Author + information
- Manoj Thangam1,
- Brian Hsi1,
- Yelin Zhao2,
- Mark Warner1,
- Salman Arain3,
- Konstantinos Charitakis4,
- Abhijeet Dhoble5,
- Nils Johnson6,
- Richard Smalling7,
- H.V. Anderson8,
- Pratik Doshi1 and
- Prakash Balan9
- 1University of Texas Medical School at Houston, Houston, Texas, United States
- 2University of Texas Health Science Center at Houston, houston, Texas, United States
- 3University of Texas Health Sciences Center at Houston, Houston, Texas, United States
- 4UTH, Houston, Texas, United States
- 5University of Texas Health Science Center, Houston, Texas, United States
- 6McGovern Medical School at UTHealth, Houston, Texas, United States
- 7UTHealth/Memorial Hermann Heart and Vascular Institute, Houston, Texas, United States
- 8University of Texas McGovern Medical School, Houston, Texas, United States
- 9University of Texas Health and Science Center, Houston, Texas, United States
Cardiac arrest (CA) is associated with high mortality. Currently, no validated method for predicting mortality post CA exists. A prognostication tool may improve medical decisions and resource allocation. We created and validated a novel scoring system (Cardiac Arrest Survival Score) for predicting survival to discharge post CA.
A retrospective evaluation of a large metropolitan CA registry (n=3952) was performed. The group was randomly divided into a development (cohort A, n=2635) or test (cohort B, n=1317) dataset in a 2:1 ratio. Univariate analysis (UA) examined baseline parameters in relation to mortality. Variables with p-value <0.2 underwent multivariate analysis (MA). A weighted score was formed by doubling each odds ratio (OR) rounded by 0.5. The sum of weighted values equals CASS.
UA of Cohort A revealed 5 significant markers: age > 75, unwitnessed arrest, home arrest, delayed CPR, and non-shockable rhythm. MA confirmed significance of all 5 factors. The area under the ROC for cohort A was 0.7193 and the goodness-of-fit test showed a p-value of 0.3599 . The OR and weighted points for each identified variable are shown in Table 1. The weighted score showed similar predictability in the both cohorts. These data were used to generate an in-hospital mortality curve correlating with CASS scores (Graph 1).
|Variable||OR||Standard Error||p-value||95% Confidence Interval||Points|
|Age > 75||1.606||0.175||< 0.001||1.298 - 1.988||3|
|Unwitnessed Arrest||1.947||0.208||< 0.001||1.579 - 2.401||4|
|Home Arrest||1.278||0.119||0.008||1.065 - 1.533||2.5|
|Delayed CPR||1.352||0.135||0.003||1.112 - 1.645||2.5|
|Non-Shockable Rhythm||3.813||0.350||< 0.001||3.186 - 4.564||8|
A novel score for predicting in-hospital mortality post ROSC after cardiac arrest was created and validated using a large registry.
CORONARY: Acute Myocardial Infarction