OBJECTIVE: Supplemented Spetzler-Martin grading (Supp-SM), which is the combination of Spetzler-Martin and Lawton-Young grades, was validated as being more accurate than stand-alone Spetzler-Martin grading, but an operability cutoff was not established. In this study, the authors surgically treated intermediate-grade AVMs to provide prognostic factors for neurological outcomes and to define AVMs at the boundary of operability. METHODS: Surgically treated Supp-SM intermediate-grade (5, 6, and 7) AVMs were analyzed from 2011 to 2018 at two medical centers. Worsened neurological outcomes were defined as increased modified Rankin Scale (mRS) scores on postoperative examinations. A second analysis of 2000-2011 data for Supp-SM grade 6 and 7 AVMs was performed to determine the subtypes with improved or unchanged outcomes. Patients were separated into three groups based on nidus size (S1: < 3 cm, S2: 3-6 cm, S3: > 6 cm) and age (A1: < 20 years, A2: 20-40 years, A3: > 40 years), followed by any combination of the combined supplemented grade: low risk (S1A1, S1A2, S2A1), intermediate risk (S2A2, S1A3, S3A1, or high risk (S3A3, S3A2, S2A3). RESULTS: Two hundred forty-six patients had intermediate Supp-SM grade AVMs. Of these patients, 102 had Supp-SM grade 5 (41.5%), 99 had Supp-SM grade 6 (40.2%), and 45 had Supp-SM grade 7 (18.3%). Significant differences in the proportions of patients with worse mRS scores at follow-up were found between the groups, with 24.5% (25/102) of patients in Supp-SM grade 5, 29.3% (29/99) in Supp-SM grade 6, and 57.8% (26/45) in Supp-SM grade 7 (p < 0.001). Patients with Supp-SM grade 7 AVMs had significantly increased odds of worse postoperative mRS scores (p < 0.001; OR 3.7, 95% CI 1.9-7.3). In the expanded cohort of 349 Supp-SM grade 6 AVM patients, a significantly higher proportion of older patients with larger Supp-SM grade 6 AVMs (grade 6+, 38.6%) had neurological deterioration than the others with Supp-SM grade 6 AVMs (22.9%, p = 0.02). Conversely, in an expanded cohort of 197 Supp-SM grade 7 AVM patients, a significantly lower proportion of younger patients with smaller Supp-SM grade 7 AVMs (grade 7-, 19%) had neurological deterioration than the others with Supp-SM grade 7 AVMs (44.9%, p = 0.01). CONCLUSIONS: Patients with Supp-SM grade 7 AVMs are at increased risk of worse postoperative neurological outcomes, making Supp-SM grade 6 an appropriate operability cutoff. However, young patients with small niduses in the low-risk Supp-SM grade 7 group (grade 7-) have favorable postoperative outcomes. Outcomes in Supp-SM grade 7 patients did not improve with surgeon experience, indicating that the operability boundary is a hard limit reflecting the complexity of high-grade AVMs.