Numerous efforts have been made to identify useful markers for the prediction of adverse postoperative events. Taking into account standard protocols of the enhanced recovery program which warrants discharge of bariatric patients by the second postoperative day 12, highlights the importance of obtaining early SIR markers, predicting postoperative leaks with sufficient. To our knowledge, this is the first study to investigate whether the POD-1 RDW, MPV, PCT, PLR, and NLR are predictive of gastric leaks in patients undergoing LSG.
SIR markers have been studied separately as prognostic and predictive tools in various diseases. Kilincalp et al.13 reported that NLR, PLR, and MPV were useful and easily available biomarkers in the screening of colorectal cancer as well as in postoperative follow-up. AUCs of these parameters were 0.921, 0.853, and 0.717, respectively. Yucel & Ustun7 evaluated the use of CBC parameters, including RDW, MPV, PCT, PLR, and NLR, in the prediction of the severity of preeclampsia. In a ROC curve analysis of the NLR, PLR and RDW AUC, no statistically significant difference was found (p = 0.636, 0.104 and 0.36, respectively). However, the AUC of MPV and PCT were determined as 0.641 and 0.712, respectively, showing statistical significance (p = 0.028 and 0.001). Authors concluded that MPV or PCT may be clinically useful markers in the prediction of severe preeclampsia. In the current study, RDW and MPV did not differ between the leakage and control groups in a univariate analysis (p = 0.754 and 0.355). Therefore, they were not selected for a ROC analysis. Furthermore, the AUC of PCT and PLR was found to be 0.637 and 0.662, respectively, indicating that they do not possess clinical utility for patients undergoing LSG.
The NLR is an emerging biomarker gaining attention across many fields. It can be easily calculated from the differential CBC test, and it is also simple and inexpensive.14 NLR provides information on both immune and inflammatory pathways that could make it a potential marker for predicting intrabdominal infections, such as acute appendicitis, gastrointestinal leakage from suture lines. 6,10,14,15 In a study with a total of 100 patients who underwent open colorectal procedures, Cook et al.6 reported that the ROC curve analysis of POD-1 NLR suggested cutoff of 9.3 for the prediction of overall complications, with an area under the curve of 0.66. At this cut-off point, sensitivity, specificity, and likelihood ratio were 66%, 69%, and 2.12, respectively. In a study with 789 patients, 88.6% of which underwent Roux-en-Y gastric bypass vs 11.4% LSG, POD-1 NLR ≥ 10 was found to be significantly associated with adverse postoperative 30-day outcomes, including a higher incidence of overall complications and major complications, readmission rate, and reoperation rate.10 This NLR threshold was chosen by the authors after referring to a study by Cook et al. 6, as the cut-off value of POD-1 NLR in bariatric surgery was unexplored previously.
In the current study, we found that a POD-1 NLR cut-off of 3.6 with an AUC of 0.911 produced a sensitivity of 80%, specificity of 92%, likelihood ratio of 10, and accuracy of 90%. Of note, this threshold serves to predict postoperative gastric leaks after LSG and not overall complications. Furthermore, when interpreting the results of this study, is should be taken into account that the positive and negative predictive values may alter depending on the prevalence of disease. The prevalence of the disease in this study was determined to be 12.4%, as 36 out of 290 patients experienced gastric leak. As is expected, the lower the leakage rate is, the higher the negative predictive value is, and consequently, the lower the positive predictive value is. In a sense, calculating the NLR less than 3.6 in a low prevalence setting confidently rules out the presence of gastric leak, whereas the NLR of ≥ 3.6 is may be due to a false positive rather than a true positive.
The main limitations of this study are related to its design, in that it is retrospective in nature, although the data was collected prospectively. Also, the study represents a single center’s outcomes, which could be a concern for its generalizability. Another limitation could be the patient selection method for control group. The authors attempted to minimize selection biases which arise from this design with a matching method, however an imbalance between the two groups might still exist. For instance, although there was no statistical difference between the two groups in terms of gender distribution, the proportion of male patients was higher in the leakage group than the control group. In this context, an epidemiological study showed that the gender differences may influence laboratorial results, including PLT, MPV, and PCT.16 However, in the current study, we did not find differences between male and female patients in any of studied laboratorial parameters (p > 0.05). Similarly, Furuncuoglu et al.17 found that CBC parameters could be significantly affected by BMI status. In their study, participants were representative of several BMI groups. When we grouped patients according to a BMI below and above 40kg/m2, we found no statistical significance in the CBC parameters (p > 0.05).
In conclusion, both PCT and PLR on POD-1 were not found to be as clinically reliable parameters in this cohort. However, NLR with a cut-off value of 3.6 may be a useful predictor of postoperative gastric leak risk in patients undergoing LSG. This observation might enable timely, targeted intervention, and may be particularly useful when evaluating patients scheduled for early discharge postoperatively. Further prospective studies with larger sample sizes are warranted to better elucidate the utility of the NLR in bariatric surgery.