Objective-To compare measured and predicted oxygen consumption ((V) over do
t o(2)) in children with congenital heart disease.
Design-Retrospective study.
Setting-The cardiac catheterisation laboratory in a university hospital.
Patients-125 children undergoing preoperative cardiac catheterisation.
Interventions-(V) over dot o(2) was measured using indirect calorimetry; th
e predicted values were calculated from regression equations published by L
indahl, Wessel et al, and Lundell et al. Stepwise linear regression and ana
lysis of variance were used to evaluate the influence of age, sex, weight,
height, cardiac malformation, and heart failure on the bias and precision o
f predicted (V) over dot o(2). An artificial neural network was trained and
used to produce an estimate of (V) over dot o(2) employing the same variab
les. The various estimates for (V) over dot o(2) were evaluated by calculat
ing their bias and precision values.
Results-Lindahl's equation produced the highest precision (+/-42%) of the r
egression based estimates, The corresponding average bias of the predicted
(V) over dot o(2) was 3% (range -66% to 43%). When (V) over dot o(2) was pr
edicted according to regression equations by Wessel and Lundell, the bias a
nd precision were 0% and +/-44%, and -16% and +/-51%, respectively. The neu
ral network predicted (V) over dot o(2) from variables included in the regr
ession equations with a bias of 6% and precision +/-29%; addition of furthe
r variables failed to improve this estimate.
Conclusions-Both regression based and artificial intelligence based techniq
ues were inaccurate for predicting preoperative ire, in patients with conge
nital heart disease. Measurement of (V) over dot o(2) is necessary in the p
reoperative evaluation of these patients.