Background/Purpose: Milnacipran (MLN) has demonstrated its benefit in treating patients with fibromyalgia. Fibromyalgia is a persistent pain condition that aggregates numerous symptoms (pain, fatigue, sleep disturbance, quality of life impairment…). The aim of this study is to find patient baseline characteristics in fibromyalgia clinical trials that influence MLN treatment effect.

Methods: Two methods were used to address this problem: the first is knowledge-oriented and consisted of post-hoc analyses describing the relationships between the clinical outcome and baseline factors. The second is a data mining analysis which takes into consideration relationships between the outcome and the candidate influencing factors. Analysis was performed using KEM (Knowledge Management and Extraction) algorithm. Data were analyzed from phase 3 placebo-controlled clinical trials. In total, 75 variables including total scores, sub-scores and items of baseline scales and demography were selected as candidate influencing factors. Continuous variables were categorized into 3 categories defined according to the 33.3% and 66.6% percentage limits of values in the whole sample. Outcomes were the response on the composite criterion, improvement in both pain, and PGIC and on the two components separately. Treatment effect was measured using odds-ratio.

Results:

More than 2500 patients were investigated from three clinical phase III clinical trials (FMS-031, MD-02 and GE-302). Different variables corresponding to potential patient profiles led to a significant increase in odds-ratio measuring the treatment effect versus placebo. For the 100mg/d dosage, 8 variables were identified which increase the odds-ratio beyond 2. The most three relevant variables are “unable to work due to FMS” OR=2.63 [2.15,3.19], “MFI-General Fatigue” OR=2.22 [1.81,2.72], and “FIQ-Physical functioning” OR=2.21 [1.80,2.70]. For the 200mg/d dosage, 18 such variables were identified. The most three relevant are “FIQ-Stiffness” OR=3.45 [2.78,4.30], “VAS-Pain(CRF)” OR=3.12 [2.50,3.90], and “FMS Duration” OR=2.99 [2.39,3.74].

For all the identified variables, the odds-ratio increase is coherent across the composite responder criterion and the two components Pain and PGIC. The identification of “fibromyalgia duration” as influencing factor in the exploratory KEM analysis was confirmed in a post-hoc analysis that compared the response rates in three subgroups: short, medium and long fibromyalgia duration. This analysis evidenced higher response rates and a higher odds-ratio in the “medium fibromyalgia duration” group. Furthermore, the “high fibromyalgia duration” group seemed to benefit more from the 200mg/d dosage.

Conclusion: KEM analysis allowed the identification of baseline characteristics associated with higher odd-ratios suggesting a greater treatment effect. Thus patient profiling can be very useful for clinicians to individualize drug therapy and advice. The coherence across the primary endpoint and its two components confirmed the robustness of the primary endpoint in the milnacipran fibromyalgia development program.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/which-baseline-characteristics-influence-the-response-to-milnacipran-in-patient-with-fibromyalgia/