Re: Confounding by Indication in Retrospective Studies of Intracerebral Hemorrhage: Antiepileptic Treatment and Mortalit
- PDF / 130,850 Bytes
- 2 Pages / 595.276 x 790.866 pts Page_size
- 75 Downloads / 221 Views
LETTER TO THE EDITOR
Re: Confounding by Indication in Retrospective Studies of Intracerebral Hemorrhage: Antiepileptic Treatment and Mortality Andrew M. Naidech • Matthew B. Maas • Eric M. Liotta • James C. Guth • Rebecca M. Bauer • Rajeev K. Garg • Stephan U. Schuele • Thomas P. Bleck
Published online: 3 January 2013 Ó Springer Science+Business Media New York 2012
To the Editor, We were interested to read the article by Battey and colleagues [1] regarding the analysis of datasets of patients with intracerebral hemorrhage (ICH). It is asserted that patients who survive at least 5 days are more suitable for analysis for the effect of seizure medications (now the preferred term [2]). This technique, however, is itself likely to lead to a biased dataset, and such a selection criterion is particularly inappropriate to evaluate seizure medications. The analysis does not take the duration of treatment or dose of seizure medications given into account, but it is likely to be important. One day of phenytoin (while a comatose patient undergoes EEG monitoring) is not likely to be the same as 14 days of phenytoin, which is why we took care to show that longer duration of treatment is associated with more complications in patients with ICH [2], and subsequent worse functional outcomes. More intense therapy is also associated with more complications and worse outcome in patients with subarachnoid hemorrhage [3]. Similar to the metric of pack-years for smoking exposure instead of a dichotomous one, analysis utilizing total dose exposure is important and fundamental to investigations of adverse drug effects. Furthermore, in this A. M. Naidech (&) M. B. Maas E. M. Liotta J. C. Guth S. U. Schuele Departments of Neurology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA e-mail: [email protected] R. M. Bauer Departments of Anesthesiology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA R. K. Garg T. P. Bleck Department of Neurology, Rush University Medical Center, Chicago, IL, USA
manuscript, seizure medications are grouped together, but each medication is likely to have a different safety profile and each may have a different effect on complications and outcomes. Five days after ICH symptom onset seems to be an arbitrary time to assess the impact of seizure medications. We [2] and others [4] have found that seizures after ICH usually occur within 24 h of symptom onset. Restricting the analysis to those who have survived at least 5 days will introduce bias by excluding patients most likely to have a seizure as a marker of disease, or who die as a consequence of seizures [5]. Mortality after ICH usually has a nonneurologic cause after a few days [6], so it is unclear why seizure medications should affect mortality. If one were to wait until 5 days after ICH symptom onset to randomize patients in a clinical trial of seizure medications, there would be relatively few events to study. The authors emphasize that their initial results suggested that seizure medications are li
Data Loading...
