Michelle Christina Johansen, MD

Yeh SJ, Tang SC, Tsai LK, and Jeng JS. Pathogenetical Subtypes of Recurrent Intracerebral Hemorrhage: Designations by SMASH-U Classification System. Stroke. 2014

Intracerebral hemorrhage (ICH) composes a high percentage of admissions to neurointensive care units and they remain a major cause of mortality. Not all hemorrhages are created equal and Dr. Yeh et.al investigate the data surrounding different bleed subtypes in an effort to offer insight into functional outcomes and mortality. Their study employed the SMASH-U classification method which separates ICH etiologies into Structural, Medication related, Amyloid Angiopathy, Systemic Disease, Hypertension and Undetermined. They focused much discussion on the two most common etiologies, Hypertension and Amyloid. They state that the approaches to these distinct patient populations differs and that other classification systems, such as those that use location of bleed are not sufficient to distinguish between these two groups.



The researchers took advantage of their access to the National Taiwan University Hospital Stroke Registry which documents all patients who had a stroke (based on head CT and ICD codes) within two weeks of admission. They obtained data from 4,578 acute ICH patients, classified the patients into the six SMASH-U types and then analyzed the outcomes of first ever ICH cases versus recurrent ICH cases similarly stratified. They excluded trauma or tumor related, subdural/epidural/subarachnoid hemorrhage and stroke hemorrhagic transformation resulting in thorough examination of 3,785 cases. Two raters independently assigned classifications and if after discussion no consensus was reached, they were labeled Undetermined. The bleed was designated Structural if there were vascular lesions in the area of ICH; Medication related if INR≥2, new oral anticoagulant within three days or use of heparin/thrombolytic agent; Amyloid related if the ICH met Boston Criteria; Systemic disease related if there were evidence of thrombocytopenia, liver cirrhosis or non-medication induced coagulopathy and Hypertension induced according to guidelines laid out by the original SMASH-U paper. Notably the investigators added renal failure in consideration of systemic disease. 

The most common etiologies were hypertension and amyloid followed by systemic disease, undetermined, structural and medication related. Of the 185 cases of recurrent ICH classified, the etiology was different in 34. Among the 44 amyloid cases, 31 recurred with the same etiology but 10 were classified as hypertension related. Seventy eight of the 93 recurrent hypertension classified ICH were the same etiology but 7 were re-classified as amyloid related. The main reason for reclassification appears to be the location of the bleed (lobar versus deep structures). The authors appropriately point out that location alone is not sufficient to determine etiology and that amyloid angiopathy can occur in younger patients calling into question the Boston classification scheme. 

The importance of accurate classification is not only deciding treatment but also reflects patient survival.  In evaluation of results, those with Systemic Disease related (51-62%) or Medication induced (49-60%) had worse survival while those with Structural etiology (5-6%) had the best survival.  Interestingly, this article suggests that those with amyloid related hemorrhages (17-27%) fared worse than those with hypertension induced ICH (12-18%) as contrasted to the similar survival curve demonstrated in the SMASH-U study by Meretoja et al. 

While confounders such as the addition of renal failure to systemic disease must be acknowledged, there are many considerations raised for the treating neurologist when assessing this data. Our first year of training is usually dedicated to learning general medicine and it is all too easy to forget this realm with increasing specialization. The importance of appropriate recognition and treatment of comorbid illness such as ITP or cirrhosis is only confirmed by this data. The worse survival curve in patients with amyloid is also something to ponder. How does one decide the appropriate treatment for a patient with Afib and suspected amyloid angiopathy? In the era of rapid expansion and development of oral anticoagulants, many with unestablished methods of reversal, does this preclude the above patients from these medications? While ICH unfortunately remains a staple admission to neurologic intensive units, perhaps consideration of etiology will lead to appropriate and more accurate therapy and assessment of risk.