Media story Full headline: Tonix Pharmaceuticals Completes Clinical Phase of PREVAIL Proof-of-Concept Study of TNX-102 SL for the Treatment of Fibromyalgia-Type Long COVID https://finance.yahoo.com/news/tonix-pharmaceuticals-completes-clinical-phase-110000281.html
Interesting that this other trial also involves a muscarinic receptor antagonist: Winsantor Phase 2 small fiber neuropathy drug (WST-057) trial news.
Recent trial of this SL form of muscle relaxant cyclobenzaprine (Flexeril) for FM. https://pubmed.ncbi.nlm.nih.gov/37165930/
Here's the abstract: Abstract Objective: To evaluate the efficacy and safety of TNX-102 SL, a once-nightly sublingual formulation of cyclobenzaprine, in reducing pain in patients with fibromyalgia (FM). Methods: RELIEF was a double-blind, randomized, placebo-controlled trial. Overall, 503 patients received TNX-102 SL 2.8 mg for 2 weeks, followed by 5.6 mg for 12 weeks (248 patients), or matching placebo (255 patients). The primary end point was change from baseline at week 14 in the weekly average of daily pain scores. Secondary end points included Patient Global Impression of Change (PGIC) scores, Fibromyalgia Impact Questionnaire Revised (FIQR) scores, Patient-Reported Outcomes Measurement Information System (PROMIS) Sleep Disturbance and Fatigue scores, and daily sleep quality. Safety was assessed by adverse event (AE) reporting. Results: Reduction in daily pain from baseline at week 14 was significantly greater with TNX-102 SL (least squares [LS] mean change -1.9 [95% confidence interval (95% CI) -2.1, -1.7]) versus placebo (LS mean change -1.5 [95% CI -1.7, -1.3]; P = 0.01). TNX-102 SL was not associated with significant improvement in PGIC at week 14 but was associated with improvements in FIQR scores, PROMIS scores, and daily sleep quality. Overall, 59.7% of patients receiving TNX-102 SL and 46.3% receiving placebo reported treatment-emergent AEs; the most common were oral hypoesthesia (17.3% with TNX-102 SL versus 0.4% with placebo), oral paresthesia (5.6% versus 0.4%, respectively), and product taste abnormal (4.4% versus 0.4%, respectively). Conclusion: In this phase III, randomized, controlled trial of patients with FM, treatment with TNX-102 SL was associated with significant reductions in daily pain and was safe and well tolerated. Trial registration: ClinicalTrials.gov NCT04172831.
So they report the mean decrease over time in both groups rather than comparing the means of the groups directly. To me that suggests they are trying to hide a null result [EDIT: but would be interesting to hear what others think: is this analysis justifiable?]
I'm not a statistician. But, to me, the level of matching at baseline is important when reporting on the difference in mean change. If the placebo group was substantially worse at baseline, then we might think that a decrease in that group might have a lot of 'reversion to the mean' impacting on the result. If the mean and variance at baseline is similar, then maybe it doesn't matter if the comparison is between the final results of the two groups or the change relative to baseline of the two groups. Table 1 tells us that the baselines of the primary outcome, Diary Pain Scores, in the two groups were very similar (mean +- standard deviation): Treatment group: 6.1 +-1.1 Placebo group: 6.0 +-1.0 That looks good. There were large numbers of participants in each group, so that increases confidence in the result too. The reported change in the mean scores is substantially larger than the baseline difference in the groups (-1.9 in the treatment group; -1.51 in the placebo group). 95% confidence intervals around those final results are reported. Figure 2 with the changes over time does show a clear difference between the two groups. They did use standard error for the error bars, which increases the appearance of separation of the two group's results, but it probably does make for a clearer chart. I haven't read the rest of the paper in detail, but I don't think the researchers had a null result to hide. The primary outcome was a subjective PROM, so the effectiveness of the double blinding might be something to look at; it looks as though the drug did cause some changed mouth sensations in some people. There is one issue with delaying the final assessment: That might introduce a point where investigator bias could affect recorded outcome, and it reduces the chance of adverse effects of the drug affecting outcomes. But the interim results are consistent with the final outcomes. I haven't read the paper closely, but I thought the study looks to have been done carefully and without over-claiming.
Thanks. I think I also misread as the p-value of 0,01 does seem to refer to a difference between groups.
