Predictive Value of Lymphocyte Percentage and CRP Level for Early Detection of Deep Surgical Site Infection Following Posterior Lumbar Spinal Surgery

Objective: To investigate the predictive value of laboratory predictors, such as C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), white blood cell (WBC) count, and WBC differential count, for the early diagnosis of deep surgical site infection (SSI) after posterior lumbar spinal surgery. We also sought to determine the diagnostic thresholds for these markers. Methods: A total of 243 patients participated in the study: 11 patients who developed deep SSI after lumbar spinal surgery (SSI group) and 232 non-SSI patients as controls (non-SSI group). White blood cell (WBC) count, WBC differential count, CRP level, and ESR were determined 1 day before surgery and on postoperative day (POD) 1, POD3, and POD7. The diagnostic thresholds for these markers were determined with the receiver operating characteristic curve. Results: CRP, ESR, and WBC were signicantly higher in the SSI group than in the non-SSI group on POD3 and POD7 ( P < 0.05). The lymphocyte percentage was signicantly lower in the SSI group, compared with the non-SSI group, on POD3 ( P < 0.05). Analysis of the receiver operating characteristic curve revealed that lymphocyte percentage < 11.5% on POD3 (sensitivity 90.9%, specicity 75.4%, area under the curve [AUC] 0.919), and C-reactive protein level > 26 mg/L on POD7 (sensitivity 90.9%, specicity 87.7%, area under the curve [AUC] 0.954) were signicant laboratory predictors for the early detection of SSI. Conclusion: Lymphocyte percentage < 11.5% on POD3 and CRP levels > 26.5 mg/L on POD7 are reliable predictors for SSI after posterior lumbar spinal surgery.

identify the most signi cant laboratory predictors for the early detection of deep SSI and to determine the appropriate cut-off values for these predictors using the receiver operating characteristic (ROC) curve. Age, gender, operating time, and intraoperative blood loss were evaluated and recorded. CRP, ESR, WBC count, and neutrophil and lymphocyte ratio before surgery and on postoperative day (POD) 1, POD3, and POD7 were collected and analyzed. CRP and ESR were measured by latex agglutination, and WBC count was measured with an automatic cell counter. Neutrophil and lymphocyte percentage were determined by analyzing the WBC count. The reference intervals for blood cell analysis and the cut-off for CRP were determined based on the National Standard of the People's Republic of China WS/T 404.9-2018.

Methods
Statistical analysis was performed with SPSS 24.0 for Windows (SPSS, Inc., IBM) and Graphpad Prism 7 (GraphPad Software, La Jolla, CA, USA). Data are expressed as mean ± standard deviation (SD).
Continuous variables are presented as means and standard deviations. Categorical variables are summarized as the number and percentage of the total study population. Normally distributed continuous variables were compared using the two-sided independent t-test. Categorical variables were analyzed with the Chi-square test. We determined appropriate diagnostic cutoffs for the markers selected with the ROC curve. Statistical signi cance was de ned as P < 0.05.

Demographics and patient information
Among 451 patients who underwent posterior lumbar spinal surgery, the nal sample consisted of 243 patients: 11 patients who developed deep SSI (SSI group) and 232 who did not (non-SSI group). The SSI group included 7 male and 4 female patients; the non-SSI group included 85 female and 116 male patients. The mean age of patients in the SSI group was 55.82 ± 20.89 years; the mean age of patients in the non-SSI group was 55.55 ± 14.43 years. Operating time was 223.64 ± 90.36 minutes in the SSI group (range, 120-420), and 189.74 ± 80.70 minutes (range, 60-450) in the non-SSI group. Intraoperative blood loss was 600.00 ± 303.32 mL (range, 200-1000) in the SSI group and 442.11 ± 251.21 mL (range, 100-1000) in the non-SSI group. There were no signi cant difference between the groups in age, sex, operating time, or intraoperative blood loss (P > 0.05) ( Table 1).

Outcomes in the SSI group
An SSI was con rmed only when the surgeon diagnosed SSI, conducted debridement, and obtained a positive microbiological culture. Among 11 patients diagnosed with SSI, 10 patients underwent debridement, and instrumentation was removed in 8 cases. The other 1 patient received antibiotic treatment without surgery. Among the 10 patients who underwent re-operation, the bacterial species detected at the surgical site was identi ed as Staphylococcus aureus in 4 cases, Klebsiella pneumonia in 3 cases, Pseudomonas aeruginosa in 2 cases, and unknown in 1 case. Among the patients treated with antibiotics, the bacterial species identi ed by blood culture was S. aureus. The timeline for the onset of infection after surgery is shown in Table 2. All patients recovered after undergoing surgery or receiving treatment with antibiotics ( Table 2). Biochemical markers There were no signi cant differences between groups in the level of any chemical marker measured before surgery. On POD 1, the WBC count in the SSI group (15.65 ± 6.05)*10 9 /L was signi cantly higher than that in the non-SSI group (12.07 ± 3.95) *10 9 /L ( t = -2.51; p = 0.01), but there was no signi cant difference between groups in CRP, ESR, neutrophil percentage, or lymphocyte percentage (Fig. 1).
On POD 3, there was no signi cant difference in WBC count between the SSI group (12.55 ± 3.78) *10 9 /L and the non-SSI group (10.29 ± 3.65) *10 9 /L ( t = -1.69; p = 0.096) (Fig. 2). The neutrophil percentage in the SSI group (0.84 ± 0.07) was signi cantly higher than that in the non-SSI group (0.75 ± 0.10; t = -2.77, p = 0.007  The biochemical markers used most widely for the early diagnosis of SSI are CRP, ESR, total WBC count, and differential WBC count. Several studies have reported the usefulness of these biomarkers and found that the combined use of normal CRP, ESR, and WBC levels reliably predicts the absence of infection after lumbar spinal surgery [14.15] . Changes in ESR and CRP after surgery may be used to distinguish between infected and non-infected patients. Larsson et al. 17 were the rst to report changes in CRP level after lumbar posterior surgery; the authors observed that CRP level returned to baseline (< 10 mg/L) within 21 days after surgery. However, Takahashi et al 18 . reported that postoperative CRP level peaked on day 2, and Aono et al. 19 reported that CRP level peaked on day 4. Postoperative CRP levels peaked between POD2 and POD4. Iwata et al. 16 reported that a CRP level > 10 mg/L at 4 days postoperatively was useful for the de nitive diagnosis of SSI. Kyu et al. 20 reported that the observation of abnormal CRP levels three days postoperatively should cause the clinician to be highly suspicious of infection.
It is believed that CRP levels remain elevated or increase further at POD4, suggesting that SSI is more likely to occur. 16.18.22 The results showed that the level of CRP had increased signi cantly on POD 3 in the SSI group and remained high on POD7. In the non-SSI group, CRP levels had increased signi cantly on POD3, then decreased slightly by POD7 (Fig. 1D). In both the SSI and the non-SSI group, ESR levels had increased signi cantly at POD3, but were then observed to have decreased slightly on POD7 (Fig. 1E). ESR increased to peak levels at POD5, followed by a slow and irregular decreasing trend 17 .
Therefore, CRP appears to be more suitable than ESR for evaluating infection.
Our results showed that CRP levels had increased in all patients on POD1. In non-SSI patients, decreases in CRP levels were observed on POD3 and POD7. However, in the SSI group, CRP levels were higher on POD3 and decreased on POD7. We determined the diagnostic cutoff for CRP by using the ROC curve. If CRP levels are > 26 mg/L on POD7, infection should be highly suspected, and the antibiotic regimen should be changed or increased in dose.
Several factors have been reported to affect postoperative CRP levels. These factors include blood loss, preoperative CRP levels, surgical approaches and the segment on which spinal surgery was performed. For example, surgery in the lumbar region is associated with higher postoperative CRP levels than surgery in other areas. 21 . 11 . Given the uncertainty of the use of CRP as a diagnostic tool for SSI, additional laboratory predictors are needed to distinguish between infected and non-infected patients at the postoperative stage.
WBC counts and lymphocyte counts were rst reported to be helpful for the early diagnosis of surgical wound infections after lumbar surgery by Takahashi et al. 23  Our results showed that, in all patients, neutrophil count increased and lymphocyte count decreased on POD1. In non-SSI patients, neutrophil count began to decrease on POD3, which is also when lymphocyte count began to increase. However, in infected patients, a further increase in neutrophil count and decrease in lymphocyte count were detected on POD3 (Fig. 1B,C). We therefore chose to determine the diagnostic cutoff for lymphocyte percentage using the ROC curve. If lymphocyte percentage was less than 11.5% at POD3, infection was highly suspected, as in studies performed previously. We believe that POD4 lymphocyte count is more useful than POD7 lymphocyte count because the data can be obtained earlier during the course of postoperative recovery. POD4 lymphocyte count appears to be an important indicator of SSI after posterior lumbar surgery.
In this study, we analyzed the sensitivity and speci city of ve laboratory markers for the early detection of SSI. Two laboratory markers were found to have acceptable levels of sensitivity and speci city. The rst indicator was CRP level > 26 mg/L on day 7 after surgery. For this marker, sensitivity and speci city were 90.9% and 87.7%, respectively, and the AUC was 0.954. The other indicator was lymphocyte percentage < 11.50% on the third day after surgery. For this marker, sensitivity and speci city were 90.9% and 75.4%, respectively, and the AUC was 0.919. If the lymphocyte percentage was < 11.50% on POD 3, or the CRP level is > 26 mg/L on POD 7, clinicians should check the surgical wound more carefully. If necessary, imaging diagnostic tools such as CT and MRI may be used.
Our research has several limitations. First, this was a retrospective study. Therefore, there may be inherent bias associated with patient selection and missing patient information. Second, chronic SSI patients may have been included in the non-SSI group. Patients were assigned to groups based on the results of the hospitalization exam. The actual number of SSI cases may therefore have been underestimated. Third, the number of SSI patients included in our samples was small; there may be a selective bias. Fourth, in this study, we investigated the predictive value of biochemical markers in patients who underwent posterior lumbar spinal surgery only. Therefore, in future studies, we will expand the study population as well as the duration of follow-up.

Conclusion
In conclusion, a lymphocyte percentage < 11.50% at 3 days postoperatively and CRP levels > 26 mg/L at 7 days postoperatively may be valuable markers for the early diagnosis and control of SSI following lumbar surgery. These predictors appear to have high sensitivity and can be measured early during the course of postoperative recovery.     ROC analysis for lymphocyte ratio at POD3 (A) and for CRP at POD7 (B).