Total of 1000 ml or more during the first 6 h. However, the decision regarding haemorrhagic reexploration was made by the surgeon in charge. Chest tube outputs were used as a measure of blood loss, and our analysis was based on the total volume of loss during the first 24 h of the patient's stay in the ICU. Post-operative acute kidney injury (AKI) was defined according to the consensus RIFLE criteria (risk, injury, failure, loss of function, and end-stage renal disease) using the maximal change in serum creatinine and estimated glomerular filtration rate during the first 7 post-operative days compared with preoperative baseline values [22].Statistical analysisprevalence of a covariate between treatment groups (Fig. 1) [23, 24]. Finally, the significance within the models was evaluated with the Wald test, whereas the strength of the association of variables with post-operative outcomes was estimated by calculating the odds ratio (OR), the unstandardised coefficient and 95 confidence intervals (CIs). The model was calibrated using the Hosmer-Lemeshow goodness-of-fit test, as well as residual diagnostics (deviance and degrees of freedom of values). Model discrimination was evaluated by using the area under the ROC curve. All tests were two-sided with the level set at 0.05 for statistical significance. Statistical analysis was performed using IBM SPSS version 22.0 software (IBM, Armonk, NY, USA).ResultsStudy populationClinical data were prospectively recorded and tabulated using Microsoft Excel
PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/9282946 software (Microsoft, Redmond,
(2-methoxy-5-methylphenyl)boronic acid 2-(2-Aminoethoxy)-5-chloropyridine hydrochloride WA, USA). Continuous data are reported as mean and standard deviation or median and interquartile range (IQR), as appropriate. Nominal variables were reported as counts and percentages. Fisher's exact test, 2 test and the Mann-Whitney U test were used for univariate analysis. No attempt to replace missing values was made. Multivariate analysis was performed using logistic and linear regression. The area under the receiver operating characteristic (ROC) curve was used to represent
PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/25116583 the regression probabilities. Since January 2012, PCC has been used systematically as first-line therapy in coagulation management, completely replacing FFP, and this PCC use was consistently followed. Therefore, the PCC study group was matched with the historical series of patients who received FFP before this time point. Because the study groups (i.e., the PCC and the FFP groups) significantly differed in a number of baseline and operative variables, a propensity score was calculated by logistic regression to estimate the probability of being assigned to each of the study treatments. The propensity score was calculated in a non-parsimonious way, including all 31 pre-operative and operative variables listed in Tables 1 and 2. The obtained propensity score was used for adjusted analysis in the overall series and for one-to-one propensity score caliper matching. The caliper width chosen was 0.2 times the standard deviation of the logit of the propensity score (i.e., 0.01). Propensity score was used as a covariate, along with the treatment method, in the multivariate analysis model for each outcome end-point. After the propensity score matching was performed, differences between the two groups were assessed. Absolute standardised differences were estimated to evaluate the pre-match and post-match imbalance, and a standardised difference <0.1 was considered a negligible difference in the mean orAmong the 3454 included patients (mean age 68.0.