LAUSR

Background: Sphenopalatine ganglion (SPG) stimulation enhances collateral flow, stabilizes blood-brain barrier, and showed evidence of benefit in patients with confirmed cortical involvement (CCI) when started 8-24h after onset in the ImpACT-24B randomized trial. To characterize SPG stimulation benefit magnitude, we derived number needed to treat (NNT) values based on shifts over all levels of 3 month global disability. Methods: From the distribution of the 7-level modified Rankin Scale (mRS) at 3m in SPG- and sham-stimulation CCI patients, NNT to benefit (NNTB) and NNT to harm (NNTH) values were derived by automated (algorithmic min-max) and expert generation of joint outcome distribution tables. For dichotomized mRS outcomes, net NNT values were derived directly from absolute risk differences. Results: Among 520 patients with confirmed cortical infarction ineligible for thrombolysis, 244 were treated with SPG and 276 with sham stimulation. NNT values for dichotomized and shift mRS outcomes are shown in the Table. Of the 6 possible binary cutpoints on the mRS, 4 showed more favorable outcome with SPG stimulation. The dichotomized endpoint with the greatest group difference was ambulatory and capable of bodily self-care (mRS 0-3), 62.3% vs 51.1%, NNTB 8.9. Across all 6 individual possible dichotomizations of the mRS, the NNTB ranged from 8.9 to -166.7. For shifts by 1 or more levels across all 6 transitions of the mRS, the biologically most plausible NNTB was 5.7 (IQR 5.6-6.5), NNTH 34.5 (IQR 30.3-40.0), and net NNTB 6.8 (IQR 6.5-7.7), These values correlated closely with the automatically derived net NNTB of 5.9. Conclusions: The findings of this pivotal trial indicate that, out of every 1000 CCI patients treated with SPG stimulation, 146 patients will have a less disabled 3-month outcome, including 76 more who will be functionally independent. SPG stimulation can substantially improve the outcome of thrombolysis-ineligible acute ischemic stroke patients.