Loading...
Predicting neurological prognosis after cardiac arrest remains an area of uncertainty for many neurologists and intensivists. Although several tools may be helpful — including the clinical exam, neurophysiology, and serum/CSF biomarkers — these have their limitations and usually are helpful only in the rare cases of clear hopelessness or clear rapid improvement.
To examine the feasibility and predictive potential of MRI for more uncertain cases, researchers prospectively enrolled 51 cardiac arrest survivors who underwent diffusion-weighted MRI within 7 days after cardiac arrest. The authors retrospectively identified apparent diffusion coefficient thresholds associated with three predefined outcome categories:
Group 1: death or vegetative state at 6 months, along with (a) the absence of motor response, pupillary reflexes, or bilateral cortical responses at 72 hours, or (b) vegetative state at 1 month
Group 2a: survival with good neurological outcome at 6 months
Group 2b: survival with poor neurological outcome at 6 months
The authors conclude that, if these findings are confirmed, MRI could be a useful adjunct to other prognostic tools in comatose survivors of cardiac arrest.
Wijman CAC et al. Prognostic value of brain diffusion-weighted imaging after cardiac arrest. Ann Neurol 2009 Apr; 65:394.
Comment
Most studies of neuroimaging for predicting neurological outcome of cardiac arrest have been hindered by small size and poor methodology. Unfortunately, this study also has several flaws. The numerical ordering of the groups is not intuitive, and each group is small (e.g., 13 patients in group 2a and 6 patients in group 2b). Thus, although MRI appears to aid in predicting survival, whether imaging characteristics can be identified that can separate those who just survive from those who survive with meaningful neurological improvement remains to be seen.
The article contains lengthy descriptions of four cases with neuroimaging, an inclusion I find surprising, as it does not provide any statistical weight or importance. Furthermore, the study has several sources of potential bias: Data are sparse regarding the neurological examinations; coma was inadequately defined as “no eye opening to voice and inability to follow commands,” and, most importantly, a self-fulfilling prophecy of early withdrawal of care likely biased the results, as in other studies. Although treating clinicians were encouraged to follow the 2006 American Academy of Neurology practice parameter when considering when to withdraw care, the manuscript leaves unclear whether this was followed. More important, for patients in whom these strict criteria were not met, it is unclear how long they were allowed to live. The authors rightfully suggest that MRI should not be the only tool used for prognostication in patients with cardiac arrest, and they encourage further study. Fortunately, a more recent study of 80 truly comatose cardiac arrest patients has been conducted (Radiology 2009; May 6).