A novel approach for identifying serological markers indicative of surgical-site infection following spine surgery: Postoperative lymphopenia is a risk factor.

BACKGROUND

Postoperative lymphopenia (PL) after spine surgery is reported to be an indicator of surgical-site infection (SSI). PL without SSI is often encountered, resulting in a treatment dilemma. We focused on PL, so as to improve the accuracy of detecting SSI.

METHODS

In total, 329 patients underwent spine surgery, including nine patients presenting with SSI. The complete blood cell counts, differential counts, and C-reactive protein (CRP) level were measured pre-surgery and on postoperative days 2, 7, and 14. The relationships between PL and SSI were evaluated, and PL and non-PL conditions were compared among all cases. We then divided the patients into two groups: PL and non-PL, and determined the useful serological markers using receiver operating characteristic curves.

RESULTS

Sixty-one patients presented with PL, including four with SSI. However, PL was not directly suggested as a biomarker of SSI (p = 0.067). We revealed PL as a risk factor for SSI (p = 0.004, Odds ratio: 7.54). Among all cases, the lymphocyte count and CRP level differed significantly between the PL and non-PL groups at all perioperative time-points. The white blood cell count, neutrophil count, and CRP levels on postoperative day 7 significantly differed between the SSI and non-SSI cases in the PL group. The area under the curve (AUC) for CRP was greater than that of the other parameters. The neutrophil count was only effective as a marker in the non-PL group. The combination of two cutoff values (CRP: 3.7 mg/dL (PL group) and neutrophil count: 6172/μL (non-PL group)) presented high specificity (87.2%) and sensitivity (88.9%), while only one cutoff value (CRP: 2.9 mg/dL) had a specificity of 77.4% and sensitivity of 77.8%.

CONCLUSIONS

Approximately one-fifth of patients developed PL post-surgery, which was a risk factor for SSI, with constant high inflammation. Grouping based on PL and establishing diagnostic cutoff values are more appropriate than establishing only one cutoff value for overall cases.