Rheumatology 2000;39:l366–l373

Eridence of cerebral hypoperfusion in

scleroderma patients
M. Cutolo, F. Nobili1, A. Sulli, C. Pizzorni, M. Briata, F. Faelli,
P. Vitali1, G. Mariani2, F. Copello3, B. Seriolo, C. Barone and
G. Rodriguez1
Diririon of Rheumatology, lClinical Neuvophyriology kevrice, D
2 o
Nucleav Medicine kevrice, Depavtment of Intevnal Medicine, Unirevrity w
of Genora and 3Depavtment of 0ccupational Medicine, k. Mavtino nl
oa
Horpital, Genora, Italy de
d
fro
Abstract m
htt
0bjectirer. To investigate regional cerebral blood flow by 99mTc-hexamethylpropylenamine- ps
oxime (HMPAO) single photon emission computed tomography (SPECT) in a series of ://
40 patients (mean age 58.5±ll.5 yr) affected by systemic sclerosis (SSc) in comparison with ac
age-matched healthy controls. ad
e
Methodr. Subjects affected by concomitant severe pathologies that might interfere with the mi
interpretation of the SPECT results were excluded. SPECT findings were correlated with the c.
severity of peripheral microvascular involvement, as assessed by nailfold ou
videocapillaroscopy (NVC). Whenever possible, patients underwent magnetic resonance p.
imaging (MRI) of the brain. co
Rerultr. Twenty-one SSc patients (52%) showed hypoperfusion in two or more regions of m/
rh
interest (ROIs) at the SPECT analysis. MRI was available in l4 of these patients, and was eu
shown to be altered in eight of them (57%). One patient with both abnormal SPECT and m
abnormal MRI was affected by mild cognitive impairment. Transcranial Doppler sonography at
was normal in all but one of these patients with hypoperfusion. Nineteen patients exhibited a ol
normal brain SPECT scan, but the MRI was shown to be altered in 3/l2 of them (25%). No og
y/
significant differences were found between the group of SSc patients showing hypoperfusion
art
and those showing a normal SPECT scan regarding age, the duration of disease, the icl
presence of vascular risk factors or damage of other organs typically involved in the disease, e/
and the severity of peripheral microvascular involvement (NVC). 39
Conclurionr. Focal or diffuse cerebral hypoperfusion was found in more than half of the /1
neurologically asymptomatic SSc patients studied, paralleling the incidence of altered brain 2/
13
MRI. The hypoperfusion was not linked to ageing and possibly reflects the cerebral location 66
of the microangiopathic process characterizing the disease. /1
78
kEY WORDS: Single photon emission computed tomography (SPECT), Magnetic resonance
42
imaging (MRI), Videocapillaroscopy, Systemic sclerosis, Cerebral blood flow. 08
by
gu
es
Systemic sclerosis (SSc) is a diffuse connective extracellular matrix leads to thickening of the skin and t
tissue disease characterized by proliferation of the on
to interstitial fibrosis of the organs, the severity of 27
vascular tissue, obliterative microvascular lesions and which is generally correlated [4]. The microangiopathy
diffuse interstitial fibrosis. Several organs are therefore seems to be the main pathogenic marker of
affected, with a severity of involvement that is often the disease, although macrovascular lesions have also
heterogeneous [l]. Current theories hold that the been reported [5]. The peripheral nervous system is
vascular lesions are the consequence of endothelial often involved during SSc, usually in the form of
damage; however, the early mechanisms remain to mononeuro- pathy or multiple neuropathy. The
be established [2, 3]. With increasing severity of the involvement of the brain vessels has been reported
disease, the accumulation of less frequently and it has usually been thought to be
related to concomitant arterial hypertension, renal
Submitted l3 January 2000; revised version accepted 27 June failure or other systemic complications of the disease
2000. Correspondence to: M. Cutolo, Division of Rheumatology, [6].
Department of Internal Medicine, Viale Benedetto XV No. 6, However, beyond the rare cases of acute vasculitis
l6l32 of the major cerebral vessels, there is increasing
Genova, Italy.
evidence of
1366 © 2000 British Society for Rheumatology
 Cerebral hypoperfusion in SSc patients 1367

SSc patients suffering from central nervous system crystal [l7]. The sensitivity of the cylindrical-shaped,
(CNS) involvement, most often in the form of transient low-energy, high-resolution lead collimator was
ischaemic attacks and cognitive impairment [5, 7, 8– l90 c.p.s./MBq (7.0 c.p.s./Ci) with a point source in air,
l2]. These events are not clearly related to common with a spatial resolution at l40 keV of 99mTc <8.5 mm
vascular risk factors or to the failure of other organs at the centre of rotation and 6.3 mm in the peripheral
or systems. Recently, calcinosis of the small vessels regions (full width half maximum).
in several brain areas, including the cerebral cortex, SPECT acquisitions were performed 30–90 min
has been reported in two autopsy SSc cases [l3]. after the i.v. injection of 740–925 MBq of freshly pre-
Therefore, the hypothesis may be advanced that pared 99mTc-HMPAO (Ceretec, Amersham Medical,
vascular involvement in the brain occurs to some Amersham, Uk). Sensory input was minimized whilst
extent in SSc patients, in the form of macro- or the tracer was injected in a quiet, dimly lit room, the D
microangio- pathy or both [5, l0]. We therefore patient lying on a reclining chair (eyes closed and o
w
investigated regional cerebral blood flow (rCBF) in a ears unplugged). nl
large number of SSc patients by 99mTc- Sixty-four axial slices parallel to the anterior– oa
hexamethylpropylene amine oxime (HMPAO) single posterior commissure (AC–PC) line, l.67 mm thick, de
photon emission computed tomo- graphy (SPECT). In were recon- structed on a l28×l28 matrix (l pixel=l.67 d
addition, the SPECT findings were correlated with the mm) using a 2-dimensional Butterworth-filtered back- fro
m
severity of the peripheral micro- vascular involvement, projection (cut-off 0.80 cm, order l0), according to the htt
as assessed by nailfold videocapil- laroscopy (NVC) method described by Minoshima et al. [l8] and ps
and, whenever possible, by magnetic resonance corrected for attenuation with Chang’s first-order ://
imaging (MRI) of the brain. method, using an attenuation coefficient of 0.l5/cm ac
[l9]. SPECT data were analysed on two transaxial ad
sections (8.3 mm thick: five slices summed), as e
mi
Patients and methods identified with the aid of a widely used anatomical c.
atlas [20]. The first (lower) section was constructed by ou
Patientr summing the five most intermediate of all the slices p.
During a period of 3 yr, 47 consecutive patients (46 parallel to the AC–PC line, which included the thalami co
(Fig. l, left). The second (upper) section was m/
women, one man, mean age 58.5±l2 yr, range 32– rh
77 yr) were recruited. The patients had been referred constructed by summing the first five transaxial slices eu
to the Division of Rheumatology of the Department of tangential to the lower border of the cingulate gyrus m
Internal Medicine of the University of Genova, and (Fig. l, right), as identified in the coronal section at the at
had SSc according to the American Rheumatism level of the thalami. In each section, the cortical ol
ribbon was automatically delimited similarly, as og
Association criteria [l4]. Patients with either limited or y/
diffuse cutaneous involvement were considered for suggested by Mountz et al. [2l], and was then divided
art
inclusion in the study [4]. into twelve 30° sectors to identify twelve regions of icl
Clinical and pathological conditions associated with interest (ROIs) in each section. For each hemisphere, e/
SSc were investigated. Exclusion criteria included there were two frontal, three temporal and one 39
severe or uncontrolled arterial hypertension; occipital area in the lower section, and three frontal, /1
two parietal and one occipital area in the upper 2/
uncontrolled diabetes mellitus; relevant renal, 13
respiratory or hepatic failure; and severe anaemia. section. Irregular but symmetrical ROIs were hand- 66
Seven patients matched one or more of these criteria drawn around the thalami and the basal ganglia in the /1
and therefore were excluded. lower section (Fig. l, left). Therefore, 28 ROIs were 78
The remaining 40 patients (39 women and one analysed in each patient: l0 frontal, six temporal, four 42
parietal and four occipital areas, the two thalami and 08
man, mean age 58.7±ll.5 yr, range 32–75 yr) by
entered the study after they had given informed the two basal ganglia. gu
consent. The dura- tion of the disease was derived Counts in the ROIs were normalized by computing es
from the clinical medical history and an interview. The ROI/cerebellar ratios (mean count per pixel in each t
major clinical features of these patients are shown in ROI divided by mean count per pixel in the whole on
cerebellum). If cerebellar asymmetry was greater than 27
Table l. Twenty-one patients in this series had
already been included in a previous study that used l0%, the cerebellar hemisphere with the higher count
the planar xenon-l33 method [l5, l6]. Twenty per was considered as the reference region.
cent of the patients included in the study showed 0thev examinationr
associated mild hypertension (diastolic blood
pressure <l05 mmHg), slight hyper- All patients were scheduled to undergo MRI of the
cholesterolaemia (total serum cholesterol <250 mg/dl) brain, which was performed with a 0.5-Tesla super-
or low-degree anaemia (Hb level <l2 g/l). conductive apparatus (MR 5000; Esaote Biomedica,
Genova, Italy) with the following parameters: bicom-
missural paraxial planes; DP—(TR/TE=2000/30 ms),
Methodr Tl-weighted (660/20) and T2-weighted
99m
Tc-HMPA0 kPECT. The SPECT equipment we (2500/l20) sequences. The Tl-weighted sequences
used (Ceraspect; Digital Scintigraphics, Waltham, were also per- formed after i.v. injection of 0.l5
Massachusetts, USA) acquired cerebral perfusion mmol/kg of gado- linium-diethylene-
images of 99mTc-HMPAO by means of an annular tetraaminopentaacetate (Gd-DTPA)
1368 M. Cutolo et al.

D
o
w
nl
oa
de
d
fro
m
htt
ps
://
ac
ad
e
mi
c.
ou
p.
co
FIG. l. 99mTc-HMPAO brain SPECT images. ROI drawings in the superficial cortex and (left) around the thalami and the basal m/
ganglia on two transaxial sections parallel to the AC–PC line. (Left) Lower section, crossing the thalami. The deep nuclei rh
ROIs are hand-drawn. (Right) Upper section, tangential to the lower border of the cingulate gyrus. In both sections the ROIs eu
around the cortical ribbon have been drawn automatically. The common visual scale is shown in the rectangle below; m
decreasing flow values are indicated by the range of tones from right to left. L, left cerebral hemisphere. at
ol
og
(Magnesvit; Schering, Berlin, Germany). MRI was capillary loops and the loss of capillaries are the most y/
performed within a l-month period after SPECT useful features for the diagnosis and follow-up of art
examination. micro- angiopathy [22]. Therefore, patients were icl
Duplex scanning of the neck vessels was classified as showing an early (E), active (A) or late e/
performed in all patients, and transcranial Doppler 39
(L) pattern, according to recently validated criteria /1
sonography (TCD) was performed in patients with an [23]. These criteria include the following: (early 2/
abnormal SPECT scan within l month after the pattern) few giant capillaries, few capillary 13
abnormal scan. TCD was performed in a supine haemorrhages, no evident loss of capillaries; (active 66
position with a 2 MHz probe through the temporal pattern) frequent giant capillaries, frequent capillary /1
78
windows of either side and the occipital window. The haemorrhages, moderate loss of capillaries, some 42
Doppler signal was subjected to automatic on-line avascular areas; (late pattern) irregular enlargement 08
fast Fourier transform with a Multi-Dop X device of the capillaries, few or absent giant capillaries, by
(German Vasculab, Sipplingen, Germany). All absence of haemorrhages, severe loss of capillaries gu
vessels of the Willis circle were searched for and with extensive avascular areas, ramified/ bushy es
capillaries. t
insonated at various depths whenever possible. NVC on
was carried out with a cold light videocapillaro- scope 27
and analogue image analysis software (Videocap; DS Contvolr
Medigroup, Milan, Italy). Each subject had been
Twenty-two subjects (l6 females, six males, mean age
indoors for a minimum of l5 min before the nailfold
6l±8.7 yr, range 44–84 yr) out of 52 volunteers were
was examined and the room temperature was 20–
enrolled as controls on the basis of criteria that com-
22°C. The nailfolds of all l0 fingers were examined prised normal values for serum glucose, creatinine,
in each patient, after a drop of immersion oil had blood urea nitrogen, complete blood count, and
been placed on the nailfold bed to improve urinalysis. Mild hypertension and
resolution. Fingers affected by recent local trauma hypercholesterolaemia were accepted in 22% of the
were not analysed. The NVC examination was subjects. Previous or present neurological,
performed by the same opera- tors, without psychiatric, metabolic or severe cardio- vascular
knowledge of the patient’s clinical condi- tion and disorders and the use of drugs other than anti-
characteristics. During routine clinical NVC analysis, hypertensive agents were exclusion criteria. The
changes in the shape and arrangement of the controls were informed about the aim of the study and