Constructing a nomogram for the recurrence of tibial osteomyelitis after debridement surgery based on platelet to lymphocyte ratio at admission: a dual-center retrospective study from China Authors: Rongbin Lu, Shaohuai Xia, Gaoyong Deng, Wencai Li, Zhao Huang, He Ling Published: 2024 ABSTRACT AbstractAimTo construction of a nomogram to predict the prognosis of patients with tibial osteomyelitis after debridement surgery.MethodThis study used binary logistic regression analysis to identify clinical independent predictive factors, and then used R language to construct a nomogram.ResultThe results showed that WBC (3.36[1.77–6.37])、LYM (0.29[0.15–0.54])、BAS (2.19[1.22–3.91])、FIB (0.27[0.14–0.54])、ALB (0.16[0.09–0.3])、TBIL (0.41[0.21–0.83])、D-Dimer (1.77[1.03–3.03])、CRP (0.43[0.2–0.91])、ESR (3.08[1.6–5.79]) and PLR (0.47[0.24–0.9]) were independent predictive factors. Good prediction performance with modest errors was shown by the nomogram in both the training and validation groups.ConclusionWBC、LYM、BAS、FIB、ALB、TBIL、D-Dimer、CRP、ESR and PLR are independent predictors of the prognosis of patients with tibial osteomyelitis after debridement surgery. The constructed nomogram can help clinicians assess and treat patients early, benefiting more patients.Keywords: Tibial osteomyelitis, Debridement surgery, Prognosis, Nomogram Introduction Osteomyelitis is a destructive bone disease characterized by the gradual degradation of the bone and the corresponding inflammatory response, which can potentially damage surrounding normal tissues [1]. In Zalavras’ study [2], the incidence of osteomyelitis in open fractures ranged from 2 to 50%. Diabetic foot patients often experience osteomyelitis in the foot [3]. Osteomyelitis is generally difficult to cure, with a recurrence rate of up to 48% in chronic cases [4]. Early diagnosis is key to successful treatment of osteomyelitis. Current treatment modalities for osteomyelitis primarily involve surgery and antibiotic therapy [5]. However, antibiotic therapy can have difficulty reaching the site of infection, making surgical debridement the mainstay of successful surgical treatment for osteomyelitis, aiming to reduce infection, remove necrotic tissue, and decrease the bacterial burden at the infection site [6]. Due to the complexity of each patient’s condition, personalized surgical and medical plans are often required [7], and a single surgery may not always yield satisfactory results. Having a clinical tool to assist clinicians in predicting the risk of recurrence in patients would be of significant importance in formulating subsequent treatment plans for patients. Pre-admission blood parameter tests are readily available to clinicians and provide intuitive tools that reflect the patient’s status. In addition to the gold standard of bone biopsy results and imaging examinations, laboratory evaluations (including erythrocyte sedimentation rate, complete blood cell count, C-reactive protein, and blood culture results) can also actively predict osteomyelitis [8]. The elevation of relevant inflammatory markers can also be clinically correlated [9]. Nomograms are widely used in medical prognosis prediction, including gynecology [10], surgery [11, 12], and critical care medicine [13], among others. The use of nomograms is user-friendly and easily accessible, and it can provide important assistance to clinicians in estimating the course of the disease [13]. Therefore, this study aims to collect pre-admission routine blood tests and biochemical examination results from patients and construct a predictive model for the recurrence of tibia osteomyelitis within three months after surgical debridement based on the platelet to lymphocyte ratio (PLR) at admission. The goal is to provide a non-invasive, early, and convenient prediction tool to help clinicians make timely decisions, utilizing the simplest approach possible. Materials and methods Patient section This is a retrospective cohort study conducted at two large medical centers in China. The study collected data from patients diagnosed with tibia osteomyelitis between February 2015 and January 2023. The inclusion criteria were as follows: (1) Patients diagnosed with tibia osteomyelitis based on at least one of the following criteria: (a) Standard imaging examinations such as X-rays, computed tomography, or magnetic resonance imaging showing classic signs of osteomyelitis, such as bone necrosis, periosteal reaction, or bone destruction; (b) Positive bacterial tissue cultures from tibia biopsies; (c) Confirmation of the presence of pus during surgery. (2) The primary surgical approach for all patients was debridement, including intralesional curettage, intramedullary reaming, or radical debridement. The exclusion criteria were as follows: (1) Inability to obtain admission blood test results; (2) Diagnosis of osteomyelitis in multiple long bones; (3) Recent use of anticoagulants or antiplatelet drugs; (4) Coexistence of immune or hematological disorders; (5) Absence of surgical treatment; (6) Segmental bone resection as the primary surgical approach. The eligible patients from the First Affiliated Hospital of Guangxi Medical University were included as the training group, while those from the Affiliated Hospital of Guilin Medical University were used as the external validation group. Since this is a retrospective study, it was approved by the ethics committee in accordance with national laws and institutional consent, and patient informed consent was not required. In this study, the patients’ personal identifiable information was anonymized to ensure privacy and confidentiality. Data collection and definition This study collected baseline clinical data and laboratory test results including complete blood count, blood biochemistry, and coagulation function from the patients. The baseline clinical data consisted of gender, age, height, weight, body mass index (BMI), history of hypertension, history of diabetes, history of hepatitis, history of tuberculosis, history of alcohol consumption, and history of smoking. A history of hypertension was defined as a prior diagnosis of hypertension or a systolic blood pressure (SBP)/diastolic blood pressure (DBP) of 140/90 mmHg or higher upon admission. A history of diabetes was defined as a prior diagnosis of diabetes or random blood glucose level greater than 11.1 mmol/L, or fasting blood glucose (FBG) level greater than 7 mmol/L. The baseline data from the two centers were compared to validate the comparability of the data between the two centers. Additionally, this study collected laboratory test results from patients upon admission before surgery, including white blood cell count, red blood cell count, hemoglobin count, platelet count, mean corpuscular volume, absolute lymphocyte count, absolute neutrophil count, absolute monocyte count, absolute eosinophil count, absolute basophil count, platelet hematocrit, fibrinogen, albumin, aspartate aminotransferase, alanine aminotransferase, alanine aminotransferase, prothrombin time, activated partial thromboplastin time, total bilirubin, creatinine, urea, D-dimer, C-reactive protein, erythrocyte sedimentation rate, international normalized ratio, and more. The following formulas were used to calculate relevant standardized ratios: platelet-to-lymphocyte ratio (PLR) = platelet count/lymphocyte count; systemic immune-inflammation index (SII) = (neutrophil count × platelet count)/lymphocyte count; systemic inflammation response index (SIRI) = (neutrophil count × monocyte count)/lymphocyte count; neutrophil-to-lymphocyte ratio (NLR) = neutrophil count/lymphocyte count; aspartate aminotransferase-to-neutrophil ratio (ANRI) = aspartate aminotransferase count/neutrophil count; gamma-glutamyl transferase-to-platelet ratio (GPR) = gamma-glutamyl transferase count (GGT)/platelet count; D-dimer-to-albumin ratio (DAR) = D-Dimer count/albumin count; monocyte-to-lymphocyte ratio (MLR) = monocyte count/lymphocyte count. Outcome In this study, the occurrence of recurrence within 3 months after debridement surgery in patients with tibia osteomyelitis was considered as the outcome event. Recurrence was defined as a positive result, while no recurrence was defined as a negative result. The gold standard for determining the recurrence of tibia osteomyelitis in this study was based on relevant imaging results including X-rays, computed tomography, or magnetic resonance imaging with classic signs of osteomyelitis, such as bone necrosis, periosteal reaction, or bone destruction or intraoperative collection of biological tissue for examination including microbiology and histology. Skilled radiologists, pathologists, and clinicians evaluated the relevant imaging data and histopathological specimens and provided diagnostic reports to determine the postoperative recurrence of tibia osteomyelitis. Statistical analysis In this study, baseline data comparison between the two centers was performed using SPSS 21.0 (SPSS Inc., Chicago, IL). Subsequently, the data from the training and validation groups were combined, and the clinical data and laboratory test results were compared between the groups based on the occurrence of recurrence. For continuous data, normality was assessed using the Shapiro-Wilk test. If the data followed a normal distribution, it was presented as mean ± standard deviation, and group comparisons were made using independent one-way analysis of variance (ANOVA). If the data did not follow a normal distribution, it was presented as median (25th percentile, 75th percentile), and group comparisons were made using the Kruskal-Wallis test. Categorical data were described using frequencies (percentages), and group comparisons were made using the chi-square test or Fisher’s exact test. A two-sided p-value less than 0.05 was considered statistically significant. The optimal cutoff values, sensitivity, specificity, and other results of factors showing statistical differences in group comparisons were calculated using SPSS 21.0. GraphPad Prism 9.5.0 was used to generate receiver operating characteristic (ROC) curves for factors with significant differences. The clinical factors were categorized using the optimal cutoff values, and then binary logistic regression analysis was performed to evaluate the independent predictive factors for recurrence. The independent predictive factors were plotted as a nomogram using R Studio (version 4.2.2). The predictive performance of the model was evaluated using ROC curves and the area under the curve (AUC). The calibration plot was used to calculate the average error of the model, and the decision curve analysis (DCA) was performed to assess the clinical utility of the model. All tests were considered statistically significant if the two-sided p-value was less than 0.05. Result Relationship between clinical factors and postoperative recurrence in debridement A total of 501 patients diagnosed with tibial osteomyelitis were collected from two large medical centers for this study. However, 90 patients were excluded due to the aforementioned factors, resulting in a final inclusion of 411 patients for a retrospective analysis, as illustrated in Fig. 1. Among them, there were 252 patients in the training set and 159 patients in the validation set. Flow chart of this study The baseline data of the two groups were compared. In the training group, there were 50 female participants (19.84%), with an average age of 49 years (34–61), a body weight of 60 kg (53.88-70), a height of 165 cm (160–170), and a BMI of 22.69 (20.2–25.1). Among them, 35 patients (13.89%) had a history of hypertension, 41 patients (16.27%) had a history of diabetes, 61 patients (24.21%) had a smoking history, and 39 patients (15.48%) had a history of alcohol consumption. No statistically significant differences were observed in the baseline data between the two groups (p > 0.05), indicating comparability of the data (Table 1). Table 1Baseline data table between training group and validation groupVariableTraining group(n = 252)Validation group(n = 159) p Sex0.822 Female50 (19.84%)33 (20.75%) Male202 (80.16%)126 (79.25%)Hypertension0.704 No217 (86.11%)139 (87.42%) Yes35 (13.89%)20 (12.58%)Diabetes0.402 No211 (83.73%)128 (80.5%) Yes41 (16.27%)31 (19.5%)Hepatitis1.000 No249 (98.81%)157 (98.74%) Yes3 (1.19%)2 (1.26%)Tuberculosis1.000 No250 (99.21%)158 (99.37%) Yes2 (0.79%)1 (0.63%)Smoke0.288 No191 (75.79%)113 (71.07%) Yes61 (24.21%)46 (28.93%)Alcoholism0.169 No213 (84.52%)142 (89.31%) Yes39 (15.48%)17 (10.69%)Weight(kg)60 (53.88,70)61 (56,65)0.616Height(cm)165 (160,170)165 (158,169)0.620BMI22.69 (20.2,25.1)22.65 (21.02,24.16)0.751Age49 (34,61)50 (41,60)0.141 Baseline data table between training group and validation group The data from the two groups that were collected before debridement surgery were combined and divided based on the presence or absence of postoperative recurrence in debridement. The group with recurrence was defined as the “Recurrence group,” while the group without recurrence was defined as the “No Recurrence group.” A comparison between the two groups revealed that MCV, WBC, NEU, LYM, ALB, D-Dimer, CRP, PLT, ESR, SII, SIRI, PLR, ANRI, DAR, NLR, and MLR exhibited statistically significant differences among the different groups of reccurence (p < 0.05) (Table 2). To facilitate visual comparison, violin plots were created for the factors that showed differences (Fig. 2). The optimal cutoff values, sensitivity, and specificity were calculated for each differentiating factor (Table 3). ROC curves were plotted, and the area under the curve (AUC) was calculated (Fig. 3). The results showed that the optimal cutoff value for MCV was 73.48, with a sensitivity of 0.745 and a specificity of 0.438. For WBC, the optimal cutoff value was 8.87 × 109/L, with a sensitivity of 0.588 and a specificity of 0.688. The optimal cutoff value for NEU was 6.49 × 109/L, with a sensitivity of 0.662 and a specificity of 0.635. LYM had an optimal cutoff value of 1.995 × 109/L, with a sensitivity of 0.814 and a specificity of 0.451. The optimal cutoff value for ALB was 38.15 g/L, with a sensitivity of 0.854 and a specificity of 0.581. D-Dimer had an optimal cutoff value of 329, with a sensitivity of 0.577 and a specificity of 0.716. The optimal cutoff value for CRP was 9.445, with a sensitivity of 0.526 and a specificity of 0.695. PLT had an optimal cutoff value of 317 × 109/L, with a sensitivity of 0.583 and a specificity of 0.548. ESR had an optimal cutoff value of 36, with a sensitivity of 0.711 and a specificity of 0.675. SII had an optimal cutoff value of 711.102, with a sensitivity of 0.8 and a specificity of 0.454. SIRI had an optimal cutoff value of 1.925, with a sensitivity of 0.835 and a specificity of 0.574. PLR had an optimal cutoff value of 135.179, with a sensitivity of 0.828 and a specificity of 0.402. ANRI had an optimal cutoff value of 2.643, with a sensitivity of 0.372 and a specificity of 0.779. DAR had an optimal cutoff value of 9.292, with a sensitivity of 0.646 and a specificity of 0.671. NLR had an optimal cutoff value of 3.243, with a sensitivity of 0.717 and a specificity of 0.594. Finally, MLR had an optimal cutoff value of 0.345, with a sensitivity of 0.559 and a specificity of 0.574. Table 2Baseline data table for comparison of recurrence group and no recurrence groupVariableNo recurrence (n = 260)Recurrence (n = 151) p Sexual0.9 Female53 (20.38%)30 (19.87%) Male207 (79.62%)121 (80.13%)Hypertension0.082 No231 (88.85%)125 (82.78%) Yes29 (11.15%)26 (17.22%)Diabetes0.883 No215 (82.69%)124 (82.12%) Yes45 (17.31%)27 (17.88%)Hepatitis0.212 No255 (98.08%)151 (100%) Yes5 (1.92%)0 (0%)Tuberculosis1 No258 (99.23%)150 (99.34%) Yes2 (0.77%)1 (0.66%)Smoke0.531 No195 (75%)109 (72.19%) Yes65 (25%)42 (27.81%)Alcoholism0.172 No220 (84.62%)135 (89.4%) Yes40 (15.38%)16 (10.6%)Weight (kg)61 (54.5,70)61 (55,68.25)0.784Height (cm)165 (159,170)165 (158.5,170)0.429BMI22.66 (20.77,25.07)22.66 (20.74,24.51)0.391Age50 (40,61)49 (39.5,60)0.752RBC(1012/L)4.82 ± 1.034.86 ± 0.950.721Hb(g/L)120 (102,134.5)120 (103.3,133.3)0.966HCT0.38 (0.33,0.43)0.38 (0.33,0.44)0.721MCV(fl.)77.37 (66.6,85.04)81.72 (72.73,88.47)< 0.001MCHC(g/L)323.3 (311.6,332.3)320.5 (310.3,330.8)0.197WBC(109/L)7.7 (6.11,9.23)9.21 (7.97,10.38)< 0.001NEU(109/L)5.39 (3.83,7.56)7.33 (5.25,8.66)< 0.001LYM(109/L)1.93 (1.38,2.42)1.51 (1.12,1.94)< 0.001MONO(109/L)0.63 (0.48,0.8)0.59 (0.45,0.77)0.352EOS(109/L)0.2 (0.11,0.37)0.21 (0.11,0.39)0.644BAS(109/L)0.04 (0.03,0.08)0.04 (0.02,0.06)0.077PCT0.26 (0.2,0.34)0.25 (0.2,0.31)0.221PT(S)11.6 (10.6,12.5)11.5 (10.7,12.57)0.881APTT(S)32.2 (29.6,34.8)32.6 (30.02,35.4)0.174FIB(g/L)4.01 (3.16,5.15)4 (3.22,4.93)0.871ALB(g/L)38.32 ± 7.0932.26 ± 5.7< 0.001GGT(U/L)47 (24,103)45 (28.38,78.5)0.883ALT(U/L)23 (14,39)24 (14,36)0.74AST(U/L)25 (16,35)23 (17,34)0.81TBIL(umol/L)7.6 (5.3,10.1)6.85 (4.7,9.53)0.048Cr(umol/L)69 (54,121.5)68.5 (54.75,104)0.93BUN(mmol/L)4.61 (3.55,5.73)4.69 (3.59,6.22)0.406D.Dimer(ng/ml)245 (119.75,455)395 (187.5,587.5)< 0.001CRP(mg/L)7.82 (5.5,10.44)15 (7.18,21)< 0.001PLT(109/L)300.7 (224.55,403.15)336 (249.9,435.7)0.034ESR(mm)25 (11,47)45 (31.5,75.5)< 0.001SII827.89 (504.37,1546.85)1405.82 (824.27,2727.81)< 0.001SIRI1.64 (1.05,3.02)2.62 (1.47,4.61)< 0.001PLR161.14 (112.3,264.79)219.34 (148.37,319.42)< 0.001ANRI4.47 (2.88,6.63)3.51 (2.13,5.23)0.002GPR0.15 (0.07,0.3)0.13 (0.07,0.24)0.207DAR6.42 (3.08,12.43)11.64 (5.82,17.92)< 0.001NLR2.89 (1.96,4.47)4.53 (2.89,7.04)< 0.001MLR0.32 (0.23,0.46)0.36 (0.26,0.54)0.021 Baseline data table for comparison of recurrence group and no recurrence group The violin plot Table 3Single binary logistic regression analysis resultsindepentvarOR[95%CI] p Sexual0.666 Male1 Female1.12[0.67–1.88]Hypertension0.508 No1 Yes0.82[0.45–1.49]Diabetes0.827 No1 Yes0.95[0.57–1.56]Hepatitis0.629 No1 Yes0.57[0.06–5.54]Tuberculosis0.902 No1 Yes0.86[0.08–9.56]Smoke0.807 No1 Yes0.95[0.61–1.47]Alcoholism0.336 No1 Yes0.74[0.4–1.37]Weight(kg)0.629 ≤ 60.21 > 60.21.1[0.74–1.65]Height(cm)0.118 ≤ 163.61 > 163.61.39[0.92–2.09]BMI0.862 ≤ 22.751 > 22.750.97[0.65–1.44]Age0.721 ≤ 48.81 > 48.81.08[0.72–1.61]RBC(1012/L)0.08 ≤ 4.841 > 4.841.44[0.96–2.17]Hb(g/L)0.992 ≤ 118.651 > 118.651[0.67–1.5]HCT0.636 ≤ 0.381 > 0.381.1[0.73–1.66]MCV(fl.)0.898 ≤ 77.761 > 77.760.97[0.65–1.46]MCHC(g/L)0.941 ≤ 320.731 > 320.730.98[0.65–1.48]WBC(109/L)< 0.001 ≤ 8.781 > 8.784.11[2.68–6.29]NEU(109/L)< 0.001 ≤ 6.541 > 6.543.51[2.31–5.35]LYM(109/L)< 0.001 ≤ 1.801 > 1.800.39[0.26–0.6]MONO(109/L)0.026 ≤ 0.651 > 0.651.6[1.06–2.42]EOS(109/L)0.106 ≤ 0.251 > 0.250.7[0.46–1.08]BAS(109/L)0.002 ≤ 0.051 > 0.051.92[1.28–2.89]PCT0.134 ≤ 0.281 > 0.280.73[0.48–1.1]PT(S)0.014 ≤ 11.701 > 11.701.67[1.11–2.52]APTT(S)0.22 ≤ 32.521 > 32.520.77[0.51–1.17]FIB(g/L)0.011 ≤ 4.131 > 4.131.7[1.13–2.57]ALB(g/L)< 0.001 ≤ 36.101 > 36.100.12[0.07–0.19]GGT(U/L)< 0.001 ≤ 65.341 > 65.342.27[1.46–3.53]ALT(U/L)< 0.001 ≤ 28.721 > 28.720.37[0.23–0.58]AST(U/L)< 0.001 ≤ 28.431 > 28.432.85[1.82–4.46]TBIL(umol/L)< 0.001 ≤ 9.491 > 9.490.19[0.11–0.33]Cr(umol/L)< 0.001 ≤ 87.171 > 87.172.3[1.48–3.56]BUN(mmol/L)< 0.001 ≤ 4.941 > 4.940.34[0.21–0.53]D-Dimer(ng/ml)< 0.001 ≤ 385.821 > 385.822.42[1.6–3.67]CRP(mg/L)< 0.001 ≤ 20.351 > 20.352.41[1.49–3.9]PLT(109/L)0.903 ≤ 335.111 > 335.110.97[0.64–1.48]ESR(mm)< 0.001 ≤ 40.301 > 40.304.05[2.64–6.22]SII< 0.001 ≤ 1590.061 > 1590.062.69[1.75–4.13]SIRI0.003 ≤ 3.151 > 3.151.96[1.27–3.05]PLR0.005 ≤ 224.321 > 224.321.81[1.2–2.74]ANRI0.015 ≤ 4.921 > 4.920.59[0.38–0.9]GPR0.062 ≤ 0.231 > 0.230.66[0.42–1.02]DAR< 0.001 ≤ 11.941 > 11.942.47[1.62–3.78]NLR< 0.001 ≤ 4.481 > 4.483.07[2-4.71]MLR0.052 ≤ 0.431 > 0.431.52[1-2.33] Single binary logistic regression analysis results the ROC plot of clinical factors Clinical data, including factors representing intergroup differences, was incorporated into a univariate binary logistic regression analysis. The results indicated that the following were risk factors for recurrence after debridement surgery for tibial osteomyelitis: WBC [4.11 (2.68, 6.29), p < 0.001], NEU [3.51 (2.31, 5.35), p < 0.001], LYM [0.39 (0.26, 0.6), p < 0.001], MONO [1.6 (1.06, 2.42), p = 0.026], BAS [1.92 (1.28, 2.89), p = 0.002], PT [1.67 (1.11, 2.52), p = 0.014], FIB [1.7 (1.13, 2.57), p = 0.011], ALB [0.12 (0.07, 0.19), p < 0.001], GGT [2.27 (1.46, 3.53), p < 0.001], ALT [0.37 (0.23, 0.58), p < 0.001], AST [2.85 (1.82, 4.46), p < 0.001], TBIL [0.19 (0.11, 0.33), p < 0.001], Cr [2.3 (1.48, 3.56), p < 0.001], BUN [0.34 (0.21, 0.53), p < 0.001], D-Dimer [2.42 (1.6, 3.67), p < 0.001], CRP [2.41 (1.49, 3.9), p < 0.001], ESR [4.05 (2.64, 6.22), p < 0.001], SII [2.69 (1.75, 4.13), p < 0.001], SIRI [1.96 (1.27, 3.05), p = 0.003], PLR [1.81 (1.2, 2.74), p = 0.005], ANRI [0.59 (0.38, 0.9), p = 0.015], DAR [2.47 (1.62, 3.78), p < 0.001], NLR [3.07 (2, 4.71), p < 0.001] (Table 3). When factors that were significant in the univariate analysis were included in a multivariate binary logistic regression analysis, WBC [OR = 3.36 (1.77–6.37), p < 0.001], LYM [OR = 0.29 (0.15–0.54), p < 0.001], BAS [OR = 2.19 (1.22–3.91), p = 0.008], FIB [OR = 0.27 (0.14–0.54), p < 0.001], ALB [OR = 0.16 (0.09–0.3), p < 0.001], TBIL [OR = 0.41 (0.21–0.83), p = 0.012], D-Dimer [OR = 1.77 (1.03–3.03), p = 0.038], CRP [OR = 0.43 (0.2–0.91), p = 0.027], ESR [OR = 3.08 (1.6–5.79), p < 0.001], PLR [OR = 0.47 (0.24–0.9), p = 0.022] were identified as independent predictors (Table 4). Table 4Multivariate binary logistic regression analysis resultsVariableOR[95%CI] p WBC(109/L)< 0.001 ≤ 8.781 > 8.783.36[1.77–6.37]LYM(109/L)< 0.001 ≤ 1.801 > 1.800.29[0.15–0.54]BAS(109/L)0.008 ≤ 0.051 > 0.052.19[1.22–3.91]FIB(g/L)< 0.001 ≤ 4.131 > 4.130.27[0.14–0.54]ALB(g/L)< 0.001 ≤ 36.101 > 36.100.16[0.09–0.3]TBIL(umol/L)0.012 ≤ 9.491 > 9.490.41[0.21–0.83]D-Dimer(ng/ml)0.038 ≤ 385.821 > 385.821.77[1.03–3.03]CRP(mg/L)0.027 ≤ 20.351 > 20.350.43[0.2–0.91]ESR(mm)< 0.001 ≤ 40.301 > 40.303.08[1.6–5.79]PLR0.022 ≤ 224.321 > 224.320.47[0.24–0.9] Multivariate binary logistic regression analysis results Nomogram construction and evaluation To further validate the predictive power of various factors on the recurrence of debridement surgery in patients with tibial osteomyelitis, a nomogram model was constructed using R studio software based on the results of multivariate analysis (WBC, LYM, BAS, FIB, ALB, TBIL, D-Dimer, ESR, PLR). We used the data from the training group to build the nomogram, as shown in Fig. 4. The predictive performance of this nomogram was validated by drawing the ROC curve for the training group (Fig. 5A), with a C-index of 0.825, indicating good predictive performance. The calibration curve for the training group (Fig. 6A) suggested an average error of 0.034, and the decision curve analysis (DCA) for the training group (Fig. 7A) indicated that there was a good clinical benefit for threshold probabilities between 0.01 and 0.66. The model was also validated using the data from the validation group; the ROC curve was plotted (Fig. 5B), showing a C-index of 0.843, the calibration curve (Fig. 6B) indicated an average error of 0.038, and the DCA (Fig. 7B) showed good clinical benefit for threshold probabilities between 0.01 and 0.92. It is clear that the nomogram demonstrated good predictive performance in both the training and validation groups, with small errors and the potential to benefit a large portion of clinical patients. The nomogram of clinical factors A ROC curve of training set data; B ROC curve of validation set data A Calibration curve of training set data; B Calibration curve of validation set data A DCA curve of training data set; B DCA curve of validation data set Discussion Accurately estimating the likelihood of recurrence in patients is a major challenge in devising treatment plans for those with osteomyelitis. Apart from the surgical approach, changes in preoperative blood indicators are one of the more readily available pieces of information during a patient’s clinical course and can directly reflect the current status of the patient. This study, through retrospective analysis of patients’ preoperative blood indicators, provides clinicians with a simple and effective tool for planning treatment strategies. Our results showed that WBC, LYM, BAS, FIB, ALB, TBIL, D-Dimer, ESR, and PLR were independent predictors of tibial osteomyelitis recurrence. There have been few studies on the recurrence of tibial osteomyelitis in the past. Xiaohua Wang et al. [14] found that the recurrence rate of tibial infection (16.22%) was higher than that of femur (8.70%), and the recurrence rate of post-traumatic osteomyelitis (14.66%) was significantly higher than that of hematogenous osteomyelitis (2.41%). Repeated surgery (≥ 3 times), post-traumatic osteomyelitis, and primary internal fixation are risk factors for recurrence of infection. Yalikun A et al. [15] proposed that Pseudomonas aeruginosa infection, bone exposure, number of previous surgeries (> 3 times), and diabetes were associated with postoperative recurrence or reinfection of tibial osteomyelitis. Our results are the first to propose a predictive relationship between serological indicators such as inflammatory factors and the recurrence of tibial osteomyelitis. The treatment of osteomyelitis requires a multifaceted approach. In addition to surgical intervention, antibiotic therapy is an indispensable component of the treatment course. Debridement is often part of the initial treatment, and surgical debridement following antibiotic therapy can significantly shorten hospital stays, reduce treatment costs, and significantly decrease the incidence of postoperative complications [8]. Open bone injuries often lead to complications of osteomyelitis. When bone is exposed to the external environment, the bone cells and the organic extracellular matrix components that produce bone are ideal colonization targets for microbes, especially Staphylococcus aureus [16]. Once established, S. aureus can produce a biofilm, making the infection persistent [17, 18]. Additionally, S. aureus can produce a variety of toxins, which may encourage adherence of organisms to tissues leading to re-colonization and reinfection [19], which may be one of the reasons why osteomyelitis is difficult to cure completely and prone to recurrence. Many studies have confirmed that abnormalities in white blood cell count (WBC) are very common in patients with infections [20, 21]. Research by Rimmer et al. suggests that different trajectories of WBC count increase can lead to higher mortality rates in patients with infections [22]. In burn patients, the initial white blood cell count has excellent predictive ability for subsequent bloodstream infections [23]. Beyond standard hematologic inflammatory markers used in the clinic, the platelet-to-lymphocyte ratio (PLR) serves as a standardized inflammatory predictor in various aspects. Wang et al., through research on PLR and other related inflammatory factors, constructed a nomogram to predict the probability of pneumonia related to stroke and ICU admission in patients with lung infections [24]. Peng et al. built a more accurate prognostic model than the current staging system for nasopharyngeal carcinoma patients based on PLR and other inflammatory factors [25]. Apart from common blood markers, composite indicators should also be valued in our clinical work. Nomograms constructed by combining traditional blood markers and composite indicators will be more beneficial for the formulation of treatment plans for patients. This study has several strengths: (1) It is a dual-center retrospective study from China, which has developed a model through a training set and validated it with an external validation set. The model is reliable and simple to apply. (2)Routine blood and biochemical tests are commonly performed at the time of patient admission. This study aims to collect relevant data upon admission to achieve very early prediction and potentially benefit more patients. However, there are also limitations to this study: (1) It does not track postoperative and long-term incidence rates in patients. (2) Future prospective studies could be introduced to further validate the results of this retrospective study. Therefore, the current study cannot draw any conclusions about the efficacy of treatments (such as the application of local antibiotics, hyperbaric oxygen therapy [26], platelet-rich plasma therapy [27], etc.) and the nomogram obtained from our study should not be used to guide clinical decisions in isolation because it does not analyze functional prognosis. Conclusion Patients admitted with tibial osteomyelitis demonstrate that WBC, LYM, BAS, FIB, ALB, TBIL, D-Dimer, CRP, ESR, and PLR are nomogram constructed based on the relevant blood markers and PLR can assist clinicians in gauging the likelihood of early recurrence following tibial osteomyelitis surgery. This facilitates early prevention and treatment, benefiting a larger patient population. Acknowledgements Not applicable. Abbreviations Body mass index Systolic blood pressure Diastolic blood pressure Fasting blood glucose Platelet-to-lymphocyte ratio Systemic immune-inflammation index Systemic inflammation response index Neutrophil-to-lymphocyte ratio Aspartate aminotransferase-to-neutrophil ratio Gamma-glutamyl transferase-to-platelet ratio gamma-glutamyl transferase count D-dimer-to-albumin ratio Monocyte-to-lymphocyte ratio White blood cell count Red blood cell count Hemoglobin Mean corpuscular volume Mean corpuscular hemoglobin concentration Neutrophils Lymphocyte Monocyte Eosinophils Basophil Erythrocyte sedimentation rate Albumin Aspartate aminotransferase Alanine aminotransferase Total bilirubin Prothrombin time Activated partial thrombin time Fibrinogen Creatinine Blood urea nitrogen Odds ratio Confidence interval The area under the curve Decision curve analysis Receiver operating characteristic Author contributions HL and ZH designed this study. RBL and SHX are responsible for writing articles, conducting statistical analysis, reviewing articles, and creating images. GYD and WCL are responsible for collecting data and conducting statistical analysis. All authors reviewed the manuscript. All authors read and approved the final manuscript. Funding The authors received no financial support for the research, authorship, and/or publication of this article. Data availability The data used to support the findings of this study are available from the corresponding author upon request. Declarations Ethics approval and consent to participate This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the the First Affiliated Hospital of Guangxi Medical University Ethics Review Committee ( 2024-E209-01). Exemption of informed consent from patients with the consent of the First Affiliated Hospital of Guangxi Medical University Ethics Review Committee. All authors have read, approved, and consented to participate in the final manuscript. Consent for publication Not Applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Rongbin Lu and Shaohuai Xia contributed equally to this work. References Associated Data Data Availability Statement The data used to support the findings of this study are available from the corresponding author upon request.