Stroke: a clinical prediction rule may help identify patients at high risk of dying

Clinical bottom line (level 2b)

  1. One in ten patients with an acute stroke die in the next 3 months
  2. The risk is increased with
    • impaired consciousness
    • urinary incontinence
    • dysphagia
    • low body temperature
    • hyperglycaemia without a history of diabetes
  3. A clinical prediction rule based on these 5 features may help identify patients at high risk of dying in the next 30 days.
Wang et al: Journal of Clinical Epidemiology 2001; 54 : 766-773
Expires November 2003

The study

Retrospective cohort study with objective outcomes, adjusted for confounding factors, validated in an independent set of patients.

Setting: teaching hospital, Australia

440 patients (aged mean 70, 57% male)

Excluded if
  • no clear final diagnosis or recorded as acute cerebrovascular accident
  • intracerebral haemorrhage or subarachnoid haemorrhage


Factors studied:
  • Age, sex, stroke risk factors, previous atherosclerotic disease, body temperature, consciousness level, dysphagia, urinary or faecal incontinence, body temperature
  • impaired consciousness
  • dysphagia
  • urinary incontinence
  • body temperature < 36.5 C
  • hyperglycaemia without a clinical history of diabetes



    A Cox regression analysis was used to identify independent prognostic factors for mortality.

    100% followed for 30 days
    Outcomes studied:
  • death
  • score 11 or more: death
  • score < 11: death
    • A clinical prediction rule was derived from 233 randomly-selected patients and then validated in the remaining 217 patients
      • impaired consciousness: 5
      • dysphagia: 3
      • urinary incontinence: 4
      • body temperature < 36.5 C: 2
      • hyperglycaemia but no history of diabetes: 2

    The evidence

    outcome time to outcome number of patients/total number %
    (95% CI)    		 NNF
    (95% CI)
    death 30 days 45/440 10.2%
    (7.4% to 13%)    		 10
    (8 to 14)
    score 11 or more: death 30 days 13/19 68%
    (48% to 89%)    		 1
    (1 to 2)
    score < 11: death 30 days 10/198 5.1%
    (2.0% to 8.1%)    		 20
    (12 to 50)

    prognostic factor for
    death     		time to outcome adjusted RR
    (95% CI)    		 NNF+
    (95% CI)
    impaired consciousness 30 days 4.6
    (2.3 to 9.6)     		3
    (1 to 8)
    dysphagia 30 days 3.1
    (1.8 to 5.6)     		5
    (2 to 12)
    urinary incontinence 30 days 4.2
    (2.4 to 7.3)     		3
    (2 to 7)
    body temperature < 36.5 C 30 days
    ( to )     		11
    (-98 to 3)
    hyperglycaemia without a clinical history of diabetes 30 days 1.9
    (1.2 to 2.9)     		11
    (5 to 49)

    Comments

    1. The clinical prediction rule needs to be prospectively validated in another population.

    Citation

    1. Wang Y, Lim LL, Levi C, et al: a prognostic index for 30-day mortality after stroke. Journal of Clinical Epidemiology 2001; 54 : 766-773
    Search Terms: from ACP Journal Club other articles noted
    Contributor: Chris Ball, November 2001
    Reviewer:

    Clinical Question.
    Patient acute stroke
    Intervention or Exposure clinical features
    Outcome death at 30 days