RT Journal
A1 Xian Y, Liang L, Peterson ED
T1 BLeeding risk with ischemic stroke therapy—reply
JF JAMA
JO JAMA
YR 2012
FD October 3
VO 308
IS 13
SP 1318
OP 1318
DO 10.1001/2012.jama.11667
UL http://dx.doi.org/10.1001/2012.jama.11667
AB 
In Reply:  Dr Kiechl and colleagues raise concerns that patients treated with warfarin in our study may have had lower bleeding risk (ie, a favorable treatment selection bias). Our data, however, do not support this assumption. Among all patients with stroke treated with warfarin prior to hospitalization, those treated with tPA actually had more severe strokes (median National Institutes of Health stroke scale scores: 14 vs 9) and higher predicted risk of sICH (median predicted sICH rate: 5.0% vs 3.1%) (both P <.001). This may explain why our results differed from prior smaller studies if those studies did not fully account for this adverse treatment selection. While we could not assess individual physicians' experience with the data, our model accounted for hospital-level clustering. Furthermore, the overall sICH rate in patients treated with warfarin (5.7%) is comparable with rates reported in nonselective patient populations outside of randomized controlled trials (5.2%).1 Thus, it is unlikely that patient or hospital-level treatment selection biased the results.