Purpose Ochsner Clinic Foundation in New Orleans, LA is a certified comprehensive stroke center which includes a hub and spoke type of telestroke system with over 30 referral hospitals. The purpose of this study is to determine the incidence, possible correlative factors, and clinical significance of different types of hemorrhages in patients who are being considered for and treated with mechanical thrombectomy (MT).

Materials and methods 195 consecutive patients were retrospectively analyzed from 1/1/12–7/1/14. Inclusion criteria were presentation with acute stroke signs, CT perfusion/CT angiography showing proximal MCA/intracranial ICA occlusion. Four patients had no follow-up imaging. An interventional neuroradiologist (JM) evaluated imaging for the initial presence of basal ganglia involvement on presenting CT or CTP. Hemorrhages were subgrouped by type (HI1, HI2, PH1, PH2). Combined groups of HI and PH were used when greater sample size was required. Risk factors and presenting signs/symptoms were obtained by retrospective chart review. Discharge and follow-up modified Rankin Scale (mRS) and the National Institute of Health Stroke Scale (NIHSS) were also found on chart review. Evaluation was performed using linear and logistic regressions.

Results The overall rate of hemorrhage was 53/191, including HI1 (n = 16), HI2 (n = 15), PH1 (n = 7), PH2 (n = 13), and no hemorrhage (n = 138). Thrombectomy was associated with an increased rate of HI (OR 2.3, p = 0.03). Although the rate of PH in MT patients (11/72) was greater than the rate of PH in nonthrombectomy patients (9/123), this was not significant (p = 0.09). The rates of PH2 were 8/72 (thrombectomy) and 5/123 (nonthrombectomy).

The following correlated with HI in MT patients: transfer from another institution (OR = 1.3, p = 0.028), dysarthria as presenting symptom (OR = 1.4, p = 0.01), obesity (OR = 2.2, p = 0.0003), ETOH abuse (OR = 1.4, p = 0.012), and BG involvement (OR = 1.35, p = 0.003). In all patients combined, basal ganglia involvement was associated with HI (OR = 6.5, p = 0.009), PH (OR = 10.1, p = 0.034), and death (OR = 5.3, p = 0.037).

The presence of HI in MT patients has a statistically significant correlation with higher 90 day NIHSS (p = 0.03). PH correlates with poor outcomes in MT patients on both 90 day NIHSS (p = 0.01) and 90 day mRS (p = 0.008). BG involvement did not have a statistically significant correlation with 90 day NIHSS (p = 0.5), 90 day mRS (p = 0.46), or death (OR = 1.3, p = 0.77) in MT patients. BG involvement did have statistically significant correlations with 90 NIHSS (p = 0.04) and 90 day mRS (p = 0.006) in patients treated with TPA alone.

Conclusion MT is associated with an increased rate of HI hemorrhages but not PH hemorrhages. HI hemorrhages have no correlation with death or mRS. mRS may be too insensitive to detect deficits relating to HI hemorrhages, NIHSS should be used in studies investigating clinical outcomes relating to HI hemorrhages. BG involvement may correlate with both HI and PH hemorrhages in patients treated with and without MT. However, having BG involvement does not seem to correlate with a worse prognosis in MT patients with any type of hemorrhage.

Disclosures G. Bennett: None. M. Al Hasan: None. G. Vidal: 3; C; Penumbra. Q. Luo: None. J. Milburn: None.