DISCUSSION
Our study is the first study evaluating the role of serum markers in the early detection of AL after gyne-oncological operations. There have been many studies showing the possible role of serum PCT and CRP in the diagnosis of AL after elective colorectal surgeries6. In the study of Alvara Garcia-Granero et al., in case of major AL, PCT and CRP were proved as reliable predictors on POD 3 to 5 with the best value for PCT on POD 5 was 0.31 ng/ml (AUC = 0.86) (Sensitivity = 100%, specifity = 72%, positive predictive value = 17%, negative predictive value = 100%) 7. Also, in a recent meta-analysis, it was concluded that PCT is a useful negative test for AL after elective colorectal surgery with negative predictive values ranging from 95% to 100% (highest AUROC was 0.88 on POD 5). But, as an isolated test, poor positive predictive values of up to 34%, limit its use in predicting AL8. Different from elective colorectal surgeries, in the surgical concept of gynecological malignancies (especially in advanced ovarian cancer), debulking procedures including both upper and lower abdomen and peritonectomy were needed to reach zero macroscopic residual disease to get overall survival advantage9,10. So, we wanted to analyze these markers whether they were also beneficial in the early diagnosis of AL after gyne-oncological surgeries. And, a statistically significant difference was found between the albumin values ​​on POD 3 (p=0.028), on POD 4 (p=0.045) and the platelet values on POD 1 (p<0.001). Pre-operative hypoalbuminemia is a well-known risk factor for AL, but few studies evaluated the role of peri-operative albumin on AL after intestinal surgeries11,12. In the study of Shimura et al., it was shown that lower average levels of serum albumin on POD 1 and POD 3 (HR = 4.49, 95% CI = 0.77–29.58; p  = 0.0955) and higher average levels of serum leukocytes on POD 1 and POD 3 (HR = 5.62, 95% CI = 0.76–115.34; p  = 0.0952) were independent risk factors for AL11. Similarly, in our study, lower levels of albumin on POD 3 and 4 were associated with AL. In the study of Margarson et al., after giving intravenous hypertonic albumin solution, serum albumin concentrations decrease significantly faster in septic patients than in healthy controls and that was explained by increased vascular leak of the albumin13. But, it is unclear whether hypoalbuminemia is the cause or result of the AL.
Interestingly, in our study, it was shown for the first time in the literature that platelet count was associated with AL as the mean platelet values were lower on POD 1 in the AL group. In the study of Dewitte et al, possible mechanisms of blood platelets in sepsis pathophysiology were reviewed14. According to this, beyond their roles in haemostasis, platelets are now accepted as active actors of immune respone playing role in host defence and tissue integrity. This interesting relation of low platelets with AL found in our study should be further evaluated in prospective trials.
Although it was not statistically significant (p>0.05), in our study, median PCT values on POD 8 to 10 were higher in the AL group. And, PCT values increased later compared to other studies in the literature7,15. Intraabdominal sepsis may have delayed in patients with AL due to our local peroperative guideline which was different from ERAS protocol16. Most likely reason for this late increase of PCT in patients with AL may be antibiotic supression during our hospitalization period. In accord to this, Charles et al. showed that empirical antibiotic therapy was associated with a greater decline in PCT following the onset of sepsis between day 2 and 317.
In contrast to previous studies, we could not find a statistically significance between AL and postoperative CRP levels7,18. In a systematic review made by Singh et al., it was shown that serum CRP level on POD 3, 4, and 5 had comparable diagnostic accuracy for the prediction of AL with a AUROC of 0.81, 0.80, and 0.80, respectively18. The derived CRP cut-off values were 172 mg/L on POD 3, 124 mg/L on POD 4 and 144 mg/L on POD 5. And these corresponded to a negative predictive value of 97% and a positive predictive value of 21-23% 18. In the study of Smith et al., for the first time in the literature, an association of biomarker trajectory was assessed, instead of isolated daily values. And the trend of CRP for the first 5 days following surgery, appeared to be highly accurate for diagnosing AL, with a daily rise of 50 units had a sensitivity of 91% and a negative predictive value of 99.3%19. Our results showed that PCT and CRP did not work to predict AL before clinical onset of sepsis in patients operated for gyne-oncological surgeries. After ROC analysis, the best cutt-off point for PCT was determined to be 0.11 ng/mL on POD 9. Although this statistical value was not useful for the early clinical management, it could be used in patients who had vague symptoms and indeterminate radiologic imaging. Our low number of positive cases (5 cases of AL) may have prevented us from observing possible effect of PCT and CRP on the early days of leakage.
Our study had some limitations. First, the study had retrospective nature comprising consecutive patients managed on the basis of local clinical guideline. Second, patient cohort included heteregoneus type of malignancies and surgeries. On the other hand, PCT, previously shown to predict AL in colorectal surgery, has been shown for the first time to be useless in gynecological oncology. Also, different from literature, the increase of PCT seen in AL appeared very soon that should be supported with further studies.
In conclusion, serum PCT and CRP concentrations were not found to be helpfull for the early diagnosis of AL in patients operated for gyne-oncological malignancies. Low levels of albumin and platelets in the first days after the operation may be clue for a possible AL.