Overlapping Structural and Functional Brain ... - Wiley Online Library

ARTHRITIS & RHEUMATISM Vol. 65, No. 12, December 2013, pp 3293–3303 DOI 10.1002/art.38170 © 2013, American College of Rheumatology

Overlapping Structural and Functional Brain Changes in Patients With Long-Term Exposure to Fibromyalgia Pain Karin B. Jensen,1 Priti Srinivasan,1 Rosa Spaeth,1 Ying Tan,2 Eva Kosek,3 Frank Petzke,4 Serena Carville,5 Peter Fransson,3 Hanke Marcus,6 Steven C. R. Williams,7 Ernest Choy,8 Olivier Vitton,9 Richard Gracely,10 Martin Ingvar,3 and Jian Kong1 Objective. There is vast evidence to support the presence of brain aberrations in patients with fibromyalgia (FM), and it is possible that central plasticity is critical for the transition from acute to chronic pain. The aim of the present study was to investigate the relationship between brain structure and function in patients with FM. Methods. Functional connectivity of the brain during application of intermittent pressure–pain stimuli and measures of brain structure were compared between 26 patients with FM and 13 age- and sex-

matched healthy controls. Magnetic resonance imaging (MRI) was performed to obtain high-resolution anatomic images and functional MRI scans of the brain, which were used for measurements of pain-evoked brain activity. Results. FM patients displayed a distinct overlap between decreased cortical thickness, decreased brain volumes, and decreased functional regional coherence in the rostral anterior cingulate cortex. The morphometric changes were more pronounced with longer exposure to FM pain. In addition, there was evidence of an association between structural and functional changes in the mesolimbic areas of the brain and the severity of comorbid depression symptoms in FM patients. Conclusion. The combined integration of structural and functional measures allowed for a unique characterization of the impact of FM pain on the brain. These data may lead to the identification of early structural and functional brain alterations in response to pain, which could be used to develop markers for predicting the development of FM and other pain disorders.

Supported in part by the pharmaceutical company Pierre Fabre through financing of a placebo-controlled drug intervention study (EudraCT database no. 2004-004249-16). Dr. Jensen’s work was supported by the COFAS Marie Curie Postdoctoral Program. Dr. Kong’s work was supported by NIH grants K01-AT003883, R21AT004497, and R01-AT006364 (from the National Center for Complementary and Alternative Medicine) and R03-AT218317 (from the National Institute on Drug Abuse). 1 Karin B. Jensen, PhD, Priti Srinivasan, MSc, Rosa Spaeth, BS, Jian Kong, MD, MS, MPH: Massachusetts General Hospital, Harvard Medical School, and Athinoula A. Martinos Center for Biomedical Imaging, Boston, Massachusetts; 2Ying Tan, PhD: Massachusetts General Hospital, Harvard Medical School, and Athinoula A. Martinos Center for Biomedical Imaging, Boston, Massachusetts, and Southwest University for Nationalities, Chengdu, China; 3Eva Kosek, MD, PhD, Peter Fransson, PhD, Martin Ingvar, MD, PhD: Karolinska Institutet, Stockholm, Sweden; 4Frank Petzke, MD: University of Göttingen, Göttingen, Germany; 5Serena Carville, PhD: UK Age Research Forum, London, UK; 6Hanke Marcus, MD: University Hospital of Cologne, Cologne, Germany; 7Steven C. R. Williams, PhD: King’s College, London, London, UK; 8Ernest Choy, MD, FRCP: Cardiff University School of Medicine, Cardiff, UK; 9Olivier Vitton, MD: Pierre Fabre, Labe`ge, France; 10Richard Gracely, PhD: University of North Carolina, Chapel Hill. Dr. Gracely has received speaking fees from SimPar 2013 in Milan, Italy (less than $10,000) and owns stock in Algynomics, Inc. Address correspondence to Karin B. Jensen, PhD, MGH Psychiatric Neuroimaging, Building 120, 2nd Avenue, Charlestown, MA 02129. E-mail: [email protected]. Submitted for publication June 2, 2013; accepted in revised form August 20, 2013.

Chronic pain is a common health problem with limited treatment options. As a result, patients often experience persistent pain for decades, with inadequate access to effective pain relief and high rates of comorbid depression (1). In spite of the high prevalence of chronic pain, the impact of long-term exposure on the human brain is still poorly investigated. Fibromyalgia (FM) is a condition of widespread musculoskeletal pain, soft tissue tenderness, fatigue, and sleep disturbances (2). Studies have suggested that the presence of long-term localized pain is the most preva3293

3294

JENSEN ET AL

lent precursor of developing widespread pain (3) and FM (4). Moreover, over long-term followup, FM remains chronic, with low probability of full recovery (5). In patients with FM, there is evidence of central nervous system aberrations (6,7), such as augmented responses to experimental pain stimuli (8,9), changes in restingstate functional connectivity (10,11), and altered function of brain neurotransmitters (12–15). In addition, it has been demonstrated that FM is associated with brain morphologic changes, such as decreased gray matter in the insula, the rostral anterior cingulate cortex (rACC), and the medial prefrontal cortex, when compared to that in age-matched healthy controls (16–20). Neuroimaging has significantly contributed to our understanding of the brain mechanisms associated with FM (21). Nevertheless, the linkage between brain structure, brain function, and clinical symptoms, which could potentially elucidate the pathologic processes leading to development of the disorder, is still missing. Analysis of a combination of several different brain measures, in the same group of patients, may represent a critical step in elucidating the central mechanisms of FM and the negative impact of long-term exposure to pain. To that end, we designed the present study to investigate measures of brain morphology and functional regional homogeneity in FM patients as compared with healthy control subjects. We also investigated how the duration of FM symptoms and the severity of comorbid depression symptoms might affect the morphology and regional homogeneity of the brain within the patient group. We hypothesized that brain regions previously implicated in FM pathology would display both functional and morphometric changes when compared with the same brain regions in healthy controls. We also hypothesized that the altered brain structure and functional changes observed in FM patients would be associated with the duration of FM symptoms and the severity of comorbid depression symptoms.

PATIENTS AND METHODS Participants. A total of 92 female patients with FM (ages 25–55 years; mean ⫾ SD age 44 ⫾ 8.2 years) were enrolled. Nine of these patients were excluded from the functional magnetic resonance imaging (fMRI) analyses due to image artifacts or incidental findings of intracranial anomalies, leaving 83 FM patients available for analysis. A total of 13 healthy female subjects (ages 24–48 years; mean ⫾ SD age

34 ⫾ 8.6 years) were available as controls. Therefore, due to the relatively large number of patients, each healthy control was age- and sex-matched with 2 FM patients, yielding a total of 26 FM patients for analysis (ages 24–48 years; mean ⫾ SD 38 ⫾ 6.8 years). All patients in this study were investigated as part of the baseline assessments in a pharmacologic fMRI clinical trial that included 3 sites: England, Sweden, and Germany. The same brain imaging assessments were also performed in healthy controls at all 3 sites, obtained from subjects who were consecutively recruited throughout the study. The study was approved by the local ethics committee at each of the 3 participating sites, and all participants gave their written informed consent. Parts of the present data set were used in a previous study by Jensen et al in 2010, in which all 92 available patients with FM were analyzed (22), and in 2009, in which 11 of 26 FM patients and all 13 controls from the present study were included (23). Neither of these previously reported studies included analyses of brain morphology or regional functional connectivity. Screening. Before inclusion, all FM patients and healthy controls were carefully screened for the presence of inclusion or exclusion criteria. All FM patients were recruited from a primary care setting, and at the beginning of the screening visit, an experienced pain physician (EK, FP, or EC) confirmed the FM diagnosis, using the American College of Rheumatology 1990 classification criteria for FM (24). The inclusion criteria required that FM patients have a level of pain intensity that corresponded to a score of at least 40 on a 100-mm visual analog scale (VAS) for pain during the week prior to screening. Exclusion criteria included, for example, a severe psychiatric disorder, suicide risk, or history of substance, drug, or alcohol abuse. The study physicians used the Beck Depression Inventory (BDI) in combination with a clinical interview to determine whether or not patients were severely depressed. The data in the present study were collected as the baseline measure in a pharmacologic randomized controlled trial, and therefore patients’ medications or other treatments were strictly limited. All treatments that could influence the patients’ pain perception were prohibited, i.e., antidepressants (e.g., selective serotonin reuptake inhibitors, tricyclic agents), analgesics (e.g., tramadol, codeine, dextropropoxyphene), and strong opioids, including patches, anticonvulsants, centrally acting relaxants, joint injections, trigger/tender point injections, biofeedback, and transcutaneous electrical nerve stimulation. The duration of the period off medication (required to ensure complete washout) was dependent on the specific pharmacologic characteristics of each treatment (e.g., 4 weeks for fluoxetine, 2 weeks for amitriptyline). Nonsteroidal antiinflammatory drugs were used as rescue medication, and zolpidem was the only sleep aid permitted (both being controlled by the study physician). Pain stimulation. Experimental pain stimuli were applied to the thumbnail using a standardized method, in which a computer-controlled stimulator applies pressure via a 1-cm2 hard rubber probe (25). The thumb is inserted into a cylindrical opening and the probe applies pressure to the nail bed. This

BRAIN CHANGES IN LONG-TERM FIBROMYALGIA

pressure–pain method is commonly used in FM studies, because sensitivity to pressure is a diagnostic criterion for FM and elicits a deep pain sensation that is clinically valid in this group of patients. Brain imaging. MR images of the brain were collected at each of the study sites using 1.5T scanners (in London, a General Electric HDx scanner; in Stockholm, a General Electric Twinspeed Signa Horizon scanner; in Cologne, a Philips scanner). Multiple T2*-weighted, single-shot, gradientecho, echo-planar imaging (EPI) sequences were used to acquire blood oxygen level–dependent contrast images. The following parameters were used: repetition time 3,000 msec (35 slices acquired), echo time 40 msec, flip angle 90°, field of view 24 ⫻ 24 cm, 64 ⫻ 64–pixel matrix, slice thickness 4 mm with a 0.4-mm gap, and sequential image acquisition order. Visual distraction during scans was minimized by placing a blank screen in front of the patient’s field of view. Highresolution T1-weighted structural images were acquired in coronal orientation for analysis of brain morphology. Parameters were as follows: spoiled 3-dimensional gradient-recalled sequence, repetition time 24 msec, echo time 6 msec, flip angle 35°, and voxel size 0.9 ⫻ 1.5 ⫻ 0.9 mm3. The scanning procedure was standardized between sites with the use of written forms outlining oral instructions and practical training sessions. Group sizes were matched between sites in order to ensure that there was an equal number of patients and controls from every site. In line with the literature on multisite neuroimaging data (26,27), our statistical analyses provided valid data to indicate that the site factor had no significant effect on any of the brain data (as determined by analysis of variance of the extracted structural and functional brain data from the 3 sites). Procedures. Behavioral data. The day before the MRI scans, each participant was calibrated for subjective pain ratings by receiving 1 ascending and 1 randomized series of pressure stimuli. Pressures (duration of 2.5 seconds each) were delivered at 30-second intervals. Participants were instructed to rate the intensity of the pain by putting a mark on a 0–100-mm VAS, ranging from “no pain” to “worst imaginable pain.” Each participant’s pain threshold (VAS score ⬎0 mm) and stimulation maximum (VAS score ⬎60 mm) was used to compute 5 different pressure intensities within the range of the threshold and maximum. In total, 15 stimuli were delivered in a randomized order, and a polynomial regression was used to determine the pressure at which each individual patient’s pain intensity was rated a score of 50 mm on the VAS, hereafter referred to as the P50. The BDI was used to quantitatively assess symptoms of comorbid depression. The BDI is a 21-item measure of the severity of depression symptoms, and it has been extensively validated for use in both medical and mental health populations (28). A score of 0–9 indicates minimal depression, a score of 10–18 indicates mild depression, a score of 19–29 indicates moderate depression, and a score of 30–63 indicates severe depression. Functional imaging. Experimental pain during fMRI scanning was induced by applying the individually calibrated P50 pressures (pressure intensity eliciting a VAS pain score of 50 mm) as well as nonpainful pressures. Stimuli were randomly

3295

distributed over the scanning time. The time interval between stimuli had a mean stimulus-onset asynchronicity of 15 seconds (range 10–20 seconds). Each participant received at least 2 runs (total scan duration 16 minutes). Participants were instructed to focus on the pressures on the thumb and to not use any distraction or coping techniques. Statistical analysis. Behavioral data. Possible correlations between the duration of FM symptoms, pain ratings, and sensitivity to pressure–pain were determined using partial correlation analyses, which were controlled for physical age and depression scores. These calculations were performed using SPSS statistical software, version 18.0. Neuroimaging data. Cortical thickness. Reconstruction of the cortical surface was performed using the FreeSurfer program, version 5.1 (http://surfer.nmr.mgh.harvard.edu). FreeSurfer uses a series of computationally intensive steps on the T1-weighted structural volumes to estimate the gray–white interface (29–31). These steps include computing of Talairach transforms, intensity normalization, skull stripping, tessellation of the gray matter–white matter boundary, automated topology correction, automatic volume labeling, and white matter segmentation. Any inaccuracies in the reconstruction of the white and pial surfaces of the brain in individual subjects were manually corrected before calculating cortical thickness. The cortical thickness measure was computed as the distance between the pial and white surfaces at each point across the cortical mantle. Group analyses were performed by normalizing each subject’s data to the FreeSurfer average atlas, distributed as a part of the FreeSurfer tool set. The maps of cortical thickness were smoothed using a Gaussian full-width half-maximum kernel of 10 mm. A generalized linear model was calculated at each vertex on the surface. In order to determine the relationship between cortical thickness and clinical scores in patients, thickness was regressed on a vertex-by-vertex basis against scores of FM duration and depression, while controlling for physical age. For these analyses, the initial vertexwise threshold of significance was set at P values less than 0.001 (uncorrected). MRI volumes. The automated procedure for labeling different brain structures, and for obtaining brain volumetric measures, is described in detail by Fischl and colleagues (30). This procedure assigns a neuroanatomic label to each voxel in an MRI volume, based on probabilistic information automatically estimated from a manually labeled training set, including both gray and white matter. This technique has been shown to be comparable in accuracy to manual labeling. The automatic segmentations were also visually inspected for accuracy. In the present study, the results are reported as the total brain volume (supratentorial), volumes of the cortical and subcortical gray matter, volume of white matter, and volumes of the predefined regions of interest (ROIs), including the rACC (12,23), the amygdala (32,33), and the lateral orbitofrontal cortex (lOFC) (17,34). The ROIs chosen were based on findings from previous FM neuroimaging studies, and each was defined using standardized anatomic labels provided in the FreeSurfer program and the Automatic Anatomical Labeling in SPM.

3296

Statistical analyses for volumetric measures, extracted from FreeSurfer’s automated segmentation, were performed using SPSS software, version 18.0. Group differences between FM patients and controls in the volumes of cortical and subcortical gray matter and cortical and subcortical white matter, as well as in the segmentation volumes of the predefined ROIs, were analyzed using independent-samples t-tests. In order to estimate the relationships between neuroanatomic volumes from the predefined ROIs and behavioral measures (depression scores and pain scores), partial correlation analyses were performed. Analyses assessing the correlations between ROI volumes and FM duration were controlled for age and depression, and, conversely, analyses assessing the correlations between ROI volumes and depression were controlled for age and FM duration. Functional connectivity. The method used for assessing functional connectivity in this study, the regional homogeneity (ReHo) method, is a model-free estimation of the synchrony of spontaneous fMRI signal oscillations within neighboring voxels. This method was recently proven to be sensitive to disease-related alterations in the brain, as demonstrated in several different disorders (35–37). Other functional connectivity assessment methods, such as seed-based methods (which focus on connectivity between a particular seed and other brain regions) (38) and independent components analysis (which focuses on identifying different components of brain networks) (39), are used to identify key networks of the brain. The ReHo method, on the other hand, measures the local synchrony of different brain regions, facilitating the identification of key brain regions rather than networks. This unique character of the ReHo method is important when investigating the relationship between specific brain structures and function. During our fMRI experiments, the pain stimulus was administered intermittently for a total of 16 minutes. We then applied the ReHo method as a model-free approach to estimate the regional coherence during the whole 16 minutes of the scan. All preprocessing and ReHo analyses were performed using the DPARSFA (Data Processing Assistant for Resting State fMRI, Advanced) (40–42). Preprocessing included correction for slice timing (using the middle slice as a reference), realignment (motion correction), normalization into Montreal Neurological Institute (MNI) space using an EPI template, linear trend removal, and temporal filtering of results by 0.01–0.08 Hz. Thereafter, the ReHo, which was assessed as the Kendall’s coefficient concordance (KCC) (43), was calculated using a cluster consisting of 27 voxels. At a given voxel, ReHo was defined as the KCC for the concordance between the time series of this voxel with the time series of its 26 nearest neighbors. The resulting map of ReHo values at each voxel was then divided by the global mean value within the whole-brain mask. This result was then smoothed with a kernel of 8 mm. Ultimately, a 2-sample t-test, in which FM patients were compared with controls, and 2 separate regression analyses were performed, using the second-level analysis function in SPM8 (Wellcome Trust Centre for Neuroimaging, London, UK) and MatLab 7.4 (Mathworks) software. Consistent with

JENSEN ET AL

Table 1. Characteristics of the patients with fibromyalgia (FM) compared with healthy controls*

Variable Age, years Pressure at VAS 50, kPa Duration of FM symptoms, years Weekly pain intensity, VAS score, mm BDI depression score

FM patients (n ⫽ 26)

Healthy controls (n ⫽ 13)

38 ⫾ 7 341 ⫾ 133 11 ⫾ 6

34 ⫾ 9 572 ⫾ 186 NA

72 ⫾ 14

NA

21 ⫾ 11

NA

T score

P†

⫺1.8 4.14

0.08 0.001

* The pressure at VAS 50 represents the pressure required to evoke pain that corresponds to a rating of 50 on a 0–100-mm visual analog scale (VAS) for pain intensity. The duration of FM symptoms is based on patients’ subjective report. The weekly pain intensity refers to patients’ rating of the average clinical pain intensity over the last week, on a 0–100-mm VAS. The Beck Depression Inventory (BDI) depression score refers to self-rated symptoms of depression, in which scores of 0–9 ⫽ minimal depression, 10–18 ⫽ mild depression, 19–29 ⫽ moderate depression, and 30–63 ⫽ severe depression. Values are the mean ⫾ SD. NA ⫽ not applicable. † Determined by independent-samples t-test.

the findings in previous studies, the initial voxelwise threshold of significance was set at P values less than 0.005 (uncorrected), with 20 contiguous voxels used for predefined ROIs, and at P values less than 0.05 when corrected for family-wise error (FWE) at the cluster level. For non-ROI brain regions, the voxelwise threshold of significance was set at P values less than 0.001 (uncorrected), and at P values less than 0.05 when FWE-corrected at the cluster level.

RESULTS Between-group effects. Behavioral outcomes. A 2-sample t-test revealed a significant difference in pressure–pain sensitivity (i.e., the P50) between FM patients and healthy controls (T score 4.14, 37 df; P ⬍ 0.001 by 2-tailed t-test), thus validating that the FM patients required significantly lower amounts of pressure than did controls to experience comparable pain intensities (Table 1). Neuroimaging outcomes. Cortical thickness. A vertexwise whole-brain analysis of cortical thickness generated significance maps for groups of vertices with significant differences between FM patients and controls. This analysis revealed several regions of significantly lower cortical thickness in FM patients compared with controls. These regions included the left rACC, which was part of our a priori hypothesis, as well as other regions beyond the territory of the hypothesis (the left superior frontal gyrus, right superior temporal gyrus,


3297

Table 2. Differences in cortical thickness and brain functional connectivity between patients with fibromyalgia (FM) and healthy controls* MNI coordinate

Cortical thickness: controls ⬎ FM rACC Superior frontal Middle temporal Superior temporal Middle temporal Fusiform Cortical thickness: FM ⬎ controls Superior parietal ReHo: controls ⬎ FM rACC

Laterality

No. of vertices

Cluster size, voxels

x

y

z

P†

Left Left Left Right Right Right

136 85 48 250 106 60

– – – – – –

⫺6 ⫺7 ⫺56 55 63 37

36 ⫺3 ⫺14 ⫺22 ⫺16 ⫺25

11 62 ⫺20 4 ⫺16 ⫺22

0.00041 0.00046 0.00095 0.00024 0.00037 0.00082

Right

181

–

26

⫺61

44

0.00021

–

756

⫺4

49

⫺11

Left

0.05

* Values are the results from FreeSurfer analyses of cortical thickness and regional homogeneity (ReHo) analyses of functional regional coherence. The rostral anterior cingulate cortex (rACC) was a predefined brain region of interest, and therefore we did not control for multiple comparisons outside of the rACC-labeled brain region. Anatomic location is given in Montreal Neurological Institute (MNI) coordinates. † P values are the maximum P values for cortical thickness, and the cluster P value for ReHo.

right and left middle temporal gyrus, and right fusiform gyrus). Only one group of vertices displayed higher cortical thickness in FM patients compared with controls, located in the right superior parietal gyrus (Table 2 and Figure 1). MRI volumes. In an overall analysis of total brain volume (supratentorial), including both the cortical and subcortical structures, FM patients displayed significantly lower total brain volume than did healthy controls (T score ⫺3.6, 37 df; P ⫽ 0.001). The automated segmentation of cerebral matter performed with the FreeSurfer program allowed for more specific comparisons of white and gray matter volumes. Analyses of gray

matter volumes, both cortical and subcortical, revealed that patients had significantly lower volumes when compared with controls, both in cortical gray matter (T score ⫺3.2, 37 df; P ⫽ 0.003 by 2-tailed t-test) and in subcortical gray matter (T score ⫺2.7, 37 df; P ⫽ 0.010 by 2-tailed t-test). In addition to the differences in gray matter volumes, patients displayed significantly lower volume of cortical white matter when compared with controls (T score ⫺2.9, 37 df; P ⫽ 0.006 by 2-tailed t-test), indicating that the decreased brain volumes seen in FM patients could not be attributed solely to an effect of gray matter atrophy. We also compared the volumes of the predefined

Figure 1. Cortical thickness and brain volumes in patients with fibromyalgia (FM) compared with healthy controls. A, Measures of cortical thickness revealed lower cortical thickness in the left rostral anterior cingulate cortex (rACC) in FM patients compared with age- and sex-matched healthy controls. Measures were determined using the FreeSurfer program. A representative image (with Montreal Neurological Institute coordinates x, y, and z) is shown. B, Voxel-based measurements of brain volume revealed significant differences between FM patients and healthy controls in all 3 predefined regions of interest, including the right and left rACC, the right and left lateral orbitofrontal cortex (lOFC), and the right and left amygdala. All statistical analyses were 2-tailed. Results are expressed as the mean ⫾ SD.

3298

Figure 2. Overlap of 3 different structural and functional measures of the brain in patients with fibromyalgia (FM). Spatial overlap between measurements of brain volume (green), cortical thickness (blue), and regional functional connectivity (according to regional homogeneity analyses of functional magnetic resonance imaging [fMRI] data) (red) is evident in the rostral anterior cingulate cortex of a patient with FM. All 3 measures were decreased in FM patients as compared with healthy controls.

ROIs (as described above) between FM patients and healthy controls. These analyses revealed that, when compared with controls, patients had significantly lower volumes in the left rACC (T score ⫺4.7, 37 df; P ⫽ 0.00003), the left lOFC (T score ⫺3.79, 37 df; P ⫽ 0.0005), the right lOFC (T score ⫺2.49, 37 df; P ⫽ 0.017), and the right amygdala (T score ⫺2.75, 37 df; P ⬍ 0.009) (Figure 1). ReHo functional connectivity. Analyses of ReHo as a measure of regional functional connectivity revealed significantly lower coherence in the rACC in FM patients compared with controls. The cluster was partially located in the left cingulate cortex, but also extended into the left medial prefrontal cortex. There were no regions that showed significantly higher regional coherence in patients than in controls (Table 2 and Figure 2).

JENSEN ET AL

Within-group effects in relation to duration of FM. Behavioral outcomes. Within the FM patient group (n ⫽ 26), a partial correlation between the duration of FM and VAS ratings of clinical pain intensity was seen (r ⫽ 0.029, P ⫽ 0.892), while there was a moderate correlation between the duration of FM and sensitivity to P50 pressure–pain (r ⫽ 0.405, P ⫽ 0.050). Neuroimaging outcomes. Cortical thickness. Regression analyses were used to investigate the effects of FM duration on cortical thickness in the FM patients (n ⫽ 26) (Table 3). The middle temporal gyrus displayed a significant negative association between FM duration and cortical thickness. This negative association was also found to be significant in the rACC, a predefined ROI included in our a priori hypothesis, indicating that cortical thickness was lower with longer duration of FM (Figure 3). Several groups of vertices, specifically in the left middle frontal cortex, bilateral parietal lobes, and bilateral temporal lobes, displayed higher cortical thickness with longer duration of FM pain (Table 3). MRI volumes. A regression analysis revealed that the supratentorial brain volume, including both the cortical and subcortical gray matter areas, was significantly decreased with longer duration of FM (r ⫽ ⫺0.435, P ⫽ 0.030). In separate analyses of the cortical and subcortical gray matter, there were significant decreases in gray matter volume with longer FM duration in the subcortical area (r ⫽ ⫺0.403, P ⫽ 0.046), but there was only a trend toward significance in the cortical area (r ⫽ ⫺0.384, P ⫽ 0.058). A separate correlation analysis of FM duration and white matter volumes displayed a significant decrease in white matter volume with longer duration of FM (r ⫽ ⫺0.403, P ⫽ 0.046). The volumes of the right and left rACC were negatively correlated with the duration of FM (left rACC, r ⫽ ⫺0.44, P ⫽ 0.034; right rACC, r ⫽ ⫺0.54, P ⫽ 0.007), suggesting that rACC volumes decreased with longer FM duration (Figure 3). The left and right lOFC also displayed negative correlations with the duration of FM (left lOFC, r ⫽ ⫺0.42, P ⫽ 0.040; right lOFC, r ⫽ ⫺0.58, P ⫽ 0.001) (Figure 3). Amygdala volumes were not significantly correlated with the duration of FM; however, there was a strong trend toward a significant correlation for the left amygdala (r ⫽ ⫺0.42, P ⫽ 0.052). ReHo functional connectivity. The effect of FM duration on functional regional coherence was assessed using regression analyses of the fMRI data from the FM


3299

Table 3. Cortical thickness in relation to duration of fibromyalgia (FM) and severity of depression* MNI coordinate

Cortical thickness–FM duration Negative correlation Middle temporal rACC Positive correlation Supramarginal Rostral middle frontal Inferior temporal Superior frontal Pars opercularis Inferior temporal Precentral Caudal middle frontal Cortical thickness–depression Negative correlation Pericalcarine Fusiform Fusiform Lateral occipital Paracentral Inferior parietal Caudal middle frontal Precuneus Fusiform Positive correlation Supramarginal

Laterality

No. of vertices

Right Right

99 82

Left Left Left Left Left Right Right Right

x

Maximum P

y

z

63 5

⫺27 31

⫺16 ⫺7

0.0003 0.00097

118 118 122 40 82 120 25 41

⫺50 ⫺30 ⫺51 ⫺20 ⫺33 46 52 39

⫺52 43 ⫺63 8 22 ⫺34 ⫺4 19

45 18 ⫺5 61 12 ⫺20 8 38

0.00011 0.00015 0.00040 0.00073 0.00074 0.000082 0.00037 0.00062

Right Right Right Right Right Left Left Left Left

896 161 20 55 64 133 66 60 14

18 38 34 24 17 ⫺39 ⫺25 ⫺18 ⫺37

⫺72 ⫺62 ⫺46 ⫺86 ⫺26 ⫺70 ⫺2 ⫺77 ⫺29

5 ⫺18 ⫺14 17 39 44 45 30 ⫺25

0.0000036 0.00010 0.0007 0.00056 0.00062 0.000059 0.00018 0.00059 0.00084

Left

265

⫺54

⫺51

27

0.00032

* Analyses of correlations between cortical thickness measures and the duration of FM pain were performed, with age and depression as covariates of no interest. Comorbid depression was measured using the Beck Depression Inventory. Negative correlations indicate decreased cortical thickness with longer FM duration or higher depression scores. Conversely, positive correlations indicate increased cortical thickness with longer FM duration or higher depression scores. The anatomic locations of each cluster are given in Montreal Neurological Institute (MNI) coordinates. The rostral anterior cingulate cortex (rACC) was a predefined region of interest.

patients (n ⫽ 26), using duration of FM as a covariate of interest. The regression analysis with FM duration as a

Figure 3. Analyses of correlation between the duration of fibromyalgia (FM) pain and brain volumes. The duration of FM pain was negatively correlated with brain volumes in the right lateral orbitofrontal cortex (OFC) (r ⫽ ⫺0.58, P ⫽ 0.001 by 2-tailed t-test) and right rostral anterior cingulate cortex (rACC) (r ⫽ ⫺0.54, P ⫽ 0.007 by 2-tailed t-test). Analyses were controlled for physical age and symptoms of comorbid depression. Circles denote individual patients.

covariate revealed no significant association, either negative or positive, with functional regional coherence. Correlation analysis across all 3 imaging domains. The individually extracted functional ReHo measures in the rACC correlated significantly with rACC volumetric measures (r ⫽ 0.42, P ⫽ 0.035), suggesting that lower gray matter volumes in the rACC were associated with lower functional coherence. Moreover, there was a correlation between rACC gray matter volumes and rACC measures of cortical thickness (r ⫽ 0.31, P ⫽ 0.067), but this did not reach significance. Within-group effects in relation to severity of depression. Behavioral outcomes. Within the FM patient group (n ⫽ 26), there was a partial correlation between BDI depression scores and the duration of FM (r ⫽ 0.23, P ⫽ 0.26). However, there was no correlation between depression scores and ratings of clinical pain intensity (r ⫽ 0.10, P ⫽ 0.60). Neuroimaging outcomes. Cortical thickness. The effect of BDI depression scores on cortical thickness was

3300

JENSEN ET AL

analyzed by regression analyses. Several areas on the cortical surface displayed significant negative correlations between cortical thickness and depression scores. These regions were located primarily in the parietal and occipital cortex, indicating that cortical thinning in these regions was associated with higher BDI scores. Only one group of vertices, located in the left parietal lobe (supramarginal gyrus), displayed a positive correlation between depression and cortical thickness (Table 3). MRI volumes. There were no significant correlations between BDI depression scores and the gray matter volume of the total brain (supratentorial), including both the cortical and subcortical areas. ReHo functional connectivity. A regression analysis of functional regional coherence and BDI depression scores revealed significant changes in regional coherence in the ventral striatum/nucleus accumbens. With high levels of comorbid depression symptoms, patients displayed relatively lower regional coherence in the nucleus accumbens region of the striatum (MNI coordinates x ⫽ 21, y ⫽ 15, z ⫽ 0; cluster size 57 voxels; peak T score 3.85 [P ⬍ 0.005 by t-test, FWE corrected for cluster]). There were no significant results for the opposite analysis of increased regional coherence with high levels of depression. DISCUSSION In this study, we investigated structural and functional brain differences between FM patients and matched controls. Our results suggest that the rACC, a key region for modulation of pain, had significantly decreased cortical thickness, brain volume, and regional functional connectivity in FM patients as compared with healthy controls. The volumetric changes were more pronounced with longer exposure to FM pain. In addition, there was an association between structural and functional changes in the mesolimbic areas of the brain and the severity of comorbid depression symptoms. The present study provides unique evidence to indicate that overlapping neuroanatomic and functional brain changes can occur in patients with chronic pain. Consistent with previous studies (44,45), our study demonstrated that FM patients were significantly more sensitive to experimental pain than were healthy control subjects. In line with this behavioral difference, the analysis of regional functional connectivity in FM patients and controls revealed that the rACC coherence was lower in FM patients, as assessed using the ReHo method. ReHo measures the synchrony of spontaneous

fMRI signal oscillations within neighboring voxels, and it has been proven sensitive to the detection of diseaserelated alterations in brain function (35–37). Impaired pain-inhibitory responses in FM patients have been previously found to be associated with reduced activation of the rACC (23) and less functional connectivity between the rACC and other regions of the pain modulatory system of the brain (46). Since the rACC is a key region for descending inhibition of pain (47,48), the lower rACC coherence in the FM patients in this study provides support for the hypothesis that pathologic development of FM may be associated with dysfunction of descending pain modulation (45,49,50). Consistent with previous morphometric studies in FM (16–20), we observed that measures of cortical thickness and brain volumes were decreased in FM patients compared with healthy controls. As expected, the cortical thickness of the rACC was decreased with longer duration of FM. In line with our hypothesis, there was a negative association between FM duration and rACC volumes, suggesting that rACC volumes are decreased with increased duration of FM pain. A study from 2007 found that ␮-opioid receptor function was lower in the rACC of patients with FM (12), possibly reflecting significant atrophy in this region. Our results indicate that the rACC may be an important element in the development and/or pathophysiology of FM. In line with previous morphometric reports of increased gray matter volumes in FM patients (20), we also observed increased cortical thickness with increased duration of FM in several temporal and frontal regions, such as the superior frontal gyrus and the inferior temporal gyrus. It is possible that increased cortical thickness with longer FM duration reflects a compensatory mechanism, in which the brain may attempt to prevent the negative effects of a constant nociceptive input. It is also possible that the increase in cortical thickness in areas relating to emotional processing, e.g., medial temporal regions, reflects an increase in the affective-evaluative processing of pain, as has been suggested to occur in patients with chronic low back pain (51); however, the mechanisms are not clear. In a recent morphometric study (52), the authors found that the difference in gray matter volumes between FM patients and matched controls disappeared when the analyses were controlled for comorbid affective measures, suggesting that brain atrophy might be related more to negative affect than to pain. To address this concern, we used patients’ ratings of symptoms of depression (BDI scores) to assess the specific impact of


comorbid depression on brain function and structure in FM. A regression analysis of the fMRI data revealed a significant association between high depression scores and decreased functional coherence in the ventral striatum. The ventral striatum is part of the mesolimbic reward circuitry, and low activation has previously been associated with core symptoms of depression, such as anhedonia and reduced motivation (53). There were no brain regions that had significantly greater regional coherence with higher levels of depression. The analysis of cortical thickness revealed that there was a significant correlation between decreases in cortical thickness in 9 regions of the brain and higher depression scores, including the parietal cortex, a region that has been previously implicated in depression (53). There was no spatial overlap between the functional decrease in mesolimbic activation and decreases in brain volumes. Furthermore, we did not find any significant depression-related changes in any of the regions that were defined as ROIs (functional or structural). Taken together, our results indicate that depression symptoms are associated with cerebral changes, independent of pain. The results are consistent with our findings from an earlier neuroimaging study in which we found that the neural mechanisms for pathologic development of FM and the neural mechanisms for depression were independent (22), even though depression and pain are inherently related in the clinical context. Recent studies of causality suggest that long-term exposure to pain causes reductions in specific brain regions, and not vice versa (54,55). Herein, we found that the duration of FM was significantly correlated with brain changes; nevertheless, the sequence of these changes needs to be fully investigated in longitudinal studies. Such studies could potentially explore whether structural and functional alterations are early indicators of the development of widespread chronic pain. These findings may help in the development of predictive tools for chronic pain. The combination of functional and structural brain measures revealed that FM patients had overlapping decreases in cortical thickness, brain volumes, and regional functional coherence in the rACC. The atrophy of the rACC, as indicated by brain volumetric measures, was associated with duration of FM pain. Given the crucial role of the rACC in descending pain modulation, our result implies that disrupted endogenous pain modulation is central in the development and pathophysiology of FM.

3301

AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Jensen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Jensen, Kosek, Petzke, Carville, Fransson, Marcus, Williams, Choy, Vitton, Gracely, Ingvar. Acquisition of data. Jensen, Kosek, Petzke, Carville, Fransson, Marcus, Williams, Choy, Gracely, Ingvar. Analysis and interpretation of data. Jensen, Srinivasan, Spaeth, Tan, Petzke, Fransson, Marcus, Choy, Gracely, Ingvar, Kong.

ROLE OF THE STUDY SPONSOR This study was performed in collaboration with the pharmaceutical company Pierre Fabre. The results are derived, in part, from a placebo-controlled drug intervention study (European Clinical Trials database no. 2004-004249-16) financed by Pierre Fabre, and they reviewed and approved the manuscript prior to submission. The authors independently collected the data, interpreted the results, and had the final decision to submit the manuscript for publication. Publication of this article was not contingent upon approval by Pierre Fabre.

REFERENCES 1. Bair MJ, Robinson RL, Katon W, Kroenke K. Depression and pain comorbidity: a literature review. Arch Intern Med 2003;163: 2433–45. 2. Wolfe F. New American College of Rheumatology criteria for fibromyalgia: a twenty-year journey. Arthritis Care Res (Hoboken) 2010;62:583–4. 3. Bergman S, Herrstrom P, Jacobsson LT, Petersson IF. Chronic widespread pain: a three year followup of pain distribution and risk factors. J Rheumatol 2002;29:818–25. 4. Forseth KO, Husby G, Gran JT, Forre O. Prognostic factors for the development of fibromyalgia in women with self-reported musculoskeletal pain: a prospective study. J Rheumatol 1999;26: 2458–67. 5. Bengtsson A, Backman E, Lindblom B, Skogh T. Long term follow-up of fibromyalgia patients: clinical symptoms, muscular function, laboratory tests: an eight year comparison study. J Musculoskelet Pain 1994;2:67–80. 6. Clauw DJ. Fibromyalgia: update on mechanisms and management. J Clin Rheumatol 2007;13:102–9. 7. Yunus MB. Fibromyalgia and overlapping disorders: the unifying concept of central sensitivity syndromes. Semin Arthritis Rheum 2007;36:339–56. 8. Cook DB, Lange G, Ciccione DS, Liu WC, Steffener J, Natelson BH. Functional imaging of pain in patients with primary fibromyalgia. J Rheumatol 2004;31:364–78. 9. Gracely RH, Petzke F, Wolf M, Clauw D. Functional magnetic resonance imaging evidence of augmented pain processing in fibromyalgia. Arthritis Rheum 2002;46:1333–43. 10. Napadow V, Kim J, Clauw DJ, Harris RE. Decreased intrinsic brain connectivity is associated with reduced clinical pain in fibromyalgia. Arthritis Rheum 2012;64:2398–403. 11. Napadow V, LaCount L, Park K, As-Sanie, Clauw D, Harris RE. Intrinsic brain connectivity in fibromyalgia is associated with chronic pain intensity. Arthritis Rheum 2010;62:2545–55. 12. Harris RE, Clauw DJ, Scott DJ, McLean SA, Gracely RH, Zubieta JK. Decreased central ␮-opioid receptor availability in fibromyalgia. J Neurosci 2007;27:10000–6.

3302

13. Harris RE, Sundgren PC, Craig AD, Kirshenbaum E, Sen A, Napadow V, et al. Elevated insular glutamate in fibromyalgia is associated with experimental pain. Arthritis Rheum 2009;60: 3146–52. 14. Kadetoff D, Lampa J, Westman M, Andersson M, Kosek E. Evidence of central inflammation in fibromyalgia-increased cerebrospinal fluid interleukin-8 levels. J Neuroimmunol 2012;242: 33–8. 15. Sarchielli P, Mancini ML, Floridi A, Coppola F, Rossi C, Nardi K, et al. Increased levels of neurotrophins are not specific for chronic migraine: evidence from primary fibromyalgia syndrome. J Pain 2007;8:737–45. 16. Burgmer M, Gaubitz M, Konrad C, Wrenger M, Hilgart S, Heuft G, et al. Decreased gray matter volumes in the cingulo-frontal cortex and the amygdala in patients with fibromyalgia. Psychosomatic Med 2009;71:566–73. 17. Kuchinad A, Schweinhardt P, Seminowicz DA, Wood PB, Chizh BA, Bushnell MC. Accelerated brain gray matter loss in fibromyalgia patients: premature aging of the brain? J Neurosci 2007;27: 4004–7. 18. Lutz J, Jager L, de Quervain D, Krauseneck T, Padberg F, Wichnalek M, et al. White and gray matter abnormalities in the brain of patients with fibromyalgia: a diffusion-tensor and volumetric imaging study. Arthritis Rheum 2008;58:3960–9. 19. Robinson ME, Craggs JG, Price DD, Perlstein WM, Staud R. Gray matter volumes of pain-related brain areas are decreased in fibromyalgia syndrome. J Pain 2011;12:436–43. 20. Schmidt-Wilcke T, Luerding R, Weigand T, Ju ¨rgens T, Schuierer G, Leinisch E, et al. Striatal grey matter increase in patients suffering from fibromyalgia—a voxel-based morphometry study. Pain 2007;132:109–16. 21. Schweinhardt P, Sauro KM, Bushnell MC. Fibromyalgia: a disorder of the brain? Neuroscientist 2008;14:415–21. 22. Jensen KB, Ingvar M, Petzke F, Carville S, Fransson P, Choy E, et al. Anxiety and depressive symptoms in fibromyalgia are related to low health esteem but not to sensitivity or cerebral processing of pain. Arthritis Rheum 2010;62:3488–95. 23. Jensen KB, Kosek E, Petzke F, Carville S, Fransson P, Marcus H, et al. Evidence of dysfunctional pain inhibition in fibromyalgia reflected in rACC during provoked pain. Pain 2009;144:95–100. 24. Wolfe F, Smythe H, Yunus M, Bennett R, Bombardier C, Goldenberg D, et al. The American College of Rheumatology 1990 criteria for the classification of fibromyalgia: report of the Multicenter Criteria Committee. Arthritis Rheum 1990;33:160–72. 25. Petzke F, Clauw DJ, Ambrose K, Khine A, Gracely RH. Increased pain sensitivity in fibromyalgia: effects of stimulus type and mode of presentation. Pain 2003;105:403–13. 26. Han X, Jovicich J, Salat D, van der Kouwe A, Quinn B, Czanner S, et al. Reliability of MRI-derived measurements of human cerebral cortical thickness: the effects of field strength, scanner upgrade and manufacturer. Neuroimage 2006;32:180–94. 27. Yendiki A, Greve DN, Wallace S, Vangel M, Bockholt J, Mueller BA, et al. Multi-site characterization of an fMRI working memory paradigm: reliability of activation indices. Neuroimage 2010;53: 119–31. 28. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4: 561–71. 29. Dale AM, Fischl B, Sereno MI. Cortical surface-based analysis. I: Segmentation and surface reconstruction. Neuroimage 1999;9: 179–94. 30. Fischl B, Salat DH, Busa E, Albert M, Dieterich M, Haselgrove C, et al. Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron 2002;33:341–55. 31. Fischl B, Sereno MI, Dale AM. Cortical surface-based analysis.

JENSEN ET AL

32.

33.

34.

35.

36.

37. 38.

39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

50. 51. 52.

II: Inflation, flattening, and a surface-based coordinate system. Neuroimage 1999;9:195–207. Burgmer M, Gaubitz M, Konrad C, Wrenger M, Hilgart S, Heuft G, et al. Decreased gray matter volumes in the cingulo-frontal cortex and the amygdala in patients with fibromyalgia. Psychosom Med 2009;71:566–73. Giesecke T, Gracely RH, Williams DA, Geisser ME, Petzke FW, Clauw DJ. The relationship between depression, clinical pain, and experimental pain in a chronic pain cohort. Arthritis Rheum 2005;52:1577–84. Jensen KB, Kosek E, Wicksell R, Kemani M, Olsson G, Merle JV, et al. Cognitive behavioral therapy increases pain-evoked activation of the prefrontal cortex in patients with fibromyalgia. Pain 2012;153:1495–503. Mankinen K, Long XY, Paakki JJ, Harila M, Rytky S, Tervonen O, et al. Alterations in regional homogeneity of baseline brain activity in pediatric temporal lobe epilepsy. Brain Res 2011;1373:221–9. Paakki JJ, Rahko J, Long X, Moilanen I, Tervonen O, Nikkinen J, et al. Alterations in regional homogeneity of resting-state brain activity in autism spectrum disorders. Brain Res 2010;1321: 169–79. Qiu C, Liao W, Ding J, Feng Y, Zhu C, Nie X, et al. Regional homogeneity changes in social anxiety disorder: a resting-state fMRI study. Psychiatry Res 2011;194:47–53. Van Dijk KR, Hedden T, Venkataraman A, Evans KC, Lazar SW, Buckner RL. Intrinsic functional connectivity as a tool for human connectomics: theory, properties, and optimization. J Neurophysiol 2010;103:297–321. Biswal BB, Mennes M, Zuo XN, Gohel S, Kelly C, Smith SM, et al. Toward discovery science of human brain function. Proc Natl Acad Sci U S A 2010;107:4734–9. Chao-Gan Y, Yu-Feng Z. DPARSF: a MatLab toolbox for “pipeline” data analysis of resting-state fMRI. Front Systems Neurosci 2010;4:13. Zang Y, Jiang T, Lu Y, He Y, Tian L. Regional homogeneity approach to fMRI data analysis. Neuroimage 2004;22:394–400. Zou Q, Wu CW, Stein EA, Zang Y, Yang Y. Static and dynamic characteristics of cerebral blood flow during the resting state. Neuroimage 2009;48:515–24. Kendall M, Gibbons J. Rank correlation methods. Edward Arnold; 1990. Kosek E, Ekholm J, Hansson P. Sensory dysfunction in fibromyalgia patients with implications for pathogenic mechanisms. Pain 1996;68:375–83. Lautenbacher S, Rollman GB. Possible deficiencies of pain modulation in fibromyalgia. Clin J Pain 1997;13:189–96. Jensen KB, Loitoile R, Kosek E, Petzke F, Carville S, Fransson P, et al. Patients with fibromyalgia display less functional connectivity in the brain’s pain inhibitory network. Molecular Pain 2012;8:32. Bingel U, Lorenz J, Schoell E, Weiller C, Buchel C. Mechanisms of placebo analgesia: rACC recruitment of a subcortical antinociceptive network. Pain 2006;120(1):8-15. Fields HL. State-dependent opioid control of pain. Nat Rev Neurosci 2004;5:565–75. Kosek E, Hansson P. Modulatory influence on somatosensory perception from vibration and heterotopic noxious conditioning stimulation (HNCS) in fibromyalgia patients and healthy subjects. Pain 1997;70:41–51. Julien N, Goffaux P, Arsenault P, Marchand S. Widespread pain in fibromyalgia is related to a deficit of endogenous pain inhibition. Pain 2005;114:295–302. Apkarian AV, Sosa Y, Krauss BR, Thomas PS, Fredrickson BE, Levy RE, et al. Chronic pain patients are impaired on an emotional decision-making task. Pain 2004;108:129–36. Hsu MC, Harris RE, Sundgren PC, Welsh RC, Fernandes CR, Clauw DJ, et al. No consistent difference in gray matter volume


between individuals with fibromyalgia and age-matched healthy subjects when controlling for affective disorder. Pain 2009;143: 262–7. 53. Drevets WC, Price JL, Furey ML. Brain structural and functional abnormalities in mood disorders: implications for neurocircuitry models of depression. Brain Struc Funct 2008;213: 93–118.

3303

54. Rodriguez-Raecke R, Niemeier A, Ihle K, Ruether W, May A. Brain gray matter decrease in chronic pain is the consequence and not the cause of pain. J Neurosci 2009;29:13746–50. 55. Seminowicz DA, Wideman TH, Naso L, Hatami-Khoroushahi Z, Fallatah S, Ware MA, et al. Effective treatment of chronic low back pain in humans reverses abnormal brain anatomy and function. J Neuroscience 2011;31:7540–50.

DOI 10.1002/art.38120

Clinical Images: Bronchial stenosis in granulomatosis with polyangiitis (Wegener’s)

The patient, a 39-year-old woman, presented with a cough and progressive dyspnea during physical exertion. We diagnosed granulomatosis with polyangiitis (Wegener’s) (GPA). She developed severe proximal bronchial stenosis, lost 5 kg in 2 months, and felt weak. She had a 4-month history of epistaxis, with sinusitis and arthralgia. Pulmonary examination revealed respiratory obstruction, with no sound of breathing in the right hemithorax. Computed tomography images showed severe stenosis of the right main and intermediate bronchi (arrows in A), and bronchoscopy revealed thickening and concentric stenosis (arrows in B), corresponding to the lesions in A. Multiple perihilar excavated masses were also noted. Bronchial biopsies showed fibrinoid necrosis (lower arrow in C) and multinucleated giant cells suggesting GPA (upper arrow in C). Antineutrophil cytoplasmic antibody proteinase 3 was highly positive. There was no evidence of renal, neurologic, or gastrointestinal involvement of GPA. Myocarditis related to GPA was diagnosed. The patient received glucocorticoid and rituximab therapy, which resulted in rapidly decreased coughing and reduced dyspnea. Subglottic and proximal trachea stenoses have been described in ⬃16% of patients during the course of GPA (1). However, bronchial stenosis is a rare initial manifestation of GPA (2). Pulmonary involvement in GPA usually responds well to treatment with immunosuppressants. In contrast, medical treatment of bronchial stenosis is sometimes inefficient, and patients may require local therapy, i.e., dilatation using a rigid bronchoscope, YAG-laser treatment, or intraluminal stenting (3). 1. Hoffman GS, Kerr GS, Leavitt RY, Hallahan CW, Lebovics RS, Travis WD, et al. Wegener granulomatosis: an analysis of 158 patients. Ann Intern Med 1992;116:488–98. 2. Daum TE, Specks U, Colby TV, Edell ES, Brutinel MW, Prakash UB, et al. Tracheobronchial involvement in Wegener’s granulomatosis. Am J Respir Crit Care Med 1995;151:522–6. 3. Hernandez-Rodrı´guez J, Hoffman GS, Koening CL. Surgical interventions and local therapy for Wegener’s granulomatosis. Curr Opin Rheumatol 2010;22:29–36.

A. Audemard, MD B. Bienvenu, MD, PhD R. Magnier, MD L. Fournier, MD F. Galateau-Salle, MD, PhD N. Martin Silva, MD Caen University Hospital and University of Caen Caen, France