory or executive functions. Other cognitive
impairments include pathological oral naming, Rey figure copy, and gestural praxis. For comparison, data from the case series reported by Manyam et al. (2001) are recalled
on the right (38 cases seen in a registry and 61 cases from literature review) but may not be strictly compared for each point (in Manyam’s series, details were not provided
and patients with ‘abnormal development’ were not included, so the patients with intellectual disability, classified here among the patients with cognitive impairment, are
not comparable).
a
Among symptomatic patients.
b
Among patients with cognitive impairment.
c
Among patients with psychiatric signs.
d
Among patients with movement disorders.

Movement disorders included mostly akinetic-hypertonic and dyskinesia during the course of the disease, which was
syndrome (78.6%) with or without tremor. Motor improvement clinically indistinguishable from idiopathic Parkinson’s disease
with dopatherapy was noted in five cases (by interview of patients (Nicolas et al., 2013). For five other patients, no subjective
and their neurologist). An acute levodopa challenge was motor improvement was reported. For the remaining 12,
performed on two of these five patients: Unified Parkinson’s dopatherapy was not introduced, in most cases because
Disease Rating Scale III motor score improved by 25% Parkinsonism’s severity was mild. Hyperkinetic signs were
and 47%. The latter is the proband of the family with a less frequent and included chorea, dystonia, and orofacial
PDGFRB mutation, and presented ON/OFF fluctuations dyskinesia.
Phenotype of idiopathic basal ganglia calcification Brain 2013: 136; 3395–3407 | 3401

Gait (43.1%) and speech (39.2%) disorders were frequent and matter (P = 0.0189), the mean scores of the thalamus and the
of multiple origin, mostly due to akinetic-hypertonic syndrome and cerebellar hemispheres approaching significance (P = 0.0916 and
cerebellar signs. Pyramidal signs were present in 23.5% of symp- P = 0.06934, respectively). Cortical calcifications were located in
tomatic patients, associated only once with paresis. Four patients the occipital cortex and/or in the depth of the sulci (Fig. 3).
presented seizures, ranging from one episode to several seizures Interestingly, no patient bearing a PDGFRB mutation presented
but none presented hard-to-treat epilepsy. cortical calcification, contrary to 77% of the patients with an
The most frequently associated signs were: cognitive impairment SLC20A2 mutation (Fig. 1B). This was also the case of the
with movement disorder (20 patients), cognitive impairment with vermis calcifications, which were present in 77% of the
psychiatric signs (17 patients), cognitive impairment with gait SLC20A2 and 0% of the PDGFRB mutation carriers (Fig. 1B).
disorder (17 patients) and movement disorder with gait disorder Examining the cerebellum more closely, we noted that cerebellar
(16 patients). calcifications in the PDGFRB mutation carriers were restricted to or
Clinical onset appeared to be earlier when associated with centred on the dentate nuclei. However, cerebellar calcification of
PDGFRB mutations, compared with SLC20A2, but the difference the SLC20A2 mutation carriers seemed to be restricted to the
was not statistically significant. Seventy per cent of the SCL20A2 dentate nuclei only in one case, and more widespread in the
patients versus 40% of the PDGFRB patients were symptomatic, others. The brainstem was rarely involved; calcification was seen
but the almost significant older age of evaluation of the in the midbrain of seven patients, including one SLC20A2 muta-

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

SLC20A2 patients (P = 0.095 for symptomatic and P = 0.071 for tion carrier with faint bilateral calcifications in the upper part of
asymptomatic patients) could account at least for a part of the the midbrain, and another with similar appearance and no genetic
difference. cause. The other five (three siblings and two unrelated patients)
No patient with a SLC20A2 or a PDGFRB mutation experienced did not carry a mutation in either of the two genes and were the
seizures, and none of the SLC20A2 mutation carriers presented only patients showing calcification in the pons, in addition to all
cerebellar signs despite the presence of extensive cerebellar calci-
other locations, except for medulla (that was calcified in only one
fications. Movement disorders seemed more common in SLC20A2
of these five patients). The medulla was also calcified in an unre-
mutation carriers.
lated patient devoid of other brainstem calcification (unknown
Migraine, with or without aura, was reported in 15 patients in
genetic cause).
our study (nine classified among the symptomatic patients, and six
Among the 48 symptomatic and the 20 asymptomatic patients
classified among the asymptomatic patients).
(Supplementary Table 2), the mean age at the time the CT scans
Interestingly, four patients diagnosed with probable IBGC
were performed was not significantly different (mean  SD:
presented cerebral haemorrhage(s), which led to the fortuitous
54.5  17.8 versus 47.3  18.1), and the mean age of clinical
discovery of BGC (Supplementary Note 2).
onset of symptomatic patients was significantly lower than the
mean age at the last clinical evaluation of asymptomatic patients
Brain imaging (one-sided t-test: P = 0.027). The TCS was significantly higher in
In patients with IBGC, inter-rater agreement was acceptable symptomatic compared to asymptomatic patients {mean [95%
(weighted kappa = 0.67). Calcification was most frequently confidence interval (CI)]: 33.6 (27.3–39.9) versus 14.5
found in (besides lenticular nuclei, 100%, by definition): caudate (8.3–20.7), P = 0.0007}. Similar results were observed pertaining
nuclei, cerebellar hemispheres, and cerebral white matter (each to the location of calcification excepting midbrain, pons, and me-
55%), thalami (43%), cortex (33%), vermis (32%), midbrain dulla, which were involved only in symptomatic patients but the
(10%), pons (7%) and medulla (3%) (Fig. 1 and Supplementary number of patients exhibiting such calcification was too small to
Table 2). Calcification of the internal capsules was not included in obtain significant results (Fig. 1C and Supplementary Table 2).
the figures and the Supplementary tables because in this location it
systematically involved adjacent location and consequently had to
forego scoring.
Among the patients with a definite diagnosis of PDGFRB-related Discussion
IBGC, one patient presented calcifications under the age-specific
This study is the first systematic description of the phenotype
threshold. In order to consider the effects of age and genotype on
associated with IBGC since the discovery of the first causative
TCS, we studied both definite and probable IBGC patients, and
genes. We found that patients with SLC20A2 mutations exhibited
excluded the aforementioned patient from the statistics. The TCS
more severe calcifications than those with PDGFRB mutations.
correlated positively with age (tau = 0.28, P = 0.0009). Using a
Moreover, a different pattern of calcifications appeared between
negative binomial regression for overdispersed count data, we
found that the mean TCS increased more rapidly in patients these two groups, the absence of cortical and vermis calcification
than control subjects significantly (P 5 0.0001, Fig. 2A). in PDGFRB mutation carriers having to be confirmed in further
Patients with an SLC20A2 mutation had a greater expected TCS studies. Clinical and radiological diversity is confirmed, even
than those with a PDGFRB mutation, with respect to age at CT among patients carrying the same mutation. Importantly, the se-
scan (P = 0.0004) (Fig. 2B). In particular, the mean scores of verity of the calcifications correlated positively with age, and a
the following locations were greater in the SLC20A2 group: len- more severe TCS was found in symptomatic versus asymptomatic
ticular (P = 0.0111), caudate (P = 0.0105) and subcortical white individuals.
3402 | Brain 2013: 136; 3395–3407 G. Nicolas et al.

A 60

PDGFRB (n=
(n 14)

SLC20A2 (n= 9)
50
Unknown (n= 46)

40

Score

30

20

10

0
TCS (/80) Lencular Caudate (/10) Thalamus Subcorcal Cortex (/5) Cerebellar Vermis (/5) Midbrain Pons (/5) Medulla (/5)

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

(/10) (/10) white maer hemisphere (/10)
(/10) (/10)

B % 100%
PDGFRB (n=14)
90
SLC20A2 (n=9)

80 Unkwnown (n= 46)

70

60

50

40

30

20

10

0
LLencular
l Caudate (/10) Thalamus
l (/10) Subcorcal
b l Cortex (/5) Cerebellar
ll Vermis
i (/5) Midbrain
i (/10) Pons (/5) Medulla
ll (/5)
(/10) white maer hemisphere
(/10) (/10)

C 60
Symptomac (n=48)
Asymptomac (n=20)

50

*
40

30
Score

20

*
10 *
* *
*
* *

0
TCS (/80) Lencular Caudate Thalamus Subcorcal Cortex (/5) Cerebellar Vermis (/5) Midbrain Pons (/5) Medulla (/5)
(/10) (/10) (/10) white maer hemisphere (/10)
(/10) (/10)

Figure 1 Calcification scores of patients with IBGC. Mean scores (95% CI) (A) and percentage of IBGC patients (B) affected by a
calcification in each location (score 4 0), among the patients with a mutation in PDGFRB, SLC20A2 or with unknown genetic cause.
Mean scores (95% CI) of symptomatic and asymptomatic patients with IBGC (C, *P 5 0.05).
Phenotype of idiopathic basal ganglia calcification Brain 2013: 136; 3395–3407 | 3403

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

Figure 2 Total calcification scores. (A) Definite (PDGFRB and SLC20A2) and probable (unknown genetic cause) IBGC patients and
control subjects (white squares). Positive correlation is seen in each group and the TCS increase more rapidly in patients with IBGC than
controls (P 5 0.0001). (B) Patients with IBCG by genotype. The curves represent the TCS predictions under a negative binomial model
accounting for age and genotype for PDGFRB and SLC20A2. Grey areas represent 95% CI for the true expected TCS conditional on age
and IBGC versus control status (A) or genotype (B). Note that these indicate the statistical confidence in the estimated mean effect; they
do not encompass the population variation of observed TCS around their expectation. In B, the light grey zone is related to the SLC20A2
mutation status and the dark grey zone to the PDGFRB mutation status. Under this model, patients with a mutation in SLC20A2 exhibit
significantly more severe calcifications than those with a mutation in PDGFRB (P = 0.0004).

Clinical phenotype and mood disorders, including bipolar disorder (Trautner et al.,
1988; Konig, 1989; Lopez-Villegas et al., 1996). More generally,
Three categories of signs were present in more than half of the BGC were significantly associated with psychotic symptoms in a
symptomatic individuals: cognitive impairment, psychiatric signs population of non-demented 85-year-old individuals, but not with
and movement disorders. These signs were already predominant anxiety or mood disorders (Ostling et al., 2003). In another study, 4
in the previous case series (Manyam et al., 2001), but with an of 18 patients with BGC had mood disorders (bipolar disorder and
apparent lesser proportion of symptomatic patients exhibiting cog- depression) and six had obsessive-compulsive disorder (Lopez-
nitive (39%) and psychiatric signs (31%) compared with ours. Villegas et al., 1996). In our case series, mood disorder and psych-
Cognitive impairment is a known symptom of IBGC and other otic symptoms were the most frequent psychiatric signs, although
diseases affecting the basal ganglia (Sobrido et al., 1993–2004). we might have overestimated the proportion of mood disorders
In IBGC, some major cortico-subcortical loops may be disrupted, veritably due to IBGC, as these syndromes—especially depres-
which is consistent with the subcortical memory impairments and sion—are frequent in the general population. Nevertheless, it
dysexecutive syndromes seen in some of our patients and in other seems that patients with IBGC of each genetic cause and with no
studies (Lopez-Villegas et al., 1996). Conversely, some patients known cause may all exhibit mood disorders or psychotic signs. Of
exhibit cortical calcification, in particular within the occipital lobe. note, obsessive-compulsive disorder seems to be a rare manifest-
Other neuropsychological tests focused on visuoconstructive func- ation of IBGC and has been reported in only one SLC20A2 mutation
tions would be of interest to explore their impact. We also found carrier to our knowledge (Hsu et al., 2013).
three patients with intellectual disability in our case series, includ- As in the series by Manyam et al. (2001), akinetic-
ing one with an SLC20A2 mutation, as previously reported (Wang hypertonic syndrome was the most frequent movement disorder,
et al., 2012). We note that intellectual disability cannot be com- whereas tremor and other hyperkinetic movements were less
pared to the study by Manyam et al. (2001), in which patients common.
with ‘abnormal development’ were not included. Until now, there have been only two studies published with
Psychiatric signs seemed to be more frequent in our series than in phenotypic data about SLC20A2 mutation carriers (Wang et al.,
the previously published one (Manyam et al., 2001), despite the 2012; Hsu et al., 2013). It is interesting to note that seizures,
fact that our patients were referred mostly by neurologists. Large which were not reported in our SLC20A2 mutation carriers,
diverse psychiatric signs have been associated with IBGC, such as were described in only two children, who were also affected by
psychotic signs (Geschwind et al., 1999; Le Ber et al., 2007; Nicolas intellectual disability/developmental delay (Wang et al., 2012).
et al., 2012), personality disorders, obsessive-compulsive disorder, Clinical onset of other patients was primarily during adulthood.
3404 | Brain 2013: 136; 3395–3407 G. Nicolas et al.

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

Figure 3 Cortical calcifications on the CT scans of the patients with SLC20A2 mutations, mainly occipital (black arrows), and/or in the
depth of the sulci (white arrows): Patient 1, Family 752 (A and B); Patient 2, Family 752 (C); Patient 1, Family 1038 (D); Case 1352
(E); Case 485 (F and G); Case 424 (H); Case 413 (I and J). Absence of cortical calcification in two PDGFRB patients (K and L, Family 870)
despite severe calcification of the basal ganglia (severe calcification of both lenticular nuclei, confluent with severe calcification of both
caudate nuclei (upstream of these slides), in K and L, and severe calcification of both thalami in K, faint in L).

These children carried a maternally-inherited missense variant, in In the present study, few clinical differences appeared between
addition to the paternally-inherited loss-of-function mutation, PDGFRB and SLC20A2 mutation carriers, which could be con-
leading to the hypothesis that a biallelic contribution could explain firmed in further studies using larger populations. Few patients
a more severe and an earlier phenotype (Wang et al., 2012). We are reported with PDGFRB mutations. Because of that and of
found that one third of our patients who were affected by an the high diversity in age of clinical onset in both genetic groups,
SLC20A2 mutation were asymptomatic. Publicly available data they cannot systematically be compared in terms of ages. The
suggest great intra-familial and inter-familial diversity for proportion of symptomatic individuals in both groups may be in-
SLC20A2 mutations. For example, in the Family BJ-IBGC (Wang terpreted carefully, since the majority of PDGFRB patients
et al., 2012), only one of nine mutation carrier exhibited one of belonged to one family and the mean age of last evaluation of
the symptoms listed by Manyam et al. (2001), whereas the nine PDGFRB patients tended to be lower than SLC20A2 patients.
mutation carriers of the American Family F1 exhibited at least one Migraine or cephalalgia is sometimes reported in IBGC cases and
symptom, particularly among the movement disorders (Geschwind family reports. Caution is needed concerning the exact prevalence
et al., 1999; Hsu et al., 2013). of cephalagia among patients with IBGC, because this symptom
Phenotype of idiopathic basal ganglia calcification Brain 2013: 136; 3395–3407 | 3405

was not systematically investigated in previous reports, and be- epidemiological study with exploration of the other causes of
cause of the large prevalence of migraine and cephalalgia in the BGC, and the absence of threshold-value to exclude calcification
general population. It was not reported in the previous case series which may be related to normal aging. Our criteria of probable
(Manyam et al., 2001), and it was present in four of the nine IBGC excluded the patients with a TCS under the age-specific
patients in the SD-IBGC family (SCL20A2 mutation) (Dai et al., threshold in order to exclude phenocopies. Such a criterion was
2010). Migraine was reported as a predominant sign in 15 of our stringent. Besides this, the age-specific radiological penetrance is
patients. Nevertheless, in the family affected by a PDGFRB muta- unknown, and one 46-year-old patient with a PDGFRB mutation
tion, migraine did not segregate with IBGC. It would be necessary presented a TCS under the age-specific threshold. This finding
to perform a systematic case control study about cephalalgia in may support an incomplete radiological penetrance for calcifica-
IBGC to address this issue. tions related to a PDGFRB mutation. Nevertheless, in this patient,
calcification was seen in more than just the lenticular nuclei (cere-
bellum). As a consequence, this stringent threshold, designed to
Imaging
select the probands, must be used with care when studying the
This study allows for the identification of differences in brain cal- segregation of genetic variants of unknown penetrance in
cification between individuals with a PDGFRB, an SLC20A2, and pedigrees.
no known mutation. In particular, the expected TCS was higher in As expected, though never before demonstrated, the TCS cor-

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

SLC20A2 mutation carriers, with respect to age. No patient with a related positively with age in patients with IBGC, suggesting an
PDGFRB mutation exhibited cortical or vermis calcifications, con- increase of the severity of IBGC-related brain calcifications during
trary to a striking majority of the SLC20A2 mutation carriers. aging. To our knowledge, this question has not yet been ad-
Expression data as given for example by the microarray or in dressed with a longitudinal cohort study. Interestingly, an increase
situ hybridization analyses of human brains, the Allen brain atlas in the volume of lenticular calcifications during the follow-up
(Sunkin et al., 2013), are useful tools in the corroboration of these period was recently observed in 35 of 49 patients with hypopara-
findings (da Silva et al., 2013). Expression data about these two thyroidism (Goswami et al., 2012). In previous reports of IBGC,
genes in the occipital cortex are not consistent between the ana- brain calcifications were mostly described in terms of gross loca-
lysed brains and further studies would be necessary. Conversely, it tion and, rarely, quantified by volumetric studies, which might
seems that the cerebellar expression of PDGFRB is mainly re- have been biased by brain atrophy (Manyam et al., 2001).
stricted to the cerebellar nuclei, whereas the expression of Moreover, volumetric studies did not take into account the density
SLC20A2 seems to be more widespread in the cerebellum, includ- of the calcifications (Manyam et al., 2001; Goswami et al., 2012),
ing the vermis. This observation is in accordance with the location which seems particularly problematic to us for faint calcification.
and the extension of the calcifications identified in our series. Our visual rating scale is likely not to be affected by cerebral
Brainstem calcifications are rare. To our knowledge, only one pa- atrophy, is easily applicable and adaptable for the everyday use
tient with an SLC20A2 mutation and calcification in the brainstem by radiologists and neurologists, reflects the visual severity of
was reported, but the precise location was not mentioned (Hsu the calcifications and has satisfactory inter-rater agreement.
et al., 2013), and we only found one SLC20A2 mutation carrier Livingston et al. (2013) used a scale designed to classify calcifica-
with midbrain calcification. Five patients with probable IBGC had tions seen in CT scans of patients with diverse disorders,
impressive calcifications in the pons. The causative gene(s) of such among the following categories: spot, rock, blush and linear.
a pattern of calcification is unknown. It (they) might have a dif- This study provided for a noteworthy description of brain neuroi-
ferent pattern of brain expression compared to SLC20A2 and maging in these disorders (e.g. Aicardi-Goutieres syndrome),
PDGFRB, which seem to have relatively weak expression in this including neuroimaging features other than calcification alone.
area (Allen brain atlas). Another hypothesis would be that a causa- Among the patients with no known diagnosis, two families
tive gene might have a similar pattern of expression, but with (seven patients) with calcification resembling dominant Fahr’s dis-
biallelic mutations (recessive pattern of inheritance), which could ease were classified apart from the others in light of the calcifica-
explain a more severe imaging phenotype with calcifications in the tion patterns. We chose to elaborate our own scale designed
pons. specifically to describe calcification related to IBGC, and using a
The mean severity of the calcifications was higher in symptom- semi-quantitative method.
atic compared to asymptomatic individuals. Interestingly, none of We found that 15% of the control patients presented BGC, all
our SLC20A2 mutation carriers exhibited cerebellar signs, despite but one 440 years of age, suggesting that the presence of BGC in
the presence of severe cerebellar calcifications. Nonetheless, cere- an individual 540 years of age may not be considered as physio-
bellar syndrome has already been reported in SCL20A2 mutations logical. The first large studies, performed before the 1990s, had
carriers, even in patients without cerebellar calcifications (Hsu initially estimated the prevalence of BGC to be 0.6–0.8%
et al., 2013). Taken together, these data suggest that the severity (Murphy, 1979; Cohen et al., 1980; Harrington et al., 1981) to
of BGC, which are a mandatory criterion for IBGC diagnosis, helps 1.1% (Konig, 1989). The thickness of the slices and the quality of
predict the symptomatic status, but is not sufficient. Other bio- the CT scans, not comparable with current technology, could
markers would be necessary to better correlate with the clinical explain why punctate lenticular calcifications would have been
diversity of IBGC. largely underestimated. More recent studies had contradictive
The prevalence of IBGC is unknown. Three main reasons could results, probably due to diverse methods (Bordignon and Arruda,
explain this: the unknown clinical penetrance, the absence of 2002; Eskandary et al., 2005). Interestingly, a recent study found
3406 | Brain 2013: 136; 3395–3407 G. Nicolas et al.

a BGC frequency of 20.5% in unselected patients from Japan, and reinvestigated in the future, as suggested by our data. As in many
the older group (465 years) had a higher frequency than the other neurogenetic disorders, the clinical penetrance is incomplete,
younger (31% versus 13%) (Yamada et al., 2013). In accordance rendering both genetic counseling and future therapeutic trials
with our results and the previously mentioned work, the studies challenging.
focusing on older individuals showed that BGC is a common find-
ing on systematic brain CT scans (Ostling et al., 2003; Simoni
et al., 2008). Moreover, our study presents the first semi-quanti- Acknowledgements
tative data of brain calcification in an unselected series of controls,
We are grateful to Tracey Avequin for her help in editing the
with interesting results concerning the prevalence and proposed
manuscript and to Séverine Jourdain and Véronique Hannier for
age-specific thresholds, which have to be replicated in other
their technical help.
studies.
The French IBGC study group are: Patrick Ahtoy, Mathieu
Anheim, Jérôme Augustin, Xavier Ayrignac, Françoise Bille-Turc,
Genetics Dominique Campion, Boris Chaumette, Michel Clanet, Luc
We studied more patients with a mutation in PDGFRB (due to the Defebvre, Gilles Defer, Nathalie Derache, Mira Didic, Franck
identification of a large family) than in SLC20A2. Nevertheless, the Durif, Emmanuel Flamand-Roze, Guillaume Fromager, Maurice
Giroud, Alice Goldenberg, Olivier Guillin, Lucie Guyant-

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

latter gene may be more frequently involved in familial IBGC.
Indeed, a previous study reported SLC20A2 mutations in 12 of Maréchal, Didier Hannequin, Cécile Hubsch, Snejana Jurici, Pierre
29 (41%) of families with IBGC (Hsu et al., 2013). Considering Krystkowiak, Pierre Labauge, Antoine Layet, Isabelle Le Ber,
only families with IBGC, we identified one family with a PDGFRB Thibaud Lebouvier, Romain Lefaucheur, David Maltête, Olivier
mutation, two with an SCL20A2 mutation and 12 remained un- Martinaud Donald Morcamp, Gaël Nicolas, Ozlem Ozkul,
explained. If adding the cases with a sporadic presentation, we Jérémie Pariente, Cyril Pottier, Philippe Rondepierre, Olivier
found three probands with a PDGFRB mutation and six with an Rouaud, Brigitte Sallé, Mathilde Sauvée, Stéphane Schaeffer,
SLC20A2 mutation, where 29 did not have a genetic alteration in Christel Thauvin-Robinet, Catherine Thomas-Antérion, Christine
these two genes, underlining the need to pursue our effort to find Tranchant, Aude Triquenot, Yvan Vaschalde, Marc Vérin,
further causative genes. Christophe Verny, Marie Vidailhet and David Wallon.
Some patients with a sporadic presentation were classified
as having definite IBGC. It was not possible to ascertain the ab-
sence of heredity for all of them, because of the lack of brain
imaging in their first-degree relatives. Indeed, the absence of
neuropsychiatric signs in first-degree relatives by history is not Funding
sufficient to exclude IBGC, because of an incomplete clinical This work was supported by grants from the French Ministry of
penetrance. Health to the CNR-MAJ and sponsored by the University Hospital
of Rouen and has received funding from the program
Investissements d’avenir ANR-10-IAIHU-06.
Conclusion
The diagnosis of IBGC is based first on neuroradiological findings
and a significant proportion of individuals are asymptomatic. These Supplementary material
observations, together with the high frequency of fortuitous dis-
Supplementary material is available at Brain online.
covery of BGC of unknown significance on brain CT scans, made
IBGC, formerly called Fahr’s disease, controversial as a distinct
disorder with clinical significance. The ongoing identification of
the molecular bases of IBGC demonstrates the high value of References
IBGC as a patent phenotypical marker. It also confirms the genetic Baba Y, Broderick DF, Uitti RJ, Hutton ML, Wszolek ZK. Heredofamilial
heterogeneity of the disorder (Oliveira et al., 2004). The number brain calcinosis syndrome. Mayo Clin Proc 2005; 80: 641–51.
of affected individuals with definite IBGC and advanced age is still Bordignon KC, Arruda WO. CT scan findings in mild head trauma: a
insufficient to determine the exact clinical penetrance of the dis- series of 2,000 patients. Arq Neuropsiquiatr 2002; 60: 204–10.
Cohen CR, Duchesneau PM, Weinstein MA. Calcification of the basal
ease. We show here that the mean severity of the calcifications is
ganglia as visualized by computed tomography. Radiology 1980; 134:
higher in symptomatic versus asymptomatic individuals, but the 97–9.
quantification of the calcifications seems insufficient to predict da Silva RJ, Pereira IC, Oliveira JR. Analysis of gene expression pattern
the clinical severity. A next step will be to investigate asymptom- and neuroanatomical correlates for SLC20A2 (PiT-2) shows a molecu-
atic and symptomatic patients of advanced age with the same lar network with potential impact in idiopathic basal ganglia calcifica-
tion (“Fahr’s Disease”). J Mol Neurosci 2013; 50: 280–3.
mutation to identify other biomarkers of the clinical severity of
Dai X, Gao Y, Xu Z, Cui X, Liu J, Li Y, et al. Identification of a novel
the disease. Moreover, the radiological penetrance, thought to genetic locus on chromosome 8p21.1-q11.23 for idiopathic basal
be near 100% at the age of 50 (Sobrido et al., 1993–2004) be- ganglia calcification. Am J Med Genet B Neuropsychiatr Genet 2010;
cause of the neuroradiological definition of the disease, has to be 153B: 1305–10.
Phenotype of idiopathic basal ganglia calcification Brain 2013: 136; 3395–3407 | 3407

Delacour A. Ossification des capillaires du cerveau. Ann Med Psychol Murphy MJ. Clinical correlations of CT scan-detected calcifications of the
1850; 2: 458–61. basal ganglia. Ann Neurol 1979; 6: 507–11.
Eskandary H, Sabba M, Khajehpour F, Eskandari M. Incidental findings in Nicolas G, Guillin O, Borden A, Bioux S, Lefaucheur R, Hannequin D.
brain computed tomography scans of 3000 head trauma patients. Surg Psychosis revealing familial idiopathic basal ganglia calcification. Gen
Neurol 2005; 63: 550–3. discussion 3. Hosp Psychiatry 2012, in press [Epub ahead of print].
Geschwind DH, Loginov M, Stern JM. Identification of a locus on Nicolas G, Pottier C, Maltete D, Coutant S, Rovelet-Lecrux A, Legallic S,
chromosome 14q for idiopathic basal ganglia calcification (Fahr dis- et al. Mutation of the PDGFRB gene as a cause of idiopathic basal
ease). Am J Hum Genet 1999; 65: 764–72. ganglia calcification. Neurology 2013; 80: 181–7.
Godefroy O, Azouvi P, Robert P, Roussel M, LeGall D, Meulemans T. Oliveira JR, Spiteri E, Sobrido MJ, Hopfer S, Klepper J, Voit T, et al.
Dysexecutive syndrome: diagnostic criteria and validation study. Ann Genetic heterogeneity in familial idiopathic basal ganglia calcification
Neurol 2010; 68: 855–64. (Fahr disease). Neurology 2004; 63: 2165–7.
Godefroy O, Roussel M, Leclerc X, Leys D. Deficit of episodic memory: Ostling S, Andreasson LA, Skoog I. Basal ganglia calcification and psych-
anatomy and related patterns in stroke patients. Eur Neurol 2009; 61: otic symptoms in the very old. Int J Geriatr Psychiatry 2003; 18:
223–9. 983–7.
Goswami R, Sharma R, Sreenivas V, Gupta N, Ganapathy A, Das S. Sarazin M, Berr C, De Rotrou J, Fabrigoule C, Pasquier F, Legrain S, et al.
Prevalence and progression of basal ganglia calcification and its patho- Amnestic syndrome of the medial temporal type identifies prodromal
genic mechanism in patients with idiopathic hypoparathyroidism. Clin AD: a longitudinal study. Neurology 2007; 69: 1859–67.
Endocrinol (Oxf) 2012; 77: 200–6. Schottlaender LV, Mencacci N, Koepp M, Hanna M, Hardy J, Lees AJ,
Harrington MG, Macpherson P, McIntosh WB, Allam BF, Bone I. The et al. Interesting clinical features associated with mutations in the
significance of the incidental finding of basal ganglia calcification on SLC20A2 gene (abstract). Eur J Neurol 2012; 19 (Suppl 1): 22–89.

Downloaded from http://brain.oxfordjournals.org/ at Dalhousie University on June 30, 2014

computed tomography. J Neurol Neurosurg Psychiatry 1981; 44: Simoni M, Pantoni L, Pracucci G, Palmertz B, Guo X, Gustafson D, et al.
1168–70. Prevalence of CT-detected cerebral abnormalities in an elderly Swedish
Hsu SC, Sears RL, Lemos RR, Quintans B, Huang A, Spiteri E, et al. population sample. Acta Neurol Scand 2008; 118: 260–7.
Mutations in SLC20A2 are a major cause of familial idiopathic basal Sobrido MJ, Hopfer S, Geschwind DH. Familial idiopathic basal ganglia
ganglia calcification. Neurogenetics 2013; 14: 11–22. calcification. In: Pagon RA, Bird TD, Dolan CR, editors.
Konig P. Psychopathological alterations in cases of symmetrical basal GeneReviewsTM [Internet]. Seattle, WA: University of Washington;
ganglia sclerosis. Biol Psychiatry 1989; 25: 459–68. 1993–2004.
Kostic VS, Lukic-Jecmenica M, Novakovic I, Dobricic V, Brajkovic L, Sunkin SM, Ng L, Lau C, Dolbeare T, Gilbert TL, Thompson CL, et al.
Krajinovic M, et al. Exclusion of linkage to chromosomes 14q, 2q37 Allen Brain Atlas: an integrated spatio-temporal portal for exploring
and 8p21.1-q11.23 in a Serbian family with idiopathic basal ganglia the central nervous system. Nucleic Acids Res 2013; 41: D996–1008.
calcification. J Neurol 2011; 258: 1637–42. Trautner RJ, Cummings JL, Read SL, Benson DF. Idiopathic basal ganglia
Le Ber I, Marie RM, Chabot B, Lalevee C, Defer GL. Neuropsychological calcification and organic mood disorder. Am J Psychiatry 1988; 145:
and 18FDG-PET studies in a family with idiopathic basal ganglia 350–3.
calcifications. J Neurol Sci 2007; 258: 115–22. Ver Hoef JM, Boveng PL. Quasi-Poisson vs. negative binomial regression:
Lemos R, Oliveira M, Oliveira J. Reporting a new mutation at the how should we model overdispersed count data? Ecology 2007; 88:
SLC20A2 gene in familial idiopathic basal ganglia calcification. Eur J 2766–72.
Neurol 2013; 20: e43–4. Wang C, Li Y, Shi L, Ren J, Patti M, Wang T, et al. Mutations in
Liu X, Jian X, Boerwinkle E. dbNSFP: a lightweight database of human SLC20A2 link familial idiopathic basal ganglia calcification with phos-
nonsynonymous SNPs and their functional predictions. Hum Mutat phate homeostasis. Nat Genet 2012; 44: 254–6.
2011; 32: 894–9. Yamada M, Asano T, Okamoto K, Hayashi Y, Kanematsu M, Hoshi H,
Livingston JH, Stivaros S, van der Knaap MS, Crow YJ. Recognizable et al. High frequency of calcification in basal ganglia on brain com-
phenotypes associated with intracranial calcification. Dev Med Child puted tomography images in Japanese older adults. Geriatr Gerontol
Neurol 2013; 55: 46–57. Int 2013; 13: 706–10.
Lopez-Villegas D, Kulisevsky J, Deus J, Junque C, Pujol J, Guardia E, et al. Yamada N, Hayashi T. Asymptomatic familial basal ganglia calcification
Neuropsychological alterations in patients with computed tomography- with autosomal dominant inheritance: a family report [Article in
detected basal ganglia calcification. Arch Neurol 1996; 53: 251–6. Japanese]. No To Hattatsu 2000; 32: 515–9.
Manyam BV. What is and what is not ‘Fahr’s disease’. Parkinsonism Zeileis A, Kleiber C, Jackman S. Regression models for count data in R.
Relat Disord 2005; 11: 73–80. J Stat Softw 2008; 27: 1–25.
Manyam BV, Walters AS, Narla KR. Bilateral striopallidodentate calcino- Zhang Y, Guo X, Wu A. Association between a novel mutation in
sis: clinical characteristics of patients seen in a registry. Mov Disord SLC20A2 and familial idiopathic basal ganglia calcification. PLoS One
2001; 16: 258–64. 2013; 8: e57060.