Skip to main content

Table 2 Independent association between on-admission fibrinogen and clot burden represented by the clot area: summary of multivariate analysis

From: Fibrinogen consumption is related to intracranial clot burden in acute ischemic stroke: a retrospective hyperdense artery study

Independents

GMR (95 % CI)

p value

On-admission fibrinogen (by 2.718-fold)a

0.639 (0.424–0.964)

0.033

Men (vs. women)

1.453 (1.105–1.911)

0.008

History of carotid stenosis >50 %

1.485 (0.972–2.269)

0.068

Proximal vs. distal middle cerebral artery (MCA)

2.445 (1.770–3.376)

<0.001

Proximal MCA vs. “non-MCA” artery

1.535 (1.039–2.268)

0.038

  1. Ln-transformed clot area values were analyzed and results are presented as geometric means ratio (GMR) with 95 % confidence intervals by unit or level change in an independent analysis
  2. The initial general linear model fitted to ln(clot area) included all independents selected from the variables depicted in Table 1 (except for symptom severity on-admission [NIHSS], acute treatment, final infarct volume, control image finding and clinical outcome) based on a trend towards univariate association with this outcome (p < 0.1) [ln(fibrinogen), sex, history of carotid stenosis, affected blood vessel (proximal or distal middle cerebral artery, or “non-middle cerebral artery”), age and prior use of antiplatelets]. Variables from this full model were then successively removed (age p = 0.562, prior antiplatelet use p = 0.115) in the order of the highest p value, if p > 0.100 (backward elimination). Two-term interactions between ln(fibrinogen) and each of the other effects remaining in the model were tested, but were insignificant with p > 0.500 and excluded. The final model is shown
  3. aSince on-admission fibrinogen was ln-transformed (to achieve normality of residuals), the “effect” of on-admission fibrinogen is presented as GMR by 2.718-fold increase