Arthritis & Rheumatism, Volume 60,
October 2009 Abstract Supplement

The 2009 ACR/ARHP Annual Scientific Meeting
Philadelphia October 16-21, 2009.

Anxiety and Depression Subselects Fibromyalgia Patients: A Cluster Analysis with Treatment Implications

de Souza1,  Juliana Barcellos, Marchand2,  Serge, Ware3,  Mark A., Shir3,  Yoram, Fitzcharles4,  MA

University of Sherbrooke, QC,
Univ, QC,
McGill University, Montreal, QC,
MGH, Montreal, QC


Successful management of fibromyalgia (FM) remains a challenge. Identification of parameters that could direct treatment approaches would be advantageous. We have previously identified 2 patient subgroups, without or with comorbid depression and anxiety, by hierarchical cluster analysis using the fibromyalgia impact questionnaire (FIQ), a well validated quality of life (QOL) assessment for FM. In the present study we report the validation of these subgroupings on a large population of FM patients in a tertiary care setting.


The FIQs for 132 FM patients attending a tertiary care multidisciplinary pain clinic were analysed as follows: (A) patients were assigned to the respective cluster, FM-Type I (without depression and anxiety) and FM-Type II (with depression and anxiety), using the classification coefficient previously published (Souza et al., Rheum. Int. 2008) and; (B) the cluster analysis was reapplied with this new sample. To confirm the number of clusters with this new cluster analysis (B), we analysed progressive changes in the agglomeration coefficient. Positive and negative predictive values for clustering according to mood disorder, as well as specificity and sensitivity for both models were calculated (A and B). Additional measures of pain and QOL included the McGill Pain Questionnaire (MPQ), Health Assessment Questionnaire (HAQ), Pain Catastrophizing Scale (PCS) and Pain Disability Index (PDI)


Tertiary care FM patients were grouped into Type-I (n=28) and Type-II (n=104) groups with Type-II patients demonstrating higher values for pain, fatigue, stiffness, morning tiredness, anxiety, depression on the FIQ. Repeat analysis (B), identified 2 clusters dependent on mood. The inter-model analysis showed a sensitivity of 0.96, and specificity of 0.71; with positive predictive value of 0.41 and negative predictive value of 0.99. Type-I vs Type-II did not differ regarding duration of pain (p=0.90) or tender point count (p=0.11), but Type II were younger (47 vs 53 yrs p<0.01), and reported higher values for MPQ, HAQ, PCS and PDI (all p<0.01).


We have shown that the clustering of FM patients remained valid in a tertiary care setting, often considered to represent patients with more severe symptoms. There is therefore little doubt that different FM profiles exist, and may be identified by a single, comprehensive instrument, the FIQ. These clusters groups have implications regarding treatment approaches. Future studies should take these different patient groups into consideration in order to focus treatment interventions applicable to distinct patient groups.

To cite this abstract, please use the following information:
de Souza, Juliana Barcellos, Marchand, Serge, Ware, Mark A., Shir, Yoram, Fitzcharles, MA; Anxiety and Depression Subselects Fibromyalgia Patients: A Cluster Analysis with Treatment Implications [abstract]. Arthritis Rheum 2009;60 Suppl 10 :1403
DOI: 10.1002/art.26477

Abstract Supplement

Meeting Menu