Use of diagnostic subtraction angiography for ischemic stroke (US DUTCH study) Regional variation and time-trend among medicare beneficiaries

Introduction: There are no guideline recommendations for DSA in the ischemic stroke work-up. We studied the rate of DSA in ischemic stroke, the recent time-trend, hospital variation and associated factors. Methods: This is a retrospective cross-sectional study among Medicare fee-for-service beneficiaries with ischemic stroke admitted between 2016 and 2020 in the United States. ICD-10 codes were used to determine ischemic stroke diagnosis and procedure codes for thrombectomy and DSA. Hospital trends and factors associated with DSA performance were analyzed in hospitals with DSA capacity. Results: 7.373 (0.7 %) of the 1,085,644 ischemic stroke patients, had a DSA for diagnostic purposes. In the patients that were admitted to a hospital with DSA facility, the following factors showed the strongest association with DSA: younger age (aOR=0.81 [95 % confidence interval (CI):0.81-0.83]), thrombectomy rate in that hospital (aOR=2549 [95 %CI:610-10663]), transfer (aOR=1.41[95 %CI:1.34-1.50]) and carotid disease (aOR=5.8 [95 %CI:5.6-6.1]). There was large variation in the hospital DSA rate, varying from 0.07 % to 11.1 %. Of the variance of DSA rates, 15 % was attributed to the residual effect hospital propensity to perform DSA. The top decile of hospitals with the highest DSA rate, performed DSA's in >2.3 % of patients, compared to the 0.6 % median. There was no change in DSA rates over time. Conclusion: DSA is used infrequently in acute ischemic stroke patients and did not change between 2016 to 2020. Hospital variation in DSA use was however large, and not solely explained by patient and facility factors.