After eliminating age dependence for the à froid
group (p = 0.0596), age dependence also proved to be
insignificant for the à chaud group (p = 0.062). The
absence of significance for age dependence in this
analysis can be explained by reference to the relatively
small number of patients with an à chaud diagnosis
(n = 274) in relation to the number of patients with an
à froid diagnosis (n = 1091). Age dependence was
therefore investigated in a model limited to the group
with an à chaud diagnosis. Here we found a significant
age dependence (p = 0.047), such that patients above
the age of 40 were estimated to have a 2.5 higher risk
of postoperative bleeding than patients below and up
to the age of 40, with a 95% CI of 1–6.2.