Delta neutrophil index for predicting mortality in trauma patients who underwent emergent abdominal surgery: A case controlled study

Background Delta neutrophil index (DNI) can be used as a biomarker for infection to predict patient outcomes. We aimed to investigate the relationship between DNI and clinical outcomes in trauma patients who underwent abdominal surgery. Materials and methods We retrospectively analyzed injured patients who underwent emergent abdominal surgery in the regional trauma center of Wonju Severance Christian Hospital between March 2016 and May 2018. Patient characteristics, operation type, preoperative and postoperative laboratory findings, and clinical outcomes were evaluated. Logistic regression analysis was performed for risk factors associated with mortality. Results Overall, 169 patients (mean age, 53.8 years; 66.3% male) were enrolled in this study, of which 19 (11.2%) died. The median injury severity score (ISS) was 12. The non-survivors had a significantly higher ISS [25(9–50) vs. 10(1–50), p<0.001] and serum lactate level (9.00±4.10 vs. 3.04±2.23, p<0.001) and more frequent shock (63.2% vs 23.3%, p<0.001) and solid organ injury (52.6% vs. 25.3%, p = 0.013) than the survivors. There were significant differences in postoperative DNI between the two groups (p<0.009 immediate post-operation, p = 0.001 on postoperative day 1 [POD1], and p = 0.013 on POD2). Logistic regression analysis showed that the independent factors associated with mortality were postoperative lactate level (odds ratio [OR] 1.926, 95% confidence interval [CI] 1.101–3.089, p = 0.007), postoperative sequential organ failure assessment score (OR 1.593, 95% CI 1.160–2.187, p = 0.004), and DNI on POD1 (OR 1.118, 95% CI 1.028–1.215, p = 0.009). The receiver operating characteristics curve demonstrated that the area under the curve of DNI on POD1 was 0.887 (cut-off level: 7.1%, sensitivity 85.7%, and specificity 84.4%). Conclusions Postoperative DNI may be a useful biomarker to predict mortality in trauma patients who underwent emergent abdominal surgery.


Introduction
The current biomarkers for diagnosis of sepsis or infections include white blood cell (WBC) count, lactic acid, procalcitonin, and C-reactive protein (CRP) [1][2][3]. The release of immature neutrophils into the bloodstream during infection or sepsis leads to an elevation of the immature/total granulocyte ratio which is defined as neutrophil 'left-shift'. This granulocytic 'leftshift' or increase in immature granulocyte (IG) rate is commonly used as a diagnostic marker of infection or sepsis in the clinical setting. However, it is difficult to accurately measure IG using a microscopic examination of blood smears, and its diagnostic value remains controversial [4,5]. Technological advances in an automated cell analyzer have enabled the acquisition of the delta neutrophil index (DNI) using leukocyte differentials obtained from two independent channels-the myeloperoxidase channel and the lobularity/nuclear density channel. The DNI is calculated as the difference between leukocyte differentials measured in these two channels, which reflects the proportion of circulating IG [6]. Several studies reported that DNI was associated with disease severities of sepsis or septic shock and mortality in patients with various infectious conditions such as bacteremia, pneumonia, and peritonitis [7][8][9][10][11]. Moreover, recent studies showed that DNI was associated with the severity and prognosis of non-infectious inflammation-related diseases, such as acute myocardial infarction, pulmonary embolism, upper gastrointestinal hemorrhage, and cardiac arrest [12][13][14][15]. However, there are few studies about the use of DNI in trauma patients. Therefore, the aim of this study was to evaluate the usefulness of DNI as a predictor of mortality in trauma patients who underwent emergent abdominal surgery.

Patient selection and data collection
The study was approved by the institutional review board of Wonju Severance Christian Hospital (IRB no. CR319077). All data were fully anonymized before access and IRB waived the requirement for informed consent. Among 6291 injured patients who were admitted in the regional trauma center of a tertiary university hospital between March 2016 and May 2018, 173 patients who underwent emergent abdominal surgery were enrolled in this study. After exclusion of four patients who died within six hours of admission, the final study population was 169 (Fig 1).
The primary end-point was to evaluate the effectiveness of DNI to predict postoperative mortality in injured patients who underwent emergent abdominal surgery. The secondary end-point was to compare DNI with other biomarkers for prediction of mortality and to access the cut-off level of DNI. The demographic and clinical characteristics of the patients such as age, sex, injury mechanism, injury severity score (ISS), associated injury (abbreviated injury scale, AIS �3), initial shock, diagnosis, gastrointestinal (GI) perforation, solid organ injury, serum lactate, and sequential organ failure assessment (SOFA) score on intensive care unit (ICU) admission were retrospectively reviewed. based hematologic analyzer uses two independent WBC counting methods-a myeloperoxidase (MPO) channel and a lobularity/nuclear density channel. DNI value was calculated using the following formula: DNI = (leukocyte subfraction assayed using the MPO channel of a cytochemical reaction)-(leukocyte subfraction assayed using the nuclear lobularity channel based on reflected light beam measurements) [6].

Statistical analysis
Continuous variables were presented as the mean±standard deviation or the median values (ranges), and comparative analysis was performed using a Student's t-test. Categorical variables were analyzed by the Chi-square test and Fisher's exact test. To identify the independent risk factors for mortality, a multivariate analysis was performed using logistic regression. A receiver operating characteristics (ROC) curve was constructed, and the Youden Index method was used to find the optimal cut-off values for lactate, DNI, and SOFA score to predict mortality. All statistical analyses were performed using SPSS 20.0 (IBM, Armonk, NY, USA). Statistical significance was accepted for p <0.05.

Performance of DNI and other laboratory markers in critically ill and injured patients who underwent emergent abdominal surgery
When the ROC curves of the postoperative lactate, SOFA score on ICU admission, and DNI on POD1 were conducted to predict mortality, area under curve (AUC) of the immediate postoperative lactate, SOFA score on ICU admission, and DNI on POD1 were 0.874 (95% CI, 0.773-0.975, p <0.001), 0.941 (95% CI, 0.898-0.984, p <0.001), and 0.887 (95% CI 0.798-0.976, p <0.001), respectively. The optimal cut-off points for the postoperative lactate, SOFA

Discussion
This study showed that DNI on the POD1 was an independent risk factor to predict the mortality in critically ill and injured patients who underwent emergent abdominal surgery. Moreover, the ROC curve for the DNI on POD1 confirmed that the optimal cut-off for predicting mortality was 7.1%, and the sensitivity (85.7%) and specificity (84.4%) were high with an AUC of 0.887. Mean DNI had a different pattern throughout the study period compared with other biomarkers such as WBC count and CRP. The mean DNI in survivors decreased after initial elevation whereas it continuously increased in non-survivors. Additionally, a recent study on DNI in patients with sepsis caused by peritonitis reported that DNI on POD3 was an independent risk factor for postoperative mortality, and the patterns of mean DNI were different between survivors and non-survivors. Moreover, the AUC for DNI was 0.88, and the optimal cut-off value was 7.8%, with a sensitivity of 77.3% and specificity of 95.9% [7]. This result is quite similar to that of our study. Indeed, in our hospital, it was possible to recognize the high probability of death in patients with elevated DNI level above 7.1% the day after surgery. DNI helped surgeons explain the patient's condition earlier to the caregiver and determine further evaluation and general ward transfer. Previous studies about surgical and medical patients showed that DNI was a useful biomarker to predict disease severity or prognosis in patients with various infections or sepsis [7,10,16,17]. However, studies about usefulness of DNI in patients who had tissue injury or hemorrhage due to trauma are limited [18]. Systematic inflammatory response syndrome (SIRS) is initiated within 30 minutes after severe injury, which is associated with an inflammatory response to hemorrhage or tissue damage rather than infections. Damage-associated molecular patterns (DAMPs) are released into the extracellular space by tissue damage that triggers an inflammatory response without infection. DAMPs activate the innate immune systems such as neutrophils, monocytes, and complements. This change can result in endothelial and organ damage and immunoparesis. As a result, a vicious cycle is triggered due to continuous inflammation and immune activation. In this process, neutrophils are activated initially by SIRS, but their bactericidal function is markedly impaired. This functional impairment is  compensated by the release of immature banded neutrophils [19]. In the present study, DNI appears to be correlated with the severity of SIRS induced by tissue damage and hemorrhage in patients with severe trauma. In other words, increase in the value of DNI in non-survivors indicated severe SIRS, which may result in poor clinical outcomes as a result of progressive persistent inflammation, immunosuppression, and catabolism syndrome [20]. A recent study on trauma patients who were admitted in ICU reported that high DNI values at 12 and 24 hours from admission were strong independent predictors of multiple organ dysfunction  syndrome (MODS). Increased DNI at 12 hours was a strong independent predictor of shortterm mortality. In addition, the cut-off value of DNI at 12 hours to predict 30-day mortality was 5.3% in that study [18], which was similar to the result in our study.
In this study, we found that although 62 (36.7%) patients had GI perforation, there was no significant difference in the mortality rate according to GI perforation (patients with GI perforation 8.1% vs. patients without GI perforation 13.1%, p = 0.319). This result has several implications. First, most patients with abdominal injuries underwent early surgery if necessary, and therefore, many patients with GI perforation might not progress to septic shock. Second, there are causes other than infection that exacerbate the condition of a trauma patient. As described above, SIRS due to tissue damage and hemorrhage was a major deterioration factor in the early stage of major trauma and was closely associated with patient severity and prognosis. Severe SIRS seemed to cause MODS and eventually lead to death [19,20].
Several studies suggested that WBC count was only a nonspecific indicator of stress such as infection, inflammation, tissue necrosis, and hemorrhage in injured patients [21,22]. Another study on patients with abdominal injury showed that the WBC count within 24 hours after admission had limited diagnostic value for predicting hollow viscus injury [23]. Similarly, the present study showed that the WBC count had poor predictability in distinguishing survivors from non-survivors over time after surgery. Although there was a significant difference in the CRP level between the two groups on POD1 and POD2, it is difficult to use CRP as a predictor of mortality, because the survivor group had higher CRP level than the non-survivor group, and CRP had a relatively slow changing pattern. A study performed in patients with multiple trauma showed that CRP increased easily and had very slow kinetics compared to other biomarkers [24]. In contrast, several studies have shown that initial lactate and lactate clearance are clinically useful in predicting mortality of trauma patients [25][26][27]. In our study, we found that postoperative lactate level was an independent risk factor for postoperative mortality and had a high predictability of mortality with an optimal cut-off level of 5.105 mmol/L (AUC 0.874, sensitivity 71.4%, and specificity: 92.6%). Taken together, DNI on POD1, postoperative SOFA score, and postoperative lactate may be used complementarily to predict patient death.
There are some limitations to our study. First, it is hard to confirm the usefulness of DNI, because of the small sample size and single institutional study. Second, data of the present study did not include serial levels of procalcitonin which is currently an important biomarker for diagnosis of infection and sepsis, because reimbursement for procalcitonin is limited within twice a week in the Korean national medical insurance. Third, there may be a selection bias in the present study, because of its retrospective nature. Despite these limitations, our study may be meaningful in that it is the first study to evaluate the usefulness of DNI to predict clinical outcome in abdominal trauma patients who underwent emergency surgery. In the future, large-scale prospective studies will be needed to confirm the results of our study.