Prediction tool indicates prognosis after cardiac arrest

Jul 9 2012

By Sarah Guy

US researchers report the development of a prediction tool that could help determine neurologic prognosis after in-hospital cardiac arrest.

Using data for over 40,000 patients who were successfully resuscitated after an in-hospital arrest, the team developed the model, which showed excellent discrimination between patients who had favorable neurologic survival, and those who did not.

The tool encompasses pre-, during-, and postarrest variables such as age, hospital location, arrest rhythm, asystole, mechanical ventilation, and sepsis.

It also allows patients and their families to make "informed treatment decisions that are most aligned with the patient's goals and values," say Paul Chan (Mid America Heart Institute, Kansas City, Missouri) and colleagues.

As reported in the Archives of Internal Medicine, the team identified 42,957 patients from 551 hospitals registered in the Get With the Guidelines Resuscitation registry who survived cardiac arrest as inpatients.

Favorable neurologic survival at discharge - the primary outcome of the study - was defined as survival without severe neurologic disability or a cerebral performance category score of 1 or 2 (indicating no-to-mild or moderate disability).

The researchers divided the cohort into derivation (n=28,629) and validation (n=14,328) subgroups, which had similar rates of neurologic survival.

After adjustment for potential confounders and exclusion of predictors that did not improve discrimination, the final predictive model included 11 variables with an overall C statistic of 0.802 (where 1=perfect discrimination).

The three variables with the greatest predictive ability (indicated by the total t statistic) were initial cardiac arrest rhythm (57.1), code duration (67.6), and pre-arrest neurologic status (51.2).

Chan and co-investigators used these results to develop the Cardiac Arrest Survival Postresuscitation In-hospital (CAPSRI) score, with higher scores for each individual predictor denoting a decreased likelihood of favorable neurologic survival.

Indeed, patients in the top decile (total CAPSRI score <10) had a 70.7% mean probability of favorable survival, compared with a 2.8% mean probability among patients in the bottom decile (CAPSRI ≥28).

"We believe this tool is simple to use, addresses a critical unmet need for better prognostication after cardiac arrest, and has the potential to enhance communication with patients and families," conclude Chan et al.

Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.