TY - JOUR T1 - Correspondence on “National reduction in cerebral arteriovenous malformation treatment correlated with increased rupture incidence” by Luther <em>et al</em> JF - Journal of NeuroInterventional Surgery JO - J NeuroIntervent Surg DO - 10.1136/jnis-2022-019866 SP - jnis-2022-019866 AU - Francis J Jareczek AU - Varun Padmanaban AU - David R Hallan AU - D Andrew Wilkinson Y1 - 2022/12/20 UR - http://jnis.bmj.com/content/early/2022/12/19/jnis-2022-019866.abstract N2 - It was with great interest that we read the paper by Luther et al1 reporting a decrease in intervention for unruptured brain arteriovenous malformations (AVMs) with a corresponding increase in the incidence of ruptured cerebral arteriovenous malformations after the publication of A Randomised trial of Unruptured Brain Arteriovenous malformations (ARUBA)2 in 2014. We commend the authors for their efforts to address an important clinical question using rigorous analytic methods. However, while the authors acknowledge the general limitations of administrative database studies, there are particular methodological limitations to studying brain AVMs and ARUBA via the Nationwide Inpatient Sample that bear mentioning—namely, the non-specificity of AVM codes, and the transition from ICD-9 (international classification of diseases, ninth revision) to ICD-10 in the US immediately following the publication of ARUBA.First, AVM diagnosis codes are particularly non-specific. In ICD-9, cerebral AVMs were included in the code 747.81, ‘anomalies of the cerebrovascular system’. As Cloft discussed previously,3 a study of patients with this code found that only 66% had a cerebral AVM,4 while the others had cavernous malformations (13%), unruptured cerebral … ER -