Skip to content

Advertisement

  • Poster presentation
  • Open Access

Is SOFA score a mortality predictor in surgical patients admitted to the postoperative acute care unit?

  • 1,
  • 1 and
  • 1
Critical Care200610 (Suppl 1) :P411

https://doi.org/10.1186/cc4758

  • Published:

Keywords

  • Risk Score
  • Surgical Patient
  • Predictive Ability
  • Emergency Surgical Procedure
  • Postoperative Care

Objective

To assess the mortality predictive value of an organ failure score (SOFA) compared with other three risk scores (POSSUM, ASA, and SAPS II) in surgical patients admitted to a postoperative acute care unit (PACU).

Methods

Prospective collection of demographic, clinical and laboratory data of 330 consecutive patients admitted to the postoperative care unit following elective and emergency surgical procedure, for a 6-month period. The ASA score was assessed preoperatively, the POSSUM score was calculated preoperatively and postoperatively, and the SAPS II and SOFA score were computed for the first 24 hours postoperatively. The outcome measure was inhospital mortality. Receiver-operating characteristic (ROC) curve analysis was used to estimate the predictive ability for inhospital mortality of the various scoring systems.

Results

Among the 330 patients admitted to the PACU, 14 were transferred to the ICU after an interval of time varying from 24 hours to 9 days. The median length of stay in the PACU was 3 (1–17) days. The area under the ROC curve (AUROC) was 0.78 for SOFA score, 0.63 for POSSUM, 0.57 for ASA, and 0.67 for SAPS II. The cutoff values were 7 for SOFA, 47 for POSSUM, 3 for ASA and 21 for SAPS II scores. The observed mortality rate was 3.6%, while expected mortality rates according to POSSUM, ASA and SAPS II scores were 6.8%, 6.2% and 5.4%, respectively.

Conclusion

Our study suggests that SOFA score might be useful to predict outcome in the PACU surgical patients. The POSSUM, ASA and SAPS II scores overestimate inhospital mortality rates for surgical patients admitted to the PACU. The use of organ failure scores might improve the power of predictive scores in these patients.

Authors’ Affiliations

(1)
Hopital Claude Huriez, Lille, France

Copyright

© BioMed Central Ltd 2006

Advertisement