Ultrasound-Guided Greater Occipital Nerve Block Across Headache Phenotypes: Outcomes and Determinants of Response in a Prospective Cohort

Giglio, M., Corriero, A., Tramacere, C. et al. Ultrasound-Guided Greater Occipital Nerve Block Across Headache Phenotypes: Outcomes and Determinants of Response in a Prospective Cohort. Pain Ther (2026). https://doi.org/10.1007/s40122-026-00881-4

Abstract

Introduction

Greater occipital nerve blocks (GONBs) are used across headache phenotypes, but short-term effectiveness and predictors of response in mixed real-world cohorts remain uncertain.

Methods

We conducted a prospective observational study of adults with chronic headache phenotypes receiving a single ultrasound-guided proximal (C2) GONB. Primary outcomes were T0 → T1 (1-month) changes in pain (NRS), headache impact (HIT-6), sleep (PSQI), and monthly attacks (crises) using paired Wilcoxon tests. We defined HIT-6 responders as ≥ 10-point reduction. Predictors of response were evaluated with multivariable logistic regression; continuous reductions were modeled with multivariable linear regression.

Results

Twenty-nine patients were included (chronic migraine n = 10; cluster n = 5; occipital neuralgia n = 7; cervicogenic n = 7; 69% female; mean age 58.2 ± 13.7 years). At 1 month, all outcomes had improved significantly: HIT-6 median 66 → 52 (Δ − 12; 95% CI − 15 to − 8; p < 0.001), PSQI 12 → 8 (Δ − 3.5; p < 0.001), monthly (crises) 24 → 12 (Δ − 14.0; 95% CI − 20.0 to − 10.0; p < 0.001), and NRS 8 → 4 (Δ − 3; p < 0.001). In the logistic model, higher baseline pain independently predicted HIT-6 response (OR 3.38, 95% CI 1.22–13.47; p = 0.017), whereas symptom duration was not associated. In multivariable linear models, cluster headache was associated with greater improvements (ΔNRS, ΔHIT-6, Δcrises), while a rheumatic comorbidity predicted smaller improvements across outcomes.

Conclusion

Across chronic headache phenotypes, a single ultrasound-guided proximal (C2) GONB was associated with clinically meaningful 1-month improvements in pain, impact, sleep, and attack frequency. Response, however, was not uniform: higher baseline pain intensity and the cluster phenotype identified the patients most likely to benefit, whereas a rheumatic comorbidity marked probable non-responders. These predictors offer a practical basis for patient selection and should be confirmed prospectively.