Ab0139 What Types of Fibromyalgia and Chronic Pain Patients Benefit From Multimodal Inpatient Pain Programs?

Background: Overlapping chronic musculoskeletal pain syndromes including fibromyalgia are heterogeneous and often treatment resistant entities carrying significant socio-economic burden. Multimodal treatment programs are recommended for refractory cases, but available places are limited and it is unclear which type of patient benefits best from such a program. Objectives: To predict the response of patients with refractory musculoskeletal pain syndromes and fibromyalgia to a 2-week multimodal pain program through an extensive set of clinical variables including immunological, psychosomatic, wearable and sleep biomarkers. Methods: Data were collected during a 2-week multimodal inpatient treatment program involving 202 patients. 78% of the patients fulfilled the criteria for fibromyalgia, 77% had a concomitant psychiatric- and 22% a concomitant rheumatic immune-mediated disorder (IMRD). The program consisted of a rheumatologic, psychiatric and pain-specialist assessment, sleep analysis via actigraphy, as well as intensive physiotherapy, cognitive behavior therapy, ergotherapy, chiropractic therapy, balneotherapy and relaxation exercises via virtual reality. Five patient phenotypes were identified using hierarchical agglomerative clustering. Subsequently, both linear and polynomial regression models were independently fitted to predict changes in the Brief Pain Inventory (BPI) Impact Score and the BPI Intensity Score following the two-week program. Results: The best-performing final model was a polynomial regression with variables related to social context (age, living in a couple, receiving disability benefits) and medications (strong vs. weak opioids, tricyclic/ tetracyclic antidepressants, mirtazapine, trazodone or neuroleptics) for predicting BPI impact. Linear regression showed a negative association of Trazodone (p=0.003), higher age (p=0.092) and strong opioids (p=0.166). Conversely, weak opioids (p=0.079), received disability benefits (p=0.146) and tricyclic/tetracyclic antidepressants (p=0.133) showed a trend towards a positive outcome. Suffering from a concomitant IMRD has no influence on the outcome of the program. The performance of the final model showed a Mean Squared Error (MSE) of 0.016, Root Mean Squared Error (RMSE) 0.127 and R^2 0.39. The poorer response in patients receiving Trazodone or strong opioids or in those with a higher age was in line with the clustering analysis. Here, the best outcomes on BPI impact after two weeks were observed in a cluster characterized by obese female individuals who were either peri- or postmenopausal, had metabolic comorbidities, lived alone, and suffered from severe sleep impairment while receiving antidepressant therapy. Another cluster, comprising younger individuals with a low BMI, alexithymia, and psychiatric conditions (primarily depression and/or anxiety), showed a significant response to virtual reality interventions during the program. Conclusion: Our models indicate that social context, along with psychotropic and opioid medication are key determinants in the short-term response to multimodal inpatient pain programs. However, the full spectrum of benefits from multimodal treatments may extend beyond pain relief and are not fully measurable within a two-week timeframe. REFERENCES: NIL. Acknowledgements: NIL. Disclosure of Interests: Thomas Hügle Roche, Novartis, Abbvie, Janssen, BMS, Galapagos, Atreon, Vtuls, Tiffany Pretat: None declared, Christopher Lovejoy: None declared, Pedro Ming Azevedo: None declared.