Prognosis in patients with severe stroke admitted to the neurological ICU is poor. The risk of mortality can be determined from clinical information available shortly after admission. A conclusion of the paper is that intubation with mechanical ventilation should be done quickly in patients with severe stroke before irreversible damage occurs.
Survival in patient population admitted to a stroke ICU:
(1) survival rate at 1 year was 33.1%
(2) about 50% of patients died in first 10 days
(3) about 60% of patients had died by 3 weeks
(4) after 2 months causes for mortality other than stroke became more important
(5) mortality was highest in the elderly patient who was comatose on admission and who required ventilation for neurological or respiratory deterioration
Variables found to be independent predictors of poor prognosis in multivariate analysis:
(1) age of patient
(2) presence of coma, as indicated by a low Glasgow coma score
(3) indication for intubation (elective for surgical or diagnostic procedure vs need to maintain oxygenation due to coma or respiratory failure)
Variable |
Use |
Value |
Regression Coefficient |
Odds Ratio |
age |
<= 65 |
0 |
0 |
|
|
> 65 |
1 |
0.9252 |
2.53 |
Glasgow coma score |
>= 10 |
0 |
0 |
|
|
< 10 |
1 |
1.1811 |
3.26 |
indication for intubation |
elective for procedure |
0 |
0 |
|
|
necessitated by coma or acute respiratory failure |
1 |
0.9358 |
2.55 |
Assumptions:
(1) Since the regression coefficients are given, it is assumed that the authors used an equation similar to the following:
logit = sum(regression of coefficients)
probability of mortality = EXP(logit) / (1 + EXP(logit))
(2) A constant is commonly included in the logit expression, but none is mentioned in the paper.
Limitations:
• Survival data for the different patient populations with each combination of risk variables would have been informative, especially showing the risk if none of the 3 factors identified were present.
Specialty: Neurology
ICD-10: ,