Post-operative complications contribute greatly to the morbidity and mortality experienced by patients following gastroesophageal surgery. The purpose of this study was to determine the diagnostic accuracy of CRP and albumin in predicting these complications. This will help doctors make better informed clinical decisions such as appropriately ordering further investigations and discharging safely.
This was a retrospective study which analysed 131 patients from St Vincent’s Public Hospital in Melbourne. ROC analysis was performed to examine the association between CRP and albumin and any occurring complications. Optimum cut-offs were determined using a Youden Index. Positive and negative predictive values were calculated based off these cut-offs. Complications were classified according to their severity as per the Clavien-Dindo grading system and their inflammatory nature. Indirect measures such as unplanned ICU admission and readmission to hospital were also analysed.
CRP demonstrated the highest diagnostic accuracy for predicting severe inflammatory complications on POD 6 (AUC: 0.849, CI 0.704 - 0.994). An optimum cut-off of 180 mg/L resulted in a sensitivity and specificity of 88.89% and 75.00% respectively. This was accompanied with a negative predictive value of 97.83%. The optimum cut-off for albumin was 24 g/L (AUC: 0.787, CI 0.629 – 0.944) on POD 6. This demonstrated poorer sensitivity (81.82%) and specificity (70.49%) compared to CRP.
CRP is a strong negative predictive marker of complications following gastroesophageal surgery. These findings suggest CRP would be useful in assisting doctors to make clinical decisions regarding further investigations and safe discharge. Other health services may benefit by determining similar cut-offs based on their own unique patient populations.