Clinical Correlate of Brain SPECT Perfusion Abnormalities in Fibromyalgia

Patients and Control Subjects

Twenty women with fibromyalgia (mean age ± SD, 48.4 ± 10.6 y [range, 25–63 y]) who fulfilled the American College of Rheumatology criteria (9) were consecutively enrolled in the study. All patients underwent a general medical assessment by 1 investigator to confirm the diagnosis. Exclusion criteria included pregnancy and concomitant major medical or psychiatric disease. No patient was receiving treatment with a strong opioid or N-methyl-d-aspartate receptor antagonist. Data on patients and healthy controls are presented in Table 1.

Pain intensity was evaluated by a Visual Analog Scale (VAS), the Questionnaire Douleur de Saint-Antoine scale (QDSA; the validated French version of the McGill Pain Questionnaire) (10), and the Tubingen Pain Behavior Scale (TBS) (11). QDSA measures the different components of the subjective pain experience, describing the sensory and the affective dimensions. The higher the QDSA score is, the worse is the condition. The TBS, evaluated by 1 investigator, rates the presence of 11 pain behaviors (e.g., groaning, slowed movements, pain-based refusal of activities) on a 0–2 scale (0, none; 1, sometimes; and 2, always).

Disability was evaluated using the Fibromyalgia Impact Questionnaire (FIQ). The FIQ is designed to measure the severity of fibromyalgia disease symptoms. It is a standardized, self-administered instrument that consists of visual analog scales and questions regarding limitations on activities of daily living. The FIQ evaluates the components of health status that are believed to be most affected by fibromyalgia. This scale comprises 20 items that assess 10 subheadings. The first 11 items make up a physical impairment scale. Each item is scored on a 4-point rating scale. Items 12 and 13 ask patients to mark the number of days they felt well and the number of days they were unable to work because of fibromyalgia symptoms. Items 14–20 are 10-cm visual analog scales, marked in 1-cm increments, on which the patient rates occupational difficulty, pain, fatigue, morning tiredness, stiffness, anxiety, and depression. Questions 1–13 are normalized to 10 points, and the total score is added to the number of points from the remaining questions. The total score of the FIQ ranges from 0 to 100, and the higher the score is, the worse is the condition (i.e., the more serious the disease).

Anxiety and depression were evaluated by the Hospital Anxiety and Depression (HAD) scale score (12). HAD is a self-screening questionnaire that consists of 14 questions, 7 for anxiety and 7 for depression. A higher score means a worse condition.

For comparison of imaging findings, 10 healthy control women without any pain, who were similar in age, were also included (52 ± 7 y; t test > 0.1; HAD subscales < 8). All of the subjects provided informed consent according to institutional guidelines.

SPECT Protocol and Statistical Analysis

Patients received an injection of 740 MBq of ^99mTc-ethylcysteinate dimer (Neurolite; BMS) and were kept at rest for 1 h, in quiet surroundings, with their eyes closed. SPECT was performed using a double-head rotating γ-camera (ECAM; Siemens) equipped with a fanbeam collimator. Data were collected in 64 projections (40-s spread) through 360°. Tomographic 3-dimensional reconstructions were performed using a filtered backprojection algorithm and Chang attenuation correction.

A voxel-by-voxel group correlation study was performed using SPM2 (Welcome Department of Cognitive Neurology, University College, London) running on MATLAB (version 6.0; MathWorks Inc.). Images were initially converted from the DICOM to the Analyze format using MRIcro (available at www.mricro.com) and transferred to SPM2. The data were then standardized on the Montreal Neurologic Institute atlas by using a 12-parameter affine transformation, followed by nonlinear transformations and trilinear interpolation. The dimensions of the resulting voxels were 2 × 2 × 2 mm. Standardized data were then smoothed by a gaussian filter (full width at half maximum, 12 mm). Correlation studies between brain perfusion and VAS for pain, FIQ total score, QDSA affective and sensory scores, and HAD total scale score and subscores were performed using a single-subject, covariates-only statistical model. Age was considered as a nuisance variable to avoid the specific effect of age on rCBF. Other clinical variables were not included as covariate variables because they were thought to characterize the disease.

Positive and negative correlations were determined. Global normalization was performed using proportional scaling. The SPM t maps were initially obtained at a height threshold of P less than 0.005, corrected for multiple comparisons for the cluster (P < 0.05). Small-volume correction was additionally used to study the clinical correlation with subcortical structures, as previously described (4). Clusters of correlation were compared with those of significant hypo- or hyperperfusion, obtained using the compare-populations 1-scan-per-subject routine between the group with fibromyalgia and the control group, which carries out a fixed-effects, simple t test for each voxel (P < 0.005, corrected for multiple comparisons for the cluster, with age as the nuisance variable). Montreal Neurologic Institute coordinates were then converted into Talairach coordinates using the Talairach Daemon server (available at http://ric.uthscsa.edu/projects/talairachdaemon.html). Results were listed by clusters, and Brodmann area (BA), k value, t score, and Talairach coordinates of peak voxels were provided for each cluster. The k value represents the number of significant voxels in the particular cluster. Spearman correlation was used to study the relationship between extracted clusters of perfusion and clinical variables. The t test was used to compare means between subgroups, and F test was used to compare variance between subgroups.

Clinical Characteristics of Subjects

According to the inclusion criteria, healthy subjects had no pain, and their HAD subscales were less than 8. By contrast, patients with fibromyalgia exhibited pain, disability, and anxiety or depression symptoms. Clinical characteristics are presented in Table 1. In particular, HAD total score and subscores were higher in patients with fibromyalgia than in healthy subjects (P < 0.0005).

Brain SPECT Comparison Between Patients and Control Subjects

Compared with healthy controls, patients with fibromyalgia exhibited posterior hyperperfusion, including of the somatosensory cortex, and hypoperfusion of the frontal, cingulate, temporal, and cerebellar cortices (Fig. 1). In particular, the temporal hypoperfusion included the polar and mediobasal cortices.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Anatomic localization of peaks of significant differences between brain SPECT of patients with fibromyalgia and healthy subjects (P voxel < 0.005; P cluster < 0.05, corrected for multiple comparisons; age as nuisance variable). Compared with healthy controls, patients with fibromyalgia exhibited posterior hyperperfusion (red), including somatosensory cortex, and hypoperfusion (green) of frontal, cingulate, temporal, and cerebellar cortices. In particular, temporal hypoperfusion included polar and mediobasal cortices. BA = Brodmann area.

Brain SPECT Perfusion Correlation

The FIQ total score was positively correlated with perfusion of the bilateral superior parietal lobule, the bilateral precuneus (BA7), the left postcentral cortex (BA7) (r = 0.6544; P = 0.0043; Talairach coordinates of the local maxima of this cluster, x = 12, y = −61, and z = 56), and the right pre- and postcentral cortex (BA4, BA6, BA1, and BA3) (r = 0.5753; P = 0.0122; Talairach coordinates of the local maxima of this cluster, x = 42, y = −19, and z = 43) (Table 2; Fig. 2). These 2 clusters of positively correlated perfusion were included in the areas of significant hyperperfusions (respectively, P cluster = 0.012, corrected for multiple comparisons, k = 927; and P cluster = 0.038, corrected for multiple comparisons, k = 684). The FIQ total score was negatively correlated with perfusion of a left anterior temporal cluster, including the inferior, middle, and superior temporal gyri (BA21 and BA38); fusiform gyrus (BA20); and uncus (BA28 and BA36) (r = −0.5617; P = 0.0143; Talairach coordinates of the local maxima of this cluster, x = −48, y = −1, and z = −15). This cluster of negatively correlated perfusion was included in the areas of significant hypoperfusions (P cluster = 0.011, corrected for multiple comparisons, k = 908).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

Anatomic localization of peaks of significant correlation between brain SPECT perfusion and FIQ total score (P voxel < 0.005; P cluster < 0.05, corrected for multiple comparisons; age as nuisance variable).

Extracted perfusion values that positively and negatively correlated with the FIQ total score were independent of other clinical variables studied (VAS pain scale, QDSA scale, TBS, and HAD total score and subscores). FIQ total score was not correlated with other clinical variables.

No positive or negative correlations were observed between brain SPECT perfusion and other clinical variables (VAS pain scale, QDSA scale, TBS, and HAD total score and subscores), even by limiting the voxel-based analysis to significant hypo- or hyperperfusions found through comparisons with healthy subjects or by using small-volume correction for subcortical structures. In particular, no clinical correlation was observed with brain SPECT perfusion of the frontal, cingulate, and cerebellar cortices.