Background: The sequential organ failure assessment (SOFA) is one of the most commonly used scoring systems to evaluate organ failure in sepsis. The trajectory of the SOFA score, a delta-SOFA (SOFAΔ), is proposed as a better indicator for predicting mortality, and potentially as an endpoint in exploratory clinical trials. However, there are some concerns about the use of SOFAΔ. For example, SOFAΔ represents only the changes in the score, and the potential value of the absolute SOFA score has not been considered. Therefore, we hypothesized that the addition of the absolute SOFA score to the SOFAΔ would improve the predictive performance for patient outcomes. Based on this theory, a new indicator, SOFAComb calculated by SOFAΔ + absolute SOFA score, was examined in this study. Methods: Data obtained from 297 patients in multiinstitutional post-marketing surveys performed during June 2014, and May 2016 were retrospectively analyzed. All patients were diagnosed as having sepsis-associated disseminated intravascular coagulopathy (DIC) and treated with antithrombin concentrate. The SOFAComb and SOFAΔ were calculated on days 2, 4, and 7, and the performance was analyzed in terms of predictive ability for 28-day mortality.Results: Of the 297 patients included in the analysis, 214 patients survived (72.1%), while 83 patients (27.9%) died. The area under the receiver operating curve (AUC) of SOFA at baseline for predicting 28-mortality was 0.679. The AUCs of SOFAΔ on day 2, 4, 7 were 0.662, 0.769, 0.815, respectively, and those of SOFAComb on day 2, 4, 7 were 0.765, 0.830, 0.866, respectively. The AUCs of SOFAComb were significantly greater at all time points compared to SOFAΔ (day 2: P <0.001, day 4: P =0.002, day 7: P <0.001). In addition, the accuracy of SOFAComb was better than that of SOFAΔ (day 2: P <0.001, day 4: P =0.067, day 7: P =0.049). Conclusions: SOFAComb is simple to calculate and provides better predictive performance compared to SOFAΔ for predicting mortality. Additional studies are needed to confirm these findings.