Chest pain: clinicians were fairly good at predicting MIs.

Clinical bottom line (level 2b)

  1. A sixth of patients with chest pain had a myocardial infarction.
  2. A MI was more likely if:
    • previous MI or angina
    • pain radiates to right shoulder
    • history of sweating or nausea
    • constant pain
    • dull, pressure-like pain
    • patients is sweating
    • third heart sound
    • pulmonary crackles
    • any Q-wave on ECG
    • ST elevation or depression on ECG
    • new T wave inversion on ECG
    • new conduction defect on ECG
  3. An MI was less likely if:
    • sharp pain
    • pleuritic pain
    • chest wall tenderness
  4. Clinicians were fairly good at predicting which patients have MI.
Tierney et al: Medical Decision Making 1986; 6: 12-17
Expires July 2003

The study

Setting: acute hospital, USA

492 patients (aged mean 57 years, ?% male) anterior chest pain as a complaint

Excluded if
  • male <30 years old
  • female <40 years old



Independent blinded reference standard, applied in all patients from a non-consecutive appropriate spectrum.
Reference standard:
  • MI diagnosed if any of:
    • CK >upper limit of normal, and CK-MB >4% total CK or LDH 1 > (or equal too) LDH 2
    • ECG- new pathological Q waves (>40 ms duration and > third amplitude of following R wave)
Diagnostic test:
  • history and physical
  • cardiac enzymes
  • ECG
  • physician's prediction of MI

The evidence

pre-test probability of MI: 12%, (95% CI: 9.5% to 15%)

diagnostic test MI no MI LR
(95% CI)		 post-test probability
physician's prediction of MI 81-100% 16 20 5.7
(3.1 to 10) 		44%
physician's prediction 51-80% 14 24 4.1
(2.3 to 7.5) 		37%
physician's prediction 21-50% 8 40 1.4
(0.69 to 2.9) 		17%
physician's prediction 9 64 0.99
(0.52 to 1.89) 		12%
physician's prediction 0-5% 11 344 0.23
(0.13 to 0.39) 		3%
total 61 431


diagnostic test MI no MI LR+
(95% CI)		 post-test probability LR-
(95% CI)		 post-test probability
patient currently sweating 17 26 4.6
(2.7 to 8.0) 		40% 0.77
(0.66 to 0.90) 		10%
third heart sound 10 22 3.2
(1.6 to 6.5) 		31% 0.88
(0.79 to 0.99) 		11%
pulmonary crackles 23 76 2.1
(1.5 to 3.1) 		23% 0.76
(0.62 to 0.92) 		10%
chest wall tenderness 9 155 0.41
(0.22 to 0.76) 		5% 1.3
(1.2 to 1.5) 		16%
any Q wave 29 52 3.9
(2.7 to 5.7) 		36% 0.60
(0.47 to 0.76) 		8%
any ST elevation 35 22 11
(7.1 to 180) 		61% 0.45
(0.34 to 0.60) 		6%
any ST depression 43 95 3.2
(2.5 to 4.1) 		31% 0.38
(0.26 to 0.56) 		5%
any T wave inversion 32 121 1.9
(1.4 to 2.5) 		21% 0.66
(0.50 to 0.87) 		9%
any conduction defect 10 26 2.7
(1.4 to 5.4) 		28% 0.89
(0.79 to 1.00) 		11%
known angina 31 112 2.0
(1.5 to 2.6) 		22% 0.66
(0.51 to 0.86) 		9%
history of associated nausea 34 129 1.9
(1.4 to 2.4) 		21% 0.63
(0.47 to 0.84) 		8%
constant pain 39 164 1.7
(1.3 to 2.1) 		19% 0.58
(0.41 to 0.82) 		8%
any coronary artery disease 49 228 1.5
(1.3 to 1.8) 		18% 0.42
(0.25 to 0.70) 		6%
dull, pressure-like pain 33 151 1.5
(1.2 to 2.0) 		18% 0.71
(0.53 to 0.94) 		9%
sharp pain 10 177 0.40
(0.22 to 0.71) 		5% 1.4
(1.2 to 1.6) 		17%
pleuritic pain 2 91 0.16
(0.039 to 0.61) 		2% 1.2
(1.2 to 1.3) 		15%
left anterior chest pain 1 65 0.11
(0.015 to 0.77) 		2% 1.2
(1.1 to 1.2) 		14%
total 61 431

Comments

  1. Raw data not given in article- numbers calculated from proportions assuming information on each clinical finding was identified in all patients (not true).

Citation

  1. Tierney WM, et al: Physicians' estimates of the probability of myocardial infarction in the emergency room patients with chest pain. Medical Decision Making 1986; 6: 12-17
Search Terms: bibliographic reference in 'Evidence-based Cardiology', publ. BMJ Publishing 1998
Contributor: Chris Ball and Musab Hayatli, July 2000
Reviewer: Kenneth Ballew

Clinical Question.
    Patient chest pain
    Intervention or Exposure clinical signs and symptoms
    Outcome diagnosis of MI