The initial idea of the present study was to discover novel infected nonunion-specific biomarkers that can help preoperatively distinguish infected nonunion from other patients with aseptic nonunion after internal fixation. A comprehensive review of the literature suggests that our study is the first to compare albumin, globulin, AGR, plasma D-dimer, plasma fibrinogen, PC, MLR, NLR, and PLR with traditional inflammatory biomarkers (WBC, CRP, and ESR) for their ability to screen infected nonunion in patients undergoing reoperation after internal fixation. Our research demonstrated that preoperative plasma fibrinogen and AGR were reliable blood biomarkers for screening infected nonunion, and the combination of AGR with plasma fibrinogen could further provided more accurate and more specific evaluation for diagnosing infected nonunion in those patients.
Albumin and globulin, easily accessible and reliable biomarkers in the basic metabolic panel, have proven to be critical markers associated with inflammation and infection.29,30 Albumin, a negative acute-phase reactant, is widely considered to be a biomarker of inflammatory and nutritional condition of the human body.31 Hypoalbuminemia is correlated with malnutrition, hepatopathy, kidney disease, and all kinds of inflammation.22,32 Globulin, another major serum protein component, consists of various immunoglobulins and acute-phase proteins.33,34 Collectively, as both decreased albumin and increased globulin played essential roles in response to inflammation and infection functions to dramatically decrease the AGR, suggesting that the AGR indicates the body’s inflammatory state more accurately.35,36 Some past studies have validated the potential role of globulin and AGR are associated with PJI and may serve as potential adjuvant biomarkers in the diagnosis of PJI.25,26 In our study, we observed that albumin, and AGR were significantly lower, while globulin was considerably higher in patients with infected nonunion. However, ROC curve analysis revealed that only AGR showed acceptable predictive value for the diagnosis of infected nonunion, with high AUC (0.845), fair sensitivity and PPV (70.27% and 74.29%), but high specificity and NPV (91.04% and 89.27%).
Many studies have indicated that systemic and local infections result in fibrinolytic activity, and coagulation-related parameters such as fibrinogen, D-dimer, and PC have been shown to be promising diagnostic biomarkers for the diagnosis of PJI in some studies.13,37−42 In our study, the AUC of plasma fibrinogen (0.805) was larger than that of plasma D-dimer (0.752) and PC (0.702). However, plasma fibrinogen had a moderate sensitivity and PPV (67.57% and 60.98%), but a high specificity and NPV (84.08% and 87.56%).
MLR, NLR, and PLR have been demonstrated as stable and cost-effective biomarkers that reflect the inflammatory response as they mediate inflammation by various biochemical mechanisms.43,44 Previous research suggested that NLR may perform better than CRP for diagnosing early PJI.45 In the contrast, most studies showed that MLR, NLR, and PLR showed limited diagnostic value in the preoperative diagnosis of infected nonunion, which was similar with our result.8,18 Our result showed that the diagnostic performance of MLR, NLR, and PLR was limited, as the AUC, sensitivity and specificity of MLR were 0.595, 31.08 and 78.26%, respectively, those of NLR were 0.687, 56.76 and 52.50%, respectively, and those of PLR were 0.725, 68.92 and 42.86%, respectively.
WBC, CRP, and ESR are the most commonly used biomarkers of infected nonunion. Unfortunately, they are usually affected by other factors such as physiological stress, treatment, and other diseases.4,46 Peripheral WBC, CRP, and ESR are frequently normal in low-grade infections and afford little value in diagnosing infected nonunion.
In subgroup analysis, we found that the diagnostic accuracy of the combination of AGR with plasma fibrinogen also performed well in infected nonunion patients who also had other comorbidities. Considering the small number of cases, a larger number of cases should be enrolled into evaluate the diagnostic accuracies of these biomarkers and reduce potential bias.
This study does have some limitations. First, all patients with suspected infected nonunion were collectively analyzed, and thus these results may not be applicable to all possible subgroups. Second, this study was retrospective, with inherent biases. Electronic medical records may contain incorrect or non-existent information for individual patients. Finally, the sample size in our study was relatively small, so we did not include data on any probable effect of antibiotic use or the different comorbidities. Therefore, multicenter, prospective, comparative studies with larger samples are required to more thoroughly determine the accuracy of these biomarkers for predicting infected nonunion.