Steinhart et al developed a model for predicting acute heart failure. This can help to identify a patient who may benefit from more aggressive management. The authors are from Saint Michael's Hospital (Toronto), University of Toronto and Massachusetts General Hospital.



(1) pre-test probability of heart failure as a decimal fraction

(2) NT-proBNP in pg/mL

(3) age in years


X =

= (5.9 * (pretest probability)) - (0.011 * (age in years)) + (2.3 * LOG10(NT proBNP) - (0.82 * (pretest probability) * LOG10(NT proBNP)) - 8


probability of acute heart failure =

= 1 / (1 + EXP((-1) * X))



• The NT-proBNP is operated on by LOG10 and then X is an exponent to e (2.71828).

• NT-proBNP was measured using the Roche Elecsys assay.

• Exclusion criteria included acute myocardial infarction, cancer or dyspnea due to respiratory disease.


To read more or access our algorithms and calculators, please log in or register.